摘要:

Abstract. Data on protective groups for the thiol function of cysteine and methods of disulfide bonds formation used in modern peptide chemistry are considered and systematised. The advantages and disadvantages of protective groups, of reagents used for cyclisation, and possible side reactions associated with them are described. The bibliography includes 119 references. I. Introduction Among other peptides and proteins occurring in nature, those containing cysteine constitute a separate group.Disulfide bridges formed between pairs of cysteine residues play an essential role in the stabilisation of tertiary and quaternary structures of the protein molecules by determining their rigid spatial orientation and giving them a unique shape required for the fulfilment of their biological functions.In low-molecular-mass peptides, all types of noncovalent contacts are weakened; therefore, disulfide bonds are of crucial importance for the structural stabilisation of a molecule and reproduction of its biological activity. Many peptides are known that contain one, two, or more disulfide bridges and take part in the regulation of physiological processes occurring in the organism.1, 2 Synthetic peptides includ- ing those with S7S bonds are widely employed in many areas of medicine and pharmacology.3, 4 Therefore, the methods used for the formation of disulfide bridges are of great practical signifi- cance.Modern peptide chemistry possesses a wide range of methods for synthesis, purification, and analysis, which allow one to obtain virtually any kind of peptides, their analogues, and even mini- proteins.However, the synthesis of peptides containing disulfide bonds still presents a challenge because of low yields at the stage of their closure. Very often, all efforts spent to synthesise linear precursors come to an overall negative result. Elucidation of the reason for such failures often goes beyond the scope of synthetic studies. The reviews devoted to the synthesis of peptides having disulfide bridges are very few in number and are often limited to particular aspects of the general problem.The present review considers modern approaches to the syn- thesis of cysteine-containing peptides, the groups employed for protection of the thiol function, the methods for their removal, and most popular reagents used for the closure of the disulfide bonds.II. Protective groups for the thiol function It is necessary that the strongly nucleophilic thiol group of cysteine be protected in the course of peptide synthesis. The protective groups for the SH function of cysteine should meet the same demands as those for the protective groups of the side-chain functions of other amino acids, viz., stability in the reactions of peptide chain elongation and deblocking of a-amino groups and retention of configuration of the asymmetric centre in the course of synthesis. The protective groups for the thiol function that have been successfully used in peptide syntheses are listed in Table 1.The choice of a protective group is determined by the procedure of subsequent cyclisation.Depending on conditions of cysteine deblocking, one may obtain free thiols or disulfides. The protec- tive groups should possess different chemical stability, which would enable the selective formation of two or more disulfide bonds if required. In order to characterise protective groups for the thiol func- tion of cysteine used in modern peptide chemistry, Hiskey 5, 6 has proposed to classify them according to the structure of the protected derivatives, viz., sulfides, thioacetals, S-acylamido- methyl thiazolidine derivatives, thioesters, and disulfides. 1. Protective groups that form sulfides and related compounds Protective groups that form this type of compounds (RSR0), are the most well-known, numerous, and popular. The benzyl group proposed by Sifferd and du Vigneaud is an example of a classical protective group that forms sulfides (see Ref. 1). Rather drastic conditions necessary for the removal of the protective benzyl group (Na in liquid ammonia or HF at 25 8C) impose certain limitations on its application. It has been observed 1, 43 that HF results in desulfurisation of the cysteine residue, which may also occur with Na/NH3.In addition, deprotection with Na/NH3 is sometimes accompanied by the cleavage of the peptide bonds; the bonds formed by the amino group of proline appear to be the most labile under these conditions.1 In this connection, it was necessary to design acid-labile protective groups by introducing electron- donor substituents into the aromatic ring of benzyl alcohols.A great number of such protective groups have been synthesised. In E V Kudryavtseva,MV Sidorova Russian Cardiological Scientific Centre, Ministry of Health Care and Medical Industry of the Russian Federation, 3-ya Cherepkovskaya ul. 15a, 121552 Moscow, Russian Federation. Fax (7-095) 414 67 86. Tel. (7-095) 414 67 16 R P Evstigneeva MV Lomonosov Moscow Academy of Fine Chemical Technology, prosp.Vernadskogo 86, 117571 Moscow, Russian Federation. Fax (7-095) 430 79 83. Tel. (7-095) 434 86 78 Received 30 October 1997 Uspekhi Khimii 67 (7) 611 ± 630 (1998); translated by R L Birnova UDC 547.466.96.07 Some peculiarities of synthesis of cysteine-containing peptides E V Kudryavtseva,MV Sidorova, R P Evstigneeva Contents I. Introduction 545 II. Protective groups for the thiol function 545 III.Methods for the formation of disulfide bonds 552 IV. Conclusion 560 Russian Chemical Reviews 67 (7) 545 ± 562 (1998) #1998 Russian Academy of Sciences and Turpion LtdTable 1. The protective groups for the thiol function of cysteine employed in peptide synthesis. Name a Formula of the protective Conditions where Conditions for removal Ref. group the protective group of the protective group is stable b Benzyl (Bzl) CH2Ph HF, 0 8C; Tl(III); B Na/NH3(liq); HF, 2, 6 ± 8 RSCl; H2/Pd; HBr/AcOH 25 8C, 30 min Methylbenzyl (Meb or Mbzl) CH2C6H4Me-p TFA; Ag(I); RSCl; B HF/MeOPh, 0 8C, 6 ± 9 60 min;c Tl(III) 3,4-Dimethylbenzyl (Dmb) TFA; Ag(I); B; RSCl HF/MeOPh, 8 0 8C, 10 min 2,4,6-Trimethylbenzyl (Tmb) TFA, 23 8C, 2 h HF/MeOPh, 0 8C, 10 30 min; TFA, D, 1 h 4-Methoxybenzyl (Mob) CH2C6H4OMe-p B; RSCl; I2 HF, 0 8C; Hg(II); 6 ± 8, 10 ± 13 Ag(I); Tl(III); (BrC6H4)3N; TFA:CH2Cl2 (1 : 1), 25 8C c 2,4,6-Trimethoxybenzyl (Tmob) B; Nu 30% TFA/scavengers,d 2, 7, 14, 15 I2; Tl(III), HF tert-Butyl But TFA; HF, 0 8C; B; I2; Ag(I) HF, 20 8C; HF/scavengers,d 2, 12, 16 0 8C; Hg(II); NpsCl; Tl(III) Diphenylmethyl CHPh2 TFA; I2;e Hg(II);e Ag(I) e TFA/MeOPh, D; HF; 5, 17 (or benzhydryl) (Dpm) Na/NH3(liq); HBr/AcOH; (SCN)2; NpsCl Di(4-methoxyphenyl)methyl CH(C6H4OMe-p)2 HCl/AcOH;e B; Tl(III) TFA, 70 8C, 2 h; TFA/MeOPh, 1, 5, 8 (Ddm) 10 min; TFA:PhOH (85 : 15), 25 8C, 2 h Triphenylmethyl (Trt) CPh3 B; Nu TFA/scavengers;d 1, 2, 5 ± 8, 18 Hg(II); Ag(I); I2; Tl(III); RSCl 4,40,400-Trimethoxytriphenyl- C(C6H4OMe-p)3 B; Nu TFA (dilute); I2 2, 7 methyl (TMTr) (1 ± 2 equiv.), 25 8C, 15 ± 30 min Pyridyldiphenylmethyl B Hg(II) pH 4; Zn/AcOH; 19 electrolytic reduction Dibenzosuberyl HCl(gas) in organic HBr/AcOH;c HF 6 solvents; TFA; B Fluorenylmethyl (Fm) HF; TFA; Tl(III) or I2 NH3/MeOH; Pip 5, 20 (stable for 1 h) 2-(2,4-Dinitrophenyl)ethyl HF; I2; Tl(III); TFA; Pip 21 (Dnpe) TFMSA b,b-Diethoxycarbonylethyl CH2CH(COOEt)2 HBr/AcOH 1 M KOH, 20 8C, 5 ± 10 min 1 (Dce) Adamantyl (Ad) TFAi Tl(III); 1MTFMSA/PhSMe 13, 22, 23 in TFA, 0 8C Xanthenyl, B dilute acids in the presence 24 R=H (Xan), of scavengers;d R=2-MeO (2-Moxan), I2; Tl(III) R=3-MeO (3-Moxan) 2-Quinolylmethyl AH FeCl3/DMF/H2O; 25 FeCl2/DMF/H2O; CuCl2/DMF/H2O Ferrocenylmethyl (Fcm) C5H5FeC5H4CH2 NaOH; NaOMe; I2; TFA; (SCN)2; RSCl; 26 N2H4; Pip Hg(II); Ag(I) Me Me CH2 Me CH2 Me Me OMe CH2 MeO MeO N (Ph)2C CH2 NO2 O2N (CH2)2 O R N H2C 546 E V Kudryavtseva,MV Sidorova, R P EvstigneevaTable 1 (continued). Name a Formula of the protective Conditions where Conditions for removal Ref.group the protective group of the protective group is stable b 2-Ferrocenylpropan-2-yl (Fc) C5H5FeC5H4CMe2 NaOH; NaOMe; I2; TFA; (SCN)2; RSCl; 26 N2H4; Pip Hg(II); Ag(I) (Z5-Cyclohexa-3,5-dienyl)tri- NaOH/Me2CO HBF4/CHCl3 27 carbonyl iron Tetrahydropyranyl (Thp) B 0.5 M AgNO3, 0 8C; TFA; 5, 6, 8 HBr/TFA; (SCN)2 Isobutoxymethyl (i-Bom) CH2OBui B TFA; HCl/AcOH; Hg(II) 5, 6, 8 p-Chlorophenoxymethyl CH2OC6H4Cl-p B TFA 5 Benzylthiomethyl (Btm) CH2SCH2Ph B HF; Hg(AcO)2/HCOOH 5, 6, 8 (80%), 25 8C, 20 min Phenylthiomethyl CH2SPh B HF; Hg(II) 5 Acetamidomethyl (Acm) CH2NHCOMe HF, 0 8C; B; H2/Pd Ag(I); Hg(II); I2; Tl(III); 2, 5, 6, 8, AgBF4; RSCl 18, 28, 29 Trimethylacetamidomethyl CH2NHCOBut HF; TFA; 1 M TFMSA; B Hg(AcO)2; TFA/MeOPh; 28 ± 31 (Tacm) I2 (10 equiv.) in AcOH; AgBF4 Phenylacetamidomethyl (Phacm) CH2NHCOCH2Ph HF; B Hg(II); I2; Tl(III); penicilline 2, 32, 33 aminohydrolase Benzamidomethyl (Bam) CH2NHCOPh AH, B Ag(OTf); Hg(II) 6,18 (2-Oxo-1-pyrrolidino)methyl B Hg(II); I2/MeOH 34 (Pym) Acetyl (Ac) COMe TFA (boiling); HBr/AcOH 0.1 M NaOMe/MeOH, 25 8C, 1, 5, 8 5 ± 10 min; N2H4/MeOH Benzoyl (Bz) COPh TFA (boiling); HBr/AcOH 0.1 M NaOMe/MeOH, 1, 5, 8 25 8C; N2H4/MeOH; 0.2 M NaOH Benzyloxycarbonyl (Z) COOCH2Ph 0.1 M NaOMe/MeOH, 25 8C; 1, 5, 8 aqueous alkali; NH3; Na/NH3(liq); TFA/PhOH, D; HBr/AcOH p-Methoxybenzyloxycarbonyl COCH2C6H4OMe-p TFA, 20 8C; 0.2N HBr/AcOH; 1, 8 (pMZ) HCl/AcOH; 0.5N NaOMe in DMF; Na/NH3(liq) tert-Butoxycarbonyl COOBut AH (TFA) 8 (Boc) p-Nitrophenoxycarbonyl COOC6H4NO2-p AH(TFA); Na/NH3(liq) 5 Phenylthiocarbonyl COSPh AH 5 Butylthiocarbonyl COSBu AH 5 Ethylcarbamoyl (Ec) CONHEt 4.5 M HBr/AcOH; 1 M HCl; Na/NH3; 4 M NH3; NH3(liq): 5, 6, 8 TFA, D MeOH (1 : 1); 1 M NaOH; NaOMe; N2H4/MeOH; Hg(II) (pH 7); Ag(I) (pH 7) N-Methoxymethylcarbamoyl CONHCH2OMe AH Na/NH3(liq); NaOMe/MeOH; 8 aqueous buffers (pH 9.5), 25 8C N-[b-(N-Acyl-N-methylamino)- CONHCH2CH2N(R)Me AH 6 ethyl]carbamoyl R=PhCH2OCO, ButOCO Alkylthio SR AH; NH4OH; NaOH b-mercaptoethanol; benzenethiol 6, 8, 35 (SEt, SPri, STrt) R=Et, Pri, CPh3 (pH 7.5 ± 9); dithiothreitol; Bu3P; oxidative sulfitolysis tert-Butylthio (SBut) SBut HCl/dioxane; Pip; TFA; Na/NH3; RSH; Bu3P; 2-pyri- 2, 6, 8, 17, RSCl; B dinesulfenyl chloride; 35 ± 37 o-nitrobenzenesulfenyl chloride Methoxycarbonylthio (Scm) SCOOMe HF; AH RSH; mild alkaline 1, 6, 8, 35, 36 reagents (pH 10.6) (N-Methyl-N-phenylcarb- SCON(Me)Ph HF RSH (especially dithiothreitol 2 amoyl)thio (Snm) and 2-pyridinethiol) (CO)3Fe + O CH2N O Some peculiarities of synthesis of cysteine-containing peptides 547Table 1, they are listed in the order of the increase in their acid lability. The most acid-labile group is S-2,4,6-trimethoxybenzyl protection,5, 7, 14 which is easily removed with 30% trifluoroacetic acid (TFA) in CH2Cl2 in the presence of phenol, thioanisole, or water (the so-called scavengers whose purpose is to neutralise the carbocations formed upon cleavage of protective groups) as well as with 5% TFA in CH2Cl2 in the presence of 3% triethylsilane or tri(isopropyl)silane. Scavengers are necessary to prevent the reverse addition of the trimethoxybenzyl group to the SH group.14 The distinctive feature of methoxy- and trimethoxyben- zyl protective groups is the possibility of their removal with heavy metal salts, e.g., Ag(I) and Hg(II), which makes it possible to selectively deblock cysteine.7, 11, 15 The mechanism of this reaction is analogous to that of splitting of trityl and acetamidomethyl groups and will be considered in more detail in subsequent sections.S-tert-Butyl protection is extremely acid-resistant.Its removal can be accomplished with HF or Hg(AcO)2 in TFA.2, 12 Quanti- tative removal of the tert-butyl group can be achieved with the use of strongly electrophilic reagents, such as 2-nitrophenylsulfenyl chloride, which binds two cysteine residues through intermediate mixed disulfides 2 (this reaction will be considered below).How- ever, this reagent cannot be used if the peptide chain contains tryptophan because side reactions may occur in the indole ring.16 The next type of protective groups includes di- and tri- arylmethyl groups. In Table 1, they are also presented in the order of the increase in their acid lability: (p-MeOC6H4)3C>Ph3C>(p-MeOC6H4)2CH>Ph2CH. The ease of removal of these protective groups under mild acidic conditions makes them attractive for solid-phase peptide synthesis, which makes use of Na-9-fluorenylmethoxycarbonyl derivatives of amino acids (the Fmoc-technique).4, 44, 45 Cleavage of these sulfides occurs faster than that of their tert-butyl and benzyl analogues.This is due to the greater stability of the carbocations formed; however, their higher reactivity can result in the reverse alkylation reaction of cysteine.Thus in the case of the triphenylmethyl protection, an equilibrium between the pro- tected and free cysteine is established in TFA in 30 min and in most cases the degree of S-detritylation does not exceed 60%.1, 8 The reaction equilibrium can be shifted to the right by addition of scavengers (especially, sulfur-containing ones) or water to the reaction mixture.1, 5, 16 In terms of its reactivity, the Ddm group occupies an inter- mediate position between the Trt and Dpm groups 5 and is much more acid-labile than the triphenylmethyl group.This permits selective removal of the Ddm group without affecting the Trt group.1,5, 8 On the other hand, the resistance of Ddm protection to heavy metal salts allows selective splitting of the Trt group in the presence of Ddm deprotection.Cleavage of the Trt group can be carried out with Hg(II) or Ag(I) ions.5 The mechanism of cleavage of sulfides with metal ions is similar to the mechanism of the acid-catalysed reaction: the difference between these mechanisms is that in the case of metal ions the last stage is irreversible and leads to a stable carbocation, which can be formed by two routes: In some cases, silver salts are more efficient than mercuric salts, because mercuric derivatives of peptides (mercaptides) are sometimes difficult to cleave.5 Metal mercaptides are usually destroyed by thio alcohols.To this end, the reaction product is treated in situ with a 20 ± 50-fold excess of a reducing reagent in an appropriate buffer, such as dithiothreitol or b-mercapto- ethanol.2, 18 The reduced product is purified by gel filtration, ion- exchange chromatography, or HPLC to remove the excess of the reducing reagents.Hg(II) ions are usually removed with hydrogen sulfide. Owing to the stabilising effect of three electron-donor methoxy groups, the TMTr protective group produces a more stable carbocation than the Trt group, which determines its specificity.The distinctive feature of this protective group is its extremely high acid lability. Deprotection is performed with dilute solutions of TFA.7 The tris(methoxyphenyl)methyl carbocation formed, like the trityl carbocation, can enter into the reverse reaction, viz., alkylation; therefore, the use of scavengers is obligatory in this case.Both Trt and TMTr protective groups may be split off together with simultaneous closure of the disulfide bonds with iodine or thallium trifluoroacetate (see Section III). The S-pyridyldiphenylmethyl group is not so popular in peptide synthesis. Only a few examples concerning this protective group are documented.19 The S-dibenzosuberyl protective group is also used rather seldom.In its chemical properties and deblock- ing conditions, it is similar to the Ddm group.6 The S-9-fluore- Ag++ R S+ R0 Ag Ag+ S2+ Ag Ag R R +S Ag Ag R++ R0 . + R S Ag R0 Ag+ +R S R0 R+ +Ag S R0, Table 1 (continued). Name a Formula of the protective Conditions where Conditions for removal Ref. group the protective group of the protective group is stable b 3-Nitro-2-pyridylthio (Npys) HF, 0 8C (in the absence RSH or other reducing 2, 38 ± 41 of thiols); TFA, 25 8C;e agents HOBT/DCC tert-Butoxycarbonylthio SCOOBut B and AH 36 (SCB) Benzyloxycarbonylthio (SZ) SCOOCH2Ph TFA; weak bases 40% HBr/AcOH; 36 strong bases Sulfonates SO3Na B Bu3P; b-mercaptoethanol, 42 dithiothreitol a The abbreviations of protective groups cited in Ref. 8 are given; b the following notations are adopted: B are bases; TFA is trifluoroacetic acid; Nu are nucleophiles; NpsCl is 2-nitrobenzenesulfenyl chloride; Pip is piperidine; TFMSA is trifluoromethanesulfonic acid; AH are acids; HOBT is N-hydroxybenzotriazole; DCC is dicyclohexylcarbodiimide; c partial removal; d scavengers are compounds that neutralise carbocations formed upon splitting of protective groups; e low stability. N O2N S 548 E V Kudryavtseva,MV Sidorova, R P Evstigneevanylmethyl protective group 5, 20 is one of the few groups that is removed under basic conditions, as a rule, this is performed by treatment with piperidine in DMF. The stability of this protective group under conditions of TFA treatment and in anhydrous HF makes possible its application in solid-phase syntheses, whereas its resistance to Tl(III) and I2 allows its selective removal in the presence of such protective groups as Acm, Trt, Meb, etc., and selective formation of disulfide bonds.Yet another example of a base-labile protective group is the 2-(2,4-dinitrophenyl)ethyl group.21 The electron-acceptor nitro groups in ortho- and para- positions make this group similar to Fm as regards its stability in the presence of strong acids and splitting by bases by a b-elimi- nation mechanism.The presence of nitro groups allows one to increase the solubility of Dnpe derivatives in polar solvents. Like Fm, this group is completely stable under conditions of classical and solid-phase syntheses, which employ Na-tert-butoxycarbonyl derivatives of amino acids (Boc-technique) 4, 45 as well as under conditions of splitting of a peptide from the polymer with HF or trifluoromethanesulfonic acid.Yet another alkali-labile protective group is S-b,b-di(ethox- ycarbonyl)ethyl group.1 It is easily split off by the b-elimination reaction upon treatment with 1 M KOH.1 The advantages of the S-adamantyl (Ad) protective group,22 although it is used rather seldom, are best manifested in compar- ison with Mob protection.The adamantyl protective group is more resistant to TFA: the degree of cleavage is only 10% over 2.5 h at 0 8C. It is quantitatively removed with a mixture of 1 M CF3SO3H and PhSMe in TFA (60 min, 0 8C) or Tl(CF3COO)3 in TFA.13, 22, 23 Xanthenyl and its 2(3)-methoxy derivatives are new protective groups for the thiol function of cysteine.24 These acid-labile groups are eliminated by dilute acids in the presence of appro- priate scavengers. An alternative procedure for their removal is the treatment with I2 or Tl(F3CCOO)3, which allows direct syn- thesis of cyclic disulfides.25 The use of a novel quinolin-2-ylmethyl protective group has been reported.25 A new procedure for the direct conversion of its derivatives into symmetrical disulfides includes treatment with FeCl3, FeCl2 , and CuCl2.25 Two new protective groups (Fc and Fcm) containing a ferrocenylmethyl fragment also deserve special mention.26 These groups are rather stable under basic conditions and in the presence of reducing reagents, such as NaBH4 and Zn in AcOH, although the Fc group is partly eliminated under these conditions.Both protective groups are rapidly split off by electrophilic reagents, e.g., thiocyanogen (SCN)2, methoxycarbonylsulfenyl chloride, and TFA. The Fc group can be also cleaved with formic acid. Iodine only partly destroys the Fc and Fcm protective groups. Removal of both these protective groups by acids is efficient only in the presence of scavengers (e.g., benzenethiol), which capture the ferrocenylcarbenium ion formed. Fcm protection, which was used in combination with Na-Boc or Na-Fmoc protections, is of greatest importance for the peptide synthesis.In this case, p-toluenesulfonic acid was used for selective removal of Boc protection.26 The (Z5-cyclohexadienyl)tricarbonyl iron complex {[(Z5-C6H7)Fe(CO)3]+} was successfully used to protect the thiol function of cysteine in the synthesis of a model peptide Ac-Cys- Gly-Ala-OEt.27 Owing to the high affinity of this complex for the SH group of cysteine, the formation of the corresponding adduct occurs rapidly and without side reactions.Protected derivatives are bright yellow coloured, which facilitates monitoring of their purification on silica gel.27 2.Protective groups that form mono- and di-thioacetals Protective groups that form mono- (S7C7O) and di-thioacetals (S7C7S) are not numerous and seldom used for cysteine protection. The protective groups of the thioacetal type, viz., tetrahydro- pyranyl, isobutoxymethyl, p-chlorophenoxymethyl, benzylthio- methyl, and phenylthiomethyl groups, are stable under alkaline conditions and can be removed by acids.The ease of their cleavage with acids decreases in the following order: acetals>monothioa- cetals dithioacetals, and is determined by the same factors as for sulfides, viz., the ease of protonation, the stability of the cation formed, and the possibility of irreversible reactions that shift the equilibrium towards the cleavage products.5 S-Monothioacetal derivatives of cysteine react with acids at the same rates as do trityl sulfides, whereas S-dithioacetal derivatives resemble Cys(Bzl) and Cys(But) under acidolysis conditions. According to the ease of their acidolysis, the protective groups are arranged in the follow- ing order:8 (p-MeOC6H4)2CH> %Me2CHCH2OCH2%Ph3C> >Ph2CH>PhCH2SCH2>PhCH2 .Cleavage of mono- and di-thioacetals is more efficient with Hg(II) and Ag(I) salts.5 The conditions for selective cleavage have been established for virtually each protective group.8 The mech- anism of cleavage of mono- and di-thioacetals by metal ions seems to be analogous to the mechanism of cleavage of sulfides.The formation of the M7S bond weakens the C7S bond. The formation of a cation is the rate-determining step of this reaction. The rate of cleavage depends of the cation stability.5 As with sulfides, regeneration of free thiols is achieved by treatment with a great excess of mercaptoalcohols or hydrogen sulfide.It is still uncertain whether selective cleavage of S-mono- thioacetals can be performed in the presence of S-dithioacetals, disulfide bonds, and other protective groups that are sensitive to heavy metal ions.5 Direct oxidation of monothioacetals into disulfides by halo- gens is inapplicable to cysteine derivatives.However, there is evidence of removal of the S-Thp group from the A-chain of synthetic insulin by its direct conversion into a symmetric disulfide using (SCN)2.5, 6 3. S-Acylamidomethyl protective groups The acetamidomethyl protective group the introduction of which can be accompanied by the formation of thiazolidine derivatives is widely employed in the synthesis of cysteine-containing peptides owing to its high stability in acidic and alkaline media and under conditions of catalytic hydrogenolysis. It is used in combination with Boc, Fmoc, and benzyloxycarbonyl protections of the a-amino group in classical and solid-phase syntheses of peptides.However, there are a few reports on the partial loss of the Acm protective group on repeated treatment of the peptide chains with HF needed for their splitting from the polymer.30 The Acm protection is stable under conditions of acid-catalysed removal of Meb, Mob, Trt, and Tmob groups. Specially selected condi- tions of deprotection of the thiol function of cysteine with heavy metal salts in appropriate solvents ensure selective removal of the S-Acm group in the presence of Bzl and But groups.Reaction with Hg(II) salts yields an intermediate peptide ± metal compound. When Hg(II) salts are used in acidic aqueous solutions, the reaction is completed within 1 ± 2 h.6 Deblocking can also take place in organic solvents, e.g., in DMF.2 O M+ + SCH2OR M SCH2OR slowly M+ + SCH2SR M SCH2SR slowly M+ROCHá2 , S M+RSCHá2 .S R is the peptide residue. Cys OH R Acm Cys OH R Hg+ Hg2+ pH 4 H2S Cys OH R Some peculiarities of synthesis of cysteine-containing peptides 549As in the case of removal of Trt protection, subsequent treatment with hydrogen sulfide or mercaptoalcohols yields free thiol groups.If Hg(II) salts are used for removal of Acm protection from the SH group of cysteine in tryptophan-contain- ing peptides, side reactions can occur.46 Thus treatment of a bis- Acm derivative of somatostatin with a large excess of Hg(AcO)2 (5 ± 20 equiv. per one Acm group) in 1%± 10% acetic acid results in substitution of hydrogen in positions 2 and 7 by the HgOAc group to form mono- and bis-mercuriacetates of tryptophan.Treatment of these derivatives with b-mercaptoethanol results in the substitution of the Hg(AcO)+ group by HOCH2CH2S to yield undesirable side products. Acetoxymercuration can be avoided by decreasing the amount of Hg(AcO)2 to 1 ± 2 equiv. and by using 50% acetic acid.46 There is evidence of the possibility of S?N and S?O migrations of the Acm protective group to the side carboxamide groups of asparagine and glutamine 2, 47 and to the hydroxy groups of serine and threonine.2, 48, 49 The migration of the Acm protective group to the side chain of glutamine (S?N migration) was observed in the deprotection of the thiol function of cysteine with I2 in MeOH. This side reaction is inhibited by an excess of glutamine as a scavenger.47 S?O Migration was observed in the deprotection of cysteine with Tl(III) and Hg(II) salts in a peptide containing several serine and threonine residues.48 This reaction is efficiently inhibited by three-carbon alcohols (e.g., glycerol) as scavengers.2, 48 The anion plays an important role in the removal of the Acm group with Ag(I) salts.The reaction occurs with silver nitrate,5 trifluoromethanesulfonate,18 trifluoroacetate,18 and tetrafluoro- borate.28 Silver acetate does not remove the Acm protection.Deprotection of cysteine with silver trifluoroacetate requires that the excess of the deblocking reagent to be twice as high as that with silver trifluoromethanesulfonate. The reaction time is also twice as long;18 therefore the use of AgOTf is preferable.Silver salts are removed by large (up to 50 equiv.) excess of b-mercaptoethanol or dithiothreitol.18 Then, the reaction mixture is purified by gel filtration. This deblocking procedure does not affect such highly sensitive amino acids as methionine, tryptophan, and tyrosine. Direct conversion of di-Acm substituted peptides into di- sulfides upon treatment with iodine is an attractive method for the removal of the Acm protective group.This method will be considered in more detail in Section III. Recently, trimethylacetamidomethyl, phenylacetamido- methyl, and benzamidomethyl protective groups, which can be regarded as derivatives of acetamidomethyl protection, have gained wide acceptance. The properties of benzamidomethyl protection are identical with those of the Acm group.6, 18 It is also sensitive to AgOTf.However, the removal of the Bam group takes more time than the removal of the Acm protection.18 Yoshida et al.28 ± 31 have proposed to use the trimethyl- acetamidomethyl group to protect the SH function of cysteine. It was found that under conditions of post-synthetic deblocking of the side functional groups of peptides with HF or NaOH, the Tacm and Acm groups, unlike the Bam group, were sufficiently stable.The authors 30 attribute the ease of introduction of the protective group and high yields of protected cysteine derivatives to salient advantages of Tacm protection. It should be mentioned that the formation of a side product (thiazolidine-4-carboxylic acid) cannot be avoided in the synthesis of S-acetamidomethyl- cysteine.30 The phenylacetamidomethyl protective group 32, 33 is stable and is split off under the same conditions as the Acm group.Its distinctive feature is the possibility to be cleaved by penicillin aminohydrolase.2 Therefore, the Phacm group is orthogonal with respect to Acm protection. First, it is possible to remove the Phacm group enzymically and then the Acm group is removed by treatment with heavy metal salts or iodine.2, 33 The 2-oxopyrrolidinomethyl group is resistant to reagents of peptide synthesis in a broad range of pH (from 0.5 to 13).In its resistance to alkalis, it surpasses the Bam group.34 Cleavage of the Pym protection with simultaneous formation of the disulfide bond is achieved by treatment with I2 in MeOH.34 4.Protective groups that form thioesters and related compounds Protective groups that form thioesters (RCOSR0) include acetyl, benzoyl, benzyloxycarbonyl, p-methoxybenzyloxycarbonyl, tert- butoxycarbonyl, and p-nitrophenoxycarbonyl groups. Dithiocar- bonates (RSCOSR0) are represented by derivatives with phenyl- thiocarbonyl and butylthiocarbonyl protective groups and N-(alkyl)thiocarbamates (RNHCOSR0), with N-ethylcarbamoyl, N-methoxymethylcarbamoyl, and N-[b-(N-acyl-N-methylami- no)ethyl]carbamoyl protective groups.A characteristic feature of thioesters (acetyl and benzoyl protections), thiocarbonates (benzyloxycarbonyl, p-methoxy- benzyloxycarbonyl, p-nitrophenoxycarbonyl protections), and thiocarbamates (N-ethylcarbamoyl and N-methoxymethyl- carbamoyl protections) is their sensitivity to nucleophiles and bases.Similarly to the corresponding esters and urethanes, thiocarbonates containing a benzyloxycarbonyl group and its analogues as well as Boc-protection are also cleaved in acidic media and with even greater ease.8 Dithiocarbonates (phenyl- thiocarbonyl and butylthiocarbonyl protections) and thiocarba- mate (N-[b-(N-acyl-N-methylamino)ethyl]carbamoyl protection) are cleaved only with acids.Alkaline treatment of thioesters can induce base-catalysed disulfide exchange, which can result in the formation of polymer mixtures and peptide parallel and antipar- allel dimers. The side reaction of b-elimination, which is accom- panied by elimination of thiobenzoic acid from the corresponding Cys(Bz) derivatives in peptides, can also take place.5 The occur- rence of this reaction depends on the solvent polarity; therefore one must avoid the addition of DMF to the reaction mixtures in the treatment of cysteine-containing peptides with bases, such as NaOMe.5 Cleavage of S-acyl groups can be achieved by treatment with hydrazine in methanol.Under these conditions, Ac protec- tion is split off rapidly as the corresponding acethydrazide without b-elimination.A peculiar side reaction was observed when the protected peptide Z-Cys(Bz)-Ser(Bz)-OMe was treated with NaOMe in MeOH.1 The serine residue was converted into an a-aminoacrylic acid residue, and the addition of the deblocked SH group to the double bond yielded products 1 and 2: N NHAc CONH2 HgOAc H AcOHg CONH2 NHAc N H AcOHg HgOAc CONH2 NHAc N H I2, MeOH Cys .. . Boc Acm Cys Acm OH Cys . . . Boc Cys OH CH COOMe CH2OBz NH CH CO CH2SBz NH Z NaOMe, MeOH 550 E V Kudryavtseva,MV Sidorova, R P EvstigneevaA serious limitation to the use of Ac, Bz, Z, pMZ, Boc, nitrophenylthiocarbonyl, butylthiocarbonyl, N-methoxymethyl- carbamoyl, Ec-, and [b-(N-acyl-N-methylamino)ethyl]carbamoyl acyl protective groups is their proneness to S?N acyl migration.Such migration takes place in the course of elongation of the peptide chain as soon as a free amino group appears.5, 8 This side reaction has been observed both in the classical and solid-phase peptide syntheses.6 When acetyl and benzoyl protective groups were used in solid-phase synthesis for protection of the thiol function of cysteine, a nearly quantitative chain growth termina- tion occurred due to the interaction of the deblocked SH group with the residual chloromethyl groups of the polymeric support after S?N migration.6 Nevertheless, Ac and Bz groups are used to protect cysteine residues in solid-phase syntheses of protected peptide fragments designed for fragment condensation.The S?N acyl migration can be suppressed if highly reactive deriv- atives (e.g., mixed anhydrides with pivalic acid) are used for the formation of the amide bond.5 In this case, the rate of aminolysis significantly exceeds the rate of S?N migration. The benzylox- ycarbonyl group also undergoes S?N migration, but this reac- tion proceeds at a much slower rate. The formation of side products can be avoided in this case if condensation is carried out with pivaloyl chloride or diphenyl phosphorochloridate.5 The p-methoxybenzyloxycarbonyl protective group is distin- guished by higher stability of the cation formed on its splitting and can therefore be removed not only by alkalis, but also by acids.However, its application is limited due to the difficulty of its introduction (the yields are about 30%) and high lability with respect to acids and alkalis.5 The N-ethylcarbamoyl and N-methoxymethylcarbamoyl pro- tective groups are resistant to acids and easily removed by basic reagents (as a rule, without racemisation and desulfurisation), they do not undergo S?N migration.Thus, they present sub- stantial practical interest.6, 8 The possibility of partial splitting of the Ec group with an excess of triethylamine during solid-phase synthesis with the use of the Boc-technique (at the stage of neutralisation of the peptidyl polymer or on repeated condensa- tion) restricts the application of this protective group.The use of S-[N-b-(N-alkoxycarbonyl-N-methylamino)ethyl]- carbamoyl protective group, yet another group of the ethyl- carbamoyl type, has been reported.6 In order to deblock com- pound 3 in which the thiol function of cysteine is protected by this group the alkoxycarbonyl (Z or Boc) group is removed by acid treatment.The resulting N-b-(methylamino)ethylcarbamoyl derivative 4 is cyclised to yield 1-methyl-2-imidazolidone 5 and the target thiol-containing peptide 6.6 5. Protective groups that form disulfides Disulfide derivatives (R ± S ± S ±R0) obtained after blocking of the SH group include symmetrical and asymmetrical cysteine di- sulfides, S-sulfonates, S-alkyl(aryl)thio, S-alkoxycarbonylthio derivatives, and sulfamides.Symmetrical disulfides are very seldom used in modern peptide chemistry for synthetic purposes. Ethylthio, isopropylthio, triphenylmethylthio, tert-butylthio, methoxycarbonylthio, (N-methyl-N-phenylcarbamoyl)thio, and 3-nitro-2-pyridinethio protective groups that form asymmetrical disulfides are good alternatives to the protective groups described above owing to their lability with respect to aliphatic and aromatic thiols.A specific feature of some of these groups (Scm, Snm, Npys) is their ability to form intra- and intermolecular disulfide bonds as soon as cysteine is deblocked.Here, the SH groups of peptide thiols play the role of a deblocking reagent (this problem will be considered in detail in Section III). S-Methoxycarbonylthio derivatives of cysteine are highly reactive compounds.6, 35, 36 Their reactivity with respect to thiols is greater than that of S-sulfonates and is comparable with that of sulfenyl thiocyanates or sulfenyl iodides, which are intermediates in the closure of disulfide bonds by the corresponding reagents (see Section III).The advantages of the Scm group are the stability of protected derivatives and the irreversibility of their reaction with thiols, unlike the reaction of sulfenyl iodides and sulfenyl thiocyanates.6 Owing to its stability in acidic media, the Scm group can be used in solid-phase syntheses using the Boc- technique.Cleavage of disulfides is performed in aqueous solutions (pH 7.5 ± 9.0) with thiols, such as b-mercaptoethanol or benzenethiol, which is a stronger nucleophile.8 The reaction with thiols is based on the thiol ± disulfide exchange. In order to shift the equilibrium in the direction of the target product, thiol must be employed in a large excess.Deblocking can also be achieved with a small excess of dithiothreitol, where the equilibrium is shifted due to the intramolecular cyclisation of the reagent, or with 1 equiv. of tributylphosphine at pH 8.8 Phosphines are specific reagents for cysteine and reduce only disulfide bonds; this reaction is selec- tive.17 Disulfides may also be cleaved by oxidative sulfitolysis.The resulting S-sulfonates are reduced with thiols or aqueous solutions of thio acids with gentle heating.8 It is known that disulfide derivatives are relatively resistant to acids. Their stability decreases in the following order: RSSBut>RSSPri*RSSEt. Thus the SBut group is cleaved by treatment with liquid HF for 1 h at 0 8C.2, 6 The SPri and SEt groups can be split off with hydrogen chloride in AcOH.6 Only the SBut group is of practical value and has been successfuly used in solid-phase synthesis using the Boc and Fmoc techniques.36, 37 The majority of mixed disulfides (Table 1) are resistant to ammonia and alkalis, which are used to split the peptides from the polymeric supports with phenolic and phenacyl anchor groups.8 Alkoxycarbonylthio (tert-butoxycarbonylthio and benzyl- oxycarbonylthio) protective groups were proposed by Nokihara and Berndt.36 Unfortunately, SCB derivatives proved to be not very stable on storage. Thus the cysteine derivative ZCys(SC- B)OMe turned into its cystine analogue within 3 months.In addition, the high lability of SCB protection with respect to alkalis restricts its application in peptide synthesis.36 The benzyloxycar- bonylthio protective group is stable in TFA.It is more resistant to alkalis that the methoxycarbonylthio (Scm) group and can be removed by treatment with strong alkalis, though this is preserved for 2 h at pH 10.6, which results in complete removal of the Scm protection. Therefore, the SZ group can be used both for the protection of the SH group of cysteine and for the selective formation of cystines.36 The N-methyl-N-phenylcarbamoylthio protective group is resistant to acids and can therefore be used in peptide syntheses with the Boc-technique. The possibility of removal of this protec- tive group by alkaline treatment has not been finally clarified.2 1 CHCONHCH CH2 NH Z S COOMe+S(CH2CHCOOH)2. CH2 2 NH2 3 CHCH2SCNH(CH2)2N CO NH Me Boc(Z) O H+ OH7 CHCH2SCNH(CH2)2N O CO NH Me H.HX 4 CHCH2SCNH(CH2)2N O CO NH Me H 5 HN NMe + O CHCH2SH NH OC 6 Some peculiarities of synthesis of cysteine-containing peptides 551Like Scm protection, the 3-nitro-2-pyridylthio and N-methyl- N-phenylcarbamoylthio protective groups may selectively gener- ate disulfide bonds in the presence of free SH groups.These groups are removed by thiolysis, which occurs in accordance with the Scheme:38, 39 It has been found39 that Npys protection is the most efficiently split off with aliphatic thiols, such as b-mercaptoethanol or mercaptoacetic acid: when the thiols are taken in a 10-fold excess over 3-nitro-2-pyridylthio derivatives, the reaction is completed within 10 min.The S7Npys bond is rather resistant to the action of aromatic thiols, 1-methylimidazole-2-thiol, and pyridine-2- thiol. In this case, selective cleavage of the O7Npys and N7Npys bonds may occur.39 Albericio et al.40 have shown that the Npys group is more stable under standard conditions of peptide synthesis using the Boc-technique (HF, TFA, and tertiary amines) in comparison with the methoxycarbonylthio group.The Npys protection for the thiol function of cysteine is absolutely stable in solvents and reagents used in solid-phase syntheses using the Boc-technique, e.g., CH2Cl2, DMF, N,N-dimethylacetamide, N-methylpyrrolidone, MeOH, trifluoroethanol, dioxane, in a mixture of 40% TFA and CH2Cl2, this also withstands treatment with HF (to which 10% anisole or p-cresol is added) for 1 h at 0 8C.Thus, the Npys group is virtually resistant under conditions of HF cleavage by the high technique.40 The low-high technique is inapplicable in this case, since the use of p-thiocresol and dimethyl sulfide results in the loss of lateral Npys groups.40 The data on the resistance of Npys protection to N-hydroxybenzotriazole are contradictory.40, 41 Partial (12%) cleavage of the Npys group was observed in the treatment with 0.1 M solution of this reagent in DMF.41 Therefore, the method of activated p-nitrophenyl and pentafluorophenyl esters was employed to form the amide bond (condensation).4, 40, 50 Matsueda et al.41 have successfully used N,N 0-dicyclohexylcarbodiimide with N-hydroxybenzotriazole to attach Boc-Cys(Npys)-OH in the synthesis of a series of peptides. The protective Npys group was found to be very labile under conditions of solid-phase synthesis using the Fmoc-technique, but was sufficiently stable in the photolytic cleavage of peptides from the polymer.40 Of interest is the report on the use of base-resistant S-sulfo- nates [Boc-Cys(SO3Na)-ONa] and [Fmoc-Cys(SO3Na)-ONa] in peptide synthesis.42 Deblocking of cysteine may be achieved by treatment with tributylphosphine, b-mercaptoethanol, and dithiothreitol.42 Splitting of the peptide from the polymer with TFA± p-cresol systems leads to partial loss of sulfonates.III. Methods for the formation of disulfide bonds Despite the intensive progress in methodology of peptide chem- istry, the formation of disulfide bonds still presents a serious problem.Its solution involves: (1) the choice of a protective group for the thiol function of cysteine and (2) the selection of a method for the formation of a disulfide bridge. These two aspects are closely interrelated, for the choice of a protective group is the principal factor that predetermines the selection of the method for the formation of the S7S bonds.This problem becomes more complicated in the case where several disulfide bonds are to be formed. The so-called `orthogonal' scheme of synthesis allows selective closure of disulfide bonds between definite cysteine residues without involvement of other protected residues or the preformed disulfide bridges. In this case, the choice of protective groups that can be selectively split off and the cyclisation procedure are of crucial importance.8 In any case, the cyclisation process is very complicated and its successful outcome depends on many factors, such as the correct choice of the solvent, temper- ature, pH of the reaction medium, concentration of reagents, type of the oxidant, and so on.The methods for cyclisation can be arbitrarily classified into three groups:2 A.Deprotection and formation of a disulfide bond starting from the corresponding SH-derivatives of peptides. B. Deprotection with simultaneous closure of the ring. C. Closure of the disulfide bond by thiolysis. This entails two possibilities: (1) The protective group of one of cysteine residues (X*) enhances the electrophilicity of the sulfur atom (activation). This facilitates the cleavage of the bond between the sulfur atom and the protective group in the thiolysis.The free SH group of the other cysteine residue is the necessary thiol. In this case, X*=X=7SCOOMe, 7SCON(Me)Ph, 3-nitro-2-pyridylthio, etc. (2) The protective groups X=CPh3, CH2 NHCOMe, But, etc., are split off with (SCN)2, 3-nitro-2-pyridinesulfenyl chloride, 4-nitrobenzenesulfonyl chloride, and other reagents to form activated asymmetrical disulfides (X*=SCN, 3-nitro-2-pyri- dylthio, 4-nitrophenylsulfonyl) that can enter into the reactions of thiol ± disulfide exchange: 1.Formation of disulfide bonds in SH-derivatives of peptides Group A methods envisage the blocking of two cysteine residues by identical protective groups that are removed under identical conditions, purification of linear polythiols, and oxidation of SH groups with any appropriate reagent providing the formation of intra- or intermolecular disulfide bonds.Oxidation of free thiols is usually carried out in highly dilute solutions at the peptide concentration 1074 ± 1075 M in order to avoid intermolecular coupling and side reactions.2, 8 Direct formation of the S7S bonds in dilute solutions strongly depends on the peptide struc- ture, particularly on the nature of the amino acids separating two cysteine residues and the number of amino acids between them.51, 52 If the number of amino acids between two cysteine residues (n) is from 4 to 20, the intramolecular S7S bonds are predominantly formed under conditions of high dilution (1073± 1074 M).At n<3, the limited chain flexibility results in dimers and polymers that are formed in addition to the peptides with the intramolecular S7S bonds.51 The formation of the disulfide bond between the neighbouring cysteine residues demands very strong oxidants, such as thallium(III) trifluoroace- tate.53 The ability to form intrachain disulfide bonds depends also on the structural stability of the disulfide loop.Thus the oxidation of peptides is facilitated by the presence in the amino acid sequence of Pro-Gly residues involved in the formation of b-folds of proteins and peptides. Many peptides (e.g., oxytocin) can be readily oxidised by atmospheric oxygen. Peptide fragments of proteins having the Pro-Pro sequence between two cysteine residues are difficult to oxidise and can form only symmetrical dimers.51 N NO2 S S CH2 RSH HNBoc CH COOH N NO2 S S R +HSCH2CH(NHBoc)COOH. SX SX SH SH deblocking [O] S S SX SX S S I2, Tl(F3CCOO)3 etc.SX SX and/or activation selective deblocking 7Y*, X* S S 7 Y*S SX* d7 d+ 552 E V Kudryavtseva,MV Sidorova, R P Evstigneevaa. Oxidation with atmospheric oxygen Atmospheric oxygen is often used to oxidise thiol groups of cysteine to yield disulfides.This reaction is usually carried out in weakly alkaline solutions at pH 7.5 ± 8.5 with vigorous stirring or by bubbling air through the reaction mixture. The required level of pH is usually maintained by adding ammonia,54, 55 or other reagents, such as ethyldiisopropylamine.55, 56 In some cases, cyclisation occurs much faster when ethyldiisopropylamine is used instead of ammonia.55 The oxidations may also be carried out in 0.1 M NH4HCO3 (pH 8.0) 57 or 0.1 M ammonium acetate buffer (pH 6.8 ± 7.0).51 Passage of pure oxygen (instead of air) through a solution of a peptide in 0.1 M NH4HCO3 did not increase the rate of cyclisation.57 In the case of peptides that aggregate in aqueous solutions, 0.5 ± 8.0 M urea or 0.1 ± 1.5 M guanidine hydrochloride are added.2, 58, 59 As a rule, the oxidation is performed at 5 ± 25 8C.The reaction time varies from several hours to several days.54 ± 57 Most probably, this reaction occurs by a free-radical mechanism. The cyclisation is usually monitored by HPLC or by a reaction for the presence of a free SH group using Ellman's reagent, 5,50-dithiobis(2-nitrobenzoic acid).60 The yields of cyclic disulfides in this oxidation are usually small (9% ± 15%).54 ± 56 There are some examples of preparation of synthetic toxins, growth factors, and enzyme inhibitors by this method.That this process is lengthy and the product yields are low are serious limitations of this method. However, its simplicity makes it an attractive procedure despite the advent of new efficient reagents.b. Oxidation with potassium ferricyanide Potassium ferricyanide is a popular cyclisation reagent. Laloo and Mahanti61 have studied the kinetics of cysteine oxidation with potassium ferricyanide and proposed a mechanism for this reaction. Using ESR spectroscopy they have found that this reaction occurs via an intermediate radical, [7SCH2C .(NH2)COO7].61 Depending on the solubility of a peptide, the oxidation is carried out in aqueous or aqueous- ethanolic solutions with high dilution (pH 8 ± 10).60, 62 Haskell- Lecevano et al.63 performed this oxidation in a saturated solution of NH4OAc containing methanol and acetonitrile. The termina- tion of the reaction is judged from the appearance of a persistent yellow colouration. Inorganic impurities are removed by ion- exchange chromatography.This method mainly yields homoge- nous reaction mixtures. Oxidation of sulfhydryl groups occurs virtually quantitatively.55 This procedure has been used to obtain oxytocin,2, 64 vasopressin analogues,62 somatostatin,2, 65 and pep- tides with two S7S bridges (endothelin-1,58 a bee venom peptide of mast cell degranulation, 66 etc.).c. Oxidation with dimethyl sulfoxide Dimethyl sulfoxide (DMSO) was used to oxidise peptide thiols for the first time in 1988.67 This compound has a number of indisputable advantages: it can be used in a wide range of pH (from 3 to 8), ensures higher reaction rates (in comparison with oxidation with atmospheric oxygen), and increases solubility of peptides.2, 59, 67 ± 71 This reaction is catalysed by acids or primary and secondary amines.The mechanism of oxidation in an acidic medium was proposed by Otaka et al.68 The limiting step of this reaction is the formation of the intermediate compound 9, which rapidly yields the disulfide 10 on reaction with the free thiol 8. The optimum pH range for the reaction with DMSO is 3 ± 8.69 In this case, the reaction rate does not depend on pH, because in this pH range both the protonation of DMSO and the formation of a thiolate anion may take place.At pH<3, methionine can undergo oxidation into the corre- sponding sulfoxide, whereas at pH>8 the rate of disulfide exchange increases. However, there are some exceptions. For example, the oxidation of peptides with an N-terminal cysteine should be conducted at pH 6 ± 7.If the amino groups of cysteine are acetylated, the pH is unimportant. Tam et al.69 interpret this phenomenon by the fact that the weakly positive a-amino group of cysteine facilitates the formation of the unstable intermediate adduct 9. In more acidic media, the a-amino group is protonated and does not exert such an effect.Oxidation with dimethyl sulfoxide is carried out at its concen- tration in the reaction mixture from 10% to 50%. An increase in DMSO concentration increases the reaction rate. Thus complete oxidation of a linear precursor of a human growth factor fragment with 10%± 30% DMSO requires 2 ± 3 h, whereas with 40%± 50% DMSO it takes only 0.5 h. The oxidation-sensitive amino acid residues of tryptophan, tyrosine, and methionine are not modified under these conditions.69, 71 A 10% DMSO solution in TFA has been used for simulta- neous splitting of acid-labile cysteine protective groups (Trt, Mbzl, Dpm) and the closure of disulfide bonds.68, 70 However, in this case the oxidation lasted several hours.Over this time, the unprotected tryptophan was strongly modified and methionine was oxidised into the corresponding sulfoxide.Oxytocin, which has no tryptophan or methionine in its amino acid sequence, was obtained in 73.5% yield from a bis(methoxybenzyl)cysteine deriv- ative by treatment with 10% DMSO in TFA in the presence of anisole (12 h, 25 8C). Human calcitonin (a-hCGRP) was obtained from a bisadamantyl derivative of cysteine under identical con- ditions (1 h , 17%).68 The resistance of the oxytocin derivatives [1,6-Cys(Acm)] and [1-Cys(Acm), 6-Cys(SH)] with respect to 10% DMSO in TFA has been studied.70 The former proved to be stable for 12 h, whereas the monosubstituted derivative [1-Cys(Acm), 6-Cys(SH)] was converted into oxytocin under identical conditions within 1 h (69%); the product was contaminated with a dimer.Tamamura et al.59 reported on the closure of the S7S bonds upon action of 10% or 50% DMSO in 1 M HCl. The thiol groups of cysteine protected with Acm of Mbzl groups had been deblocked beforehand with silver trifluoromethanesulfonate in TFA. Two chemical reactions occurred simultaneously in the DMSO± HCl system: conversion of Cys(Ag) derivatives into cysteine and oxidation of cysteine residues with DMSO.This approach was used to obtain oxytocin, urotensin II, tachiplen- sin I, and endothelin I in 72%, 80%, 79%, and 62% yields, respectively (as calculated per bis-Acm derivatives). No side reactions that could affect methionine, tyrosine, or tryptophan residues were observed under these conditions.59 Oxidation by DMSO is especially efficient in cyclisation of basic peptides that are insoluble in weakly alkaline solutions, which are required for oxidation with oxygen or K3[Fe(CN)6].This problem appeared in the synthesis of basic hydrophobic growth factors that remained insoluble in weakly alkaline solu- tions even after addition of 8 M urea.69, 71 The advantages of this method have been also exemplified in the synthesis of a natriuretic peptide (pBNP) and tachiplensin I having two S7S bonds.67 CH COOH CH2SR H2N 8 + CH COOH CH2SH H3N + Me2SOH TFA 9 S S CH2 OH 8 7Me2S Me Me CH + COOH NH3 R=Trt, Mbzl, Dpm.+ 10 +H2O CH CH2SSCH2 HOOC H3N CH + COOH NH3 Some peculiarities of synthesis of cysteine-containing peptides 553However, despite the obvious merits of this method, it has certain disadvantages, which are connected with the difficulty of removal of the excess of DMSO and dimethyl sulfide formed in this reaction.d. Oxidation with hydrogen peroxide Hydrogen peroxide is a common component of biological systems and participates in physiological oxidation processes.72 Therefore, the use of this oxidant in peptide synthesis is fully justified.Oxidation of thiols with H2O2 may occur in two ways: via generation of a thiolate anion (RS7) or a radical (RS . ):73 The oxidation performed under alkaline conditions acceler- ates the generation of the thiolate anion; therefore in this case the formation of the disulfide occurs at higher rate.74 Using oxidation of N-acetylcysteine and glutathione as exam- ples, it was shown that this reaction proceeds via several thiol and disulfide complexes with H2O2.75, 76 The important role of H2O present in the reaction mixture was emphasised, since it stabilises these complexes by creating a solvate shell around them.75, 76 The dependence of the rates of these reactions (in a phosphate buffer, pH 7.0) on the concentration ratio, K=[RSH]/[H2O2], has been studied. The formation of the complex 11 takes place at K<2.5, while that of the free disulfide 12 at K>2.5, i.e., where the concentration of the residual thiol is<1072 mol litre71.A systematic study of the applicability of hydrogen peroxide for the closure of disulfide bonds has been carried out by Kudryavtseva et al.77, 78 Table 2 lists the yields obtained in the formation of disulfide bonds in peptides of various lengths and structures under the action of both iodine and hydrogen per- oxide.78 Syntheses of some biologically active peptides, such as antigenic determinants of HIV proteins (compounds 13 ± 16, Table 2), atrial (vasodilator) (compounds 18 ± 20, Table 2) and vasoconstrictor peptides, containing oxidation-sensitive trypto- phan, tyrosine, and methionine residues demonstrate that the oxidation with H2O2 enables the preparation of the target disulfides within only 15 ± 20 min and in high yields (60% ± 85%) without formation of side products.78 The oxidation was carried out in aqueous media at pH 6.5 ± 8.0.It was also shown that H2O2 can be used for the synthesis of preparative (gram-scale) amounts of antigenic determinants of HIV proteins (13 ± 16) and oxytocin (17).78 There are other reports concerning the use of H2O2 in the synthesis of oxytocin and vasopressin.3, 79 Hydrogen peroxide was used for cyclisation of methionine-containing peptides (pH 8 ± 11), human natriuretic peptide (a-hNAP), and somatos- tatin in phosphate buffer (pH 8.5, 10 min).The reaction products were obtained in more than 90% yields and with high purity.Apparently, the use of hydrogen peroxide in cyclisation has limitations of its own. It is known, for example, that H2O2 may induce more profound oxidation of disulfides to sulfones.72 e. Oxidation with glutathione Yet another way to close S7S bonds is the use of buffers that simulate physiological conditions. Since in vivo disulfide bridges are formed under reductive conditions, and the reducing medium in cells is maintained by the glutathione system (5 mM reduced and 0.1 mM oxidised glutathione),81 buffer solutions containing reduced and oxidised glutathione at a 10 : 1 ratio are most often used for the formation of S7S bonds.2 Other ratios of reduced and oxidised forms, e.g., 1 : 1 (Ref. 82) or 2 : 1 (Ref. 83) are also employed.The reaction of peptide polythiols is conducted in dilute solutions at pH 7.3 ± 8.7 and 0.1 ± 0.5 M Tris ± HCl buffer is used for dissolution.83, 84 The oxidation of thiols with glutathione occurs in two steps.85 At first, a `reduced' peptide having free SH groups reacts with the oxidised glutathione (GSSG) to give mixed disulfides, which undergo intramolecular thiol ± disulfide exchange to form disulfide bonds in peptides and the reduced glutathione.85 The limiting step of this reaction is the reaction of the peptide with GSSG; therefore, the reaction rate depends on the concen- tration of the latter.The oxidation is normally carried out for 16 ± 72 h at 25 ± 35 8C.82, 83 The reaction product is purified by affinity chromatography2 or by HPLC.83, 84 This method was used to obtain a 48-membered neurotoxin of sea anemone (yield 22.8%),84 the trypsin inhibitor CMTI III (yield 10%),82 human epidermal growth factor (h-EGF) peptides,83 and o-conotoxin.86 All these peptides contain three disulfide bonds; their closure was performed simultaneously.This method yields predominantly natural, thermodynamically favourable structures.82 ± 86 How- ever, the procedure that employs a mixture of oxidised and reduced glutathiones is mainly limited by large molecules with a thermodynamically favourable physiologically active confor- mation and is hardly suitable for the synthesis of short peptides containing disulfide bonds. RSH RS7 O2 RS RS7 RS [RSSR] O2 RSSR 7O2 7 7O2 7 7 O¡2 RSH+H2O2 [RSH...H2O2], RSH+2H2O2 [RSH...H2O2...H2O2], [RSH...H2O2...RSH] RSSR+2H2O. 2[RSH...H2O2] [RSSR...H2O2] + 2H2O, 11 [RSSR...H2O2]+2RSH 2RSSR+2H2O, 12 R Cys. . .Cys GSSG R0 R Cys. . .Cys SG R0 + R Cys. . .Cys SG R0 R Cys. . .Cys R0 +GSH Table 2. Comparative yields of cyclic peptides during the closure of the S7S bridge with iodine (in ACOH or MeOH) and hydrogen peroxide. No. Structure Yields (%) of com- pound I2 H2O2 13 51 75 ± 85a 14 56 85 ± 90a 15 66 80 ± 87a 16 25 85 ± 90a 17 50 70 ± 75b 18 23 83 19 26 50 20 20 62 a The scatter in yields on oxidation of 0.1 ± 1.5 g of the corresponding bis- Acm derivatives of peptides; b the scatter in yields on oxidation of 2.0 ± 20.0 g of bis-Acm derivatives of oxytocin. H-LGLWGCSGKLIC-NHEt H-LGIWGCSGKHIC-NHEt H-NSWGCAFRQVC-NHEt H-WGCSGKLICTTAVPWNA-OH H-CYIQNCPLG-NH2 H-SSCFGGRIDRIGAQSGLGCNSFR-OH H-SSCFGGRIDRIGAQSGLGCNSFRY-OH H-SLRRSSCFGGRIDRIGAQSGLGCNSFRY-OH 554 E V Kudryavtseva,MV Sidorova, R P Evstigneevaf. The use of polymers Clarc and Pai have proposed an original procedure for cyclisation of linear SH-containing peptides using an EkathioxTM polymer.87 The resin (40 ± 80 m, pore diameter 100 A) contains 0.2 ± 0.4 mmoles of oxidising functional groups per gram.The oxidising polymer was removed by filtration after 4 h at 20 8C. The efficiency of the polymer was evaluated by the formation of oxytocin (yield 92%). The authors 87 note that the high yield and the absence of dimeric side products were due to pseudo-dilution (see Section III.4), where the reaction was performed on a solid phase.87 The polymer EkathioxTM does not modify tryptophan, tyrosine, or methionine residues. Moreover, it allows one to conduct oxidation in both aqueous and organic media. It is difficult to evaluate the advantages of this method, because the nature of the functional groups was not specified by the authors.An alternative approach to cyclisation using polymeric reagents was proposed by Lin Chen and Barany.88 N-Substituted 2,3-dithiasuccinimides 21 were used as oxidants to produce intramolecular disulfide bridges in the synthesis of oxytocin and deaminooxytocin performed under different con- ditions.One procedure consisted in the reaction of soluble peptides with a dithiasuccinimidoacetyl polymer, and the other, the reaction of a peptidyl polymer with a soluble derivative of 2,3-dithia- succinimide.The rate of this reaction depends on the pH of the reaction medium. The reaction kinetics has been studied in aqueous, organic, and aqueous-organic media. The disadvantage of this method is the high reactivity of dithiasuccinimide deriva- tives, which irreversibly oxidise functional side groups of amino acids.88 g. Miscellaneous reagents Other reagents, such as 5,50-dithiobis(2-nitrobenzoic acid) (Ellman's reagent 89) and a CCl4 ± NEt3 mixture,2, 7 have been used to oxidise free thiol groups of cysteine residues.Although these compounds did not gain wide acceptance for oxidation in solution, they have been used for solid-phase cyclisation and therefore will be considered below. 2. Removal of protective groups with simultaneous closure of an S ± S bridge Direct conversion of S-protected peptides into the corresponding disulfides is one of the most attractive methods of disulfide bond formation.A number of reagents are now available that can be used for such conversions. a. Iodine Iodine is the most popular reagent used in the formation of the disulfide bonds. The reactions of cysteine with Trt, Acm, Tmob, Mob, and other protective groups with iodine have been studied in great detail.Two mechanisms have been proposed to explain the conversion of a sulfide into a disulfide.5 One of them implies the formation of an intermediate sulfenyl iodide 22, and the other, of alkyl(aryl)thiosulfenyl iodide 23. Neither of these mechanisms has been favoured, although both routes have been confirmed by several authors.5 Since the intermediate sulfenyl iodides are highly reactive, the efficiency of oxidation strongly depends on the amino acid sequence of a peptide.Several procedures for oxidation of protected peptides with iodine exist, which differ in the ratios of the reaction components, solvents, reagents used to remove the excess of iodine, etc. The oxidation is carried out in both neutral solvents (MeOH or aqueous MeOH, trifluoroethanol, dioxane, CHCl3, DMF) and in acids (acetic, propionic, trifluoroacetic) or their mixtures with water.Depending on the type of the thiol function protection and the nature of the solvent, the reaction rates may vary considerably. Thus the rates of oxidation of Cys(Trt) in MeOH, acetic acid, and their mixtures with H2O are from 4 to 500 times as high as that of oxidation of Cys(Acm) in the same solvents.If oxidation is carried out in chloroform, dichloromethane, and trifluoroethanol, this difference is much higher. This allows selective oxidation of Cys(Trt) in the presence of Cys(Acm). On the contrary, the rate of oxidation of Cys(Acm) in DMF and its mixtures with water is 5 ± 10-fold higher than that of Cys(Trt).2 The excess of iodine plays an important role in this process.With equimolar amounts of the oxidant with respect to the [Cys(Acm)]2 derivative of gramicidin S, the reaction in methanol was completed within 3 h and the yield of gramicidin S was 67%.90 The use of 5 ± 15-fold excess of I2 in AcOH made it possible to obtain cyclic disulfides of various lengths and struc- tures and in sufficiently good yields.66, 91, 92 The use of 40- and 50- fold excess of I2 in AcOH was reported to result in low yields (10%) of the target products.93, 94 Using the synthesis of the cyclic disulfude 15 from the bis-S- Acm-protected fragment 593 ± 603 of HIV-2 gp41 (24) as an example (Scheme 1), the effect of the excess of iodine and of the solvent on cyclisation has been studied.55, 78, 95 With 2 ± 13-molar excess of I2 in AcOH, the yields of the target product 15 did not exceed 39%.55, 78, 95 The highest yield (66%) was obtained when + SH SH S S S S N R O O 21 +2COS:+H2NR R=CH3, CH2CO2H, CH2CO27P , P is the insoluble polymer.R=Acm, Trt, Tmob, Mob. S S MeOH 7I7, ROMe I2, MeOH 7ROMe 2I2, MeOH 72ROMe SR SR 22 S S I I 72I7 S S 23 S S I R + 15 C-Gly-Cys-Ala-Phe-Arg-Gln-Val-NH-CH-CONHEt+ H-Asn-Ser-HN S2O3H 26 H-Asn-Ser-Trp-Gly-Cys-Ala-Phe-Arg-Gln-Val-Cys-NHEt O CH2 CH SCH2 NH 25 24 H-Gly-Cys-Ala-Phe-Arg-Gln-Val-Cys-NHEt + H-Asn-Ser-Trp-Gly-Cys-Ala-Phe-Arg-Gln-Val-Cys-NHEt Acm Acm a H-Asn-Ser-Trp-Gly-Cys-Ala-Phe-Arg-Gln-Val-Cys-NHEt+ 15 24 27 H-Asn-Ser-Trp-Gly-Cys-Ala-Phe-Arg-Gln-Val-Cys-NHEt S2O3H(Acm) Acm(S2O3H) b Scheme 1 (a) 13 moles of I2 per 1 mole of the peptide in AcOH; (b) 5 moles of I2 per 1 mole of the peptide in AcOH.Some peculiarities of synthesis of cysteine-containing peptides 555oxidation was performed in aqueous methanol using 2 moles of I2 per 1 mole of the peptide.78, 95, 96 The excess of iodine is removed with sodium thiosulfate, ascorbic acid, zinc dust, and by extraction with carbon tetra- chloride.2, 55, 78, 91 ± 95 Removal of the excess of iodine with sodium thiosulfate was accompanied by modification of the cysteine residue with the thiosulfate ion 55, 78, 95, 96 (compounds 26 and 27, the exact positions of the Acm and S2O3Hgroups in compound 27 were not determined) (Scheme 1).The limitation of the use of iodine for the disulfide formation is the possibility of side reactions affecting oxidation-sensitive amino acid residues, e.g., iodination of the benzene ring of tyrosine,8 oxidation of methionine to the corresponding sulfox- ide,8 and scission of the indole ring of tryptophan with subsequent cleavage of the peptide chain.8, 55, 78, 95 ± 97 Thus the formation of a shortened peptide 25 (Scheme 1) was a result of oxidative degra- dation of tryptophan by iodine.78, 95 Since these reactions are catalysed by acids, the cleavage of the peptide chain at the tryptophan residue can sometimes be avoided if oxidation is conducted in neutral organic solvents.55, 95, 96, 98 Oxidation of tryptophan-containing peptides with iodine sometimes results in tryptophan-2-yl sulfides as a side product of the type 26.Apparently, one of the reasons for this side reaction is the formation of highly reactive intermediate sulfenyl iodides. Tryptophan-2-yl sulfides were obtained under various condi- tions.42, 95, 98 For example, in the synthesis of somatostatin from a precursor with different protective groups at cysteine residues (Acm in position 3 and Trt in position 14), treatment with iodine in 10% aqueous trifluoroethanol resulted in the formation of a side product in 65% yield.98 According to Sieber et al.,98 this reaction occurs due to the different lability of Acm and Trt groups in trifluoroethanol.2, 98 When oxidation was carried out in a MeOH/DMF mixture, tryptophan-2-yl sulfide was not formed.The distance between the cysteine and tryptophan residues plays a significant role in the formation of tryptophan-2-yl sulfides.Oxidation of model tryp- tophan-containing peptides with iodine shows that only disulfide dimers are formed if tryptophan and cysteine are separated by one or two glycine residues. If this distance is four glycine residues, the yields of tryptophan-2-yl sulfides are 40% and 70%, respec- tively.98 Alternatively, tryptophan-2-yl sulfides ware formed when sodium thiosulfate was used to remove iodine in the synthesis of a fragment of the HIV-2 protein gp41 (if Na2S2O3 is replaced by ascorbic acid, this side reaction does not take place).Blocking of the thiol function of one of cysteine residues with the thiosulfate ion (compound 26) triggers the reaction of the thiol function of the other cysteine residue on position 2 of the indole ring of tryp- tophan.55, 78, 95 No side reactions involving the indole ring of tryptophan took place during cyclisation with iodine when its Nim-formyl deriva- tive was used.58 Despite the obvious limitations of this method, there are many examples of its successful application in the synthesis of com- pounds with several disulfide bonds in particular..54, 58, 66 b.Cyanogen iodide and N-iodosuccinimide These compounds also serve as a source of the I2+ ion. The mechanism of oxidation with these reagents is identical with that of iodine. Good solubility of ICN in aqueous-organic solvent systems has been reported.99 The utility of ICN was demonstrated in the cyclisation of several Cys(Acm) peptide derivatives in aqueous methanol.A comparison of oxidation with cyanogen iodide and iodine showed that although large excess of the oxidant (from 25- to 80-fold) is required in the former case and the duration of these reactions was 24 ± 48 h, the yields were about twice as high as with I2. N-Iodosuccinimide was used in the synthesis of oxytocin and vasopressin on a polymer,100 therefore this reagent will be considered below in more detail.c. Thallium(III) trifluoroacetate A new, rather efficient oxidant, Tl(CF3COO)3, was first used in 1987 for removal of certain cysteine protective groups with simultaneous formation of the disulfide bond.9, 101 Thallium trifluoroacetate removes many protective groups of the sulfide type (Mob, Tmob, But, Trt, Xan, Ad, Meb) and acylaminomethyl derivatives (Acm, Phacm).The oxidation occurs through the formation of a highly reactive intermediate compound 28, viz., thallium(III) thiolate. At the final step of this reaction, the protective group is split off from the thiol group of the second cysteine residue with concomitant disulfide bond closure. The oxidation is carried out in trifluoroacetic acid (1 ± 2 h, 0 8C) in the presence of 2 equiv.of anisole to bind the carboca- tions formed 9, 101 or in DMF.2, 7 The use of aqueous solvents is excluded in this case, since the presence of water can bring about the formation of small quantities of cysteic acid.9 Thallium salts are readily soluble in ether and their removal does not present any problem.9 The use of Tl(CF3COO)3 gave satisfactory results in the synthesis of oxytocin, urotensin II, and human calcitonin.The yields of the products purified by gel filtration were 45%, 35%, and 11%, respectively.101 In the latter two cases, the decrease in the product yields can be explained by the presence of a trypto- phan residue in the amino acid sequences of these peptides, which undergoes considerable modification. It was shown that only 38% of tryptophan remains intact under these conditions.The use of a Trp(Mts) derivative allows one to preserve tryptophan for 60 min on treatment with Tl(CF3COO)3 at 0 8C. Methionine is partly oxidised (by *38%) to the corresponding sulfoxide with Tl(III) salts.9 The use of this reagent is efficient in the case of difficultty oxidisable thiols, where cysteine residues are adjacent as in the peptide fragment of the acetyl choline nicotine receptor (alpha 185 ± 196).53 A comparison of efficiencies of atmospheric oxygen, potas- sium ferricyanide, and thallium(III) trifluoroacetate as oxidants in some syntheses revealed that Tl(CF3COO)3 gave the highest yields.17, 53 It has been widely employed for the closure of disulfide bonds on a polymer.7, 17, 102 The high toxicity of thallium salts imposes certain restrictions on their application, which will hardly go beyond the scope of laboratory experiments.R=Boc-Ala-Gly-Cys-Lys-Asn-Phe-Phe Boc Acm RNHCHCO-Lys-Thr-Phe-Thr-Ser-NHCHCOOBut BocBut ButBut CH2 NH S CH2 S S R R Tl(F3 Ê COO)3 7R+,7CF3COO7 28 S S R Tl OOCCF3 OOCCF3 S S Tl OOCCF3 7TlF3CCOO R+, CF3COO7 S S 556 E V Kudryavtseva,MV Sidorova, R P Evstigneevad. Sulfoxides in the presence of silyl chlorides This cyclisation method was proposed in 1991 independently by Kiso 103, 104 and Fujii.105 In the presence of silyl chlorides in trifluoroacetic acid, aryl(alkyl) sulfoxides may split off Acm, Tacm, Bam, But, Meb, and other cysteine protective groups with simultaneous closure of the S7S bridges in 10 ± 30 min at 4 8C.The mechanism of this reaction can be presented in the following way:103, 104 The closure of the disulfide bond occurs through the gener- ation of the sulfonium cation 31 upon treatment of the protected peptide 30 with diphenyl sulfoxide and silyl chloride. Further, electrophilic attack of the second substituted sulfur atom in compound 31 results in the formation of an S ± S bridge and the splitting of the protective group, which is similar to the reaction of sulfenyl iodide with S-protected cysteine in the oxidation with iodine.The nature of silyl chloride seems to play a key role in the reactivity of the intermediate compound 29. The ability of the protective group split off to form stable structures similar to the immonium ion derived from the Acm group is also important.These two factors determine the rate of the reaction.104 The efficiency of various silyl chlorides (SiCl4, MeSiCl3, Me3SiCl), and Me4Si in combination with sulfoxides [PhS(O)Ph, MeS(O)Me, and MeS(O)Ph] has been studied.104 SiCl4 and MeSiCl3 were found to be the most efficient silylating reagents. There was no difference in reactivities of different sulfoxides.104 The Ph2SO± MeSiCl3 system or SiCl4 removes Acm, Mob, Meb, Bzl, and But protective groups with equal efficiency.105 Tryptophan is subject to side reactions under these conditions.Chlorination of the indole ring is the predominant reaction in the treatment of tryptophan-containing peptides with silyl chlo- ride.104, 106 The addition of scavengers (anisole and 3-methyl- indole) routinely used to prevent tryptophan modification was inefficient.104 It is therefore recommended to use tryptophan in the form of a formyl derivative.103, 104, 106 Nevertheless, successful application of this reagent has been demonstrated in the synthesis of somatostatin (yield 60%),104, 106 oxytocin (yield 56%± 69%),104 and human natriuretic peptide (hBNP) (yield 50%).104 This method was also used for regioselective formation of disulfide bonds in the synthesis of conotoxin M1 and human insulin.107 3.Closure of disulfide bonds by thiolysis or thiol ± disulfide exchange Implementation of the route C requires that the two protective groups of two cysteine residues be different in stability. This enables selective removal of one of the protective groups (Y), whereas the other group (X*) activates the sulfur atom of cysteine, thus facilitating the formation of a disulfide bond.In this case, the intermediate electrophilic derivative (7S7X* in compound 7) undergoes a nucleophilic attack by the sulfhydryl group of the second cysteine residue. This approach is used for the closure of both intra- and intermolecular S7S bonds and is particularly widely employed for the preparation of asymmetrical disulfides. Methods A and B are less efficient to this end.The reason is that simultaneous oxidation of two cysteine residues, when the for- mation of the S7S bonds between two different peptide chains is needed, gives predominantly dimers, whereas asymmetrical di- sulfides are the minor products.The peptide thiols are activated by introducing electron acceptor substituents, which enhance the ability of sulfur for the nucleophilic substitution. Activation may be accomplished either in situ, or as an individual stage with the isolation and characterisation of the intermediate product, or by selection of appropriate protective groups. In the latter case, the protective groups (e.g., of the disulfide type) must be sufficiently stable under conditions of elongation of the peptide chain and simultaneously ensure the formation of a disulfide bond by thiol ± sulfide exchange.a. Thiocyanogen (SCN)2 Thiocyanogen 2, 6 reacts both with free and protected thiols. The resulting activated peptide (X*=SCN in compound 7) reacts with the thiol group of the same or another peptide in situ to form a disulfide bond.It has been established that this reagent selec- tively reacts with Thp and Trt derivatives of peptides in the presence of Dpm and Bzl protections.5, 6 The reaction of thiocya- nogen with sulfides can be presented as a reaction of an electro- phile with a dissociated form of a sulfide.5 The rate of this reaction is determined by the degree of dissociation of the corresponding protected derivative 33, which, in turn, depends on the stability of the cation R+ formed upon splitting.Therefore, the cyclisation of Trt and Mob derivatives of peptides occurs rapidly and specifically even in nonpolar solvents, whereas Dpm sulfides react slowly and are converted into the corresponding disulfides in 60% ± 75% yields.5 The degree of dissociation of derivatives 33 may be enhanced by acid catalysis.Cleavage of Dpm and But sulfides requires stronger acids than that of Trt and Mob derivatives. If the reaction with thiocyanogen occurs at a slow rate (as is the case with Bzl derivatives), the formation of homodimeric side products is more probable.2, 5 High electrophilicity of thiocyanogen favours the occurrence of side reactions.2 Thiocyanogen is decomposed by amines (includ- ing free bases of arginine and histidine), the corresponding reactions may be carried out only in acidic or neutral media.b. Aromatic sulfenyl halides 2-Nitrobenzenesulfenyl chloride (Nps-Cl) or 2-pyridinesulfenyl chloride (PyS-Cl) activate free or protected cysteine residues. The closure of the disulfide bonds by these reagents may be illustrated by the following general scheme:2 In the case of intermolecular closure of the S ± S bonds, Y is a protective group of any type, and X is hydrogen or a protective group that can be removed by aromatic sulfenyl halides, e.g., Trt, Dpm, But, SBut, and Acm.Nps-Cl and PyS-Cl are extremely reactive, they decompose in the presence of moisture.PyS has been proposed for deprotection of cysteine protected with Trt, Dpm, Acm, But, and SBut groups and subsequent preparation of asymmetrical disulfides.108 Reac- tions with this reagent are carried out in glacial acetic acid. The reactive mixed disulfide formed reacts with a free SH group to give the target disulfide. The mixed disulfide Cys(PyS) is resistant to acids. Removal of the Boc group in the derivative Boc-Cys(PyS)- OH by treatment with acid is not accompanied by splitting of the PyS group.The activation of thiols and formation of disulfide bonds occur without removal of protective groups, such as Boc, Z, and Bzl. The PyS derivatives of peptides are formed in quantita- tive yields and can be isolated in a crystalline state. The mixed Ph S Ph+MeSiCl3 O 7ROSiMeCl2 30 RS SR 29 MeSiCl2 O + Ph S Ph 31 S +S Ph Ph SR 7R+ +PhSPh S 32 S 33 R S Cys R++7S Cys R0 S SCN R0 S S Cys+R++SCN7.SY SH S deprotection 7Y SX S SAr S ArSCl 7XCl pH 379 7ArSH Some peculiarities of synthesis of cysteine-containing peptides 557disulfides Cys(PyS) and the respective peptides are unstable in alkaline solutions and rapidly decompose to yield symmetrical disulfides and di(2-pyridyl) disulfide, (PyS)2.Free SH groups may be liberated by the action of an excess of thiols (e.g., b-mercap- toethanol).108 Nps-Cl was used to remove Trt, Dpm, Ddm, and Acm protections 109 and intermediate mixed disulfides can be isolated. Free SH-derivatives can be obtained under the action of an excess of a thiol (e.g., b-mercaptoethanol or dithiothreitol). Depro- tection and formation of disulfide bonds using Nps-Cl occur much more slowly than with PyS-Cl.108 The use of aromatic sulfenyl halides is limited because of difficulties associated with their synthesis and isolation (these compounds are easily hydrolysed even in humid air) and due to the fact that tryptophan can undergo side reactions under the action of these compounds.Peptides containing cysteine with free SH groups can be activated directly in solution or on a solid phase with di(2-pyridyl) disulfide, (PyS)2. Highly specific conversion of a Cys(Snm) derivative into a Cys(PyS) derivative under the action of 2- mercaptopyridine in chloroform is an alternative procedure for obtaining stable S-arylsulfenyl derivatives.2 And, finally, cysteine- containing peptides may react with Ellman's reagent in buffer solutions at pH 7.4 to give intermediate disulfides.The interaction of the latter with a free SH group of the other cysteine residue also results in the closure of S7S bridges.2 c. Disulfide derivatives The characteristic feature of this method is that disulfides play a dual role. On the one hand, they are used for protection of the thiol functions of cysteine because they are stable under condi- tions of a multistep peptide synthesis.On the other hand, being strong electron acceptors, they activate the thiol groups in reactions of thiol ± disulfide exchange and thiolysis and thus facilitate the closure of the S7S bonds. These groups include alkoxycarbonylthio (Scm, Sce), Snm, and Npys protections.Asymmetrical disulfides have been synthesised from peptides containing Cys(Npys) derivatives.17, 110 Derivatives containing Npys protective groups are also used to form intramolecular disulfide bonds.2 This reaction is carried out over a rather broad pH range. The reaction time varies from several hours at pH*4 to 30 min at pH 8 ± 9. However, at pH>9, disproportionation of disulfides may occur, which shifts the reaction equilibrium towards the formation of symmetrical homodimers.The probability of disulfide exchange between asymmetrical cystines formed and the free SH groups present decreases at low pH. The optimum pH values are 4.45 ± 9.35. The stability of the Npys protective group in acidic media (TFA at 25 8C and HF at 0 8C in the absence of thiols) makes it possible to use these cysteine derivatives in solid-phase syntheses using Boc- technique.This protection was successfully used for the formation of S ± S bridges in the synthesis of insulin, its analogues, and relaxin.17 Methoxycarbonylsulfenyl chloride (Scm-Cl), like aromatic sulfenyl halides, removes such protective groups as Trt, Acm, But, Ddm, and Dpm to form Cys(Scm) derivatives:111 Compound 34 reacts with free thiols to produce the disulfide bonds.The latter reaction is catalysed by mild bases, such as triethyl- amine.111 Sulfenylthiocarbonates are unstable in the presence of alkalis; therefore, experiments with these compounds must be carried out in neutral or acidic media. However, undesirable modification of tryptophan is possible.Conversion of Acm derivatives into Scm derivatives or the reaction of free SH groups with Cys(Scm) may also be carried out on a solid phase 2 with formation of both inter- and intramolecular disulfide bonds. Ethoxycarbonylsulfenyl chloride (Sce-Cl) was proposed for the formation of intramolecular disulfide bonds. This reagent is identical in structure and mechanism of action with Scm-Cl.111 The activated cysteine derivatives thus formed react with free thiol groups of cysteine to yield intramolecular disulfide bridges. The optimum results were obtained with ammonium-acetate buffer, pH 7.0. However, this reaction may also be carried out in weakly acidic solutions, but in this case the yields are lower.The rate of this reaction is very high: cyclisation is completed within 10 min, but the reaction conditions could not be found where the yields exceeded 20%± 30%.111 The application of Sce-Cl to tryptophan- containing peptides is problematic.111 Nevertheless, this reagent was used in the synthesis of somatostatin, urotensin IIA,111 and a trypsin inhibitor.56 d.Bis(tert-butyl) azodicarboxylate The closure of the disulfide bond using bis(tert-butyl) azodicar- boxylate (Boc7N=N7Boc) occurs via thiolysis.Boc ±N=N± Boc has not gained wide acceptance despite its successful applica- tion in the synthesis of asymmetrical disulfides.112, 113 The synthesis of an activated peptide derivative was carried out in DMF with 2 ± 3 equiv. of bis(tert-butyl) azodicarboxylate under argon (4 ± 12 h, 25 8C).112, 113 The resulting compound is rather stable and can be isolated.The activated derivative is dissolved in an appropriate solvent, after which the second peptide containing a free sulfhydryl group is added to form a disulfide heterodimer. BocN=NBoc is bis(tert-butyl) azodicarboxylate. e. The use of cysteine sulfoxide This method is used for the closure of both intra- and intermo- lecular S7S bonds.One of the cysteine residues is used in the form of a sulfoxide.114 The reaction occurs in a strongly acidic medium (TFA or a trifluoromethanesulfonic acid ±TFA mixture). Under these conditions, the sulfoxide is protonated, and the resulting electrophilic derivative readily reacts with the nucleophilic thiol. R1, R2, R3, R4 are amino acids. S H R4 N O R1 H H N NO2 R2 O N R3 S S R2 + R3 Cys S S R1 Cys R4 N NO2 S H X is Trt, Acm, But, Ddm, Dpm.X+MeO S R S Cl C O 7XCl RS 34 S C OMe O 34 R0SH+RSSCOOMe R0SSR+COS+MeOH S S 7Y deblocking SH N NHBoc Boc 7(BocNH)2 BocN NBoc SY SH S S+X SH HO SX SY 7Y+ H+ O S S H+ 7HX,7H2O Y is Mbzl, Ad, But; X is Acm or Mob. 558 E V Kudryavtseva,MV Sidorova, R P EvstigneevaThe use of dimethyl sulfide accelerates the reaction and favours regeneration of the SH group.The efficiency of this method was demonstrated by the synthesis of oxytocin from Cys(Acm)(O) and Cys(Mbzl) derivatives (yield 86%).114 f. The use of disulfide-containing polymers An original procedure for the synthesis of asymmetrical disulfides on polymeric supports containing disulfide fragments in the anchor group has been proposed by Ponsati et al.115 This method is based on thiol ± disulfide exchange.At the first step, the peptide thiol HSR1 reacts with the anchor group of the polymer 35 to form a disulfide peptidyl polymer 36. At the second step, compound 36 reacts with the second peptide HSR2 as a result of which the target disulfide R1SSR2 is cleaved from the polymer. An optimum anchor group for the polymer and conditions for this reaction have been selected.With X=7CH27CO*, the reaction was slow (48 h). If benzoic or nitrobenzoic acid residues were used as X, the reaction proceeded much faster, and the yields of the target heterodimers reached 60%± 70%. This method has a number of advantages. The polymer used in this reaction can be easily regenerated; the addition of the peptide to the polymer facilitates the procedure of purification of the reaction mixture from the excess of the reagents, by-products, and other impurities, which are removed by mere filtration.115 4.Closure of disulfide bonds on a polymeric support In a search for the most convenient cyclisation methods, it was suggested to form S7S bonds directly on a polymeric support.This method has a number of advantages. As oxidation requires high dilutions, a solid-phase system may be ideal in this respect. When a peptidyl polymer is swollen in an appropriate solvent, the distance between neighbouring peptide chains increases, which minimises interaction between them. This phenomenon, which is referred to as pseudo-dilution, simulates dilution in solution. The distance between molecules may be increased by decreasing the degree of the polymer substitution by the starting amino acid.Rigid fixation of molecules restricts their conformational mobility and decreases intermolecular contacts, which ensures predo- minant occurrence of intramolecular reactions.7 Oxidation on a polymeric support permits one to circumvent the problem of peptide solubility.In many cases, solvent systems the composition of which depends on cyclisation conditions do not provide complete dissolution of peptides to be oxidised, which inevitably results in the loss of the product and decrease in yields. Moreover, solid-phase synthesis significantly facilitates the workup of the reaction mixture. Thus, removal of an excess of the oxidant and side products is achieved merely by filtration.Since the oxidation reaction is often at equilibrium, this may be shifted towards the formation of the target product as a result of repeated treatment of the peptidyl polymer with a fresh portion of the oxidising mixture. The oxidation process can be automated and included into the reaction protocol after elongation of the peptide chain has been completed.However, despite the apparent simplicity of this procedure, the existence of a three-phase (peptide ± polymer ± - solvent) system presents certain problems for cyclisation. At present, there is still no deep understanding of the nature of the interaction between the peptide chains linked with the polymer and between the polymer and the peptide.Nevertheless, examples of successful cyclisation on a solid phase have been reported. The choice of a solvent is of crucial importance in this process. A solvent must afford good swelling of a peptidyl polymer in order to provide the maximum access of the oxidant to the peptide. The most popular solvents are DMF, CH2Cl2, CCl4, AcOH, trifluoro- ethanol, TFA, and N-methylpyrrolidone.2, 7, 15, 89, 100, 116 ± 119 The use of aqueous solutions should apparently be excluded with traditional polystyrene supports.As in solution, the closure of the peptide S7S bonds on a polymer may be conducted with both preliminary and simultaneous cleavage of protective groups for the sulfhydryl functions of cysteine. Removal of Hg(II) or Ag(I) ions employed for deblocking of cysteine with heavy metal salts requires the use of reagents that convert them into soluble derivatives.b-Mercaptoethanol, which yields Hg(II) salts soluble in a large amount of DMF, is especially recommended in this case.2 Systematic studies of disulfide bonds formation on polymeric supports in the synthesis of model peptides and oxytocin have been carried out.2, 89, 116 Cyclisation was conducted on polymeric supports with different anchor groups and using different reagents.Cysteine was used in the form of Cys(Fm), Cys(Trt), Cys(Acm),89 and Cys(Tmob) derivatives.14, 15 It was shown that the disulfide bond was rather stable under conditions of acid- promoted (TFA and HF) splitting of the cyclic peptide from the supports with p-alkoxybenzyl and benzhydrylamino anchor groups.89, 116 However, splitting of peptides from supports with photolabile anchor groups caused modification of the S7S bond; therefore special conditions are required in order to achieve satisfactory results.Closure of disulfide bonds was performed using atmospheric oxygen (DMF, pH 7 ± 8, 1 h), Ellman's reagent (0.1 M N-methylmorfolinium acetate in DMF, pH 7.5, 1 h), and diethyl azocarboxylate (0.5 equiv.DMF, pH 7.0, 2 h).89 It was found that the oxidation of oxytocin on a polymeric support occurred in higher yields than in solution.89 Oxidation on a polymer employed a mixture of CCl4 and NEt3 (1 : 1) in DMF or N-methylpyrrolidone.2, 7, 15 The hypothetical mechanism of this reaction is given below.7 The target oxytocin was obtained in 70% yield; no side products were detected.Oxidation of a 26-membered peptide containing free SH groups on a polymer was carried out in 1 Msolution of potassium ferricyanide in DMF. The reaction lasted 12 h and gave a cyclic disulfide in good yield.2 Methods of direct conversion of cysteine-protected peptides into disulfides [(I2, Tl(CF3COO)3)] have proved to be very efficient in oxidation on a solid phase.Standard conditions of this oxidation include treatment of cysteine-protected (Acm, Trt, Tmob) peptidyl polymers with 3 ± 10 equiv. of I2 in an appropriate solvent (TFA, AcOH, DMF, etc.) (1 ± 2 h, 25 8C) or with 1.2 equiv. of Tl(CF3COO)3 (1 ± 2 h, 0 8C). The yields of cyclic products are 60%± 90%.89, 117 Treatment of peptidyl polymers with acid-labile anchor groups [p-alkoxybenzyl or tris(alkoxy)- benzylamide] with solutions of I2 or Tl(F3CCOO)3 in TFA or AcOH results in simultaneous closure of S7S bridges and splitting of the peptide from the polymeric support.This approach was used by Barlos et al.118 for cleavage of the acid-labile X=7CH2 , CO , CO NO2 CO ; 35 P+HSR1 Y S S X 36 P+YSH R1 S S X 36 P+HSR2 R1 S S X R1SSR2+HS X P R1, R2 are the polypeptide residues.Y is 3-nitro-2-pyridyl; =(Bio-Gel P-2)7NH7(CH2)27NH7; P Et3N Cl3C S S H 7Et3N+H Et3N H S S H 7Et3N+H 7Cl7 CCl4 S7 S H 7CCl¡3 +H+ S S Cl3C S S7 Some peculiarities of synthesis of cysteine-containing peptides 5592-chlorotrityl anchor group with an AcOH: trifluoro-etha- nol :CH2Cl2 mixture (1 : 2 : 7). When this mixture was used in combination with 10 equiv.of I2 (10 ± 15 min, 20 8C), cyclic disulfides were obtained in *60% yields. Protective groups of the tert-butyl type were not removed under these conditions.118 The advantage of direct conversion of Cys(Acm) and Cys(Trt) peptide derivatives into the corresponding disulfides by treatment with Tl(III) salts over iodine has been demonstrated by several authors.89, 102, 116 Treatment of S-Acm or S-Trt derivatives of the model peptide Ac-Cys(Acm)-Pro-D-Val-Cys(Acm)-NH2 with Tl(CF3COO)3 in TFA: anisole (21 : 1) with subsequent cleavage from the polymer with HF gave a cyclic product in 91% yield, whereas after treatment with 10 equiv.of I2 in aqueous AcOH the yield was 52%.89, 116 In the latter case the decrease in the yield is unrelated to the nature of the oxidant but rather is due to the application of the AcOH± water mixture (1 : 4, v/v), which pre- vents swelling of the polystyrene-peptidyl polymer.Cyclisation on a solid phase was highly efficient in the syn- thesis of parallel dimers of oxytocin and deaminooxytocin.7, 119 The attempts to obtain these compounds in solution failed, because in dilute solutions it was monomeric products that were predominantly formed, whereas at high concentrations polymers were formed.119 Rigid fixation of peptides on a polymer has made it possible to obtain target dimers in good yields.The synthesis of a parallel dimer of deaminooxytocin is shown in Scheme 2. The CCl4 ± NEt3 mixture was used for the closure of the S ± S bond between Tmob-blocked N-terminal residues of b-mercapto- propionic acid (bMpa) after preliminary treatment of the peptidyl polymer with dilute TFA and scavengers. This procedure gave an intermediate product 37 in more than 90% yield.A mixture of the target parallel dimeric product 38 and deaminooxytocin 39 was obtained in 3 : 1 ratio after closure of the second disulfide bond with Tl(CF3COO)3 and cleavage of the reaction product from the polymer.Whether deaminooxytocin 39 results from the oxidation of an intermolecular side reaction product with Tl(F3CCOO)3 or from the dimer ± disulfide exchange still remains unclear.119 A scheme with an alternative combination of protective groups [Cys(Tmob) and bMpa(Acm)] has also been tested. At the first step of this synthesis, the disulfide bond was closed between cysteine residues.It is of note that in this case the yield of the intermediate product having only one disulfide bridge was as low as 20%. The closure of S7S bonds with N-iodosuccinimide in a DMF±CH2Cl2 mixture (1 : 1) in the synthesis of the (1 ± 29) fragment of the human growth hormone having one disulfide bond (yield 15.6%) and an 18-membered peptide apamine having two S ± S bonds (yield 11%) was reported.100 In this case, the closure of S7S bridges did not follow the conventional procedure, viz., oxidation after the completion of the linear peptide synthesis, but was rather included into the protocol of a solid-phase syn- thesis on an automatic peptide synthesiser using the Boc-techni- que.After attachment of the appropriate pair of cysteine residues, the disulfide bond between them was closed and the elongation of the peptide chain was continued.Acm and Mbzl protections were used for the blocking of the thiol functions of cysteine. The use of an analogous procedure for the synthesis of oxytocin resulted in extremely low yields of the cyclic product (*1%), although [8 Arg]vasopressin was obtained in 41% yield under identical con- ditions.The question about the reason for such a low yield of oxytocin is still obscure. In our opinion, the authors' attempts 100 to interpret this phenomenon in terms of intermolecular aggrega- tion of Cys(Acm) derivatives of oxytocin contradicts the results obtained by Albericio et al.89, 116 according to which it is cyclisa- tion leading to oxytocin on a polymer under various conditions that occurs in sufficiently high yields, which often exceed those obtained by cyclisation in solution.IV. Conclusion The methods of formation of disulfide bonds considered above usually give compounds having one or two S7S bonds. However, each of the above-described methods has limitations of its own. Most often, they are related to oxidation-sensitive amino acids: tyrosine, tryptophan, and methionine.Despite the explosive development of methods and techniques of modern peptide chemistry, no versatile cyclisation procedure exists, and each time when an S7S bond is to be formed in a new compound, special experiment is required for the selection of the most adequate protection for the thiol function of cysteine and cyclisa- tion reagent.In this context, further development of protective groups for the thiol function, elaboration of selective methods for deblocking of SH groups and specific activation of the thiol groups, and a search for new mild reagents suitable for the formation of S7S bridges are currently central tasks. Their solution assumes special importance in the synthesis of com- pounds with several intra- and intermolecular S7S bonds.At present, only few syntheses of such compounds have been carried out. Synthesis of complex molecules is a challenge of the nearest future. In this respect, methods of closing S7S bonds on a solid phase hold especially great promise. Further development and improvement in these techniques might facilitate laborious and sometimes unpredictable processes of cyclisation and make them TFA:CH2Cl2 : Et3SiH :H2O (7 : 95 : 1 : 0.5), 2613 min, 25 8C PAL PAL S-Tmob S-Acm P S-Tmob S-Acm NMP, 35 8C, 4 h 2.0 equiv.CCl4 : Et3N, 1.2 equiv. Tl(F3CCOO)3, DMF: PhOMe (19 : 1), 4 8C, 4 h TFA:CH2Cl2 : Et3SiH :H2O (95 : 5 : 0.5), 25 8C, 2 h 37 P PAL PAL S-Acm S-Acm S S PAL PAL S S P S S +bMpa-Tyr-Ile-Gln-Asn-Cys-Pro-Leu-Gly-NH2 39 bMpa-Tyr-Ile-Gln-Asn-Cys-Pro-Leu-Gly-NH2 bMpa-Tyr-Ile-Gln-Asn-Cys-Pro-Leu-Gly-NH2 38 + PAL PAL S-Acm S-Acm P SH SH Scheme 2 is the styrene copolymer with 1% divinylbenzene, PAL is the tris- (alkoxy)benzamide anchor group, NMP is N-methylpyrrolidone, bMpa is b-mercaptopropionic acid.P 560 E V Kudryavtseva,MV Sidorova, R P Evstigneevaautomatic by including this stage into the program of a peptide synthesiser.References 1. I Photaki, in The Chemistry of Polypeptides (Ed. P Katsoyannis) (New York: Plenum, 1973) 2. D Andrey, F Albericio, N A Sole,M C Munson,M Ferrer, G Barany, in Peptide Synthesis Protocols (Eds MWPennington, BMDunn) (New Jersey: Hymana Press, 1994) p. 91 3. O S Papsuevich, G I Chipens, S V Mikhailova, in Neirogipoézarnye Gormony (Neurohypophysis Hormones) (Riga: Zinatne, 1986) p. 6 4. K Alitalo, P Partanen, A Vaheri (Eds) Synthetic Peptides in Biology and Medicine (Amsterdam: Elsevier, 1985) 5. J FWMcOmie (Ed.) Protective Groups in Organic Chemistry (New York: Plenum, 1973) 6. R G Hiskey, in The Peptides Vol. 3 (Eds E Gross, J Meienhofer) (New York: Academic Press, 1981) p. 137 7. M C Munson, PhD Thesis, University of Minnesota, USA, 1993 8.G Barany, R B Merrifield, in The Peptides Vol. 2 (Eds E Gross, J Meienhofer) (New York: Academic Press, 1981) p. 3 9. N Fujii, A Otaka, S Funakoshi, K Bessho, T Watanabe, K Akaji, H Yajima Chem. Pharm. Bull. 35 2339 (1987) 10. M Yoshida,M Shimokura, Y Fyjiwara, T Fujisaki, K Akaji, Y Kiso Chem. Pharm. Bull. 38 382 (1990) 11. O A Tamamura,W Takada, Y Terakawa, H Yoshizawa, M Masuda, T Ibuka, T Murakami, H Nakashima Chem. Pharm.Bull. 43 12 (1995) 12. O Nishimura, C Kitada,M Fujino Chem. Pharm. Bull. 26 1576 (1978) 13. N Fujii, A Otaka, S Funakoshi,M Nomizu, K Akaji, H Yajima, I Yamamoto, K Torizuka, K Kitagawa, T Akita, K Ando, T Kawamoto, Y Shimonishi, T Takao Chem. Pharm. Bull. 34 613 (1986) 14. M C Munson, C Garcia-Echeverria, F Albericio, G Barany J.Org. Chem. 57 3013 (1992) 15. M C Munson, G Barany J. Am. Chem. Soc. 115 10203 (1993) 16. S N McCurdy Pep. Res. 2 147 (1989) 17. E E Bullesbach Kontakte (Darmstadt) 21 (1992) 18. N Fujii, A Otaka, T Watanabe, A Okamachi, N Tamamura, H Yajima, Y Inagaki,M Nomizu, K Asano J. Chem. Soc., Chem Commun. 283 (1989) 19. H-D Jakubke, H Jeschkeit AminosaÈ uren, Peptide, Proteine (Berlin: Academie-Verlag, 1982) 20.C Garcia-Echeverria, M A Molins, F Albericio, M Pons, E Giralt Int. J. Pept. Protein Res. 35 434 (1990) 21. M Royo, C Garcia-Echeverria, E Giralt, R Eritja, F Albericio Tetrahedron Lett. 33 2391 (1992) 22. N Fujii, A Otaka, S Funakoshi, H Yajima, O Nishimura Chem. Pharm. Bull. 34 869 (1986) 23. N Fujii, A Otaka, S Funakoshi,M Nomizu, K Akaji, H Yajima, K Kitagawa, T Akita, K Ando, T Kawamoto Chem.Pharm. Bull. 34 603 (1986) 24. Y Han, J Vagner, C M Gross, G Barany, in Proceedings of the 14th American Peptide Symposium (Abstracts of Reports) (Leiden: Escom, 1995) P061 25. H Yoshizawa, A Otaka, H Habashita, N Fujii Chem. Lett. 803 (1993) 26. A S J Stewart, C N C Drey J. Chem. Soc., Perkin Trans. 1 1753 (1990) 27.Shanlin Fu, F A Carver, L A P Kane-Maguire J. Organomet. Chem. 454 C11 (1993) 28. M Yoshida, K Akaji, T Tatsumi, S Iinuma, Y Fujiwara, T Kimura, Y Kiso Chem. Pharm. Bull. 38 273 (1990) 29. M Yoshida, T Tatsumi, Y Fujiwara, S Iinuma, T Kimura, K Akaji, Y Kiso Chem. Pharm. Bull. 38 1551 (1990) 30. Y Kiso,M Yoshida, Y Fujiwara, T Kimura, M Shimokura, K Akaji Chem. Pharm. Bull. 38 673 (1990) 31. Y Kiso,M Yoshida, T Kimura, Y Fujiwara, M Shimokura Tetrahedron Lett. 30 1979 (1989) 32. L Vesenkov, L Milev, A Dimitrova, in Proceeding of the 22nd Euro- pean Peptide Symposium (Eds C H Schneider, A N Eberle) (Leiden: Escom, 1992) p. 421 33. M Royo, J Alsina, E Giralt, U Slomczynska, F Albericio, in Proceedings of the 13th American Peptide Symposium (Eds R S Hodges, J A Smith) (Leiden: Escom, 1994) p. 116 34. O S Papsuevich, G I Aukone, S Ya Miksta, U O Kalei Zh. Obshch. Khim. 52 460 (1982) a 35. B H Rietman, R F R Peters, G I Tesser Synth. Commun. 24 1323 (1994) 36. K Nokihara, H Berndt J. Org. Chem. 43 4893 (1978) 37. R Eritja, J P Ziehler-Martin, P A Walker, T D Lee, K Legesse, F Albericio, B E Kaplan Tetrahedron 43 2675 (1987) 38. M Ueki,M Honda, Y Kazama, T Katoh Synthesis 21 (1994) 39.O Rosen, S Rubinraut, M Fridkin Int. J. Pept. Protein Res. 35 545 (1990) 40. F Albericio, D Andreu, E Giralt, C Navalpotro, E Pedroso, B Ponsati, M Ruiz-Gayo Int. J. Pept. Protein Res. 34 124 (1989) 41. R Matsueda, S Higashida, F Albericio, D Andreu, in The 29th Con- ference on Peptide Chemistry (Abstract of Reports), Osaka, 1991 p. 111 42.I Mangras, J Gostelli, E Rapp,R Nyfeler Int. J. Pept. Protein Res. 45 152 (1995) 43. E A Hallinan Int. J. Pept. Protein Res. 38 601 (1991) 44. G B Fields, R L Noble Int. J. Pept. Protein Res. 35 161 (1990) 45. B Walker, in Peptide Antigens (Ed. G B Wisdom) (Oxford: IRL Press, 1994) p. 27 46. H Nishio, T Kimura, S Sakakibara Tetrahedron Lett. 35 1239 (1994) 47. H Lamthank, H Virelizier, D Frayssinhes Pept.Res. 8 316 (1995) 48. H Lamthanh, C Roumestand, C Deprun, A Menez Int. J. Pept. Protein Res. 41 85 (1993) 49. L Lepsa, I Blaha, V Cerna, M Flegel, in 24th Symposium of the European Peptide Society (Abstract of Reports) (Edinburgh: Perkin Division, 1996) p. 126 50. MBodanszky, in The Peptides Vol. 1 (Eds E Gross, J Meienhofer) (New York: Academic Press, 1981) p. 106 51. L Moroder, G Hubener, S Gohring-Romani,W Gohring, H-J Musiol, E Wunsch Tetrahedron 46 3305 (1990) 52. P M Hardy, B Ridge,H N Rydon, F O dos S P Serrao J. Chem. Soc. 1722 (1971) 53. Hung Lam Thank, G Mourier, A Menez, P Fromageot, in Second Forum on Peptides (Colloquies Inserm) (Montroge: J Lifbey Eurotext Ltd, 1989) p. 229 54. R G Almquist, S R Kadambi, D M Yasuda, F L Weitl, W E Polgar, L R Toll Int.J. Pept. Protein Res. 34 455 (1989) 55. E V Kudryavtseva,M V Sidorova, M V Ovchinnikov, Zh D Bespalova, V N Bushuev Int. J. Pept. Prot. Res. 49 52 (1997) 56. D Le-Nguyen,D Nalis, B Castro Int. J. Pept. Prot. Res. 34 492 (1989) 57. Cui-Rong Wu, JDWade,GWTregear Int. J. Pept. Protein Res. 31 47 (1988) 58. S-I Kumagaye, H Kuroda, K Nakajima, T X Watanabe, T Kimura, T Masaki, S Sakakibara Int.J. Pept. Protein Res. 32 519 (1988) 59. H Tamamura, A Otaka, J Nakamura, K Okubo, T Koide, K Ikeda, T Ibuka, N Fujii Int. J. Pept. Protein Res. 45 312 (1995) 60. J M Stewart, J D Young Solid Phase Peptide Synthesis (New York: Pierce Chem. Co., 1984) 61. D Laloo,M K Mahanti Oxyd. Commun. 11 231 (1988) 62. A Buku, N Yamin, D Gazis Int. J.Pept. Protein Res. 33 28 (1989) 63. C Haskell-Lecevano, N D Shenderovich, Sh D Sharma, G V Nikiforovich, Mac E Hadley, V J Hruby J. Med. Chem. 38 1736 (1995) 64. A Misicka, V J Hruby Pol. J. Chem. 68 893 (1994) 65. J Rivier, R Kaiser, R Galyean Biopolymers 17 1927 (1978) 66. A Buku, P Blandina, C Birr, D Gazis Int. J. Pept. Protein Res. 33 86 (1989) 67. N Fujii, A Otaka, A Okamachi, T Watanabe, H Arai, H Tamamura, S Funakoshi, H Yajima, in Proceedings of the 20th European Peptide Symposium (Eds G Jung, E Bayer) (Berlin: Walter de Gruyter, 1989) p. 58 68. A Otaka, T Koide,A Shide,N Fujii Tetrahedron Lett. 32 1223 (1991) 69. J P Tam, Cui-Rong Wu, Wen Liu, Jing-Wen Zhang J. Am. Chem. Soc. 113 6657 (1991) 70. T Koide, A Otaka, N Fujii Chem. Pharm. Bull. 41 1030 (1993) 71. Chuan-Fa Liu, J P Tam Proc. Natl. Acad. Sci. USA 91 6584 (1994) 72. WC Schumb, C N Satterfield, R L Wentworth Hydrogen Peroxide (New York: Reinhold Publ. Co., 1958) Some peculiarities of synthesis of cysteine-containing peptides 56173. S Patai (Ed.) The Chemistry of Peroxides (Chichester: Wiley, 1983) p. 429 74. J D Roberts, MC Caserio Basic Principles of Organic Chemistry (New York: Benjamin Inc., 1964) 75. J Arroub, J Berges, Z Abedinzadeh, J Langlet, M Gardes-Albert Can. J. Chem. 72 2094 (1994) 76. Z Abedinzadeh, J Arroub,M Gardes-Albert Can J. Chem. 72 2102 (1994) 77. M V Sidorova, E V Kudryavtseva,M L Antopol'skii, M E Pal'keeva,M V Ovchinnikov, Zh D Bespalova Bioorg. Khim. 22 273 (1996) b 78. E V Kudryavtseva, Candidate Thesis in Chemical Sciences, MV Lomonosov Moscow Academy of Fine Chemical Technology, Moscow, 1997 79. L Kisfaludy, I Schon Acta Chim. Acad. Sci. Hung. 84 227 (1975) 80. Eur. P. 363 192; Chem. Abstr. 113 115 889 (1990) 81. G T Shulz, R H Schirmer Principles of Protein Structure (New York: Springer, 1979) 82. G Kupryszewski, U Ragnarsson, K Rolka, T Wilusz Int. J. Pept. Protein Res. 27 245 (1986) 83. S-Y Shin,M Shimizu, E Munekata Chem. Lett. 1997 (1994) 84. M W Pennington,W R Kem,R S Norton, B M Dunn Int. J. Pept. Protein Res. 36 335 (1990) 85. D L Rabenstein, P L Yeo J. Org. Chem. 59 4223 (1994) 86. M W Pennington, S M Festin,M L Maccecchini, in Proceedings of the 21st European Peptide Symposium (Abstracts of Reports) (Eds E Giralt, D Andreu) (Leiden: Escom, 1991) p. 164 87. B R Clarc, M Pai, in Proceedings of the 14th American Peptide Symposium (Abstracts of Reports) (Leiden: Escom, 1995) P090 88. Lin Chen, G Barany, in Proceedings of the 14th American Peptide Symposium (Abstracts of Reports) (Leiden: Escom, 1995) P119 89. F Albericio, R P Hammer, C Garcia-Echeverria, M A Molins, J L Chang, M C Munson,M Pons, E Giralt, G Barany Int. J. Pept. Protein Res. 37 402 (1991) 90. K Satoh, H Okuda, H Horimoto, H Kodama, M Kondo Bull. Chem. Soc. Jpn. 63 3467 (1990) 91. M V Ovchinnikov, Zh D Bespalova, A S Molokoedov, M I Titov, I V Revenko, V A Vinogradov, N V Korobov, S V Zhukovskii Bioorg. Khim. 14 777 (1988) b 92. S Chaturvedi, Om P Bahl Int. J. Pept. Protein Res. 35 133 (1990) 93. R Drozdz, A N Eberle J. Pept. Sci. 1 58 (1995) 94. P Fourqet, E Bahraoni, J C Fontecilla-Camps Int. J. Pept. Protein Res. 32 81 (1988) 95. E V Kudryavtseva,M V Sidorova,M V Ovchinnikov, Zh D Bespalova, V N Bushuev, R P Evstigneeva Bioorg. Khim. 22 370 (1996) b 96. M V Sidorova, E V Kudryavtseva, A S Molokoedov, M V Ovchinnikov, Zh D Bespalova Bioorg. Khim. 21 675 (1995) b 97. A Darbre (Ed.) Practical Protein Chemistry. A Handbook (New York: Wiley, 1987) 98. P Sieber, B Kamber, B Riniker,W Rittel Helv. Chim. Acta 63 2358 (1980) 99. P Bishop, J Chmielewski Tetrahedron Lett. 33 6263 (1992) 100. H Shih J. Org. Chem. 58 3003 (1993) 101. N Fujii, A Otaka, S Funakoshi, K Bessho, H Yajima J. Chem. Soc., Chem. Commun. 163 (1987) 102. W B Edwards, C G Fields, C J Anderson, T S Pajeau, M J Welch, G B Fields J. Med. Chem. 37 3749 (1994) 103. K Akaji, T Tatsumi, M Yoshida, T Kimura, Y Fujiwara, Y Kiso J. Chem. Soc., Chem. Commun. 167 (1991) 104. K Akaji, T Tatsumi, M Yoshida, T Kimura, Y Fujiwara, Y Kiso J. Am. Chem. Soc. 114 4137 (1992) 105. T Koide, A Otaka, H Suzuki, N Fujii Synlett 345 (1991) 106. K Akaji Yakugaku Kenkyn no Shinpo 11 25 (1995) 107. K Akaji, K Fujino, T Tatsumi, Y Kiso J. Am. Chem. Soc. 115 11384 (1993) 108. J V Castell, A Tun-Kyi Helv. Chim. Acta 62 2507 (1979) 109. A Fontana J. Chem. Soc., Chem. Commun. 976 (1975) 110. M S Bernatowicz, R Matsueda, G R Matsueda Int. J. Pept. Protein Res. 28 107 (1986) 111. D Le-Nguyen, J Rivier Int. J. Pept. Protein Res. 27 285 (1986) 112. S Romani, L Moroder,W Gohring, R Scharf, E Wunsch, Y A Barde, H Thoenen Int. J. Pept. Protein Res. 29 107 (1987) 113. E Wunsch, L Moroder, S Gohring-Romani, H-J Musiol, W Gohring, G Bovermann Int. J. Pept. Protein Res. 32 368 (1988) 114. H Yajima, N Fujii, S Funakoshi, T Watanabe, E Murayama, A Otaka Tetrahedron 44 805 (1988) 115. B Ponsati, M Ruiz-Gayo, E Giralt, F Albericio, D Andreu J. Am. Chem. Soc. 112 5345 (1990) 116. C Garcia-Echeverria, F Albericio, M Pons, G Barany, E Giralt Tetrahedron Lett. 30 2441 (1989) 117. Jpn. P. 06 256 385; Chem. Abstr. 122 56 591 (1995) 118. K Barlos, D Gatos, S Kutsogianni, G Papaphotiou, C Poulos, T Tsegenidis Int. J. Pept. Protein Res. 38 562 (1991) 119. M C Munson,M Lebl, J Slavinova, G Barany Pept. Res. 6 155 (1993) a�Russ. J. Gen. Chem. (Engl. Transl.) b�Russ. J. Bioorg. Chem. (Engl. Transl.) 562 E V Kudryavtseva,MV Sidorova, R P Evstign

ISSN:0036-021X    

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Abstract. The review deals with the current state of the problem of catastrophic oxidation of metals (`hot corrosion') caused by formation of a liquid corrosion product in a multi-component system. The kinetics, thermodynamics, and mechanisms of the two stages of accelerated oxidation of metals, namely, fast and super-fast stages, are considered with the copper±bismuth oxide system as an example.It is shown that the fast stage is caused by the formation of a liquid-channel grain boundary structure in the corrosion product, while the super-fast stage is caused by the high rate of dissolution of the oxide layer. The results of studies of the interaction of liquid phase with two-dimensional defects (grain boundaries) in solids are generalised. A model of catastrophic oxidation of metals is proposed.The bibliography includes 104 references. I. Introduction The creation of high-temperature corrosion-resistant materials intended for prolonged exploitation under extremely drastic conditions is a current task of materials science. Such materials are used in nuclear and fuel-based power industries, in space technology, and in other branches of industry.In most cases, composite materials based on metals, such as metal-ceramics, sintered powder compositions of metals with oxides, nitrides, and borides, as well as amorphous alloys are employed. The practical solution of problems involved in the creation of corrosion- resistant materials depends largely on the development of theo- retical concepts in the field of high-temperature oxidation of metals.By now, it has been reliably established that the rate of high- temperature oxidation of metals leading to the formation of dense scale is controlled by diffusion processes. The rate of metal oxidation is determined by bulk diffusion of the reacting ions and electron transfer through the growing scale. These aspects are described by the Wagner model1, 2 which makes it possible to find quantitative relationships between the rate of metal oxidation, the type and extent of oxide disordering, and the partial pressure of oxygen as a function of the type of conductivity of the scale (ionic or electronic).However, in certain cases the rate of metal oxidation exceeds that predicted by the Wagner model by several orders of magni- tude.This phenomenon referred to as catastrophic oxidation of metals (COM)3, 4 may be caused by the following. 1. The loss of isothermicity in the oxidation zone. The rate of metal oxidation increases continuously with temperature. This occurs if the liberated heat of the exothermic reaction cannot dissipate efficiently. Hence, the temperature of the metal surface increases, and the oxidation accelerates continuously.At a suffi- cient rate of oxygen supply, this process can result in inflammation of the metal. Inflammation of niobium, tantalum, titanium, and some other metals has been observed under extremal condi- tions.5, 6 2. The formation of volatile oxides. For example, tungsten, molybdenum, vanadium, rhenium, and osmium form volatile oxides but no protective oxide films.7 ±10 3.The formation of a small amount of a liquid phase in the oxidation product. This may be exemplified by oxidation of alloys the components of which form low-melting oxides; oxidation of metals in an atmosphere containing vapour of low-melting oxides; direct contact of metals and alloys with low-melting oxides. Even the appearance of a small amount of a liquid phase in the scale can lead to the COM.11 ± 27 In the latter case, two stages of COM, namely, fast and super- fast ones, can take place.28 ± 30 The fast stage is caused by the formation of a liquid-channel grain-boundary structure (LGBS) in the oxidation product 31, 32 and occurs by an electrochemical mechanism.The super-fast stage is caused by the high rate of dissolution of the protective oxide layer and occurs by a fluxing mechanism. The term `COM' has been first introduced by Leslie and Fontana3, 4 for an abrupt (by several orders of magnitude) increase in the rate of metal and alloy oxidation under isothermal conditions in a medium containing low-melting oxides (oxides of molybdenum, vanadium, lead, etc.).This phenomenon is also named `hot corrosion'.33 ± 38 Strictly speaking, this definition of COMrefers only to the third case discussed above.The other cases do not fit the definition completely. In the first case, metal oxidation occurs under non-isothermal conditions (a continuous increase in temperature in the reaction zone; the theory of this process has been suggested by Zel'dovich 5, 6) and is related mainly to the physics of burning.In the second case, the oxidising medium does not contain a foreign low-melting oxide. In the present work, only the third case is considered in detail with the copper ± bismuth oxide system as an example. To date, some progress has been achieved in the investigation of COM. However, many key questions remain debatable or unclear. For example, the mechanism of anomalously fast mass V V Belousov Open Joint Stock Venture `Research Institute of Steel' Dubninskaya ul. 81a, 127411 Moscow, Russian Federation. Fax (7-095) 485 43 95. E-mail: nsteel@dol.ru Received 21 March 1997 Uspekhi Khimii 67 (7) 631 ± 640 (1998); translated by S S Veselyi UDC 539.219.3:669.094.3 ± 977:541.135.4 Catastrophic oxidation of metals V V Belousov Contents I. Introduction 563 II.Fast stage of catastrophic oxidation of metals 564 III. Super-fast stage of catastrophic oxidation of metals 569 IV. Conclusion 570 Russian Chemical Reviews 67 (7) 563 ± 571 (1998) #1998 Russian Academy of Sciences and Turpion Ltdtransfer in solids involving the liquid phase and the mechanism of formation of a distributed structure (LGBS) are obscure; the interaction of a liquid phase with two-dimensional defects of a solid have been studied insufficiently. In this review, the author attempted to discuss, from common viewpoints (thermodynamic, kinetic, structural), the mechanism of theCOMcaused by the formation of a liquid corrosion product in a multi-component system.II. Fast stage of catastrophic oxidation of metals 1. Oxidation kinetics, structure and composition of the reaction products Let us consider COM with the copper±bismuth system as an example, which most fully reflects the general features character- istic of other systems.It is known26, 27 that copper coated with a thin film of bismuth (10 ± 100 mm) undergoes catastrophic oxida- tion in air at 770 8C(Fig. 1). The rate of copper oxidation does not depend on the starting state of bismuth but depends on its quantity (Fig. 2). It has been found28, 29 that the rate of copper oxidation increases with the thickness of the coating to reach a maximum at 100 mm thickness (Fig. 3). The process of COM is determined by a parabolic time law, which indicates that the process is diffusion-controlled. Dm S 2 =kt (1) Here Dm is the change in the mass of the specimen, S is the area of the surface that undergoes oxidation, k is the oxidation rate constant, and t is time.The k value depends on the coating thickness and exceeds the scale-formation constant on copper by a factor of over 1000. The temperature of 770 8C (threshold temperature for catastrophic oxidation of copper in the phase diagram of CuO ± Bi2O3, see Refs 39 ± 47) corresponds to the melting of the eutectics between Bi2O3 and Bi2CuO4.The COM process is associated with the formation of a liquid phase in the reaction products. The scale on copper consists of two layers,26 ± 29 viz., the inner dark layer and outer light layer with a distinct interface (Fig. 4). The outer layer is formed of Bi2CuO4 with minor amounts of Bi2O3, Cu2O, and CuO.The inner layer is built of CuO with minor Bi2CuO4, Bi2O3, and Cu2O. The concentrations of the latter two compounds increase abruptly near the copper surface. The distribution of elements across the scale is in quantitative and qualitative agreement with the phase distribution shown in Fig. 4. The position of layers in the scale shows that copper is oxidised predominantly due to oxygen diffusion. 2. Interaction of liquid phase with grain boundaries of the solid Model experiments carried out in order to reveal the specific features of penetration of Bi2O3 and Bi2CuO4 into the bulk of CuO formed upon copper oxidation show that intense penetration of Bi2O3 or Bi2CuO4 (in liquid state) along CuO grain boundaries occurs at 770 8C (Fig. 5). The width of intergranular layers depends on the quantity of compound that has penetrated and can vary from fractions of a micron to several microns. The 0 3 6 107 3 t /s 0.6 0.4 0.2 200 400 600 800 3 2 1 Dm S /kg m72 T /8C Figure 1.Kinetics of copper oxidation (1, 2) and temperature change (3); 1, without coating; 2, coated with a 30 mm-thick bismuth film. 1 2 3 4 0.4 0.3 0.2 0.1 0 6 12 18 24 1072 t /s Dm /g Figure 2.Kinetics of oxidation of copper with different coatings; (1), Bi2O3+Bi2CuO4; (2), Bi; (3), Bi2CuO4; and (4), Bi2O3. In all cases, the amount of bismuth per unit area of copper is the same, T=800 8C. 0.8 0.6 0.4 0.2 0 6 18 30 1072 t /s 4 3 2 1 Dm S 2 /kg72 m74 Figure 3. Kinetics of oxidation of copper coated with Bi2O3 films of different thickness at T=800 8C. Film thickness6105 /m: (1), 0; (2), 3; (3), 6; (4), 9; 105 k /kg2 m74 s71: (1), 0.012; (2), 1.6; (3), 4.9; and (4), 21.Figure 4. Microstructure of the scale on copper coated with a Bi2O3 film and oxidised in air at 800 8C; magnification is 400. 564 V V Belousovpenetration rate is high and reaches a few centimetres per minute (Fig. 6). The dependence of the penetration depth on time is linear.The activation energy of the process is 8720 kJ mol71. The linear time dependence of the depth of penetration of the liquid phase along grain boundaries and its high rate indicate that the process is not diffusion-controlled. 3. Phase formation According to the phase diagram, only one compound, namely, Bi2CuO4, is formed in the CuO ± Bi2O3 system.39 ± 47 The amount of this compound decreases upon heating above 730 8C (Fig. 7).32 The weight loss is accompanied by an exothermic effect on the DTA curve (2) and increases with temperature. The original weight is restored on cooling, but a small hysteresis appears on the TGA curve (1). The weight recovery is accompanied by additional exothermic effects corresponding to the reverse poly- morphic transformation of Bi2O3, d?g?b?a.48 ± 58 The weight loss is caused by partial dissociation of bismuth cuprate according to the reaction 2Bi2CuO4 2Bi2O3+Cu2O+0.5O2 : (2) The decomposition of polycrystalline Bi2CuO4 occurs both from the outer and inner surfaces (grain boundaries).A liquid phase is formed at grain boundaries (Fig. 8). The figurative point of the system in the phase diagram of Bi2O3±Cu2O (Fig. 9) with composition 2Bi2O3+Cu2O and a temperature of 730 8C is above the liquidus line, which corresponds to the liquid state. The volume fraction of the liquid phase increases with temper- ature. Hence, the equilibrium at the ternary diagram of Bi ±Cu ±O is shifted from the Bi2O3 ±CuO cross-section to the Bi2O3±Cu2O cross-section. The liquid phase forms a connected network of channels along Bi2CuO4 grain boundaries (see Fig. 8), thus forming an LGBS.31, 32 This implies that the thermodynamic condition of wetting, sgb>2ssl (where sgb is the surface tension of grain boundaries and ssl is the surface tension of the crystal± melt interface) is fulfilled. The exothermic effect at the DTA curve (first cycle, see Fig. 7) is probably related to wetting of grain boundaries with the liquid.Figure 5. Microstructure of a CuO ceramics following thermal treatment in contact with Bi2O3 in air (cross-section). 4 h /mm 3 2 1 0 10 20 t /s 12 6 Figure 6. Dependence of the depth of penetration of the liquid phase (h) along CuO grain boundaries on time at different temperatures. T /8C: (1), 860; (2), 900; (3), 950; and (4), 1050. 800 600 400 200 T /8C 0 1.44 2.88 4.32 5.76 7.20 1073 t /s 1 2 3 4 1 2 3 Dm /mg Figure 7.TGA (1), DTA (2), and T (3) curves for Bi2CuO4 exposed to cyclic heating and cooling in air. Figure 8. Microstructure of Bi2CuO4, quenching from 820 8C (fractured surface). 800 700 600 Bi2O3 0.5 Cu2O T /8C L b+L C a+b a+L a b Figure 9. Cross-section of the phase diagram of Bi2O3 ± CuO, pO2= 1072 Pa. Point C corresponds to 60 mol.% of Cu2O and 40 mol.% of Bi2O3, T=816 8C.Catastrophic oxidation of metals 5654. Liquid-channel grain-boundary structures Experimental results 26 ¡¾ 29, 31, 32 provide a reason to believe that the liquid phase plays an important role in the COM. This phase interacts with two-dimensional defects (grain boundaries) to form an LGBS, which affects considerably the mass transfer in the solid.In some instances, the contact of solids with liquids results in the formation of grooves where grains outcrop at the outer surface;59 ¡¾ 85 these grooves are characterised by a definite dihedral angle y (Fig. 10). The value of y is related to the surface tensions at the grain boundary and at the crystal ¡¾ melt interface by a mechanical equilibrium at the groove apex: sgb =2ssl cos y 2 .(3) A theoretical description of changes in topography of grain- boundary grooves and various transfer mechanisms is given by Mullins.60, 61 Let us briefly consider the Mullins theory of the development of a grain-boundary groove with a thermally etched groove as an example. Let a flat grain boundary outcrop at a flat surface (Fig. 10). The problem is to determine the surface profile at time t. The dependence of the chemical potential of an atom at a curved surface on the surface curvature is described by the Gibbs ¡¾ Thomson equation mi (L)=mi (1)+Oi sL, (4) where mi (L) and mi (1) are the chemical potentials of the i-th component at a surface of curvature L and at a flat surface, respectively, Oi is the partial volume of the i-th component, and s is the surface tension.Hence, the chemical potential of an atom is determined by the surface curvature at this point L=y 00 a1 a Oy 0U2a¢§3=2 . (5) The value Dm=mi (L) ¢§ mi (1) (6) reaches a maximum at x=0 and decreases at x>0 and x<0.At the same time, Hm is not equal to zero, causing a flux of atoms from the region with a higher potential (groove tip and banks) to that with a smaller potential (plateau), hence the groove should deepen and widen.The dependence of the groove dimensions on time (t) is determined by the mechanism of the atom transfer. For example, for surface diffusion the groove width (w) and depth (h) are proportional to t1/4: w=4.6(At)1/4 , (7) h=0.97(At)1/4 ctg y 2 , A= DssO2ns kT , where Ds is the surface diffusion coefficient, ns is the atom surface density, k is the Boltzmann constant, and T is temperature.For bulk diffusion w=5(Bt)1/3 , (8) h = (Bt)1/3 ctg y 2 , B= DsO2r kT , where D is the bulk diffusion coefficient and r is the bulk density. Quite a different physical picture appears if a fast diffusion pathway, such as liquid-phase diffusion, is possible. An analysis of the role of this process requires that the cases of saturated and unsaturated solutions are distinguished.The growth of a groove in a saturated solution occurs under the action of capillary forces. This process can be described using the Mullins theory, i.e. by the bulk diffusion of atoms through the liquid phase. This approach has been used repeatedly (see e.g., Refs 76 ¡¾ 78) to find that the groove width and depth change proportionally to t1/3.Calcula- tions agree satisfactorily with experimental data,76 which shows that the maximum growth rate of the groove in this mode is as high as 0.5 mm h71. This value is considerably lower than that in the case of COM, where the liquid phase interacts with the grain boundaries of a solid (see Fig. 6).Groove growth in unsaturated solution occurs under concen- tration gradient. Two extreme cases are possible. First, the process can occur in a kinetic mode (h * t, w * t), where the groove growth rate is controlled by dissolution, and diffusion transfer in the liquid phase occurs fast. Second, the process can occur in a diffusion mode (h*t 1/2, w*t 1/2), where the groove growth rate is controlled by diffusion through the liquid phase, and dissolution occurs rapidly.Calculations and experimental data 75 show that the maximum groove growth rate in the diffusion mode can be as high as several millimetres per hour, which is also much slower than in the case of COM. The possibility of attainment of high rates of the groove deepening in the kinetic mode was considered by Fradkov.81 It was shown that in narrow grooves the groove tip moves much faster than the banks, resembling the movement of scissors.The groove deepening rate (u) is related to the rate of dissolution of a flat solid surface (u0) 81 u a u0 sinOy=2U . (9) However, an estimate of v using this formula showed that it is physically unreal to reach rates of several centimetres per minute. Such a rate can only be obtained at y'0.20.Hence the Fradkov model is not applicable, as also noted by the author. The situation changes if u0 increases. For example, this can occur if the groove deepening rate is controlled by the kinetics of channel opening at the tip rather than by that of the banks' expansion. It has been established experimentally that in the case of COM, where a liquid phase interacts with grain boundaries of a solid, the rate of the channel opening at the tip is controlled by the chemical reaction CuO+Bi2O3=Bi2CuO4 .(10) It was noted above that Bi2CuO4 is thermally unstable and undergoes partial dissociation according to reaction (2). At grain boundaries, reaction (10) occurs, which [according to Eqn (2)] does not require any diffusion-controlled supply of the reagent, Bi2O3, or diffusion-controlled withdrawal of the product, Bi2CuO4.The oxygen evolved at the groove tip favours convective mixing (as the width of liquid channels exceeds considerably the free path of molecules). Thus, fast penetration of a liquid along the CuO grain boundaries occurs due to acceleration of a chemical reaction at the grain boundaries rather than in the bulk.The most significant factors that favour the development of this process are the following: the chemical potential of atoms at the grain boundaries is higher than in the bulk; a chemical reaction occurs which does not require a diffusion-controlled supply of reagents (or diffusion- controlled withdrawal of products); a liquid, thermally unstable y sgb ssl ssl y h w x Figure 10.Profile of a grain boundary groove of liquid-phase etching. 566 V V Belousovproduct is formed; reaction is accompanied by gas evolution; a local temperature gradient appears due to a local chemical reaction; and the liquid is supplied to the channel by capillary forces. The development of liquid channels occurs in the kinetic mode (see Fig. 6), hence equation (9) is still valid. However, v0 acquires the meaning of the chemical reaction rate on a solid flat surface. Estimates show that the kinetic mode of the liquid channel growth controlled by a chemical reaction can result in narrow deep channels with a dihedral angle at the tip 0<y<p at u *1 cm min71 and with u0 values 0.001 ± 1 cm min71, which are characteristic of topochemical reactions.86, 87 The activation energy of the process, 87 kJ mol71, is also typical of topochemical reactions.86 As a result of fast liquid penetration along the CuO grain boundaries, a continuous network of intergranular liquid chan- nels is formed, eventually resulting in the LGBS formation.Liquid-channel grain boundary structure can also be formed in different ways, e.g., by topochemical decomposition of ther- mally unstable Bi2CuO4.In both cases, the LGBS formation manifests itself as a sharp increase in electric conductivity caused by the onset of high ionic conductivity due to the oxygen ions.31,32 It has been shown28, 29 that an LGBS based on CuO (see Fig. 5) readily conducts the oxygen ions. The ionic conductivity reaches 1 O71 cm71 at 780 8C.The magnitude of conductivity is characteristic of superionic conductors.88 ± 92 Judging by phase composition and structure, LGBS should be attributed to heterogeneous structures. The question about the passage of electric current through heterogeneous structures has been first raised by Maxwell.93 In particular, he considered the conditions of the passage of current through a two-phase system, which is a continuous medium with randomly distributed spher- ical inclusions.Odelevskii 94 pointed out that heterogeneous structures can be organised in two ways. A feature of distributed matrix structures is that one of the two or several constituent phases always forms a bound matrix, however small volume the phase occupies. The phases constituting a statistically distributed structure do not form a bound matrix. Therefore, we shall treat LGBS as distributed matrix structures the matrix of which consists of liquid channels with a thickness from fractions of a micron to several microns localised at the grain boundaries.In some cases, the ionic conductivity increases at a certain ratio of phases that form a distributed phase.95 ± 100 For example, the ionic conductivity can increase due to boundary phenomena on hetero-junctions.98, 99 This is caused by disordering of crystal lattices in boundary regions and an increase in the concentration of mobile carriers.This effect is only characteristic of ionic conductors whose conductivity originates from the Frenkel or Shottky defects. Analogous effects are also possible if segregation of an admixture on phase boundaries produces surface transition phases of a different chemical composition with enhanced con- ductivity.100 Unlike the distributed structures mentioned above, LGBS incorporates continuous liquid channels along the grain bounda- ries.These channels can serve as diffusion paths for ions, since the activation energy of diffusion in a liquid is much lower than that in a solid.Figure 11 shows the dependence of the effective diffusion coefficient (Def) in LGBS on the volume fraction of phases calculated according to Maxwell and Hart.93, 101 The diffusion coefficient in the solid phase was assumed to be 1079 cm2 s71. For example, Def is 1076 cm2 s71 if the volume fraction of the liquid phase is 10%.Such a high diffusion permeability of LGBS is due to the presence of a highly conductive liquid matrix. Thus, the high ionic conductivity of LGBS is caused by the formation of a distributed matrix structure rather than by structural disordering as is the case in superionic conductors. 5. Mechanism of the fast stage in the catastrophic oxidation of metals The LGBS model makes it possible to explain the highCOMrates.Before discussing the mechanism of copper catastrophic oxida- tion, let us accept several key assumptions. 1. A liquid layer is formed on the surface of copper due to thermal dissociation (or melting) of Bi2CuO4. Let us assume that reaction (2) involves at least two stages: Bi2CuO4 Bi2O3+CuO , (11) CuO 0.5Cu2O+0.25O2 . (12) Each of these steps has its own equilibrium constant. Since ionic bonds predominate in all of the above oxides, it may be assumed that further dissociation into ions occurs in the melt.As an example, let us express reaction (12) in the ionic form CuO (1¡ a)Cu2+ +aCu+ + (1¡ a/2)O27 + (a/4)O2 , (13) where a characterises the degree of CuO dissociation according to reaction (12). Reaction (13), which describes an equilibrium between ions in solution, was written assuming total dissociation of CuO.In reality, the degree of dissociation is much less than unity, therefore we shall consider the liquid layer as a dilute solution. 2. As a result of interaction of grain boundaries with the liquid phase, an LGBS is formed as a product of copper oxidation. 3. Copper is oxidised with oxygen dissolved in the liquid layer.The oxygen concentration is replenished continuously by the atmospheric oxygen. 4. The diffusion of oxygen ions through the liquid layer and liquid intergranular channels (Fig. 12) is the rate-determining step of the process. 0.5 0.6 0.7 0.8 V2 0.2 0.3 0.1 1 2 Def D1 Figure 11. Effective diffusion coefficient of LGBS calculated according to Hart (1) and Maxwell (2); D1 is the diffusion coefficient in the liquid phase (*1075 cm2 s71), V2 is the volume fraction of the solid phase.Air Metal Oxide L c* c0 c1 c2 D h x d Figure 12. Topography of penetration of the liquid phase along the grain boundaries and interphase boundaries and a LGBS fragment in the catastrophic oxidation of copper. Catastrophic oxidation of metals 5675.The steady state is established locally all over the scale (where the scale is understood as the oxidation product + liquid layer). 6. The scale thickness is large in comparison with distances over which the effect of the bulk charge is felt (double electrical layer). 7. The solubility of oxygen in the metal can be neglected. Let us consider the mechanism of diffusion growth of the oxide (see Fig. 12). Let us assume the thickness of copper oxide that formed to be h at time t and its growth to occur due to the oxygen supplied to the surface through liquid intergranular channels. If the concentration of oxygen in the growing oxide is designated as c*, the mass balance equation is represented as follows c dh dt =J , (14) where J is the density of oxygen ion flux J=¢§D dc dx , (15) where D is the diffusion coefficient of oxygen ions and c is the concentration of the oxygen ions; D a RT 4F 2c , (16) where K is the partial specific electric conductivity of oxygen ions and F is the Faraday constant.Let us consider two variants of the COM model. In the first one, the liquid layer is a film of thickness D=10 ¡¾ 100 mm, D4Dcr, and D4h (Dcr'100 mm corresponds to the experimen- tally obtained film thickness at which the maximum rate of copper oxidation is observed).28, 29 It is assumed in this case that the material in the film is mixed and the concentration gradient of oxygen across the film thickness is absent.The oxygen concen- tration in the film is determined by dissociation equilibrium. However, a concentration gradient of oxygen exists in a liquid channel.The diffusion of oxygen occurs under permanent boun- dary conditions (stationary problem): O (x=0, t) = c0 , O (x=D, t)= c1, O (x=D+h, t)= c2. (17) Assuming a linear distribution of oxygen concentration in the channel, one obtains Hc a c1 ¢§ c2 h . (18) The oxygen concentration c2 is very small, so one may consider c2=0 and (taking into account the mixing in a liquid film) O1= c0.Thus, Hc a c0 h . (19) Substitution of equation (19) into (15) gives J a ¢§D c0 h . (20) The equation of mass balance (14) for thin films, taking account of (20), takes the form cl dh dt a dD c0 h , (21) (l is the mean grain size of copper oxide and d is the width of a liquid channel in LGBS) or c dh dt a Def c0 h , (22) where Def a D d l . (23) Separation of the variables in equation (22) and integration gives a parabolic law of copper oxidation h2 a kt .(24) Here k is the rate constant for copper oxidation k a 2Def c0 c . (25) The diffusion of oxygen anions along the liquid channel to the growing oxide surface is accompanied by a less intense counter- diffusion of copper cations into the film, which increases their concentration in the film.According to equation (13), this shifts the equilibrium to the left. As a result, the concentration of oxygen anions c0, and hence k, decrease. If this phenomenon is named `solution poisoning', it is obvious that the thinner the film the more pronounced this effect, and vice versa. Therefore, the dependence of k on the film thickness (for thin films, 4100 mm) can be explained by solution poisoning.This effect is insignificant and can be neglected for thick films. Estimates show that if the degree of CuO dissociation accord- ing to reaction (12) equals unity and, accordingly, the degree of Bi2CuO4 dissociation according to reaction (11) is 0.25 (see Ref 59), and the degree of dissociation of Bi2O3 and Cu2O into ions is about the same, the concentration of the oxygen anions in the liquid film is c00.03 (in atomic fractions). The rate constant for copper oxidation under a thin layer of Bi2CuO4 is calculated according to the formula ktheor a 2Def c0 c r2, (26) where r is the CuO density.At O*=0.5, Def&10710 m2 s71, and rCuO=6.456103 kg m73, we have ktheor45.261074 kg2 m74 s71, which is comparable to the experimental value kexp42.161074 kg2 m74 s71 (see Fig. 3). In the second case, the liquid layer thickness exceeds the critical value, D>Dcr and D>h. In this case, the gradient of oxygen concentration exists both in the channel and in the layer: in the layer Hc a c0 ¢§ c1 D , (27) where D=nh , n>1, i.e. Hc a c0 ¢§ c1 nh ; (28) in the channel Hc a c1 h . (29) The expressions for the flux of oxygen ions have the form: in the layer JODU a ¢§D c0 ¢§ c1 nh , (30) in the channel JOhU a ¢§Def c1 h .(31) In the steady state JODU a JOhU , (32) hence D c0 ¢§ c1 nh a Def c1 h . (33) 568 V V BelousovSubstituting (23) into (33), we obtain c0 ¢§ c1 n a d l c1 , (34) whence c1 a c0 1 a dn=l . (35) The equation of mass balance (14) for thick layers is written in the form c dh dt a Def c0 hO1 a dn=lU .(36) Integration of this expression gives a parabolic law of copper oxidation h2=kt, (37) where k a 2Def c0 c 1 1 a dn=l . (38) Thus, the number n increases and k decreases as the layer thickens, which is in accord with experimental results.28, 29 The theory is also applicable if the rate of metal oxidation is controlled by the diffusion of cations.In this case, the picture changes symmetrically, and the liquid layer is adjacent to the metal (Fig. 12). The suggested theory of COM is primarily valid in systems where oxidation results in the LGBS formation in the reaction products. As shown above, an LGBS is formed due to top- ochemical destruction or upon interaction of a liquid with grain boundaries of a solid.Therefore, the fast stage of theCOMis most probable in metal ¡¾ oxide systems in which formation of thermally unstable mixed or low-melting oxides is possible. III. Super-fast stage of catastrophic oxidation of metals 1. Oxidation kinetics In a study of the oxidation of copper under a bulk Bi2O3 layer thicker than 5 mm, it was found 28, 29 that at the instant when Bi2O3 melts (point A in Fig. 13) atmBi2O3/mCu5 20, copper plates (0.5 ¡¾ 1 mm thick) undergo super-fast (in 5 ¡¾ 10 s) oxidation in the medium of Bi2O3, thus reducing the latter. If mBi2O3/mCu < 20, only the amount of copper equal to 1/20 of the Bi2O3 mass is oxidised super-fast, while the rest of copper is oxidised much more slowly. The mBi2O3/mCu ratio at which super-fast oxidation of copper occurs depends on the partial oxygen pressure pO2: when pO2 decreases, mBi2O3/mCu decreases as well.The minimum value mBi2O3/mCu=3.6 is reached at pO2&102 Pa (Fig. 14). The literature contains controversial thermodynamic data for both the free energy of Bi2O3 formation and the dissociation pressure. According to some sources, copper reduces Bi2O3, while it does not do so according to others.102 ¡¾ 104 2.Mechanism of the super-fast stage of catastrophic oxidation of metals The high rate of copper oxidation under a layer of Bi2O3 melt at pO2=1072 ¡¾ 102 Pa (see Fig. 14) can be explained as follows. The initial mass ratio mBi2O3/mCu=3.6 corresponds to 60 mol.% of Cu2O and 40 mol.% of Bi2O3 (taking into account that all of the copper is oxidised into Cu2O and the initial amount of Bi2O3 decreases accordingly).According to the phase diagram of Bi2O3¡¾Cu2O (see Fig. 9), the composition of 60 mol.% of Cu2O and 40 mol.% of Bi2O3 and the temperature 816 8C correspond to point C on the liquidus line. If mBi2O3/mCu> 3.6, the figurative point of the system is to the left of point C, while at mBi2O3/ mCu=3.6, these points coincide.Copper is oxidised (reducing Bi2O3) into Cu2O, which in turn undergoes continuous dissolu- tion. Hence, super-fast oxidation occurs at the metal surface (without a protective oxide layer) due to a redox reaction. With- drawal of reaction products from the reaction zone probably occurs by convective diffusion. At mBi2O3/mCu<3.6, a fraction of copper which is 1/3.6 of the Bi2O3 mass is oxidised super-fast.After that, the figurative point of the system shifts to the right of C. Copper oxide formed on the copper surface is in the solid state and does not dissolve. Subsequent dissolution of copper occurs by the diffusion mecha- nism. Super-fast oxidation of the metal occurs also at higher partial oxygen pressures, so that the mBi2O3/mCu ratio increases (see Fig. 14). In this case, the mechanism of copper oxidation does not differ in principle from that described above and the increase inmBi2O3/mCu at which super-fast copper oxidation occurs is due to the fact that CuO is stable at pO2&6103 Pa,102 therefore the formation of Bi2CuO4 is possible,39 ¡¾ 42 which is in accord with experimental results.28, 29 The decrease in the amount of copper oxidised super-fast in air in comparison with the critical mass of copper oxidised in vacuo is due to the formation of Bi2CuO4 and hence depletion in the starting Bi2O3. Copper undergoes super- fast oxidation and reduces bismuth oxide to the metal, which is subsequently oxidised with atmospheric oxygen (section AB in Fig. 13). Simultaneously, Cu2O is partially oxidised to CuO.The insignificant mass decrease in the kinetic curve of copper oxida- 0 0.9 2.7 4.5 1073 t /s 12 16 Dm /mg 8 4 600 800 400 200 T /8C B A Figure 13. Kinetics of copper oxidation under a layer of Bi2O3 in air, mBi2O3=2.226 g, mCu=0.036 g. 15 20 10 5 0 mBi2O3 mCu 72 0 2 log (pO2 /Pa) Figure 14. Dependence of the critical ratio (mBi2O3/mCu) on the oxygen partial pressure (pO2).Catastrophic oxidation of metals 569tion in point A is due to partial dissociation of Bi2CuO4 according to reaction (2). Thus, super-fast copper oxidation is caused by continuous dissolution of the protective oxide film of Cu2O in a Bi2O3 melt. Oxidation occurs on a permanently bare metal surface due to a redox reaction. The critical ratio mBi2O3/mCu at which super-fast copper oxidation occurs is determined by the position of the figurative point on the liquidus line on the Bi2O3±Cu2O binary phase diagram, or on the liquidus surface in a ternary diagram if mixed oxides are formed.IV. Conclusion It should be noted that COM can occur both according to the electrochemical mechanism (diffusion of ions along liquid inter- granular channels) and according to the flux formation mecha- nism (dissolution of the protective oxide layer).The realisation of the particular mechanism depends on the mass ratio of the liquid phase (mL) and the metal (mM) and the Gibbs free energies of formation of the solid product (DG1) and the liquid phase (DG2). 1. If |DG1| > |DG2| and mL/mM 5 A, where A is the number determined by the position of the figurative point of the system on the liquidus surface corresponding to a definite composition and temperature on the phase diagram.In this case, a redox reaction is possible, while the LGBS formation is impossible; the figurative point of the system is located above the liquidus surface or coincides with it. The metal oxidation occurs at a very high rate according to the flux formation mechanism.The metal is oxidised within seconds. 2. If |DG1| > |DG2| and mL/mM < A, then both the redox reaction and the LGBS formation are possible. A fraction of the metal amounting to 1/A of the mass of the liquid phase is oxidised according to the flux formation mechanism. The figurative point of the system is then shifted below the liquidus surface. An LGBS is formed on the metal surface.Subsequent metal oxidation occurs more slowly according to the electrochemical mechanism. 3. If |DG1| < |DG2| and mL/mM 5 A, then both the redox reaction and the LGBS formation are impossible. Both the flux formation and electrochemical mechanisms are operative, there- fore, the rate of metal oxidation is much lower than that in case 1. 4. If |DG1| < |DG2| and mL/mM < A, then the LGBS formation is possible but the redox reaction is impossible.The metal fraction which amounts to 1/A of the mass of the liquid phase is oxidised as in the preceding case. The figurative point of the system is then shifted below the liquidus surface. An LGBS is formed on the metal surface. Subsequent oxidation follows the electrochemical mechanism. The general picture of COM can be represented as follows.The catastrophic oxidation of metals under a thin liquid layer no thicker than 100 mm occurs by the electrochemical mechanism and is caused by the LGBS formation in the corrosion product. In this case, the contribution of super-fast oxidation according to the flux formation mechanism is insignificant, as the metal fraction that undergoes super-fast oxidation is as small as 0.5% ± 3%.As the layer thickness increases, the contribution of super-fast oxidation becomes greater, which would change the character of oxidation and increase the rate of the process. Thus, the following conclusions can be made on the basis of the data presented above. 1. The fast stage of COM is caused by the LGBS formation in the corrosion product.The oxidation rate depends on the diffu- sion of ions along the liquid intergranular channels. 2. The super-fast stage of COM is caused by the high rate of dissolution of the protective oxide film. The oxidation occurs by a redox reaction on a continuously bare metal surface. References 1. C Wagner Z. Phys. Chem. 32 447 (1936) 2. C Wagner Z. Angew.Chem. 49 735 (1936) 3. W C Leslie, M G Fontana Trans. Am. Soc. Metals 41 1213 (1949) 4. M G Fontana Ind. Eng. Chem. 42 65 (1950) 5. Ya B Zel'dovich Zh. Eksp. Teor. Fiz. 9 159 (1939) 6. Ya B Zel'dovich Zh. Eksp. Teor. Fiz. 11 159 (1941) 7. P Kofstad High Temperature Oxidation of Metals (New-York: Wiley, 1966) 8. E A Gulbransen, G H Meier, in Proceedings of the 10th Symposium on Materials Research Vol. 561 (Washington, DC: National Bureau of Standards, 1979) p. 1639 9. E A Gulbransen, K F Andrew, F A Brassart J. Elektrochem. Soc. 110 952 (1963) 10. N D Tomashov, G P Chernova Teoriya Korrozii (The Corrosion Theory) (Moscow: Metallurgiya, 1993) 11. V V Belousov Corros. Sci. (1998) (in the press) 12. A De S Brasunas, N J Grant Trans. Am. Soc. Metals 3W 30 (1952); Chem.Abstr. 46 326a (1952) 13. S S Brenner J. Electrochem. Soc. 102 7 (1955) 14. G W Rathenau, J L Meijering Metallurgia 42 167 (1950) 15. J L Meijering, G W Rathenau Nature (London) 165 240 (1950) 16. K R Peters, D P Whittle, J Stringer Corros. Sci. 16 791 (1976) 17. M E El-Dahshan, D P Whittle, J Stringer Corros. Sci. 16 83 (1976) 18. M G Cox, B Mc Enaney, V D Scott Philos. Mag. 29 595 (1974) 19.M G Cox, B Mc Enaney, V D Scott Philos. Mag. 26 839 (1972) 20. J A Goebel, F S Pettit, G W Goward Metall. Trans. 4 261 (1973) 21. H L Logan Corrosion 15 443 (1959) 22. T T Huang PhD Thesis University of Pittsburgh, Pittsburgh, 1979 23. M E El-Dahshan, D P Whittle, J Stringer Corros. Sci. 16 77 (1976) 24. S S Brenner J. Electrochem. Soc. 102 16 (1955) 25. K Aning, Engen.Sci. Doct. Thesis Columbia University, New York, 1976 26. V N Konev, V V Belousov Zashchita Metallov 25 138 (1989) 27. V V Belousov Oxid. Metal. 38 289 (1992) 28. V V Belousov Zashchita Metallov 30 599 (1994) 29. V V Belousov Oxid. Metal. 42 511 (1994) 30. V V Belousov, B S Bokshtein Zashchita Metallov 34 36 (1998) 31. V V Belousov Elektrokhimiya 31 1343 (1995) a 32. V V Belousov J.Am. Ceram. Soc. 79 1703 (1996) 33. V V Belousov, B S Bokschtein Oxid. Metal. (1998) (in the press) 34. N Birks, G H Meier Introduction to High Temperature Oxidation of Metals (Pittsburg: E Arnold, 1983) 35. D K Gupta, R A Rapp J. Electrochem. Soc. 127 2194 (1980) 36. J A Goebel, F S Pettit Metall. Trans. 1 1943 (1970) 37. D R Chang, R Nemoto, J B Wagner, Jr Metall. Trans. 7A 803 (1976) 38. M R Wootton, N Birks Corros. Sci. 12 829 (1972) 39. J-C Boivin, O Thomas,G Tridot C. R. Hebd. Seances Acad. Sci., Ser. C 276 1105 (1973) 40. B G Kakhan, V B Lazarev, I S Shaplygin Zh. Neorg. Khim. 24 1663 (1979) b 41. Yu F Kargin, V M Skorikov Zh. Neorg. Khim. 34 2713 (1989) b 42. M P Kulakov, D Ya Lenchinenko Termochim. Acta 188 129 (1991) 43. C Changkan, H Yongle, B M Wanklyn, J B Hodby, F R Wondre J.Mater. Sci. 28 5045 (1993) 44. Y Ikeda, Y Sano, Y Bando, T Niinae, J Takada Funtai Oyobi Funmatsu Yakin 41 384 (1994); Chem. Abstr. 121 41 029 (1994) 45. C-F Tsang, J K Meen, D Ehhon J. Am. Ceram. Soc. 77 3119 (1994) 46. C Mallika, O M Sreedharan, J. Alloys Compd. 216 47 (1994) 47. B Hallstedt, D Risold, L J Gauckler J. Am. Ceram. Soc. 79 353 (1996) 48.G Gattow, H Schroeder Z. Anorg. Allg. Chem. 318 176 (1962) 49. H A Harvig Bismuthesguide: Phase Relations and Electric Properties, Utrecht, 1977 50. E M Levin, R S Roth J. Res. Nat. Bul. Stand. A, Phys. Chem. 68 189 (1964) 51. H A Harvig, A G Gerards Termochim. Acta 28 121 (1979) 52. M Tsubaki, K Kichiro Mater. Res. Bull. 19 1613 (1984) 570 V V Belousov53. H H Moebius, P Shuk, in 17 Jahrestagung Vereinigung fuÈ r Kristallorgaph Gesellschaft Geologische Wissenschaften, Leipzig, 1983 S. 12 54. L G Sillen, B Aurivillins Z. Kristallogr. 101 483 (1939) 55. L G Sillen, B Sillen Z. Phys. Chem. 49 27 (1941) 56. G Gattow Z. Anorg. Allg. Chem. 298 64 (1959) 57. H A Harvig, A G Gerards J. Solid. State. Chem., 26 265 (1978) 58. L G Sillen Ark. Kemi. Miner. Geolog., 12A 1 (1937) 59.V V Belousov, V N Konev, A K Roslik Sverkhprovodimost: Fiz., Khim., Tekhn. 3 1895 (1990) 60. W W Mullins J. Appl. Phys. 28 333 (1957) 61. W W Mullins J. Appl. Phys. 30 77 (1959) 62. H J Vogel, L Ratke Acta Metall. Mater. 39 641 (1991) 63. L G Pertsov, L A Pogosyan, B D Summ, Yu V Goryunov Kolloidn. Zh. 36 699 (1974) c 64. V V Belousov J. Am. Ceram. Soc. (1998) (in the press) 65.L A Pogosyan, Yu V Goryunov, V Yu Traskin Vestn. Mosk. Univ., Ser. 2, Khim. 5 589 (1974) d 66. K Nord-Varhaug J. Am. Ceram. Soc. 79 1147 (1996) 67. S Ramamuthy, M P Mallamaci, C B Carter Scr. Mater. 34 1779 (1996) 68. N Eustathopoulos, L Condurier, J Loud, P Desre J. Cryst. Growth. 33 (1) 105 (1976) 69. A Passerone, R Sangiorgi, N Eustathopoulos Scr. Metall. 14 1089 (1980) 70.A Passerone, R Sangiorgi, N Eustathopoulos Scr. Metall. 16 547 (1982) 71. A Passerone, R Sangiorgi Acta Metall. 33 77 (1985) 72. W W Mullins Metall. Mater. Trans. A, Phys. Metall. Mater. Sci. 26 1917 (1995) 73. A Passerone, N Eustathopoulos Acta Metall. 30 1349 (1982) 74. F N Rhines, B R Patterson Metall. Mater. Trans. A, Phys. Metall. Mater Sci. 13 985 (1982) 75.B S Bokstein, L M Klinger, I V Apikhtina Mater Sci. Eng. A, Sctruct. Mater., Prop. Mikrostruct. Proc. 203 373 (1995) 76. W M Robertson Trans. Metall. Soc. 233 1232 (1965) 77. W M Robertson J. Appl. Phys. 42 463 (1971) 78. B S Allen Trans. Metall. Soc. AIME 245 1621 (1969) 79. V V Belousov, B S Bokshtein Poverkhnost (1998) (in the press) 80. R L Kluch Trans. Am. Nucl. Soc. 15 746 (1972) 81.V E Fradkov Scr. Metall. Mater. 30 1599 (1994) 82. P L Flaitz and J Pask J. Am. Ceram. Soc. 70 449 (1987) 83. T M Shaw J. Am. Ceram. Soc. 74 2495 (1991) 84. E L Rabkin, V N Semenov, L S Shwindlerman, B B Straumal Acta Metall. 39 627 (1991) 85. D R Clarke, M L Gee, in Materials Interfaces. Atomic-Level Structure and Properties (Eds D Wolf, S Yip) (London: Chapman and Hall, 1992) p. 255 86. E A Prodan Neorganicheskaya Topokhimiya (Inorganic Topochemistry) (Minsk: Nauka i Tekhnika, 1986) 87. E A Prodan J. Thermal Anal. 29 941 (1984) 88. V N Chebotin Khimicheskaya Diffuziya v Tverdykh Telakh (Chem- ical Diffusion in Solids) (Moscow: Nauka, 1989) 89. V N Chebotin,M V Perfil'ev Elektrokhimiya Tverdykh Elektrolitov (Electrochemistry of Solid Electrolytes) (Moscow: Khimiya, 1978) 90. S Chandra Superionic Solids (Amsterdam: North Holland, 1981) 91. JW Perram (Ed.) Physics of Superionic Conduct and Electrode Materials (New York: Plenum, 1983) 92. A L Laskar, S Handra (Eds) Superionic Solids and Solid Electro- lytes ìRecent Trends (New York: Academic Press, 1989) 93. J C Maxwell, in A Treatise on Electricity and Magnetism Vol. 1 (Oxford: Clarendon Press, 1881) p. 403 94. V I Odelevskii Zh. Tekh. Fiz. 21 661 (1951) 95. B N Shklovskii, A L Efros Elektricheskie Svoistva Legirovannykh Poluprovodnikov (Electrical Properties of Alloyed Semiconductors) (Moscow: Nauka, 1979) 96. D J Bergman, in Electrical Transport and Optical Properties of Inhomogeneous Media (New York: American Physics Institute, 1978) p. 46 97. S Kirkpatrick Rev. Mod. Phys. 45 574 (1973) 98. I L Tallon,W H Robinson, S Smedley J. Phys. C 10 L579 (1977) 99. M Yan, R Cannon, H Bowen, R Coble J. Am. Ceram. Soc. 60 120 (1977) 100. C C Liang J. Electrochem. Soc. 120 1289 (1973) 101. E W Hart Acta Metall. 5 597 (1957) 102. O Kubashevskii, S B Allcock Metallurgicheskaya Termokhimiya (Metallurgical Thermochemistry) (Translated into Russian; (Moscow: Metallurgiya, 1982) 103. P Schuk, S Jakobs, H Moebius Haupt. Chem. Ges. DDR 100 (1980) 104. N A Vasil'eva, A F Maiorova Dokl. Akad. Nauk SSSR 280 385 (1985) e a�Electrochemistry (Engl. Transl.) b�Russ. J. Inorg. Chem. (Engl. Transl.) c�Colloid J. (Engl. Transl.) d�Moscow Univ. Bull. (Engl. Transl.) e�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) Catastrophic oxidation of

ISSN:0036-021X    

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Abstract. The mechanisms of the cryotropic gelation of poly(vinyl alcohol) solutions as well as the influence of the characteristics of the polymer and the conditions of the cryogenic treatment on the structure and physicochemical properties of the cryogels obtained are examined. Data on the characteristics of the freezing of concentrated poly(vinyl alcohol) solutions are presented. The influence of soluble additives, possessing different lyotropic properties, on the course of gelation in frozen systems is discussed.The possibility of the cryocracking of poly(vinyl alcohol) chains is considered. The bibliography includes 349 references. I. Introduction The cryogels based on polymer systems are macroporous hetero- phase gels formed as a result of the freezing, storage in the frozen state, and subsequent thawing of the initial solutions or colloidally disperse objects (colloid solutions, pastes, coagulates, etc.) in which preconditions for structure formation and transition into a gel are already present or are specially created (for example, by the introduction of a cross-linking agent).An obligatory condi- tion for the formation of heterogeneous gels is then the crystal- lisation (freezing) of the main bulk of the low-molecular liquid (solvent).After its thawing out, anisotropic macroporous gel-like polymeric products, called cryogels, or cryostructurates are formed.1 ±4 As in the case of the usual gels, produced in a liquid solvent, the junctions of the three-dimensional network of the cryogels can be of different nature: they can be established covalently (chemically cross-linked preparations), maintained as a result of relatively undissociated salt or coordinate-ionic bonds (ionotropic systems), stabilised by noncovalent interaction forces [thermoreversible (physical) cryogels], or they may represent a combination of these types of intermolecular contacts.4± 9 Although individual examples of the formation of gel-like structurates after the freezing ± thawing of solutions or dispersions of polymers have been known for a fairly long time (thus sponges produced by the freezing of starch paste were manufactured in Germany in the 1940s 10), this type of gelation began to be studied systematically only during the last 15 years.These investigations led to the creation of definite ideas about the principal mecha- nisms of the cryotropic gelation processes, specific effects charac- teristic of this process alone, its common features with gelation in unfrozen systems, and also the influence of various parameters of each stage of the cryogenic treatment on the properties of the materials obtained.2, 4 ± 9, 11, 12 The cryotropic gelation of solutions of poly(vinyl alcohol) (PVA) is examined in the present review.PVA-based cryogels are of great interest�both theoretical and applied. They have found extensive application in biotechnology, medicine, food industry, etc. In many respects, this interest is due to the excellent mechan- ical, diffusional, and thermophysical properties of the PVA cryogels (cryoPVAG), the availability of the polymer itself, its nontoxicity and biocompatibility, and also the relative ease of the method of formation of the gel (the PVA solution is simply frozen for a certain period and then thawed out).By varying the characteristics of the polymer employed (molecular mass, molec- ular mass distribution, the content of residual O-acyl groups, tacticity, PVA concentration in the system) the composition of the solvent, the nature of soluble additives (if these are introduced in the initial system), and also the regime in the cryogenic treatment (temperature and duration of freezing, rate of thawing, the number of refreezing cycles, etc.), it is possible to regulate within wide limits the physicochemical parameters of the final gels and their macro- and micro-structures.All these aspects will be discussed in the subsequent sections of the review. The ability of PVA to give rise to gels after the freezing ± th- awing of its solutions (here and henceforth we are dealing mainly with a polymer having an insignificant content of acyl residues in the lateral chains) was apparently first observed by Inoue.13 ± 17 He described the preparation of gel fishing baits from a mixture of a solution of PVA with fillers and aromatising agents.In order to obtain the gel, the initial suspension, poured into an appropriate mould, was initially frozen at a temperature below 75 8C for several tens of hours and then thawed out. In the scientific literature, a direct indication of such gelation appeared in the paper of Kukharchik and Baramboim,18 who noted that in the case of PVA (M=80 000) solutions with a concentration >3 g/ 100 ml the cryolytic cycles (freezing for 2 min at 778 8C and thawing at 25 8C for 15 min) resulted not only in an increase in viscosity but also in gelation.However, none of the characteristic features of such gelation of PVA and the properties of the cryogels were reported in the above investigation.18 Mention should also be made of the work of Polish inves- tigators,19, 20 who observed by the light scattering method the promoting effect of the cryogenic treatment on the association of V I Lozinsky N A Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences, ul Vavilova 28, 117813 Moscow, Russian Federation.Fax (7-095) 135 50 85.Tel. (7-095) 135 64 92 Received 16 July 1997 Uspekhi Khimii 67 (7) 641 ± 655 (1998); translated by A K Grzybowski UDC 541.64;539.3 Cryotropic gelation of poly(vinyl alcohol) solutions V I Lozinsky Contents I. Introduction 573 II. Characteristic features of the freezing of poly(vinyl alcohol) solutions 574 III. Dependence of the properties of poly(vinyl alcohol) cryogels on the characteristics of the gel forming polymer 576 IV.The influence of the conditions in the cryogenic treatment on the properties of poly(vinyl alcohol) cryogels 577 V. Dependence of the properties of poly(vinyl alcohol) cryogels on the nature and composition of the initial solvent 581 VI. Conclusion 582 Russian Chemical Reviews 67 (7) 573 ± 586 (1998) #1998 Russian Academy of Sciences and Turpion LtdPVA macromolecules in aqueous media, and also of the somewhat later study by Peppas.21 Some time later, the PVA cryogels were, as it were, `rediscov- ered'. This is probably associated with the observation of the possibility of increasing the mechanical strength of cryoPVAG as a result of the repeated freezing ± thawing of the cryogel already obtained 12, 22 ± 31 or the partial dehydration of the frozen specimen in vacuo (partial lyophilisation).32 ± 40 According to Chemical Abstracts, about 500 communications have been published on the cryotropic gelation of PVA, only a small proportion of these publications having been considered in the reviews of Nambu 11 (1990) and Peppas and Stauffer 12 (1991), as well as in our earlier review 41 (1992).The aim of the present review is to provide a more comprehensive pattern of the state of research into the given problem and also to note the problems which are still obscure.II. Characteristic features of the freezing of poly(vinyl alcohol) solutions It is well known that prolonged storage of concentrated aqueous PVA solutions at room temperature gives rise to gel formation (see, for example, Refs 42 ± 46).A similar result can be achieved faster by cooling the system (without freezing) 20, 47 or by impair- ing the thermodynamic quality of the solvent as a result of the introduction of nonsolvent additives.48 ± 52 In all cases low-melting (melting point Tm<30 ± 40 8C), mechanically weak, and as a rule almost transparent (opalescent) thermoreversible hydrogels are formed.On the other hand, if concentrated aqueous PVA sol- utions are initially frozen and are then thawed out after being maintained in the frozen state for a time, then turbid, elastic, high- melting (Tm>60 ± 80 8C) heterogeneous gels (cryogels) are formed under certain conditions instead of the transparent hydro- gels. Evidently the behaviour of the water ±PVA system subjected to cryogenic aion should affect the properties of the cryogels obtained and the kinetics of their formation. Figure 1 presents a generalised picture of the variation of the temperature of the system at all the stages of cryostructuration, i.e on freezing, storage in the frozen state, and thawing.IntervalA± B corresponds to the cooling of the specimen from the initial temperature to the temperature of the onset of crystallisation of the solvent (supercooling to a temperature Tu is frequently observed for PVA solutions).Owing to the evolution of the latent heat of crystallisation of ice, a `crystallisation plateau' is usually recorded along section C± B, but at high rates of cooling (uc>40 K min71) this plateau may not in fact be observed on the real time scale.53 A further decrease in temperature occurs between the points D and E up to the experimentally specified value (Ts).It has been shown 4, 54, 55 that the depression of the crystal- lisation temperature of the system (T0 ± Tf) for 5%± 10% aqueous solutions of PVA did not exceed 0.3 ± 0.8 8C, while the extent of supercooling is a parameter poorly reproducible with respect to temperature in view of the statistical nature of the ice nucleation processes.56 In the main, the trend is as follows: with increase in the PVA concentration in the initial solution, the supercooling temperature (Tu) falls and ice formation proper proceeds at a higher rate.The higher the rate of crystallisation, the finer are the polycrystals of the freezing low-molecular liquid produced.57 Thus the size of the ice crystals in the 7% PVA solution frozen at 710 8C is several times smaller, and in the 14% solution � smaller by 1 ± 2 orders of magnitude than in frozen pure water.58 Such microstructure of the polymer-reinforced ice affects significantly the physicomechanical characteristics of the frozen specimens 59 and an appreciable strengthening of such composite ice is observed even for a 1% PVA content in the initial solution and the ultimate uniaxial compression strength of the 10% preparation frozen at 710 8C is already an order of magnitude higher than for pure ice.After a time, this parameter slowly increases, which has been shown by studies involving the use of the NMR method,58 to be accompanied by a gradual decrease in the amount of unfrozen water.Important information about the properties of the water ± - PVA system at various negative temperatures has been provided by the study of the ESR spectra of a spin probe (4-hydroxy- 2,2,6,6-tetramethylpiperidine-1-oxyl) introduced into the initial polymer solution before freezing. It has been found 60 ± 63 that the correlation times of the rotational mobility of the stable nitroxyl radical, which depend directly on the viscosity of the micro- environment where reorientation movements of the probe take place, corresponds, for not unduly deeply frozen PVA solutions, to the rotations of the nitroxyl radical in a highly viscous but liquid medium.This shows that such polymer solutions, solidified by freezing (in particular aqueous PVA solutions), represent at least two-phase systems incorporating the crystalline phase of the frozen solvent and the phase comprising a heavily concentrated polymer solution � the so-called liquid microphase.64 ± 66 Quan- titative estimates of the volume of such liquid microphases of frozen PVA solutions have been made with the aid of a special method 60 for the analysis of ESR spectroscopic data.It has been shown that the amount of the liquid component varies from 0.86 ± 0.88 g of H2O per gramme of the polymer at 79 8C to 0.75 g of H2O per gramme of the polymer at 785 8C; the latter quantity was limiting and referred to unfrozen water. Approx- imately the same result (*0.8 ± 0.9 g of H2O per gramme of polymer) was obtained with the aid of NMR spectroscopy 67, 68 and differential scanning calorimetry.69 Taking into account the 10%± 15% error of the determination by each of these methods, such agreement should be regarded as rather good.In the region of `high' negative temperatures (from 71 to 77 8C), hysteresis phenomena are observed (Fig. 2), i.e. the amount of the liquid microphase and hence the concentration of the polymer dissolved in it depend greatly on the thermal pre- history of the system in the real (far from equilibrium) cooling or heating regimes.In particular, if fairly rapid heating of the frozen specimen is arrested at some intermediate temperature (in Fig. 2, this is74 8C), the system gradually relaxes to a phase state closer to equilibrium (transition 3, Fig. 2). Therefore, the slower the cooling of the frozen preparation or correspondingly the slower its heating during thawing, the greater is the extent to which the possible transformations of liquid water into ice are able to occur or, conversely, the closer to equilibrium is the amount of the liquid microphase. Direct observations at negative temperatures of the mobility of the unfrozen part of water in the water ±PVA system have been carried out 67, 68 with the aid of 2H NMR spectroscopy.For this purpose, the polymer was dissolved in D2O. The NMR signals of Ts Tu Tf T0 Ti A C D B E F G H t K T I Figure 1. Generalised thermogram for cryotropic gel formation; Ti=initial temperature of the system, T0=temperature at which the pure solvent is frozen, Tf=temperature at which the system is frozen, Tu=lowest supercooling temperature, Ts=temperature at which the specimen is stored in the frozen state. 574 V I Lozinskythe heavy water deuterons in the solid and liquid states differ in width by two orders of magnitude and it is therefore possible to record the spectra referring to the D2O molecules not incorpo- rated in the ice structure. The 2H NMRspectra of a 10% solution of PVA (M± 82 000), frozen to a temperature of752.2 8C, which is 56 8C lower than the crystallisation point of the pure solvent (T0=+3.8 8C), contain one symmetrical signal without quadru- pole splitting. The width of this signal depends on the temperature and in the temperature range indicated varies from several tens to *5000 Hz.The nature of the spectra then indicates the virtually isotropic reorientation of the solvent molecule, which indicates the rapid (1075 s on theNMRtime scale) rotational and translational diffusion of water in the isotropic medium; i.e.in the unfrozen liquid microphase the water molecules are subjected to interac- tions which are virtually equivalent and equally probable in all directions. On the other hand, the proportion of the solvent located near the anisotropic interfaces is, firstly, small and, secondly, it apparently exchanges rapidly with the molecules from the bulk of the liquid inclusions.On further cooling of the specimen to temperatures 60 ± 70 8C lower than its freezing point, the symmetrical Gaussian peak (with a width of *5 kHz at the level of half its height) is transformed into a two-component signal, the upper component of which has been broadened by a factor of 4, while the lower component, the so-called `pedestal', corresponds to a signal with a greatly increased spin ± lattice relaxation time. It has been suggested 67 that this character of the spectrum at the temperatures indicated is caused by the slow exchange of the 2H nuclei of the solvent and the polymer.Thus it has been shown that moderately frozen PVA solutions are heterophase objects containing the unfrozen mobile solvent, the minimum amount of which corresponds to the ratio*4 water molecules for 3 polar chain groups. Yet another effect which can occur in principle during the freezing of polymer solutions, including aqueous PVA solutions, is cryocracking , i.e.the cryogenic mechanical degradation of the chains of high-molecular compounds under the influence of the dynamic stresses arising on crystallisation of the low-molecular solvent. Berlin and coworkers 70 ± 73 were the first to draw atten- tion to the possibility of such mechanochemical transformations of polymers during the freezing ± thawing of their solutions. Somewhat later Jellinek and Fok 74 calculated the crit rates of movement of the fronts of the crystallisation of the solvent (ucr) and of the migration of the microstreams of the liquid ahead of the growing faces of the crystals, which are necessary to develop forces sufficient for the disruption of the covalent bonds of the main framework of the carbochain polymers.These calculations were performed for two types of physical action on the macromolecule: (a) impact across the chain fixed (frozen in) at both ends (which is generally speaking an extremely improbable event) and (b) the frictional shear force of the liquid stream ahead of the crystal- lisation front exerted on the chain fixed at one end.Calculation yielded the following values of ucr: *100 m s71 in the first case and*0.5 m s71 in the second.On the other hand, in practice the attainable rates for preparations with a volume of several milli- litres do not exceed 0.002 ± 0.003 m s71.75 For preparations with a volume of several microlitres, it is possible to achieve an increase in the rate by 1 ± 2 orders of magnitude during their rapid freezing (for example, in liquid nitrogen), but even such increase in the rate makes it possible merely to approach the second case.Therefore one should apparently not expect a real manifestation of the mechanical degradation of PVA for shallow freezing of its solution under the conditions employed for the formation of cryoPVAG. Thus in a study of the behaviour of PVA solutions during thawing and freezing with the aid of capillary viscometry, Kukharchik and Baramboim 18, 76 failed to observe any kind of cracking of PVA either during single or repeated freezing ± th- awing of dilute and semidilute solutions of the polymer (M=80 000).Nevertheless, facts have been described which indicate, according to the authors,77 ± 80 the occurrence of homo- lytic cleavage of PVA by the cryogenic influence; here it has also been suggested 77 that cracking is accompanied by the simulta- neous covalent cross-linking of the polymer molecules as a result of reactions between the macroradicals formed and the neigh- bouring chains.However, a thorough test of the results of the above stud- ies 77 ± 80 failed to confirm the conclusions reached in them.81, 82 In the first place, the cryoPVAG are thermoreversible gels which melt on heating (for example, on a boiling water bath).Here the viscosities of the initial solutions of PVA (cp=6%± 10%) having different molecular masses (69 000 ± 110 000) and solutions obtained from the corresponding cryogels after fusion of the latter hardly differ, which indicates the absence of intermolecular cross- linking. Capillary viscometry does not reveal reliably detectable changes in the intrinsic viscosities of the PVA preparations, provided that the initial polymer does not contain any impurities.Thus lines 4 and 5 in Fig. 3 illustrate the dependence of [Z] on the number of cycles in the cryogenic treatment of solutions of two different batches of the same brand of technical-grade PVA, for which both a decrease (line 5) and an increase (line 4) in [Z] were observed, whereas variation within wide limits of the freezing ± th- awing regimes for solutions of the same polymer purified by reprecipitation does not induce changes in the intrinsic viscosity (lines 1 ± 3, Fig. 3). Thus the possible reactive impurities contained in industrial PVA, concentrated in the liquid microphase of the frozen preparations, may be the cause of the change in the properties of the polymer recorded after the cryogenic action. 78 76 74 72 0 0.5 1.0 1.5 2.0 mw/mPVA /g g71 T /8C 1 2 3 Figure 2. Dependence of the relative amount of unfrozen water on the temperature of the frozen 10% PVA solution (mw=mass of water remaining liquid in the frozen specimen; mPVA=mass of PVA dissolved in the liquid microphase);60 (1) cooling of the specimen beyond the crystallisation plateau of pure water; (2) heating; (3) thermostatic regime at74 8C. 1 2 0 3 4 5 n 0.4 0.6 0.8 1.0 [Z] /dl g71 1, 2, 3 4 5 �1 �2 �3 Figure 3. Dependence of the intrinsic viscosity [Z] of PVA on the number of freezing ± thawing cycles (n) for a 10% PVA solution of brand 20/1. (1) ± (3) PVA purified by reprecipitation (M=81 700, DD=98%); (4) and (5) unpurified PVA (two different batches of technical-grade poly- mer).Cryogenic treatment regimes: (1), (4), and (5) uc=1.8 K min71 (freezing at715 8C and storage at this temperature for 18 h), heating rate uth=1.0 K min71; (2) uc=1.8 K min71, uth=0.3 K min71; (3) uth=180 K min71 (freezing in liquid nitrogen for 30 min and storage at715 8C for 18 h), uth=0.3 K min71. Cryotropic gelation of poly(vinyl alcohol) solutions 575What precisely these impurities are is unknown, their composi- tions and amounts are not the same in different batches of the polymer of the same brand.On the other hand, in the case of specimens of pure PVA the changes in molecular mass and manifestations of mechanical degradation are not observed either viscometrically or with the aid of high performance GPC or any other physicochemical methods � IR, ESR, etc.81, 82 One may therefore conclude, to within the sensitivity of the analytical procedures employed, that the cracking of the PVA chains on cryogenic treatment is negligible, if it does occur at all, which is in no way reflected in the cryotropic gelation of the given polymer.III.Dependence of the properties of poly(vinyl alcohol) cryogels on the characteristics of the gel forming polymer The structure and properties of PVA cryogels depend on many factors, including both the characteristics of the polymer itself and the conditions of the low-temperature gelation. One of the most important parameters determining the possibility, in principle, of the formation of cryoPVAG, is the presence in the polymer of residual O-acetyl groups.9, 24, 25, 83 In the majority of cases described, the cryogels were prepared from poly(vinyl alcohols) obtained by hydrolysing poly(vinyl acetate).It was found that, for a degree of deacetylation (DD) below 90%, the cryoPVAG are not formed under any, even the most favourable, conditions. For a content of OAc residues of 5%, very weak cryogels are obtained, and it is actually difficult to measure correctly their rheological characteristics.Only on the basis of concentrated aqueous sol- utions of PVA of highly deacylated brands (DD>97%) is it possible to form fairly strong cryogels. Bearing in mind that the bonds between the polymer chains in the gels formed are non- covalent, i.e.the intermolecular contacts in the regions of micro- crystallinity are ensured by a multiplicity of hydrogen bonds between the secondary alcohol groups of different chains,35, 84 the steric hindrance generated by the bulky OAc residues, inter- fering with the establishment of extended complementary con- tacts, should be included among the evident causes of the `poor' gelation of the insufficiently deacylated PVA preparations. The macromolecules of atactic poly(vinyl alcohols) in water are known 43, 85, 86 to tend to associate, whereupon the intermo- lecular hydrogen bonds arise mainly with participation of the OH groups of the syndiotactic (s)PVA blocks while the intramolecular hydrogen bonds are produced with participation of the OH groups of the isotactic (i) fragments.Thus the higher the content of the s-diads in the polymer employed, the stronger is the manifestation of the tendency towards phase separation in its solutions 87, 88 with formation of a microgel fraction, a gel, or even a precipitate. The cryogels based onPVAenriched in syndiotactic fragments have a higher degree of crystallinity 89 compared with specimens based on the atactic polymer of approximately the same molecular mass, which indicates the retention of the above tendency also on cryotropic gelation of this polymer.Unfortunately the cryoPVAG prepared from poly(vinyl alcohols) with an increased content of s-sections have been inadequately investigated, but there is no doubt of the interest in such studies, particularly since new ways of synthesising PVA with a controllable tacticity are being vigorously developed nowadays.It is entirent that the molecular mass of the polymer used for the formation of cryogels must influence its properties. Thus fairly weak cryoPVAG are obtained from highly concen- trated (cp=20%) solutions of PVA with a low degree of polymer- isation (M<30 000 ± 40 000),4, 11, 24, 54, 90 so that it is preferable to employ PVA brands with M>50 000 for the formation of strong cryogels.Figure 4 presents the creep curves for equiconcentrated cry- oPVAG prepared from PVA with different contents of the residual OAc groups and different degrees of polymerisation. For the close molecular masses (curves 2 and 4), the additional presence in the polymer of even small amounts (1.3%) of unhydrolysed side groups affords a cryogel with a compliance greater by a factor of 3, whilst in specimens with an almost identical degree of deacylation of PVA (curves 1, 2, and 3) less compliant (more elastic) cryoPVAG are obtained for the higher the molecular mass of the polymer.54 These and similar results 4, 24, 25, 90 ± 93 for the influence of the size of the PVA macromolecules on different characteristics of the corresponding cryogels are fully consistent with the well-known dependences of the properties of noncovalent thermoreversible gels (formed at positive temperatures) based on both PVA49, 94 ± 97 and other polymers.98 ± 101 Since the hydrodynamic dimensions of the mac- romolecules increase with increase in the chain length, the degree of overlap of the coils in solution rises and the number of topological entanglements increases, i.e.there is an increase in the number of intermolecular contacts, as a result of which a denser physical network is produced in the gel formed. Such gels are harder and have higher melting points The properties of cryoPVAG are influenced similarly also by the variation of the PVA concentration in the initial solution � the higher the concentration, the stronger and more thermostable (other conditions being equal) are the corresponding cryogels obtained.4, 9, 11, 12, 22 ± 25, 35 ± 37, 40, 54, 55, 58, 83, 84, 89, 91, 102 ± 113 It is shown in Fig. 5 that the cryogel formed on cryotreatment of a 10% PVA solution is fairly compliant and is deformed appreci- ably during the application of a shear load, whereas the cryogels obtained from a 14% and especially a 16% PVA solution exhibit only a slight creep under the same shear stresses. i-blocks ...... CH2 CH CH2 CH CH2 CH O O O H H H ... ... ... ... H H O CH2 CH CH2 CH CH2 CH O H H H O CH CH2 CH2 CH CH2 CH O H s-blocks O O 1074 J0 /Pa71 0 20 40 60 80 t /min 2 6 12 16 20 1 2 3 4 Figure 4.Creep curves for cryogels formed from 10% aqueous solutions of PVA with different molecular masses and degrees of deacylation.54, 91 (1) M=69 000, DD=98%; (2) M=90 000, DD=98%; (3) M=110 000, DD=98%, (4)M=95 000, DD=99.3%. Cryostruc- turation regime: Ts=710 8C, ts=24 h, uth=1.0 K min71. 576 V I LozinskyWhen a PVA solution freezes, the pure solvent crystallises initially (Fig. 1), while the solute is concentrated in the still liquid part of the specimen. Such cryoconcentration intensifies signifi- cantly the polymer ± polymer interactions which ultimately leads to the formation of stable junction of the three-dimensional cryogel network. Since these processes occur in a heterophase system, a heterogeneous gel is obtained after it has thawed out.Pores of different size and geometry are formed in it under the influence of the crystals of the frozen solvent. Thus the solvent plays the role of a porogen. According to electron microscopy data, the structures of the cryoPVAG formed from solutions with unequal polymer contents differ, although their structural feature such as pronounced heterogeneity, i.e. the presence of micropores and macropores, is characteristic of any PVA cryogels.The morphology of 14% and 16% cryoPVAG is illustrated in the photomicrographs in Figs 6a and 6b. Their creep curves are presented in Fig. 5 (curves 2 and 3). The high porosity of the preparations is most striking. With increase in the PVA concen- tration, a decrease in the size of the macropores and a thickening of the elements of the polymer phase of the heterogeneous gel are observed, which agrees well with the smaller compliance of the more concentrated specimen.Perhaps the only characteristic of PVA the influence of which on the properties of cryoPVAG has so far been virtually unin- vestigated is the molecular mass distribution (MMD) of the gel forming polymer. Nevertheless it is quite evident that the use of PVA with a narrower MMD should promote an increase in the strength parameters of the cryogels.IV. The influence of the conditions in the cryogenic treatment on the properties of poly(vinyl alcohol) cryogels It was already stated above that virtually all the parameters of the regimes in the cryogenic treatment influence, albeit not to the same extent, the characteristics of the cryoPVAG and the dynam- ics of their formation.These questions have been investigated in fair detail. The process temperature and the effectiveness of the heat exchange determine the possibility of the supercooling and the rate of crystallisation of the specific specimen: usually the lower the Tf (Fig. 1), the lower the probability of the supercooling, and hence the more rapid the liquid ± solid phase transition for the main bulk of the low-molecular solvent.By altering the thermal conductivity conditions (for example, by employing vessels made from different materials, or by increasing the thickness of the walls of the vessel, or by employing thermal insulation), it is possible to control the dynamic freezing and thawing parameters. As a result of such experiments, it has been found 55, 102 that the degree of supercooling and the rate of crystallisation of ice do not vary in parallel with the variation of the rate of cooling, which indicates the complexity of the dependence of the freezing regimes for concentrated aqueous PVA solutions on different process factors.Thus, for a 14% polymer solution poured into a 5 mm duralumin mould, the change in uc from 0.09 to 17.0 K min71 (i.e.almost by a factor of 190) entails a change in ucr in the range from 0.35 to 2.5 K min71 (the difference is a factor of only 7.1), while the rheological parameters of the cryogels obtained altogether change very slightly and, other conditions being equal, depend little on uc. Therefore, in comparing the properties of the cryoPVAG speci- mens formed at different Tf, the influence of the dynamics of the freezing itself may be virtually disregarded.In the main, the physical characteristics and the structure of the PVA cryogels are determined by the temperature and duration of the cryogenic action,24, 58, 91, 106, 114 provided that the preparations are thawed out under identical conditions.55, 115 Thus it has been observed that, after brief (up to 24 h) storage of the specimen in the frozen state, the lower the Tf, the less compliant the corresponding cryoPVAG.With increase in the duration of the cryogenic process, the trend changes (Fig. 7). The specimens kept at 720 8C and then for 24 h at 710 8C become harder than the preparations maintained at 730 8C. A characteristic feature is that the duration of the storage in the frozen state ceases to influence their strength properties fairly early (after *24 h), whereas in the specimens stored at 710 and 720 8C the values of J0 after cryogenic treatment for 10 days still remain unstabi- lised,91 which indicates the continuation at these temperatures of the changes associated, as already mentioned, with the slow decrease in the amount of the uncrystallised solvent.58 Like their strength, the thermophysical properties of the cryoPVAG also depend on the parameters examined above; the increase in the degree of deacylation or in the molecular mass of the polymer as well as in its concentration in the initial solution promotes an increase in the melting point of these thermorever- sible gels.11, 23 ± 25, 35 ± 37, 40, 54, 55, 83, 105, 106 A decrease in Tf and an increase in the duration of the storage of the specimens in the frozen state exert the same influence.55, 1 1074 J0 /Pa71 0 20 40 60 80 t /min 4 8 12 16 20 2 1 3 Figure 5.Creep curves for cryogels formed from aqueous solutions of PVA (M=69 000, DD=98%) with different polymer contents: (1) 10%; (2) 14%; (3) 16%.91 Cryostructuration regime: Ts=710 8C, ts=24 h, uth=1.0 K min71.a b Figure 6. Structures of cryogels formed from 14% (a) and 16% (b) aqueous solutions of PVA (M=69 000, DD=98%).54, 91. Cryostructuration regime: Ts=710 8C, ts=24 h, uth=1.0 K min71. Cryotropic gelation of poly(vinyl alcohol) solutions 577However, the most significant factors which radically affect the properties of the PVA cryogels are the regimes of the defrost- ing of the specimens.55, 106, 109, 115, 116 A detailed study of this stage of the cryostructure formation made it possible to establish the principal features of the mechanisms of this type of gel formation.It has been observed 4, 54, 55 that, when the thawing out is carried out at a rate greater than 10 K min71, then, regardless of the conditions under which the initial polymer solution was frozen and then exposed to temperatures between710 and730 8C, one obtains after such `rapid' thawing out not a gel but a viscous and turbid colloid solution.Conversely, on slow thawing out (uth40.05 ± 0.01 K min71) very elastic cryogels are formed. Figure 8a illustrates the influence of the rate of thawing of the specimens on the apparent instantaneous compliances (J0) of the PVA cryogels, whilst Fig. 8b demonstrates the dependence of the same rheological parameter on the polymer concentration in the initial solution for two series of experiments in which the thawing out was carried out at rates of 1.0 and 0.02 K min71. Figure 8 shows that the compliance of the cryogels diminishes as the thawing out of the corresponding preparations slows down; for uth=1.0, 0.1, and 0.02 K min71, J0=8.161074, 0.661074, and 0.4561074 Pa71 respectively (Fig. 8a). If the PVA solution is maintained at720 8C for 10 days and is then thawed out at a rate of 1.0 K min71, then a cryogel with J0=1.2561074 Pa71 is obtained, while in the case where the rate of thawing is lower by an order of magnitude (uth=0.1 K min71), after freezing for 1 h at 720 8C a more rigid cryoPVAG is formed.Similarly a decrease in uth makes it possible to lower the concentration of the gel forming agent without loss of the strength characteristics of the cryoPVAG (Fig. 8b). Figure 9 presents the thermomechanical curves for specimens of cryogels thawed out at different rates. It is seen from the figure that the slowly thawed out preparations are, firstly, deformed to a smaller extent by an applied load as the temperature rises and, secondly, they melt at higher temperatures on heating, i.e.are more thermostable.106 The presence of two inflection regions on the thermomechanical curve for the specimen which was thawed out fairly rapidly (curve 1, Fig. 9) indicates the presence in the structure of such a gel of at least two types of junctions differing in the thermal dissociation energies.Those of the first type melt at 35 ± 45 8C, while these of the second type melt at 60 ± 65 8C. Two endothermic peaks have been recorded in the same temperature ranges also with the aid of differential scanning calorimetry 38, 116 and the first of these, corresponding to the decomposition of the low-melting nodes, is approximately at the temperatures (20 ± 40 8C) at which melt the usual thermoreversible PVA hydro- gels formed on ageing of concentrated polymer solutions without freezing.44, 117 On the other hand, the complete fusion of the cryoPVAG, caused by the thermal dissociation of the high- melting junctions (the microcrystallinity zones), occurs at a temperature higher by tens of degrees.The melting points Tm, like the other characteristics of the gel objects, are sensitive to the previous thermal history of the specimens (Table 1).55 If the dependence of the melting point of the cryoPVAG on the polymer concentration in the system is represented in terms of the variables of the Eldridge ± Ferry equation 98 lg cp à DH 2:3RT á const , 0 50 100 150 200 t /h 2 4 6 8 10 1074 J0 /Pa71 1 2 3 Figure 7.The influence of the temperature and duration of freezing of 14% solutions of PVA (M=69 000, DD=98%) on the rheological properties of the cryogels formed.91 Ts (8C): (1) 710; (2)720; (3) 730. 1074 J0 /Pa71 b 0 0 0.4 1.2 uth /K min71 4 8 12 16 1074 J0 /Pa71 a 10 8 6 4 2 6 9 12 cp /g dl71 1 2 Figure 8.The influence of the rate of thawing of the specimen (a) and of the concentration of the polymer in the initial solution (b) on the rheological properties of cryogels formed from aqueous solutions of PVA (M=69 000, DD=98%).55 (a) cp=14%; (b) Ts=720 8C, ts=1 h, uth=1.0 (1) and 0.02 K min71 (2). 1073 D l /m 0 0.4 0.8 1.2 1.6 20 30 40 50 70 60 T /8C 1 2 3 Figure 9.Thermomechanical properties of cryogels formed from 14% solutions of PVA (M=69 000,DD=98%) at Ts=720 8C and ts=1 h and thawed out at different rates uth (K min71):106 (1) 1.0; (2) 0.2; (3) 0.02. Table 1. The influence of the conditions in the cryotropic gel formation on the melting point of cryoPVAG. Cryogenic treatment regimes Tm /8C Tf /8C tf /h uth /K min71 710 24 1.00 73.5 0.3 720 24 1.00 74.0 0.2 0.20 74.8 0.2 0.02 76.0 0.2 72 1.00 74.1 0.2 240 1.00 74.4 0.1 730 24 1.00 73.7 0.2 Note.A 14% solution of PVA with M=69 000 and DD=99.0% was used. 578 V I Lozinskyit is possible to estimate the enthalpies of fusion of the cryogels obtained under different conditions (Fig. 10). The values of DH are 682 kJ mol71 for preparations thawed out at a rate of 1.0 K min71 and 872 kJ mol71 for preparations thawed out at uth=0.02 K min71.55 Comparison of these quantities with the available data 44 for the usual PVA hydrogels obtained on ageing of 10% ±20% solutions of the polymer (M=60 000), the melting points of which are in the range 14 ± 28 8C and DH= 37 kJ mol71, shows clearly that the energy of the thermal decom- position of the junctions of the cryogel network is at least twice as high as for the usual gel, i.e.the promoting effect of the freezing of the PVA solutions on the association of the macromolecules of the polymer and its crystallisation is clearly seen. The microstructure of the PVA cryogels also depends on the dynamics of the thawing out process: the slower is the thawing out process, the lower the degree of heterogeneity of the preparations according to scanning electron microscopy data.4, 54, 55, 118 According to transmission electron microscopy data,91 the pri- mary elements of the supermolecular structure of the cryoPVAG are globular formations 0.01 ± 0.02 mm in diameter, combined into larger globular aggregates 0.2 ± 0.5 mm in diameter, which form in their turn bead-like strands of the three dimensional network of the macroporous cryogel.It is of interest that similar globular aggregates have been observed 51 also in the structure of hydrogels obtained either on ageing of concentrated polymer solutions or after the introduction of a nonsolvent. The following question is relevant: if the rate of freezing of the polymer solutions influences little the properties of the cryoPVAG, the temperature and the time of the cryogenic treat- ment exert a definite effect on the characteristics of the cryogels, and the dynamics of the thawing out process is of decisive importance for them, then what is the precise stage in the cryostructure formation process where gelation proper takes place � during freezing (where the polymer is concentrated in the liquid microphase as a result of the crystallisation of ice), during the storage of the frozen solutions at a negative temper- ature, or in the course of the thawing out process? It has been shown 55 that, when the PVA solution is initially frozen at720 8C for 1 h, is subsequently maintained at a temper- ature in the range from 70.5 to 76.0 8C, and is then rapidly (in order to rule out the influence of the thawing out dynamics) thawed out, the dependence of the rheological characteristics of the cryogels formed by means of this scheme on the temperature at which they are stored has an extremum (Fig. 11).The least complia (Fig. 11a), i.e. the most elastic, cryoPVAG is formed when the frozen system is maintained in the region of 72.5 8C (Fig. 11b): above and below this temperature, less strong cryogels are obtained. On the other hand, overall the region of the maximum effectiveness of such cryotropic gel formation is in the subzero temperature range � only several degrees below the crystallisation point of the solvent. Hence the influence of the thawing out kinetics on the proper- ties of the PVA cryogels becomes understandable: the slower and closer to equilibrium conditions is the thawing out of the speci- men, the longer it resides in the region of temperatures optimum for the formation of the gel network.For example, on rapid thawing (uth=10 K min71) the system `passes' through the range 75 ± 0 8C in only 30 s, in the moderate thawing out regime (uth=0.1 K min71) the passage takes 5 min, while in the slow regime (uth=0.02 ± 0.01 K min71) the passage takes 4 ± 8 h.Thus a slow increase in temperature during thawing out is in a certain sense equivalent to the storing of the frozen preparations for a long time at the temperatures corresponding to the formation of the strongest cryogels, when in the PVA± water system mainly the junctions of the three dimensional network of the cryoPVAG are formed. If the concentrated aqueous solutions of the polymer during their cooling to the temperatures of the most intense gelation (*72.5 8C) could be guaranteed to freeze, which is frequently prevented by supercooling effects, then in this case there would apparently be no need at all for the use of lower temperatures. In fact they are required to induce the crystallisa- tion of ice.It has been demonstrated with the aid of NMR spectro- scopy 67, 68 that when the frozen preparations are heated after the attainment of a threshold temperature range, which is 4 ± 6 K below the crystallisation point of the system, a sharp increase in the amount of the mobile solvent is observed. An additional amount of the liquid component appears precisely at these temperatures still in the solid frozen specimen in addition to the firmly bound unfreezable solvent (see also Fig. 2). By virtue of the partial `prefusion' of the ice, a medium is produced in which the macromolecules (or their segments) acquire a certain mobility, which ultimately leads to effective polymer ± polymer interactions with formation of microcrystallites performing the function of cryogel network junction knots.It is shown in Fig. 12 that such mobility is reflected in the 13C NMR spectra of the PVA± water system, while the signals of the CH(OH) and CH2 groups of PVA are detected at 66 and 44 ppm respectively. Since the polymer is dissolved in D2O in order to record the spectra, the temperature scale on the figure is 2.83 2.87 2.95 2.91 2.99 0 0.5 1.0 1.5 2.0 2.0 1000/Tm /K71 log cp 1 2 Figure 10.Relation between the melting points of cryogels and the concentration of PVA (M=69 000, DD=98%) in the initial solutions frozen at 720 8C during 24 h and thawed out at the rate of 0.02 (1) and 1.0 K min71 (2) 55 in terms of the variables of the Eldridge ± Ferry equation. 0 5 10 15 3.0 2.0 1.0 0 1074 J0 /Pa71 a G /kPa b 76 74 72 0 T /8C Figure 11.Dependence of the shear compliances (a) and the moduli of shear elasticity (b) of cryogels formed from 14% solution of PVA (M=69 000, DD=98%) on the temperature at which the frozen speci- mens were maintained in the subzero region.4, 55 Cryotropic gelation of poly(vinyl alcohol) solutions 579indicated both in absolute numbers and relative (Tf ± T0) to the melting point of the pure solvent.At Tf ±T0=77.8 8C, the resonance peaks of the carbon atoms of PVA are hardly noticeable, while at a temperature three degrees higher, at Tf ±T0=74.4 8C, fully resolvable, although weak, signals are already observed. On further increase in temperature, but still 2 ± 3 8C below the temperature of the thawing out of the specimen as a whole, the `normal' spectrum of the polymer solution in D2O is recorded.Hence it follows that, in the temperature range 5 ± 6 8C below the freezing point of the system, there is a conventional boundary of the retardation of rapid (on theNMRfrequency scale) movements of the carbochain framework of PVA. Estimation of the correlation times of the spin ± lattice relaxation of the polymer carbon atoms, which serves as a criterion of the mobility of the kinetic segments of the chain, shows that the reorientation movements of the 13C nuclei of the CH2 groups of PVA in the system frozen at subzero temperatures occur more slowly by two orders of magnitude than in the initial liquid solution.It is precisely because of the low mobility of the polymer chains that a fairly prolonged time is needed for the occurrence of the cryotropic gel formation.The passage through an extremum of the dependence of the rheological properties of the cryoPVAG on the temperature of the cryotropic structuration process in its subzero region then becomes understandable (Fig. 11). On the one hand, at temperatures above the extremum there is a significant increase in the amount of the mobile solvent, i.e.the specific concentration of the gel forming agent is reduced in the liquid microphase. This entails a decrease in the probability of effective intermolecular contacts, which are also acted against by the intensification of the thermal movement in the system. On the other hand, at temperatures below the extremum there is a rapid decrease in the amount of the `additional' mobile solvent (the polymer retains only very strongly bound water), i.e. the liquid medium in which movements leading to the mutual approach of the interacting sections of neighbouring chains may occur van- ishes.Thus, having ensured a prolonged residence of the system in the temperature range 72 to 73 8C after freezing, for example during slow thawing, it is possible to prepare even from solutions with a not very high PVA concentration a cryogel which is several times less compliant than the cryogel obtained from a significantly more concentrated polymer solution but under the conditions of a brief residence of the specimen at subzero temperatures [at uth=0.02 K min71 for an 8% PVA solution, J0=1.5561074 Pa71 (Fig. 8b), whereas at uth=1.0 K min71 for a 16% PVA solution, J0=7.961074 Pa (Fig.5) for the same freezing time tf=1 h].As regards the influence on the properties of the cryoPVAG of cryogenic treatment parameters such as the temperature and duration of freezing (Fig. 7), they are more likely to act on the process involving the formation of a microgel fraction, which then performs the function of building blocks for the formation (in the range of subzero temperatures) of a gel throughout the volume of the specimen. At any rate, the liquid colloidal systems formed were more viscous and turbid the longer the corresponding objects existed in the frozen state.54 The structure of the supermolecular aggregates entering into the composition of such a microgel fraction has not so far been investigated, although data of this kind would be very useful for the understanding of the nature of the processes under discussion.Yet another example of the cryogenic treatment which makes it possible to increase the strength characteristics of cryoPVAG is repeated freezing ± thawing of the PVA± water system. A large number of communications (see Refs 11, 12, 23 ± 25, 28, 30, 40, 69, 84, 104, 105, 108, 119 ± 134) and patents 22, 26, 27, 29, 135 ± 151 have been devoted to the study of this phenomenon and its employ- ment.Repeated cryogenic treatment (`temperature saw') pro- motes more far reaching gel formation in these systems. Thus the cryogel obtained as a result of four freezing ± thawing cycles for a solution of PVA (M=75 000 and DD=99.5%, cp=15%, Tf=720 8C, tf=24 h, thawing at 15 8C for 7 h) has a dynamic Young's modulus an order of magnitude greater 30 than that of the specimen obtained as a result of a single cycle.According to X-ray diffraction data,28, 84 such successive freezing lusmn; thawing increases the content of the crystallites and their dimensions in the cryoP- VAG. The repeated `passage' of the system through the region of temperatures corresponding to the most intense gelation both during thawing and freezing (starting from the second cycle for what are no longer solutions but cryogels) leads to the formation of new additional intermolecular contacts and to an increase in the extent of those existing beforehand.In particular, this is man- ifested by the appearance of numerous compacted regions in the structure of the cryoPVAG observed 84 even under the light microscope.Admittedly the authors 69 do not regard this as the main cause of the increase in the strength of the cryogels, attributing the main role to the intensification of the interactions (hydrogen bonding) between water molecules and the OH groups of the PVA chains with increase in the number of refreezing cycles.However, this conclusion appears controversial. It is more likely that the converse is true: as a result of the crystallisation of PVA, water is displaced from the microcrystallinity zones. It is appa- rently not permissible to disregard also the compaction of the supermolecular strands of the cryogel network by the purely mechanical mutual approach due to the action of the faces of the ice crystals growing in it.This results in an increase in the strength parameters of the cryoPVAG and an increase in their thermal stability (Tm increases).40, 105 Yet another regime in the cryogenic treatment, influencing the characteristics of the PVA cryogels, is the partial sublimation of ice from the frozen specimens. As already mentioned in the Introduction, this method 31 ± 39 makes it possible to increase the mechanical strength of the cryoPVAG, primarily as a result of the increase in the strength of the surface layers of the preparations,37 because the sublimation of water vapour takes place initially from the peripheral open sections of the frozen block and the crystal- lised solvent sublimes from increasingly deeper regions only as evacuation proceeds.It has been shown 35 ± 37 that an increase in the duration of such lyophilisation leads to a lowering of the compliance of cryogels in relation to compression, to a decrease in the tangent of the mechanical losses angle following the application of a variable-sign load, and to an increase in the tensile strength of the specimens. According to patent data 31 ± 34, 38, 39, 152 ± 162 the degree of hydration of the frozen PVA solutions subjected to such vacuum treatment can reach>90%, i.e.almost all the nonsolvate water may be removed. On subsequent immersion of the xerogel in an aqueous T /8C T0 Tf7T0 /8C 70.8 72.2 74.4 77.8 4 2 0 72 74 CH OH CH2 10 kHz Figure 12. 13C NMR spectra of a 10% solution of PVA (M=81 700, DD=98%) in D2O frozen at720 8C and then maintained for 2 h at the temperature of the recording of the spectrum.67 580 V I Lozinskymedium, it swells slowly, after which the osmotic characteristics of the preparations remain unchanged for a long time.37 Thus cryogenic action on the PVA±water system makes it possible to vary withinwide limits the properties of the noncovalent cryogels of this polymer and to modify their macroporous struc- ture.Here the dominant factor, determining the effectiveness of the cryotropic gelation of PVA, is the time during which the frozen preparation resides at `high' negative temperatures. V. Dependence of the properties of poly(vinyl alcohol) cryogels on the nature and composition of the initial solvent The previous sections dealt with the influence of the character- istics of the PVA and the freezing � thawing regimes on the cryotropic gel formation and the parameters of the cryogels obtained.Only aqueous solutions of the polymer without appre- ciable amounts of other components were then considered as the initial objects. However, there is no doubt that the nature of the solvent employed and the presence of various additives should exert a definite effect on the freezing processes of the initial PVA solutions, the volume and composition of the liquid microphases, and the range of optimum temperatures of gel formation (if this occurs), on the one hand, and the properties and texture of the materials produced, on the other.For example, 10%± 16% solutions of PVA (M=69 000, DD=99.0%) in DMSO (T0=18.4 8C) at room temperature remain liquid for a long time, whilst the freezing ± thawing of such solutions leads, as for the aqueous solutions, to the formation of cryogels.9 However, the morphologies of the cryoPVAG obtained from aqueous and dimethyl sulfoxide solutions differ significantly.107, 163 In particular a layer of an aqueous cryoPVAG 0.6 mm thick (M=74 800 and DD=99.5%, cp=20%, Tf= 720 8C, tf=10 h, thawing for 10 h at 5 8C) transmits about 40% of light at a wavelength of 570 nm, while the same layer of cryoPVAG formed from a solution inDMSOin which the organic solvent has been replaced by water transmits 76% of light, i.e.is more transparent.163 Incidentally, the formation of PVA cryogels from solutions inDMSOitself disproves the hypothesis of Nagura et al.,69 mentioned above, that water molecules are involved in the structure of the tight contacts between the polymer chains on formation of crystallites.The characteristic features of the cryo- tropic gel formation in the PVA±DMSO system are virtually the same as in the PVA± water system, although the dimensions of the molecules and the physicochemical properties of DMSO andH2O differ significantly.In the cryogel formed from DMSO, it is possible to replace fully the DMSO by water on prolonged washing. This could hardly be done if DMSO molecules play the role of stable bridges (via hydrogen bonds) between the polymer chains. The cryoprotecting properties of DMSO are well known.164, 165 The addition of 40%± 80% of DMSO to aqueous systems very greatly reduces the freezing points of the latter.If PVA solutions are prepared in water ±DMSO mixtures, they gelate on cooling long before the attainment of the freezing point,89, 95, 96, 166, 169 affording transparent thermoreversible gels, which after washing out the DMSO can be used for the prepara- tion of soft contact lenses.166 However, if the DMSO content in the initial system is either below 30 vol.% or above 90 vol.%, the corresponding PVA solutions freeze at740 8C107 and after their thawing out macroporous cryogels are obtained (here it is important to employ rapid freezing to ensure that there is sufficient time for the crystallisation of the system before the onset of gelation). The effects caused by the competition between the rates of freezing and formation of the `usual' gels have been clearly demonstrated in studies of the cryotropic gelation processes in water ± polyol ±PVA systems.170 ± 172 When polyols such as glyc- erol, ethylene glycol, and oligo(ethylene glycol) (OEG) are intro- duced into aqueous PVA solutions, they lower the critical polymer gelation concentration and increase the temperature of the sol ± - gel transition as a consequence of the impairment of the thermo- dynamic quality of the solvent.48 ± 52 However, the gelation process itself in the presence of a not unduly high content of such polyols is fairly slow and, in order to obtain the cryogels, it is therefore necessary to `accelerate it' kinetically with the aid of rapid freezing.It has been found that the rate of freezing affects the compliance of the cryoPVAG prepared from aqueous solu- tions of the polymer to which oligomeric ethylene glycols have been added: the slower the freezing of the specimen, the weaker, other conditions being equal, the cryogel obtained.172 The main cause of the decrease in the strength of such specimens is that on slow cooling the gelation of the system begins before its freezing, as a consequence of which ice crystallises in the bulk of the gel formed, disturbing its structure. On the other hand, on rapid freezing the usual gel has insufficient time to arise and for this reason the mechanisms of the cryotropic gelatiooccurring in a medium significantly more concentrated in terms of the polymer than the initial solution `work' subsequently in the frozen hetero- phase system. The nature of the influence of the addition of ethylene glycol (EG) and its oligomers on the rheological and thermophysical properties of the corresponding cryoPVAG is illustrated in Fig. 13. An increase in the content of EG in the initial polymer solution leads to less hard and readily fusible cryogels (glycerol and propylene glycol act similarly), but, starting with triethylene glycol, an increase in the concentration of the cosolvent promotes the strengthening of the preparations and the growth of their thermal stability (incidentally,the addition of sucrose has the same effect on the properties of cryoPVAG 173). Diethylene glycol occupies an intermediate position: for a content of the additive up to *5%, weak cryogels are obtained, the cryogel being strengthened with increase in the concentration of diethylene glycol.It has been supposed that the strengthening effect of the OEG is caused by the liquid ± liquid phase separation in the 65 70 75 80 1 2 4 Tm /8C b 10 1 0.1 log J0 a 1 2 3 4 5 0 5 10 15 20 25 Polyol concentration (%) Figure 13. Rheological (a) and thermophysical (b) characteristics of cryogels formed at 720 8C (ts=24 h, uth=0.2 K min71) from 10% solutions of PVA (M=81 700, DD = 98%) in the presence of added polyols of the ethylene glycol series: (1) ethylene glycol; (2) diethylene glycol; (3) triethylene glycol; (4) PEG-400; (5) PEG-1000.172 Cryotropic gelation of poly(vinyl alcohol) solutions 581unfrozen inclusions, whereas in the crystallisation of ice the system falls (on the phase diagram) into the region of thermody- namic incompatibility (liquid two-phase water ±PVA±OEG sys- tems are well known 174) owing to the increase in the concentration of PVA and of the low-molecular polyol.As a result of this the gel forming agent is additionally concentrated in its `own' (i.e. PVA- enriched) liquid microphase. On the other hand, an increase in the concentration of the polymer usually leads to the strengthening of the cryogel and to an increase in its melting point, which has been observed experimentally.The influence of electrolytes on the cryotropic gelation of the polymer has been investigated.54, 175 ± 177 It was found that, in terms of the nature of their effect on the properties of the cryoPVAG formed from water ± salt (MX) ±PVA systems (Fig. 14), cations and anions can be arranged in a sequence corresponding to the positions of the corresponding ions in Hofmeister's lyotropic series, i.e. chaotropic ions (Li+, Br7, SCN7, etc.), which destabilise hydrogen bonding, prevent to some extent the formation of strong cryoPVAG, whilst antichao- tropic ions (Cs+, F7, OAc7, etc.) promote the strengthening of the cryogels.Neutral electrolytes of the type of NaCl, which have a weak influence on hydrogen bonds, exhibit a small strengthening effect with increase in concentration as a consequence of phenom- ena involving the salting-out of the polymer. Thus the composition of the solvent, i.e. its cryoscopic proper- ties, the affinity for PVA, the effect of the medium components on hydrogen bonding, and also the salting-out and salting-in effects influence the cryotropic gelation of the polymer and hence the parameters of the cryogels obtained.VI. Conclusion By virtue of their high porosity and at the same time effective mechanical strength as well as stability in any biological media, the materials based on cryoPVAG have found extensive applica- tions in various branches of biotechnology.Poly(vinyl alcohol) cryogels proved to be excellent carriers for the immobilisation of biological macromolecules 159, 178 ± 187 and cells 4, 41, 188 ± 265 and not only of microorganisms, which can stand quite well the cryogenic treatment, but also of certain animal cells.266 ± 268 Since the cryoPVAG are stable up to temperatures of 70 ± 80 8C, the corresponding carriers have been used for the immobilisation of both mesophilic cells and thermophilic microorgan- isms,210, 232, 237, 242, 245, 251 which functioned for more than a year at 60 ± 65 8C without loss of activity.242, 251 This is a record- breaking feature for biocatalytic systems entrapped in thermor- eversible gel matrices.It has also been proposed that polyol cryoprotectors can be introduced for the simultaneous protection of cultures sensitive to freezing and in order to increase the strength characteristics of the gel carrier.170, 172, 252 Solid nutritional media have been developed on the basis of PVA cryogels for the culturing of microbial 269 and plant cells, in particular the ginseng biomass.270 The use of such media makes it possible to replace the valuable nutrient gel former agar-agar by a synthetic polymer produced on a large scale.CryoPVAG coatings have found applications in the food industry for frozen fish and meat.271 ± 275 Chromatographic materials have been created from microgranulated cryoPVAG (0.2 ± 0.3 mm),276 while artificial bait for sporting and amateur fishing has been made from larger (2 ± 3 mm) coloured granules or gel in the form of small fish-like objects.13 ± 17, 277 An interesting aspect of the employment of mixed cryogels (PVA+polyelectrolyte-type polymer) is the creation of mechanochemical actuators elements simulating the work of muscle (the studies in this field are being particularly vigorously prosecuted in Japan 127, 132, 278 ± 292).There have been scientific publications and a multiplicity of patents describing a wide variety of versions of polymer systems, employed in medicine, based on cryoPVAG(Refs 28, 29, 31 ± 34, 90, 92, 93, 110, 120, 129, 131, 135, 137 ± 143, 146 ± 149, 154 ± 161, 266, 268, 293 ± 310) containing both PVA and its mixtures with other high-molecular-mass compounds and constituting compositeswith disperse fillers�inorganic (of the type of titanium dioxide) or organic (for example collagen fibres,303, 310) etc.These gels have been used successfully as sub- stitutes for the cartilage tissue,11 as protheses for surgery within the mouth,11, 28, 29, 34, 148 as materials for ophthalmology 139 and den- tistry,147 as protective coatings for wounds and burns,31, 135, 154 as cooling dressing, and cold helmets which do not secrete water for the prevention of hair loss in cancer chemother- apy,31, 137, 138, 155 ± 158, 293 and as special purpose catheters.141 The application of cryoPVAG as the gel medium for the tuning and calibration of NMR tomographs 120, 140, 143 and as systems for the controlled release of medicinal substances has been described.12, 110, 129, 131, 142, 149, 161, 294, 297, 298, 302, 307, 309 Materials for cosmetics,311, 312 etc.are being developed. Macroporous PVA membranes formed by cryogenic structure formation procedures may apparently find application.123, 124, 153 The cryostructure formation process in PVA solutions has been proposed also for the preparation of artificial ice and snow 59, 313 and for the fixation of soils which have thawed out.314 There is no doubt that the successful application of the above materials in medicine and other branches of science and technol- ogy 145, 315 ± 333 must be based on the fundamental knowledge of the mechanisms of the cryotropic gelation of PVA and of the influence of the polymer characteristics and also of the freezing ± thawing parameters on the properties and structures of the cryogels formed under these conditions.All these factors promote their further comprehensive investigation. Studies on cryoPVAG are concerned with the mechanisms of gel formation,334 ± 336 the macroscopic properties of cryoPVAG and materials based on them,337 ± 343 and the structures of gels formed under different conditions,344 ± 347 on the one hand, and fine effects associated with the different states of the solvent molecules in such hetero- phase polymer systems,348, 349 on the other.This is in fact the reason why attention has been concentrated the review on these problems. The interested reader may become acquainted in detail with the practical applications of cryoPVAG in the literature quoted. References 1.V I Lozinsky, E S Vainerman, S V Rogozhin Colloid Polym. Sci., 259 1198 (1981) 2. S V Rogozhin, E S Vainerman, V I Lozinsky Dokl. Akad. Nauk SSSR 263 115 (1982) a 3. USSR P. 1 008 214; Byull. Izobret. (12) 131 (1983) 4. V I Lozinsky, Doctoral Thesis in Chemical Sciences, A N Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences, Moscow, 1994 1074 J0 /Pa71 0 0.1 0.2 0.3 0.4 cMX /mol litre71 6 4 2 1 2 3 4 5678 Figure 14.Rheological properties of cryogels formed at 720 8C (ts=24 h, uth=0.2 K min71) from 14% solutions of PVA (M=69 000, DD=98%) in the presence of low-molecular electrolytes MX at different concentrations 177 (1) NaSCN; (2) NaBr; (3) LiCl; (4) NaCl; (5) KCl; (6) NaOAc; (7) CsCl; (8) NaF. 582 V I Lozinsky5. V I Lozinsky, E S Vainerman, S V Rogozhin, in 3-ya Vsesoyuzn.Konf. po Khimii Nizkikh Temperatur (Tez. Dokl.), Moskva, 1985 [The Third All-Union Conference on the Chemistry of Low Temperatures (Abstracts of Reports), Moscow, 1985] p. 18 6. V I Lozinsky, E S Vainerman, S V Rogozhin, in The 31st IUPAC Symposium on Macromolecules `MACRO'87' (Abstracts of Reports), Merseburg, 1987 Vol. 4, p. 171 7. V I Lozinsky, E S Vainerman, S V Rogozhin, in The International Conference `Networks'91' (Abstracts of Reports), Moscow, 1991 p. 50 8. V I Lozinsky,M N Korneeva, E S Vainerman, S V Rogozhin Dokl. Akad. Nauk SSSR 270 101 (1983) a 9. S V Rogozhin, V I Lozinsky, E S Vainerman, L V Domotenko, A M Mamtsis, S A Ivanova,M I Shtil'man, V V Korshak Dokl. Akad. Nauk SSSR 278 129 (1984) a 10.M Rikhter, Z Augustat, F Shirbaum, in Izbrannye Metody Issledo- vaniya Krakhmala (Selected Methods for the Investigation of Starch) (Moscow: Izd. Pishch. Promst, 1975) p. 14 11. M Nambu Kobunshi Ronbunshu 47 695 (1990) 12. N A Peppas, Sh R Stauffer J. Contr. Release 16 305 (1991) 13. Jpn. P. 47-012 854; Chem. Abstr. 77 153 121 (1972) 14. BRD P. 2 146 129; Chem. Abstr. 77 62 982 (1972) 15.Jpn. P. 48-030 462; Chem. Abstr. 81 38 398 (1974) 16. Jpn. P. 48-030 463; Chem. Abstr. 81 38 397 (1974) 17. US P. 3 875 302; Chem. Abstr. 77 62 982 (1972) 18. M M Kukharchik, N K Baramboim Vysokomol. Soedin., Ser. B 14 843 (1972) b 19. A Labudzin'ska Polimery 12 411 (1967) 20. A Labudzin'ska, A Ziabicki Kolloid-Z. Z. Polym. 243 21 (1971) 21. N A Peppas Makromol. Chem. 176 3433 (1975) 22. Jpn. P. 57-130 543; Chem. Abstr. 98 108 337 (1983) 23. M Watase, K Nishinari, M Nambu Cryo-Letters 4 197 (1983) 24. M Watase Nippon Kagaku Kaisi 973 (1983) 25. M Watase Nippon Kagaku Kaisi 1254 (1983) 26. Eur. P. 107 055; Chem. Abstr. 101 78 891 (1984) 27. Jpn. P. 61-021 178; Chem. Abstr. 104 171 578 (1986) 28. M Nagura, M Nagura, H Ishikawa Polym. Commun. 25 313 (1984) 29.Jpn. P. 60-199 453; Chem Abstr., 104 56475 (1986) 30. M Watase, K Nishinari Makromol. Chem., 186 1081 (1985) 31. M Watase,K Nishinari J. Polym. Sci., Polym. Phys. Ed. 23 1803 (1985) 32. Jpn. P. 57-190 072; Chem. Abstr. 98 129 328 (1983) 33. Eur. P. 0 095 892; Chem. Abstr. 100 74 018 (1984) 34. Jpn. P. 58-036 630; Chem. Abstr. 99 58 899 (1983) 35. Jpn. P. 58-215 424; Chem.Abstr. 100 109 154 (1984) 36. M Watase, K Nishinari, M Nambu Polym. Commun. 24 52 (1983) 37. M Watase, K Nishinari Polym. Commun. 24 270 (1983) 38. K Nishinari, M Watase, K Ogino,M Nambu Polym. Commun. 24 345 (1983) 39. Eur. P. 058 497; Chem. Abstr. 98 108 337 (1983) 40. Eur. P. 070 986; Chem. Abstr. 98 129 328 (1983) 41. V I Lozinsky, A V Vakula, A L Zubov Biotekhnologiya (4) 5 (1992) 42.S N Ushakov Polivinilovyi Spirt i ego Proizvodnye [Poly(vinyl alco- hol) and Its Derivatives] (Moscow, Leningrad: Izd. Akad. Nauk SSSR, 1960) Vols 1, 2 43. J C Pritchard Poly(vinyl alcohol). Basic Properties and Uses (London: Gordon and Breach, 1970) 44. R Maeda, T Kawai, R Kasiwagi Kobunshi Kagaku 13 193 (1956) 45. M Matsumoto, J Ohyanagi J. Polym. Sci. 26 389 (1957) 46. S P Papkov,A K Dibrova Vysokomol.Soedin., Ser. A 25 630 (1983) b 47. L Z Rogovina, G L Slonimskii, L S Gembitskii, E A Serova, V A Grigor'eva, E N Gubenkova Vysokomol. Soedin. 15 1256 (1973) b 48. K Arakawa Bull. Chem. Soc. Jpn. 31 842 (1958) 49. S A Glikman, L A Roth, L J Khomutov, E N Gubenkova, J J Feigelson J. Polym. Sci., Part C., Polym. Symp. Ed. 16 2001 (1967) 50. E Pines, W Prins Macromolecules 6 888 (1973) 51.E S Obolonkova, L Z Rogovina, N A Dmitrieva, E M Belavtseva, G L Slonimskii Kolloidn. Zh. 36 284 (1974) c 52. P Molynex Water-Soluble Synthetic Polymers. Properties and Behaviour Vol. 2 (Boca Raton, FL: CRC Press, 1984) p. 119 53. N S Pushkar', A M Belous, Yu A Itkin, V I Vishnevskii, L F Rozanov Nizkotemperaturnaya Kristallizatsiya v Biologicheskikh Sistemakh (Low-Temperature Crystallisation in Biological Systems) (Kiev: Naukova Dumka, 1977) p. 78 54. L V Domotenko, Candidate Thesis in Chemical Sciences, A N Nesmeyanov Institute of Organoelement Compounds, Academy of Sciences of the USSR, Moscow, 1986 55. L V Domotenko, V I Lozinsky, E S Vainerman, S V Rogozhin Vysokomol. Soedin., Ser. A 30 1661 (1988) b 56. F Franks, in Effects of Low Temperatures on Biological Membranes (Eds G J Morris, A Clarke) (New York: Academic Press, 1981) p. 3 57. B J Luyet, in Cryobiology (Ed.HT Meryman) (New York: Academic Press, 1966) p. 115 58. V I Lozinsky, E S Vainerman, L V Domotenko, A L Blyumenfel'd, V V Rogov, E N Barkovskaya, E I Fedin, S V Rogozhin Kolloidn. Zh. 51 685 (1989) c 59. USSR P. 1 649 219; Byull. Izobret.(18) 145 (1991) 60. O I Mikhalev,M Serpinski, V I Lozinsky, P V Kapanin, I I Chkheidze,M V Alfimov Cryo-Letters 12 197 (1991) 61. O I Mikhalev,M Serpinski, V I Lozinsky, P V Kapanin, I I Chkheidze,M V Alfimov, in 5-ya Vsesoyuzn. Konf. po Khimii Nizkikh Temperatur (Tez. Dokl.), Moskva, 1991 [The Fifth All-Union Conference on the Chemistry of Low Temperatures (Abstracts of Reports), Moscow, 1991] p. 104 62.M Serpinski, Candidate Thesis in Physicomathematical Sciences, N N Semenov Institute of Chemical Physics, Russian Academy of Sciences, Moscow, 1992 63. O I Mikhalev, V I Lozinsky, in The 29th Annual Meeting of the Society of Cryobiololy (Abstracts of Reports), Ithaca, 1992 p. 6 64. G B Sergeev, V A Batyuk,M B Stepanov, B M Sergeev Dokl. Akad. Nauk SSSR 213 891 (1973) a 65.G B Sergeev, V A Batyuk Usp. Khim. 45 793 (1976) [Russ. Chem. Rev. 45 (1976)] 66. V A Batyuk, in Khimiya Nizkikh Temperatur i Kriokhimicheskaya Tekhnologiya (The Chemistry of Low Temperatures and Cryochem- ical Technology) (Moscow: Moscow State University, 1987) p. 63 67. D G Gusev, V I Lozinsky, E S Vainerman, V I Bakhmutov Magn. Reson. Chem. 28 651 (1990) 68.D G Gusev, V I Lozinsky, V I Bakhmutov, in 5-ya Vsesoyuzn. Konf. po Khimii Nizkikh Temperatur (Tez. Dokl.), Moskva, 1991 [The Fifth All-Union Conference on the Chemistry of Low Temperatures (Abstracts of Reports), Moscow, 1991] p. 53 69. M Nagura, T Hamato, H Ishikawa Polymer 30 762 (1989) 70. A A Berlin Dokl. Akad. Nauk SSSR 110 401 (1956) a 71. A A Berlin, E A Penskaya Dokl. Akad.Nauk SSSR 110 585 (1956) a 72. A A Berlin Usp. Khim. 27 94 (1958) 73. A A Berlin, E A Penskaya, G I Volkova J. Polym. Sci. 56 477 (1962) 74. H H G Jellinek, S Y Fok Makromol. Chem. 104 18 (1967) 75. G Blond Cryobiology 25 61 (1988) 76. M M Kukharchik, N K Baramboim Nauchn. Tr. Moskovsk. Tekhnol. Inst. Legk. Promsti (41) 66 (1977) 77. V E Gul', L I Bulatnikova, O N Belyatskaya, E V Ermakova, L B Bykova Vysokomol.Soedin., Ser. A 18 118 (1976) b 78. V E Gul', E V Ermakova, L I Bulatnikova, O N Belyatskaya, A M Vasserman, E A Penskaya, A M Solov'ev Vysokomol. Soed., Ser. B 20 43 (1978) b 79. E V Popova, L I Bulatnikova, O N Belyatskaya, V E Gul' Vysokomol. Soedin., Ser. A 24 517 (1982) b 80. L I Bulatnikova, E V Popova, O N Belyatskaya, V E Gul' Vysokomol.Soedin., Ser. B 25 614 (1983) b 81. V I Lozinsky, L V Domotenko, E S Vainerman, A M Mamtsis, S V Rogozhin Polym. Bull. 15 333 (1986) 82. V I Lozinsky, V K Kulakova,M V Budanov, E S Vainerman, S V Rogozhin, in 4-ya Vsesoyuzn. Konf. po Khimii Nizkikh Temper- atur (Tez. Dokl.), Moskva, 1988 [The Fourth All-Union Conference on the Chemistry of Low Temperatures (Abstracts of Reports), Moscow, 1988] p. 206 83. M Watase, K Nishinari Makromol. Chem. 190 155 (1989) 84. F Yokoyama, I Masada, K Shimamura, T Ikawa, K Monobe Colloid Polym. Sci. 264 595 (1986) 85. ENagai, SKuribayashi,MShiraki,MUkita J.Polym. Sci. 35 295 (1959) 86. K Fujii Macromol. Rev. 5 431 (1971) 87. K Ogasawa, T Nakajima, K Yamaura, S Matsuzawa Progr. Colloid Polym. Sci. 58 145 (1975) 88. K Yamaura, H Katoh, T Tanigami, S Matsuzawa J.Appl. Polym. Sci. 34 2347 (1987) 89. K Yamaura, K-i Karasawa, T Tanigami, S Matsuzawa J. Appl. Polym. Sci. 51 2041 (1994) Cryotropic gelation of poly(vinyl alcohol) solutions 58390. A Takamura, M Arai, F Ishii Yakugaku Zassi 107 233 (1987); Chem. Abstr. 106 219 473 (1987) 91. V I Lozinsky, E S Vainerman, L V Domotenko, A M Mamtsis, E F Titova, E M Belavtseva, S V Rogozhin Colloid Polym.Sci. 264 19 (1986) 92. K Morimoto, A Nagayasu, S Fukanoki, K Morisaka, S-H Hyon, Y Ikada Drug Develop. Ind. Pharm. 16 13 (1990) 93. A Takamura, F Ishii Yakugaku Zassi 111 45 (1991); Chem. Abstr. 114 214 333 (1991) 94. WStoks, H Berghmans J. Polym. Sci., Part B, Polym. Phys., 29 609 (1991) 95. M Ohkura, T Kanaya, K Kagi Polymer 33 3686 (1992) 96.M Ohkura, T Kanaya, K Kagi Polymer 33 5044 (1992) 97. T Kanaya,M Ohkura, K Kagi,M Furusaka,M Misawa Macromolecules 27 5609 (1994) 98. J E Eldridge, J D Ferry J. Phys. Chem. 58 992 (1954) 99. A Todd Nature (London) 191 567 (1961) 100. S P Papkov Studneobraznoe Sostoyanie Polimerov (The Gel-like State of Polymers) (Moscow: Khimiya, 1974) p. 191 101. Yu S Lipatov Kolloidnaya Khimiya Polimerov (The Colloid Chemistry of Polymers) (Kiev: Naukova Dumka, 1984) p. 282 102. L V Domotenko, A M Mamtsis, V I Lozinsky, E S Vainerman, S V Rogozhin, in Teoriya i Prikladnye Voprosy Kriobiologii (Tez. Dokl. 2-i Vsesoyuzn. Konf.), Khar'kov, 1984 [The Theory and Applied Problems of Cryobiology (Abstracts of Reports at the Second All-Union Conference), Khar'kov, 1984] Vol. 1, p. 29 103. L V Domotenko, V I Lozinsky, A M Mamtsis, E S Vainerman, S V Rogozhin, in Vysokomolekulyarnye Soedineniya (Tez. Dokl. 22-i Vsesoyuzn. Konf.), Alma-Ata, 1985 [High-Molecular Com- pounds (Abstracts of Reports at the 22nd All-Union Conference), Alma-Ata, 1985] p. 175 104. F Urushizaki, H Yamaguchi, H Mizumachi Yakugaku Zassi 106 491 (1986); Chem. Abstr. 105 98 338 (1986) 105.M Watase, K Nishinari Makromol. Chem. 189 871 (1988) 106. V I Lozinsky, L V Domotenko, E S Vainerman, S V Rogozhin Vysokomol. Soedin., Ser. A 31 1805 (1989)b 107. K Yamaura, M Itoh, T Tanigami, S Matsuzawa J. Appl. Polym. Sci. 37 2709 (1989) 108. X Li, C Liu, W Chen, Z Yuan Gaofenzi Xuebao (5) 519 (1989); Chem. Abstr. 114 83 629 (1991) 109. S-H Hyon, W-I Cha, Y Ikada Kobunshi Ronbunshu 46 673 (1989) 110.Y Wang, C Zheng Wuli Huaxue Xuebao 6 509 (1990); Chem. Abstr. 113 153 628 (1990) 111. K Yamaura, T Tanigami, N Hayashi, N Kosuda, S Okuda, YTakemura,MItoh, SMatsuzawa J. Appl. Polym. Sci. 40 905 (1990) 112. G Huang,Y Feng, J He, Z Lu Gaodeng Xuexiao Huaxue Xuebao 12 564 (1991); Chem. Abstr. 115 258 064 (1991) 113. G Huang, Y Feng, J He Gaofenzi Xuebao (3) 316 (1992); Chem. Abstr. 120 135 745 (1994) 114. T Hatakeyama, A Yamaguchi, H Hatakeyama Eur. Polym. J. 23 361 (1987) 115. L V Domotenko, V I Lozinsky, A M Mamtsis, E S Vainerman, S V Rogozhin, in Protsessy Studneobrazovaniya v Polimernykh Sistemakh (Tez. Dokl. Vsesoyuzn. Simp.), Saratov, 1985 [Gel- Formation Processes in Polymer Systems (Abstracts of Reports at the All- Union Symposium), Saratov, 1985] Vol. 1, p. 123 116. M Watase, K Nishinari, in The 31st IUPAC Symposium on Macromolecules `MACRO'87' (Abstracts of Reports), Merseburg, 1987 Vol. 4, p. 136 117. A Packter,M S Nerukar Eur. Polym. J. 4 685 (1968) 118. V I Lozinsky, E F Titova, L V Domotenko, E S Vainerman, E M Belavtseva, S V Rogozhin, in 4-ya Vsesoyuzn. Konf. po Khimii Nizkikh Temperatur (Tez. Dokl.), Moskva, 1988 [The Fourth All- Union Conference on the Chemistry of Low Temperatures (Abstracts of Reports), Moscow, 1988] p. 155 119. M Watase Nippon Shokuhin Kogyo Gakkaishi 31 38 (1984); Chem. Abstr. 100 137 666 (1984) 120. I Mano, H Goshima, MNambu, MIio Magn. Reson. Med. 3 921 (1986) 121. Sh Ogawa, H Sugii, H Seki, K Iwamoto, M Seno Sen'i Gakkaishi 45 382 (1989); Chem. Abstr. 112 21 606 (1990) 122.T Hirai, Y Asada, T Suzuki, S Hayashi, M Nambu J. Appl. Polym. Sci. 38 491 (1989) 123. T Hirai, K Hanaoka, T Suzuki, S Hayashi Kobunshi Ronbunshu 46 613 (1989) 124. M Tasaki, T Kida, T Homma Kobunshi Ronbunshu 46 733 (1989) 125. F Urushizaki, H Yamaguchi, K Nakamura, S Namajiri, K Sugibayashi, Y Morimoto Int. J. Pharm. 58 135 (1990) 126. T Hirai Kobunshi 40 524 (1991); Chem.Abstr. 117 112 327 (1992) 127. M Nagura, Y Arai, Y Ohkoshi Sen'i Gakkaishi 48 127 (1992); Chem. Abstr. 117 91 268 (1992) 128. T Hirai,H Muruyama, T Suzuki, S Hayashi J. Appl. Polym. Sci. 45 1449 (1992) 129. N A Peppas, J E Scott J. Contr. Release 18 95 (1992) 130. Sh R Stauffer, N A Peppas Polymer 33 3932 (1992) 131. B J Ficek, N A Peppas J. Contr. Release 27 259 (1993) 132.O Kolosov,M Suzuki, K Yamanaka J. Appl. Phys. 74 6707 (1993) 133. O Ariga, T Kubo, Y Sano J. Ferment. Bioeng. 78 200 (1994) 134. M Hirai, T Hirai, T Ueki Macromolecules 27 1003 (1994) 135. Jpn. P. 57-087 979; Chem. Abstr. 98 182 658 (1983) 136. Jpn. P. 06-032 960; Chem. Abstr. 120 325 000 (1994) 137. Jpn. P. 59-115 378; Chem. Abstr. 101 157 719 (1984) 138. Jpn. P. 59-115 379; Chem.Abstr. 101 137 072 (1984) 139. Eur. P. 194 598; Chem. Abstr. 106 134 741 (1987) 140. BRD P. 3 614 142; Chem. Abstr. 106 55 979 (1987) 141. Jpn. P. 62-254 773; Chem. Abstr. 109 125 450 (1988) 142. Jpn. P. 62-205 035; Chem. Abstr. 108 62 437 (1988) 143. Eur. P. 228 692; Chem. Abstr. 107 183 607 (1987) 144. Jpn. P. 63-046 240; Chem. Abstr. 108 222 866 (1988) 145. Jpn. P. 63-111 040; Chem.Abstr. 109 130 500 (1988) 146. Jpn. P. 63-115 561; Chem. Abstr. 110 63 794 (1989) 147. Jpn. P. 63-119 748; Chem. Abstr. 110 101880 (1989) 148. Jpn. P. 63-249 567; Chem. Abstr. 111 160296 (1989) 149. Jpn. P. 03-041 020; Chem. Abstr. 114 235 088 (1991) 150. Jpn. P. 06-145 363; Chem. Abstr. 121 110 496 (1994) 151. Jpn. P. 58-047 492; Chem. Abstr. 99 4100 (1983) 152. Jpn. P. 45-013 709; Chem. Abstr. 73 88692 (1970) 153. Jpn. P. 52-008 071; Chem. Abstr. 86 177 348 (1977) 154. Jpn. P. 56-045 439; Chem. Abstr. 95 133 192 (1981) 155. Jpn. P. 56-073 938; Patent Abstr. Jpn., Sec. C 7 44 (1981) 156. Jpn. P. 57-139 705; Patent Abstr. Jpn., Sec. C 8 74 (1982) 157. Jpn. P. 58-195 558; Chem. Abstr. 100 73 792 (1984) 158. Jpn. P. 59-025 873; Chem. Abstr. 101 43 640 (1984) 159.Jpn. P. 60-180 587; Chem. Abstr. 104 33 075 (1986) 160. Jpn. P. 63-097 170; Chem. Abstr. 109 237 090 (1988) 161. Jpn. P. 02-088 519; Chem. Abstr. 113 848 72 (1990) 162. Jpn. P. 05-125 200; Chem. Abstr. 119 121 567 (1993) 163. S-H Hyon, W-I Cha, Y Ikada Polym. Bull. 22 119 (1989) 164. H T Meryman Cryobiology 8 173 (1971) 165. N S Pushkar', M I Shrago, A M Belous, Yu V Kalugin Krioprotektory (Cryoprotectors) (Kiev: Naukova Dumka, 1978) 166.US P. 4 663 358; Chem. Abstr. 106 157 622 (1987) 167. M Watase, K Nishinari Polym. J. 21 567 (1989) 168. M Watase, K Nishinari Polym. J. 21 597 (1989) 169. T Shiga, K Fukumori, Y Hirose, A Okada, T Kurauchi J. Polym. Sci., Part B, Polym. Phys. 32 85 (1994) 170. Russ. P. 2 070 901; Byull. Izobret. (36) 107 (1996) 171.V I Lozinsky, E V Solodova, A L Zubov, L N Kurkovskaya, in International Conference on Low Temperature Chemistry (Abstracts of Reports), Moscow, 1994 p. 87 172. V I Lozinsky, E V Solodova, A L Zubov, I A Simenel J. Appl. Polym. Sci. 58 171 (1995) 173. K Yamaura, M Mitsuishi,M Fukuda, T Tanigami, Y Hoei, Sh Matsuzawa J. Appl. Polym. Sci. 56 653 (1995) 174. I Inamura, K Toki, T Tomae, T Araki Polym.J. 16 658 (1984) 175. V I Lozinsky, L V Domotenko, E S Vainerman, S V Rogozhin Vliyanie Dobavok Rastvorimykh Veshchestv na Reologicheskie Svoistva Kriogelei Polivinilovogo Spirta [Influence of Admixtures of Soluble Compounds on the Rheological Properties of Poly(vinyl alcohol) Cryogels] Article deposited at the VINITI, No. 6178-V88, Moscow 1988 176. V I Lozinsky, A L Zubov, E F Titova Enzyme Microbiol.Technol. 18 561 (1996) 177. V I Lozinsky, L V Domotenko, A L Zubov, I A Simenel J. Appl. Polym. Sci. 61 1991 (1996) 178. O Ariga, T Yamakawa, H Fujimatsu, Y Sano J. Ferment. Bioeng. 68 293 (1989) 584 V I Lozinsky179. Jpn. P. 02-167 077; Chem. Abstr. 114 38 446 (1991) 180. W-I Cha, S-H Hyon, Y I Ikada Kobunshi Ronbunshu 48 425 (1991) 181. E Kokufuta, E Jinbo Macromolecules 25 3549 (1992) 182.TYamauchi,EKokufuta,YOsadaPolym.Gels Networks 1 247 (1993) 183. T Hayashi, S-H Hyon, W-I Cha, Y I Ikada J. Appl. Polym. Sci., 49 2121 (1993) 184. Jpn. P. 05-017 656; Chem. Abstr. 119 119 048 (1993) 185. S M Gautier, P E Phillippe, L J Blum Anal. Lett. 27 2055 (1994) 186. VILozinsky,FMPlieva,A L Zubov Biotekhnologiya (1 ± 2) 32 (1995) 187.F M Plieva, Candidate Thesis in Chemical Sciences, N A Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences, Moscow, 1997 188. Eur. P. 060 052; Chem. Abstr. 98 87 656 (1983) 189. V I Lozinsky, E S Vainerman, S V Rogozhin, in Inzhenernaya Enzimologiya (Tez. Dokl. 4-i Vsesoyuzn. Konf.), Kiev, 1983 [Engi- neering Enzymology (Abstracts of Reports at the Fourth All-Union Conference), Kiev, 1983] Vol. 2, p. 54 190. V I Lozinsky, E S Vainerman, S V Rogozhin, L V Gavrilenko, I N Semenova, S N Egorov, in Biosintez Fermentov Mikroorganiz- mami (Tez. Dokl. 3-i Vsesoyuzn. Konf.), Kobuleti, 1986 [Biosyn- thesis of Enzymes by Microorganisms (Abstracts of Reports at the Third All-Union Conference), Kobuleti, 1986] p. 20 191. Jpn. P. 60-183 095; Chem.Abstr. 104 10 107 (1986) 192. S Hashimoto, K Furukawa, H Hama J. Jpn. Sewage Water Assoc. 23 16 (1986); Chem. Abstr. 105 102 036 (1986) 193. Jpn. P. 61-139 385; Chem. Abstr. 106 182 100 (1987) 194. USSR P. 1 351 974; Byull. Izobret. (42) 89 (1987) 195. USSR P. 1 400 071; Byull. Izobret. (20) 111 (1988) 196. USSR P. 1 405 310; Byull. Izobret. (23) 109 (1988) 197. USSR P. 1 505 972; Byull. Izobret. (33) 119 (1989) 198. USSR P. 1 559 716; Byull. Izobret. (15) 148 (1989) 199. USSR P. 1 559 719; Byull. Izobret. (15) 149 (1989) 200. USSR P. 1 589 627; Byull. Izobret. (36) 114 (1990) 201. Russ. P. 2 036 095; Byull. Izobret. (15) 31 (1995) 202. R Zh Manolov, V I Lozinsky, I M Tavobilov, E S Vainerman, S I Bezborodova, S V Rogozhin, F M Bezborodov Biotekhnol.Biotekhn. 2 3 (1987) 203. O Ariga, H Takagi, H Nishizawa, Y Sano J. Ferment. Technol. 65 651 (1987) 204. Jpn. P. 62-186 995; Chem. Abstr. 107 204 598 (1987) 205. Jpn. P. 62-146 589; Chem. Abstr. 107 152 935 (1987) 206. R Zh Manolov, I M Tavobilov, V I Lozinsky, E S Vainerman, E F Titova, E M Belavtseva, S I Bezborodova, S V Rogozhin, A M Bezborodov Prikl. Biokhim. Mikrobiol. 24 514 (1988) 207. Jpn. P. 63-226 286; Chem. Abstr. 111 149 519 (1989) 208. L V Belousova, V I Lozinsky, E S Vainerman, S V Rogozhin, S N Egorov, N S Egorov, in Fermenty Mikroorganizmov (The Enzymes of Microorganisms) (Ed. V G Debabov) (Moscow: VNIISENTI, 1989) Vol. 2, p. 224 209. V I Lozinsky, N G Faleev, A L Zubov, S B Ruvinov, T V Antonova, E S Vainerman, V M Belikov, S V Rogozhin Biotechnol.Lett. 11 43 (1989) 210. S D Varfolomeev, E I Rainina, V I Lozinsky, S B Kalyuzhnyi, A P Sinitsyn, T A Makhlis, G P Bachurina, I G Bokova, O A Sklyankina, E V Agafonov, in Physiology of Immobilised Cells (Eds J AM de Bont, J Visser, B Mattiasson, J Tramper) (Amsterdam: Elsevier, 1989) p. 325 211. Jpn. P. 01-281 195; Chem. Abstr. 112 145 065 (1990) 212. Jpn. P. 01-228 471; Chem.Abstr. 112 194 476 (1990) 213. Eur. P. 3 031 222; Chem. Abstr. 111 170 085 (1989) 214. V I Lozinsky, A L Zubov, V K Kulakova, E S Vainerman, S V Rogozhin Biotekhnologiya (5) 32 (1990) 215. G P Alebyan, E N Arzumanov,MV Mkrtchyan, P V Tozalakyan, V I Lozinsky, E S Vainerman, S V Rogozhin Biotekhnologiya (6) 29 (1990) 216. M N Filimonova, G K Gabdulina, I F Zabbarova Prikl. Biokhim.Mikrobiol. 29 706 (1990) 217. A L Zubov, V I Lozinsky, E S Vainerman, S V Rogozhin, in Fermenty Mikroorganizmov i Degradatsiya Biopolimerov (The Enzymes of Microorganisms and Degradation of Biopolymers) (Ed. V G Debabov) (Moscow: VNIISENTI, 1990) p. 111 218. Jpn. P. 02-211 872; Chem. Abstr. 114 20 177 (1991) 219. V I Lozinsky, in Inzhenernaya Enzimologiya (Tez. Dokl. 7-go Vsesoyuzn.Simp.), Moskva, 1991 [Engineering Enzymology (Abstracts of Reports at the Seventh All-Union Symposium), Moscow, 1991] p. 33 220. A P Sinitsyn, E I Rainina, V I Lozinsky, S D Spasov Immobilizo- vannye Kletki Mikroorganizmov (Immobilised Cells of Microor- ganisms) (Sofia: Sofia University, 1991) p. 154 221. H Myoga, H Asano, Y Nomura, H Yoshida Water Sci. Technol., 23 1117 (1991) 222.Jpn. P. 03-072996; Chem. Abstr. 115 56 913 (1991) 223. Jpn. P. 03-146128; Chem. Abstr. 115 153 916 (1991) 224. BRD P. 4 027 218; Chem. Abstr. 115 227 115 (1991) 225. P Remmers, K-D Vorlop DECHEMA Biotechnol. 5 939 (1992) 226. Eur. P. 472 249; Chem. Abstr. 117 72 475 (1992) 227. Eur. P. 472 250; Chem. Abstr. 117 55 600 (1992) 228. BRD P. 4 027 219; Chem. Abstr. 117 55 604 (1992) 229.BRD P. 4 027 221; Chem. Abstr. 117 55 602 (1992) 230. BRD P. 4 027 223; Chem. Abstr. 117 55 603 (1992) 231. Jpn. P. 04-335 892; Chem. Abstr. 118 232 391 (1992) 232. S D Varfolomeev, E I Rainina, V I Lozinsky Pure Appl. Chem. 64 1193 (1992) 233. S G Velizarov, E I Rainina, A P Sinitsyn, S D Varfolomeev, V I Lozinsky, A L Zubov Biotechnol. Lett. 14 291 (1992) 234. H Asano, H Myoga, M Asano,M Toyao Water Sci.Technol. 26 1037 (1992) 235. H Asano, H Myoga, M Asano,M Toyao Water Sci. Technol. 26 2397 (1992) 236. A L Simonian, E I Rainina, V I Lozinsky, I E Badalian, G E Kha-chatrian, S Sh Tatikian, T A Makhlis, S D Varfolomeev Appl. Biochem. Biotechnol, 36 199 (1992) 237. OANikitina, S S Zatsepin, SVKalyuzhnyi, EIRainina,A L Zubov, S D Varfolomeev, V I Lozinsky Mikrobiologiya 62 477 (1993) 238.O Ariga, N Kato, T Sano, Y Nakazawa, Y Sano J. Ferment. Bio- eng. 76 203 (1993) 239. K Furukawa, A Ike, S L Ryu,M Fujita J. Ferment. Bioeng. 76 515 (1993) 240. C Hatanaka, T Haraguchi, S Ide Kitakyushu Kogyo Koto Senmon Gakko KenkyuHokoku 26 107 (1993); Chem. Abstr. 119 70 472 (1993) 241. V I Lozinsky, in The 35th IUPAC International Symposium on Macromolecules `MACRO'94' (Abstracts of Reports), Akron, 1994 p. 519 242. E I Rainina, MA Pusheva, AMRyabokon, N P Bolotina, V I Lozinsky, S D Varfolomeev Biotechnol. Appl. Biochem. 19 321 (1994) 243. Jpn. P. 06-099 186; Chem. Abstr. 121 65 232 (1994) 244. V V Fokina, A Yu Arinbasarova, A L Zubov, V I Lozinsky, K A Koshcheenko Prikl. Biokhim. Mikrobiol. 31 213 (1995) 245.A M Ryabokon',M A Pusheva, E N Detkova, E I Rainina Mikrobiologiya 64 776 (1995) 246. M Okazaki, T Hamada, H Fujii, A Mizobe, Sh Matsuzawa J. Appl. Polym. Sci. 58 2235 (1995) 247. M Okazaki, T Hamada, H Fujii, A Mizobe, Sh Matsuzawa J. Appl. Polym. Sci. 58 2243 (1995) 248. S Ledakowicz, U Becker, W-D Decker Catal. Today 24 65 (1995) 249. A Muscat, U Prusse, K-D Vorlop, in Immobilized Cells.Basics and Applications (Eds R H Wijffels, R Buitelaar, C Bucke, J Tramper) (Amsterdam: Elsevier, 1996) p. 55 250. V Fokina, N Susina, A Arinbasarova, A Zubov, V Lozinsky, K Koshcheenko, in Immobilized Cells. Basics and Applications (Eds R H Wijffels, R Buitelaar, C Bucke, J Tramper) (Amsterdam: Elsevier, 1996) p. 90 251. A M Ryabokon,M V Kevbrina, M A Pusheva, A L Zubov, V I Lozinsky, E I Rainina, in Immobilized Cells.Basics and Appli- cations (Eds R H Wijffels, R Buitelaar, C Bucke, J Tramper) (Amsterdam: Elsevier, 1996) p. 106 252. V I Lozinsky, A L Zubov, T A Makhlis, in Immobilized Cells. Basics and Applications (Eds R H Wijffels, R Buitelaar, C Bucke, J Tramper) (Amsterdam: Elsevier, 1996) p. 112 253. E J T M Leenen, V A P M Dos Santos, J Tramper, R H Wijffels, in Immobilized Cells.Basics and Applications (Eds R H Wijffels, RBuitelaar,CBucke, J Tramper) (Amsterdam: Elsevier, 1996) p. 205 254. S Bielecki, R Bolek, in Immobilized Cells. Basics and Applications (Eds R H Wijffels, R Buitelaar, C Bucke, J Tramper) (Amsterdam: Elsevier, 1996) p. 472 Cryotropic gelation of poly(vinyl alcohol) solutions 585255. W M Rostron, D C Stuckey, A A Young, in Immobilized Cells.Basics and Applications (Eds R H Wijffels, R Buitelaar, C Bucke, J Tramper) (Amsterdam: Elsevier, 1996) p. 703 256. Th Willke, K-D Vorlop, in Immobilized Cells. Basics and Applications (Eds R H Wijffels, R Buitelaar, C Bucke, J Tramper) (Amsterdam: Elsevier, 1996) p. 718 257. E I Rainina, I E Badalian,O V Ignatov,A V Fedorov,A L Simonian, S D Varfolomeev Appl.Biochem. Biotechnol. 56 117 (1996) 258. E I Rainina, E N Efremenko, S D Varfolomeev A L Simonian, J R Wild Biosens. Bioelectron. 11 991 (1996) 259. V I Lozinsky, F M Plieva, in Bioencapsulation'V (Eds H Dautzenberg, D Poncelet) (Potsdam: BRG, 1996) T3 260. U Prusse, F Bruske, J Breford, K-D Vorlop, in Bioencapsulation'V (Eds H Dautzenberg, D Poncelet) (Potsdam: BRG, 1996) T7 261.U Jahnz, T Willke, K-D Vorlop, in Bioencapsulation'V (Eds H Dautzenberg, D Poncelet) (Potsdam: BRG, 1996) T30 262. E Rainina, S Harvey, E Efremenko, I Badalian, J Wild, in Bioen- capsulation'V (Eds H Dautzenberg, D Poncelet) (Potsdam: BRG, 1996) T33 263. C Reimann, P Wittlich, T Willke, K-D Vorlop, in Bioencapsula- tion'V (Eds H Dautzenberg, D Poncelet) (Potsdam: BRG, 1996) P4 264.E Rainina, V Lozinsky, J Wild, in Bioencapsulation'V (Eds H Dautzenberg, D Poncelet) (Potsdam: BRG, 1996) P33 265. U Prusse, B Fox,M Kirchhoff, F Bruske, J Breford, K-D Vorlop Biotechnol. Techn. 12 105 (1998) 266. L Lelli, N Barbani, A Bonaretti, P Guisti, P Marchetti, R Giannarelli, R Navalesi, P Viacava J. Mater. Sci., Mater. Med. 5 887 (1994) 267.K Ito, N Minoura, S Aiba, M Nagura, M Kodama J. Biomed. Mater. Res. 28 459 (1994) 268. M G Cascone,M Laus, D Ricci, R S del Guerra J. Mater. Sci., Mater. Med. 6 71 (1995) 269. USSR P. 1 213 069; Byull. Izobret. (7) 146 (1986) 270. USSR P. 1 438 236; Byull. Izobret. (42) 105 (1988) 271. USSR P. 683 705; Byull. Izobret. (33) 15 (1979) 272. L P Zuikova, E I Izyumova, G G Korneva Rybn.Khoz. (1) 70 (1977) 273. V E Shapovalov, E V Yakush,M N Vakhrusheva, V F Mikhaleva, N A Gradov Rybn. Khoz. (11) 66 (1982) 274. USSR P. 971 209; Byull. Izobret. (41) 16 (1982) 275. L I Bulatnikova, EVPlisko,OI Belyatskaya Plast. Massy (7) 52 (1983) 276. R Murakami, H Hachisako, K Yamada, Y Motozato J. Chromatogr. 678 180 (1994) 277. Russ. P. 2 054 254; Byull. Izobret. (5) 8 (1996) 278.Jpn. P. 62-115 064; Chem. Abstr. 108 23 066 (1988) 279. US P. 4 717 800; Chem. Abstr. 108 151 520 (1988) 280. Jpn. P. 63-223 045; Chem. Abstr. 110 96 865 (1989) 281. M Suzuki Kobunshi Ronbunshu 46 603 (1989) 282. T Shiga, Y Hirose, A Okada, T Kurauchi Kobunshi Ronbunshu 46 709 (1989) 283. Jpn. P. 03-017 133; Chem. Abstr. 114 166 034 (1991) 284. T Shiga, Y Hirose, A Okada, T Kurauchi J.Appl. Polym. Sci. 44 249 (1992) 285. Jpn. P. 04-072 339; Chem. Abstr. 117 50 368 (1992) 286. MYonese,KBaba,MKondo,HKishimoto Chem. Lett. 199 (1992) 287. Jpn. P. 04-053 843; Chem. Abstr. 117 132 442 (1992) 288. T Shiga, Y Hirose, A Okada, T Kurauchi J. Intel. Mater., Syst. Struct. 4 553 (1993) 289. T Shiga, Y Hirose, A Okada, T Kurauchi J. Appl. Polym. Sci. 47 113 (1993) 290.M Suzuki, O Hirasa Adv. Polym. Sci. 110 241 (1993) 291. N Tsunemoto, M Suzuki Polym. Gels Networks 2 247 (1994) 292. Y Mori, H Tokura,M Yoshikawa Kagaku Kogaku Ronbunshu 20 625 (1994); Chem. Abstr. 121 232 602 (1994) 293. B J Ficek, N A Peppas, in Biomaterials for Drug and Cell Delivery (Proceedings of the International Symposium), London, 1994 p. 223 294. Jpn. P. 57-006 403; Chem. Abstr. 96 223 289 (1982) 295. H S Shin, S Y Kim, Y M Lee J. Appl. Polym. Sci. 65 685 (1997) 296. Y Honda, S Y Kim, M Nambu, N Naoi, E Sakaue Doc. Ophthalmol. 63 205 (1986) 297. Jpn. P. 61-218 517; Chem. Abstr. 106 90 182 (1987) 298. Jpn. P. 62-053 662; Chem. Abstr. 107 102 658 (1987) 299. Jpn. P. 62-230 718; Chem. Abstr. 108 137 898 (1988) 300. MKuriaki, K Nakamura, J Mizutani Kobunshi Ronbunshu 46 739 (1989) 301. Jpn. P. 01-149 717; Chem. Abstr. 112 84 206 (1990) 302. A Takamura, F Ishii, H Hidaka J. Contr. Release 20 21 (1992) 303. P Giusti, L Lazzeri,N Barbani, P Narducci, A Bonaretti,M Palla, L Lelli J. Mater. Sci., Mater. Med. 4 538 (1993) 304. C H Lee, K J Lee, A J Park, Y A Shin Arch. Pharm. Res. 16 43 (1993) 305. Jpn. P. 05-111 531; Chem. Abstr. 119 56 259 (1993) 306. R S del Guerra,M G Cascone, N Barbani, L Lazzeri J. Mater. Sci., Mater. Med. 5 613 (1994) 307. M G Cascone, L Di Silvio, B Sim, S Downes J. Mater. Sci., Mater. Med. 5 770 (1994) 308. L Lazzeri, N Barbani, M G Cascone, D Lupinocci, P Giusti, M Laus J. Mater. Sci., Mater. Med. 5 862 (1994) 309. M G Cascone, B Sim, S Downes Biomaterials 16 564 (1995) 310. L Lazzeri Trends Polym. Sci. 4 249 (1996) 311. Jpn. P. 61-236 711; Chem. Abstr. 106 55 665 (1987) 312. Jpn. P. 61-236 712; Chem. Abstr. 106 55 666 (1987) 313. Jpn. P. 02-233 793; Chem. Abstr. 114 49 311 (1991) 314. USSR P. 1 705 500; Byull. Izobret. (2) 131 (1992) 315. Jpn. P. 60-177 066; Chem. Abstr. 104 34 695 (1986) 316. Jpn. P. 61-233 038; Chem. Abstr. 106 177 963 (1987) 317. U Prusse, S Horold, K -D Vorlop Chem. Ing. Techn. 69 100 (1997) 318. US P. 4 575 519; Chem. Abstr. 104 188 588 (1986) 319. Jpn. P. 61-021 178; Chem. Abstr. 104 171 578 (1986) 320. Jpn. P. 62-045 637; Chem. Abstr. 107 8576 (1987) 321. Jpn. P. 01-270 867; Chem. Abstr. 112 180 964 (1990) 322. Jpn. P. 02-086 838; Chem. Abstr. 113 155 140 (1990) 323. Jpn. P. 02-112 407; Chem. Abstr. 113 99 313 (1990) 324. Jpn. P. 02-212 523; Chem. Abstr. 114 25 265 (1991) 325. Jpn. P. 02-232 291; Chem. Abstr. 114 45 625 (1991) 326. USSR P. 1 680 720; Byull. Izobret. (36) 110 (1991) 327. WO PCT 93-06 154; Chem. Abstr. 119 182 036 (1993) 328. C Cristallini, L Lazzeri, M G Cascone, G Polacco, D Lupinacci, N Barbani Polym. Int. 44 510 (1997) 329. Jpn. P. 06-220 228; Chem. Abstr. 122 83 204 (1995) 330. Jpn. P. 06-256 616; Chem. Abstr. 122 107 660 (1995) 331. Eur. P. 583 170; Chem. Abstr. 121 42 802 (1994) 332. J Byun, Y M Lee, Ch-S Chao J. Appl. Polym. Sci. 61 697 (1996) 333. Y M Lee, S H Kim, Ch-S Chao J. Appl. Polym. Sci. 62 301 (1996) 334. V I Lozinsky, in The 35th IUPAC International Symposium on Macromolecules `MACRO'94' (Abstracts of Reports), Akron, 1994 p. 520 335. J Rault, in Hydrogen Bond Networks (Eds M-C Bellissent-Funel, J C Dore) (Amsterdam: Kluwer Academic, 1994) p. 441 336. M Liu, R Cheng, R Qian J. Polym. Sci., Part B, Polym. Phys. 33 1731 (1995) 337. V I Lozinsky, A L Zubov, V K Kulakova, E S Vainerman, S V Rogozhin, in Sintez, Struktura i Svoistva Setchatykh Polimerov (Tez. Dokl. Vsesoyuzn. Konf.), Zvenigorod, 1988 [The Synthesis, Structure, and Properties of Network Polymers (Abstracts of Reports at the All-Union Conference), Zvenigorod, 1988] p. 160 338. S Ikoma, E Nomoto, H Yokoi Kobunshi Ronbunshu 47 1001 (1990) 339. T Takigawa, H Kasihara, T Masuda Polym. Bull. 24 613 (1990) 340. V I Lozinsky, A L Zubov, V K Kulakova, E F Titova, S V Rogozhin J. Appl. Polym. Sci. 44 1428 (1992) 341. M Yonese, K Baba, H Kishimoto Polym. J. 24 395 (1992) 342. A L Zubov, E A Titova, V I Lozinsky, in The 35th IUPAC Inter- national Symposium on Macromolecules `MACRO'94' (Abstracts of Reports), Akron, 1994 p. 521 343. K Ito, N Minoura,M Nagura Polymer 36 2579 (1995) 344. V I Lozinsky, A L Zubov, E A Titova Enzyme Microbiol. Technol. 20 182 (1997) 345. H H Trieu, S Qutubuddin Colloid Polym. Sci. 272 301 (1994) 346. H H Trieu, S Qutubuddin Polymer 36 2531 (1995) 347. M Kobayashi, I Anao, T Ishii, Sh Amiya Macromolecules 28 6677 (1995) 348. T Hatakeyama R. Progr. Polym. Phys. Jpn. 24 211 (1981) 349. Y Feng Sichuan Daxue Xuebao, Ziran Kexueban 26 170 (1989); Chem. Abstr. 113 98 578 (1990) a�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) b�High Polym. (Engl. Transl.) c�Colloid J. (Engl. Transl.) 586 V I Lo

ISSN:0036-021X    

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Abstract. The main types of heterogeneous catalysts used for hydrogenation, the methods for their preparation, and the struc- ture and chemistry of their surfaces are considered, as well as the catalytic activity and the mechanism of action in the hydro genation of unsaturated and aromatic compounds, of CO, and of carbonyl compounds and in the hydrorefining of fuels. Chief attention is paid to supported Ni catalysts, to the methods for their preparation and physicochemical studies, and to the devel- opment of novel catalytic systems through modification.A novel type of catalyst for hydrogenation, viz. metal carbides, is described. Some aspects of the mechanochemical treatment of hydrogenation catalysts, including in situ methods, are discussed. Sulfide catalysts for hydrotreating are also discussed in detail.The bibliography includes 340 references. I. Introduction Catalytic hydrogenation includes a large group of addition reactions of hydrogen at unsaturated bonds C=C, C=O, C:N, C:C etc., as well as hydrogenolysis (cleavage and reduction) of C7C, C7O, C7S, and other bonds. In addition to heterogeneous catalysts applied in large-scale processes, homogeneous metal-complex and hydride catalysts are also used (mostly, in fine organic synthesis).Enantioselective hydrogenation catalysts, basically metal-complex compounds of Ni, Co and other metals with optically active ligands, are also being increasingly used. In 1991, of 1040 catalysts for oil processing produced by world industries the number of hydrogenation catalysts was 380.1 However, modern technologies require more effective and less expensive hydrogenation catalysts.They can be designed only by development of novel scientific approaches to the preparation of supported catalysts and by the use of new physicochemical techniques. Supported nickel catalysts of hydrogenation are the most widely used ones; they are followed by catalysts based on noble metals.Binary Ni7Mo, Ni7W, Co7Mo and Co7W catalysts are extensively employed in the processes of hydrotreating and hydrodesulfurisation. Metal carbides appear to be new promising catalysts. This paper reviews certain problems related to the prepara- tion, composition, and mechanism of action of heterogeneous catalysts of hydrogenation of unsaturated compounds (including aromatics), hydrogenation of CO and carbonyl compounds, as well as to the processes of hydrotreating.Primary attention is given to supported Ni catalysts, which have found increasing use in recent years not only in large-scale processes but also in fine organic synthesis. Catalysts based on noble metals (Pt, Pd, Ru, Rh, Ir), supported sulfide catalysts of hydrotreating and some novel hydrogenation catalysts, such as metal carbides, which have been extensively studied in recent years, are also the subjects of discussion.Some examples are given of the use of mechanochem- istry techniques for preparing catalysts allowing the synthesis of environmentally safe motor fuels.2 MD Navalikhina Institute of High Temperatures, Russian Academy of Sciences, Izhorskaya ul. 13/19, 127412 Moscow, Russian Federation. Fax (7-095) 485 99 22 O V Krylov N N Semenov Institute of Chemical Physics, Russian Academy of Sciences, ul. Kosygina 4, 117977 Moscow, Russian Federation. Fax (7-095) 938 21 56. Tel. (7-095) 939 71 68 Received 25 August 1997 Uspekhi Khimii 67 (7) 656 ± 687 (1998); translated by V D Gorokhov UDC 541.128 Heterogeneous hydrogenation catalysts MD Navalikhina, O V Krylov Contents I.Introduction 587 II. General characterisation of hydrogenation catalysts 588 III. Nickel hydrogenation catalysts 593 IV. Catalysts based on noble metals 600 V. The mechanism of hydrogenation on metal catalysts 603 VI. Carbide catalysts 604 VII. Mechanochemical treatment of hydrogenation catalysts 606 VIII. Sulfide catalysts for hydrotreating 607 IX.Conclusion 611 At present, the most widespread are the following hydro- genation catalysts: Process Catalyst Hydrogenation of alkenes supported Ni, Cu, Pt, Pd, Ru, and Rh Hydrogenation of alkynes supported Ni, Pd, Ni7Pd, Co7Pd, and Cu7Pd Hydrogenation of aromatic supported Ni, Pd, Rh, Pt, hydrocarbons and Co7Mo, Raney Ni Hydrorefining of motor fuels Ni, Ni7Mo, and Co7Moon Al2O3 and kieselguhr Hydrogenation of carbonyl skeletal and supported Ni and Co, compounds supported Pd, Rh, and Pt, and Cu7Cr7O, Cu on kieselguhr Hydrotreating Ni7Mo, Ni7W, Co7Mo, Co7W, (hydrodesulfurisation) and Ru on Al2O3 and carbon, Mo andW carbides Hydrogenation of niriles supported Ni, Pd, and Pt Hydrogenation of CO and CO2 supported Ni, Cu, Co, Fe, Pt, (methanation) and Rh, Pt mesh,MoandWcarbides Russian Chemical Reviews 67 (7) 587 ± 616 (1998) #1998 Russian Academy of Sciences and Turpion LtdThis review is confined to heterogeneous hydrogenation catalysts and leaves out consideration of numerous homogeneous systems, as well as Raney catalysts (Ni, Co, Cu, and others) and co-precipitated catalysts.II. General characterisation of hydrogenation catalysts 1.Conventional hydrogenation processes and catalysts The first industrial hydrogenation process used a nickel catalyst discovered by Sabatier in 1897.3, 4 Other well known hydrogena- tion catalysts were proposed later: Group VI ± VIII metals, including platinum group metals, both on supports and as components of complex oxide systems, as well as Raney metal catalysts. The results of Ipat'ev's studies of catalytic hydrogenation reactions under pressure dating back to the beginning of this century have served as the basis for the development of processes that are widely used in modern industries.Many specific hydro- refining catalysts have been developed that were obtained by impregnation or co-precipitation (Al7Co7Mo, Ni7Mo, sulfide Ni7W7S catalysts, etc.) and proved to be resistant to catalytic poisons present in crude oil (S-, N-, and O-containing com- pounds).5 Production of catalysts for the synthesis of organic com- pounds is rather limited.Relatively large-scale industrial proc- esses of this type are hydrogenation of fats, 2-ethylhexenal, acetylene (for its removal from ethylene), hydrogenation of benzene to cyclohexane required for the manufacture of capro- lactam, and some others.Hydrogenation catalysts are increas- ingly used for the production of drugs and physiologically active compounds. Important properties of this category of catalysts are high selectivity, extended lifetime, and the fact that they meet environmental requirements. Such characteristics are basically displayed by supported metal (Ni, Pt, Pd, and Ru) catalysts, and co-precipitated Cu7Cr and Ni7Cr catalysts.Hydrogenation of benzene to cyclohexane may be used as an example illustrating the trends for the use of particular catalysts depending on the strictness of requirements for the final product purity.4, 6 Cyclohexane of high purity (99.88% and above) is used for the manufacture of caprolactam and fibres based on it, and for nylon-6.In the USA alone the cyclohexane output in 1992 was 1.3636106 m3 (Ref. 7). Since the beginning of this century benzene has been tradi- tionally hydrogenated on nickel (Raney Ni, impregnation cata- lysts) and on supported platinum group metals. In early industrial processes carried out under high pressure and at high temperature, the main reaction was accompanied by side reactions: cracking with production of n-hexane, isomerisation of cyclohexane to methylcyclopentane, isomerisation and cracking with the forma- tion of lower hydrocarbons.4 A two-stage scheme gives the best results in the production of high-purity cyclohexane: in the first stage, a supported Ni catalyst is used (30% ± 45% Ni/Al2O3), while in the second stage, a supported Pt catalyst (0.2% ± 1.0% Pt/Al2O3) is used.Here, the purity of cyclohexane is 99.57% ± 99.88%.8 An important contribution to the development of hydro- genation catalysts has been made by D V Sokol'skii and his associates.9, 10 In our country, much attention was given to co-precipitated Ni7Cr catalysts enabling the production of high-purity cyclo- hexane in a few stages in tubular reactors provided the techno- logical requirements are strictly observed.Such catalysts were prepared by precipitation of Ni and Cr nitrates with a sodium carbonate solution at room temperature. This was followed by filtration, washing of the precipitate to remove nitrate ions, drying, calcination, pelleting, and activation in a flow of hydro- gen.11 The contents of Ni and Cr2O3 were 48 mass%± 63 mass% and 37 mass%± 52 mass %, respectively.The disadvantages of this and other techniques for preparing Ni7Cr catalysts are their multistage character, cumbersome filtration equipment, large volumes of waste water from the washing stage, etc. In this connection, a sewage-free method was proposed for producing Ni7Cr catalysts 11 that included the stage of mixing Ni(NO3)2 and (NH4)2Cr2O7 solutions with subsequent drying and calcination of the resulting mixture by spraying it over a heated inert packing in the fluidised bed reactor system at 310 ± 340 8C, pelleting, and reduction.A multistage synthesis of Ni7Cr and co-precipitated Al7Co7Mo, Al7Ni7W and Ni7W catalysts used in oil refining 5, 12 is complicated, and the technology of catalyst prep- aration is also complex.Besides, these catalysts are insufficiently active and require large quantities of active additives. Even at a high Ni percentage in Ni7Cr catalysts the active surface does not exceed 25 ± 30 m2 g71, while their thermostability is preserved only up to 250 ± 280 8C,9, 11 which imposes definite restrictions on technological processes.The method of impregnation of supports possessing specific structures and properties yields more active Al7Ni7Mo and Al7Co7Mo catalysts. 2. Modern trends in the development and application of new hydrogenation catalysts In the 1980s, studies were undertaken aimed at developing supported Ni hydrogenation catalysts the activity of which and other parameters were not inferior to those of skeletal catalysts.Such properties are displayed, for example, by Euro-Ni-1 catalyst (25% Ni/SiO2) developed by a large group of investigators headed by Coenen.13 ± 15 This catalyst has been prepared by homogeneous precipitation in which part of the nickel forms hydrosilicate on the support. The specific surface of the active metal reaches 113 ± 182 m2 g71 (Ref. 15). Homogeneous precipitation of Ni was carried out in the presence of urea on a water-suspended microspherical silica gel, which was prepared by direct flame pyrolysis of SiCl4.14 In the presence of SiO2, the pH of an aqueous solution of Ni(NO3)2 .6H2O does not exceed 5.3 ± 5.5, whereas in its absence the pH reaches 5.8 during precipitation of nickel hydroxide and remains constant.The existence of a strong interaction between Ni and SiO2 in the preparation of the catalyst is evident from the difficulty of reduction of Ni species, especially in comparison with the non- supported Ni(OH)2. Complete reduction of the Euro-Ni-1 cata- lyst requires temperatures above 400 8C, up to 600 8C.14 In the deposition of Ni on SiO2, the development of the surface may take place due to the formation of intermediate nickel hydrosilicate.14 However, this depends on the conditions, Ni : SiO2 ratio, and the reactivity of SiO2.Along with Euro-Ni-1, in recent years novel technologies allowed preparation of Ni/SiO2 catalysts with a lower metal content (10 mass%± 15 mass %) for the hydrogenation of acetone.16 In these catalysts, the specific surface of Ni does not exceed 10 ± 30 m2 g71. Other attempts aimed at increasing the specific surface of Ni on supports with a concomitant decrease in its content down to 4%± 10% have also been undertaken.17, 18 To this end, use was made of both the impregnation technique and organometallic complexes, modifying the surface of SiO2 and Al2O3 supports, in organic solvents.According to the available data,18 ± 20 the active surface of Ni may reach 124 and 197 m2 g71 at its content of 4.8% and 10%, respectively. In the 1980s, great interest arose in Russia in the so-called `sorptional' catalysts.19 ± 21 The method of their preparation con- sisting of binding of organometallic complexes to the surface of the support, is rather complex and expensive.Nonetheless, this technique may be successfully used for the preparation of Pd catalysts with a unique mechanical strength, which is achieved by proper selection of supports, while the size and structure of metal particles may be controlled.21, 22 In recent years, public concern about environmental problems has become especially acute in many countries.For example, the 588 M D Navalikhina, O V KrylovUS `Clean Air Act', of 1992 rigorously limits the content of carcinogenic aromatic compounds, sulfur and other noxious compounds in motor fuels (both hydrocarbon- and alcohol- based).23 ± 32 New requirements stimulated the rapid development of refining of various fuels thus making the role of hydrogenation catalysts very important.To meet all normative requirements under the conditions of a total ban on the use of tetraethyllead in gasoline, it is necessary to restructure the entire oil refinery sector; specifically, to replace reforming by other processes of high-octane gasoline produc- tion.2, 24, 25, 28 The size of the required investment amounts to hundreds of billion dollars.30 There are doubts whether this task may be coped with at all.5 Nowadays, along with studies aimed at replacing reforming by alkylation, skeletal isomerisation, and synthesis of oxygenates allowing an increase in the octane number of gasoline, hydro- genation of aromatic and unsaturated hydrocarbons is extensively studied (to remove their excess from petroleum fractions), and the appropriate catalysts of hydrorefining of motor fuels are being developed and improved.The content of aromatic hydrocarbons in gasoline and diesel fuel fractions can vary within broad limits. Fractions from direct petroleum refining contain from 20 vol.% to 40 vol.% of aro- matic compounds, while diesel fractions after cracking range from 40 vol.% to 70 vol.%. Hydrotreating of directly distilled gasoline involves mainly co- precipitated Al7Co7Mo catalysts capable of removing sulfur.These catalysts ensure only partial hydrogenation of aromatic compounds 5, 12, 31 and are insufficiently active in general for the hydrogenation of aromatic and unsaturated compounds. In some cases, Ni7W7Al or Ni7Mo7Al catalysts are more suitable for this purpose. In major modern catalytic upgrading technologies, the directly distilled gasoline is supplemented with products of secon- dary processing of heavier fractions, which are even more enriched in aromatic hydrocarbons and alkenes (gasoils of catalytic crack- ing, hydrocracking, reforming, pyrolytic gasoline).Therefore, catalytic hydrodearomatisation and hydrogenation of unsatu- rated hydrocarbons, which are also used for alkylation, have become highly important for the production of high-quality and ecologically clean reformulated gasoline, jet and diesel fuels.In the preparation of jet fuels by hydrodearomatisation of kerosene, the co-precipitated Ni7Wsulfidised catalysts proved to be the best of the known Co7Mo, Ni7Mo, and Ni7W systems.5, 12 In such catalysts, the content of tungsten reaches 20 mass%± 30 mass %; hence, they are rather expensive and their practical application is limited.In recent time, much attention has been given to new types of supported Co7Mo, Ni7Mo, and Ni7W catalysts, as well as to the catalysts based on noble metals.5, 12, 31 ± 40 Sulfur-resistant catalysts are widely used in both single- and two-stage processes for hydrogenation of aromatics.31, 38 ± 40 It was found 31 that the Ni7W/Al2O3 catalyst is poisoned by sulfur more intensely than Ni7Mo/Al2O3.Therefore, two-stage hydro- refining involves the Ni7Mo catalyst in the first stage (mainly, hydrogenation of sulfur compounds), whereas in the second (`final' hydrogenation) stage low-percentage Ni7W catalyst e.g., (6%Ni+20%WO3)/Al2O3 is used.39 Since the late 1980s a number of companies dealing with hydroprocessing of motor fuels have been developing new Ni7W catalysts on supports.31 ± 35, 39 A common shortcoming of all proposed catalysts is the rather high content ofWO3; furthermore, effective hydrotreating on these catalysts requires high pressures (up to 10 MPa) and temperatures (up to 360 8C).35, 39 Thus, at present the most attractive and inexpensive catalysts for the hydrogenation of unsaturated compounds (including aromatics) are supported Ni-containing catalysts modified with WorMo oxides.The nickel-based catalysts are also successfully used in other processes, e.g., for the hydrogenation of CO into methane and other hydrocarbons.41 ± 43 Most catalysts of CO methanation are also prepared on supports (Al2O3, Cr2 O3, kieselguhr, Mg-con- taining spinels). In addition to nickel, the role of active component may be played by Co, Pt, and Pd; however, the resulting systems are less active than the modified Ni-containing catalysts.Inves- tigation ofCOmethanation has become topical due to attempts to adapt this process for the large-scale production of clean gaseous fuels through gasification of low-grade coals,44 as well as for the realisation of chemical-energetic transfer and the use of heat from nuclear power stations and other industrial fields.45 Effective methanation of CO has been found to require 46 thermally and mechanically stable Ni catalysts possessing the largest possible specific surface.These requirements can be met using appropriate supports supplemented with structure modifiers and promoters. Until the 1980s, the long-term operation of a catalyst was believed to be associated with a sufficiently high content of an active metal (up to 35 mass %).For this reason, preference was given to co-precipitated rather than to impregnated catalysts.46 However, the results of recent studies disprove this belief. At present, there is a trend for the use of modified low-percentage Ni catalysts (0.8 mass%± 8.0 mass% Ni) on supports both for the hydrogenation of unsaturated and aromatic compounds 34, 39 and for the methanation of CO.47 ± 49 3.Catalysts for the hydrotreating of motor fuels According to the estimates of experts,50 the oil reserves will substantially decrease and their quality will worsen by the year 2000. As pointed out above, further use for motor transport of fuels with excess levels of aromatic compounds and alkenes may result in air pollution hazardous to humans.23, 25 Upgrading technologies for motor fuels (gasoline, diesel, and kerosene distillates) involve a number of catalytic processes that ensure the transformation of directly distilled crude oil or prod- ucts of petroleum processing into high-quality fuels meeting rigorous norms for the levels of toxic admixtures, primarily reactive aromatics and alkenes.The actual content of aromatic hydrocarbons in the diesel fractions of petroleum is considerably higher than the mentioned norms and reaches 80%± 95%. In the 1980s, they were mostly removed by extraction, while nowadays preference is given to catalytic methods of dearomatisa- tion.5, 12, 31, 32, 35 ± 37 This improves both the environmental and operational parameters of diesel fuels.Thus, the decrease in the content of aromatic compounds from 46.5% to 30% increases the cetane index of diesel fuel to the allowed level (from 31 ± 34 to 40 ± 46).51 Many reports 31, 37, 40, 50 describe new catalytic processes for dearomatisation of different types of diesel fuel. The Arosat and Synsat processes developed by the Lummus Co.make it possible to reduce more than 50% of aromatic components of raw materials by the combined application of two catalysts designed by the Criterion Co. in a stationary layer under a relatively low pressure.32 The resulting diesel fuel meets the European standards (content of aromatics, 20% ± 25%; cetane index, 40 ± 45). The catalysts of the Criterion Co.offer four ways for the decrease in the content of aromatic compounds in diesel fuel.32 1. Single-stage hydrogenation under high pressure (103 atm) and a moderate flow rate (1.0 h71). When supported catalysts (Ni, Ni7W, or Ni7Mo on Al2O3) are used, the level of aromatic compounds in raw materials can be>50%. 2. Hydrocracking under high pressure (100 ± 140 atm) at a moderate flow rate with the use of Ni-containing catalysts on different supports (Al2O3, aluminosilicates, zeolites).This one- or two-stage process is most popular. 3. Two-stage hydrogenation under pressure ranging from medium to high (83 ± 105 atm.) with the use of supported noble metal catalysts in the second stage. The pressure is determined by the conditions of the first stage aimed at maximum removal of heteroatom-containing components by drastic hydrotreating on Al2O3-supported catalysts (Co7Mo, Ni7W, and other).Heterogeneous catalysts of hydrogenation 5894. Single-stage hydrogenation under low pressure and small flow rate (the Synsat process). In this process, the above-men- tioned combination of Ni-containing catalysts on supports is used.The catalysts are placed in the reactor in such a way that the products obtained in the first catalytic layer improve their characteristics upon passage through the second layer. Under a pressure of 43 atm and a flow rate of 0.5 h71 the degree of saturation of aromatic hydrocarbons present in the diesel fraction reaches 40%± 50%. When it is necessary to reduce the level of aromatics below 10% (for preparing a high-quality jet fuel), it is sufficient merely to increase the pressure.Realisation of the Synsat process requires relatively modest expense. Modern catalytic processes for hydrorefining diesel fuels are more efficient, flexible and economical than the extraction used earlier for the increase in cetane index. The efficiency and simplicity of hydrodearomatisation are provided by the use of catalysts that in most cases refer to supported nickel contacts.Nishijima et al.52 reported a two-stage process of hydro- refining of middle and heavy distillates with using new catalysts. Heteroatom-containing compounds are removed in both stages. In the first stage, hydrogenation takes place, whereas in the second stage the remaining aromatic component undergoes hydrocrack- ing.Fundamental physicochemical studies allowed the develop- ment of two types of industrial catalysts on supports for single- and two-stage hydrorefining. The composition of the catalysts obtained, predominantly Ni-containing ones, is presented in Table 1. Catalysts were modified in order to make them suitable for hydrogenation or hydrocracking.Changes in the character of hydrogenation were brought about by the use of specific supports. As regards their hydrogenation activity in the presence of sulfur- containing compounds, the Al2O3-supported catalysts form the following series: Ni7W>Ni7Mo>Co7Mo. Physicochemical studies pointed to a different behaviour of Ni7W and Ni7Mo catalysts on Al2O3. They possess higher hydrogenation activity at a smaller and larger degree of sulfida- tion of tungsten and molybdenum, respectively.Consequently, the choice of a Ni-containing catalyst for a single-stage hydro- treating depends on the content of sulfur in the raw material. Similar results were obtained earlier 31 in the studies of Ni7W and Ni7Mo catalysts for hydrodearomatisation of diesel fuel fractions containing up to 70% of aromatic compounds.The ratios of the rate constants for hydrogenation on Ni7Mo and Ni7W catalysts were 100 : 90 (1.7% sulfur) and 100 : 225 (0.14% sulfur). The trend for maximum decrease in the content of sulfur in motor fuels of all types (down to 0.01% ± 0.05%) will favour wider use in the near future of two-stage hydrotreating in which the first stage provides for profound hydropurification of fuel from sulfur, while in the second stage aromatic compounds undergo hydro- genation on efficient Ni7W catalysts.At present, platinum catalysts (0.6%Pt/Al2O3) are used in such processes at a residual sulfur concentration of 1.5 ppm.31, 40 In the pressure range of 4.5 ± 6.0 MPa, at temperatures 320 ± 400 8C, and at flow rates of 0.75 ± 1.0 h71, the degree of hydrogenation of the aromatic component is 60%± 70%. The Haldor Topsùe Co.has also developed single- and two- stage hydrotreating refinement processes 40 for producing high- quality diesel fuels (the so-called City Diesel) with the levels of aromatic compounds below 10 mass% and sulfur below 5 ppm. In the two-stage process, the first stage is based on the use of common sulfidised Co7Mo, Ni7Mo, and Ni7W catalysts, while the second stage relies on the Pt- and Pd-zeolite catalysts.The latter are more resistant to sulfur than the bimetallic catalysts Pt7Re/Al2O3, Pt7Sn/Al2O3, Pd/LaY, and others. The catalyst designed for the final hydrogenation preserves its activity at levels of sulfur up to 1000 ppm and nitrogen up to 50 ppm.After the first stage, the diesel oil contains 10 ppm of sulfur, less than 1 ppm of nitrogen and 18 vol.% of aromatic compounds, and the cetane index is 52. After the second stage, the nitrogen content remains at the same level, the sulfur content drops below 1 ppm, the level of aromatics decreases down to 4 vol.%, and the cetane index is 55.5. It is seen that the most significant removal of aromatic compounds is achieved in the second stage.According to earlier estimates,36, 50 by the year 2000 the predominant type of fuel produced will be low-sulfur fuel (sulfur content 0.05 mass% instead of 0.1 mass%± 0.3 mass%) with a reduced content of aromatics. To achieve this, it was recom- mended: (1) to restrict the content of high-boiling components in the diesel fuel since sulfur is most difficult to remove from fractions with b.p.>315 8C; (2) to conduct hydrotreating refinement at relatively low temperature, reduced flow rate, and high hydrogen pressure; (3) to remove aromatic compounds on supported Ni7Mo or Ni7Wcatalysts. These predictions are somewhat at variance with the critical need for additional resources of diesel and jet fuels.53 The proportion of the diesel fuel and gasoline consumed is known to shift in favour of the former.54 The need for the jet fuel also increases, while its production by direct distillation and hydrotreating refinement of crude oil fractions cannot meet the demand.55, 56 For this reason, in some countries specialists pro- pose to raise the limiting boiling point of diesel fuels to 360 ± 370 8C55 and to include in the diesel fuel some gasoline fractions, as well as secondary products of petroleum refining (gas oils from hydrocracking, fracturing, coking, etc.) that will unavoidably lead to a poorer quality of the final products.In gasoils from the secondary oil refining, the content of aromatic compounds reaches 46%± 87%, while the cetane index does not exceed 25 ± 39.Hydrogenation of such feed stocks improves their quality but makes them as expensive as synthetic liquid fuels. Hydrogenation of coal is equally expensive, while the products of its liquefaction contain up to 54% of aromatics.54 The ways of expanding the production of high-quality diesel and jet fuels have been discussed.53 It was pointed out that, in addition to the above methods, the processes of motor fuel production from alkenes (the MOGD{ process of Mobil Co.) and from synthesis gas formed upon coal gasification (the Sasol process), are promising as well as the processes based on natural gas conversion (SMDS { of Shell Co.and the Mossgas process) and gasification of any hydrocarbon raw material.57 ± 66 Selective oligomerisation of light alkenes to heavy isoalkenes over zeolite-based catalysts (ZSM-5, ZSM-12, H-offretites and Table 1.Catalysts for the two-stage hydrotreating of diesel fuels.52 Active Support Content of active Process compon- components (mass %) ent CoO NiO MoO3 WO3 Co ±Mo g-Al2O3 0 ± 6 ± 10 ± 30 ± hydrogenation USYa±Al2O3 0 ± 5 ± 10 ± 25 ± hydrocracking Ni ±Mo g-Al2O3 ± 0 ± 6 10 ± 30 ± hydrogenation SiO2±Al2O3 ± 0 ± 3 0 ± 15 ± hydrocracking TiO2±Al2O3 ± 0 ±3 0± 15 ± " USYa±Al2O3 ± 0 ± 5 10 ± 25 ± " ZSM-5 ±Al2O3 ± 0 ±3 0± 15 ± " Mordenite ± ± 0 ± 3 0 ± 15 ± " ±Al2O3 Ni ±W g-Al2O3 ± 0 ± 6 ± 16 ± 48 hydrogenation USYa±Al2O3 ± 0 ± 5 ± 16 ± 40 hydrocracking aUSY is an ultrastable zeolite of Y type. { Mobil Olefins to Gasoline and Distillates.{ Shell middle distillate synthesis. 590 M D Navalikhina, O V Krylovother) (the MOGD process) allows preparation of gasoline or middle distillates with upgraded characteristics after the hydro- genation stage in which noble metals on supports are commonly used as catalysts.57, 58 After hydrogenation, the cetane index increases from 33 to 52, while the levels of aromatic compounds and sulfur become 2%± 4% and 0.002%, respectively.These parameters are much better than those of diesel fuel from petroleum or coal (hydrogenation), while they remain inferior to the parameters of fuels from the synthesis-gas produced by the Fischer ± Tropsch synthesis. For instance, low-temperature hydrogenation of CO on a Ru catalyst can yield diesel fuel that is predominantly composed of n-paraffins (72.4%) and is free of aromatics and sulfur.After the hydrostabilisation stage (removal of alkenes), the cetane index of the product is 90 and it may be used as a high-cetane component of the diesel fuel.66 As regards its characteristics, the jet fuel produced by the MOGD process differs favourably from the existing commercial fuels of petroleum origin. It contains as little as 4% of aromatic compounds (instead of 25%), 1% of alkenes (instead of 5%), and less than 0.002% of sulfur (instead of 0.3% ± 0.4%), while the proportion of paraffins and isoparaffins is 96%.However, the aircraft engines are continuously upgraded with more and more rigorous requirements for the fuel. Catalytic hydroprocessing of raw oil is a method widely used for producing high-quality aircraft fuels with low sulfur content, increased thermal and oxidation stabilities, and upgraded envi- ronmental characteristics.67, 68 Earlier, hydrotreating of directly distilled fractions was mostly carried out 12 on Co7Mo7Al catalysts at 350 ± 400 8C, 30 ± 50 atm, and a raw material flow rate of 2 ± 5 h71.The process of single-stage hydrorefinement of kerosene fractions reduced the level of sulfur, while the decrease in the level of aromatic compounds and alkenes was insignificant. The need for the hydrotreating of gasoils of secondary origin with a high content of aromatic compounds and alkenes, using the technology based on Co7Mo7Al catalysts that is popularly used in Russia, has necessitated an increase in the pressure in the system to 140 atm.To reduce the pressure down to 35780 atm, the products of secondary refining were mixed with directly distilled fractions so that the percentage of aromatic compounds was 34%± 36%.56 However, even under a pressure of 83 ± 85 atm, hydrotreating on the Co7Mo7Al catalyst was not sufficiently effective. The resulting product contained 18%± 19% of aromatic compounds, which is considerably higher than the level admissible for jet fuels. Two-stage processes in which hydrodesulfurisation (e.g., on a Co7Mo7Al catalyst) and hydrodearomatisation (e.g., on sup- ported Ni7Moor Ni7Wcatalysts) are conducted separately (see above) are efficient.In Western technologies, in the second stage the sulfur-free raw materials are hydropurified from aromatic hydrocarbons and alkenes on Pt- or Pd-containing catalysts since the exhaustive hydrogenation on Ni catalysts with oxide additives of the MoO3 or WO3 type requires elevated pressures.40 Studies of benzene hydrogenation and thiophene hydrogeno- lysis on Ni7W7Al catalysts have shown 69 that both processes take place in the same active centres that are poisoned with hydrogen sulfide.The rate of benzene hydrogenation may be increased 30 ± 70-fold under non-steady-state conditions and after pretreatment of the catalyst with hydrogen.The results obtained made it possible to develop a process of dearomatisation of petroleum fractions in a non-steady-state regime. The production of environmentally clean gasoline is much more problematic. The products of catalytic reforming yielding a high-octane gasoline contain large quantities of aromatic hydro- carbons and alkenes.For this reason, reforming should be replaced by other processes. Even under the conditions of catalytic cracking, current technologies yield on average 29.7% of aromatic compounds (which is 20%± 25% higher than the required level) and more than 5 mass%of alkenes (up to 30%). In order to produce high-octane gasoline, attempts are made to use processes of skeletal isomerisation, alkylation, synthesis and addition of oxygenates.Hydrodearomatisation and hydro- genation of alkenes as gasoline components attracted less atten- tion, though these processes will in the near future be among the leading methods in the technologies of oil refining targeted at the production of environmentally clean motor fuels.23, 24, 29, 70 In these processes a special role is played by supported Ni-containing catalysts.The possibility of their upgrading and the relatively low production cost (by an order of magnitude lower than that of Pt- or Pd-containing hydrogenation catalysts) make them rather attractive subjects for further studies. 4. CO Hydrogenation catalysts Like hydrorefining, catalytic hydrogenation of CO and CO2 is highly interesting from the environmental viewpoint and may be used for large-scale production.Carbon mono- and di-oxides are contained in the exhaust gases from power stations and different chemical industries. The simplest reaction of CO hydrogenation (methanation) CO+3H2 CH4+H2O, DH298=7206.4 kJ mol71 (1) allows an energetically advantageous production of a valuable fuel, viz.methane, that is used further in gas turbines. In connection with the depletion of natural gas resources studies are being undertaken as to the possibility of the environmentally clean coal gasification with subsequent CO methanation.44, 71 ± 73 In the first decade of the 21st century, important facilities for coal gasification and methanation are expected to be put into oper- ation.72 ± 74 This requires the development of new technologies and active catalysts.This review leaves out other reactions of CO hydrogenation that yield various products (paraffins, alkenes, alcohols,75 alde- hydes, etc.) depending on the catalysts applied and the conditions of the process. Catalytic methanation finds applications in chemo-energetic cycles where atomic (nuclear reactors) or solar energy is used.} 74, 76 The thermochemical cycle includes the exothermic reaction of methanation (1) and endothermic reactions of con- version of the methane produced CH4+H2O CO+3H2 , (2) CH4+CO2 2CO +2H2 , (3) which are conducted in reactors with heat withdrawal or supply, respectively.These reactions require active and highly selective catalysts. Specifically, only the catalysts inhibiting side reactions in which CO can be consumed are suitable for methanation: CO+H2O CO2 +H2 , DH298=741.5 kJ mol71, (4) 2CO CO2+C, DH298=7171.7 kJ mol71.(5) Common and widespread Ni catalysts for reactions (1) ± (3) are rapidly deactivated, and it was proposed to replace them by noble metals on supports.74 The major difficulties are associated with carbonisation in reactions (1) ± (3), particularly, in the methanation at low H2 :CO ratios.On noble metals, the carbon- isation is considerably less pronounced, especially in an excess of water vapour due to the reaction C+H2O CO+H2 . (6) Reaction (6) is best accelerated by Ru catalysts, while Rh is more suitable for methanation.74 Thermodynamic calculations show that it is advantageous to obtain methane in reaction (1) at temperatures not exceeding 550 ± 600 8C.The yield of CH4 is increased at higher pressure and H2 :CO ratio. At low H2 :CO ratios and temperatures up to 600 8C, the common Ni and Ru catalysts may be replaced by an expensive rhodium catalyst (0.5%Rh/Al2O3).74 The activity of metals on supports in the methanation reaction strongly depends on the mode of the catalyst preparation and } See also V N Parmon Catal.Today 35 153 (1997). Heterogeneous catalysts of hydrogenation 591conditions of its pretreatment and reduction. In this respect, Ni and Co catalysts have been studied in most detail. In the methanation reactions, supported catalysts containing Ni on metal oxides (MgO, Cr2O3, ThO2, ZrO2, MoO3, SiO2, Al2 O3), promoted by Cu (0.1% ± 0.3%), promoted by alkali metals, and Ni sulfide on Al2O3 are the most efficient.75 The specific activity of Group VIII metals on Al2O3 supports decreases in the following series: Ru>Ni>Co>Rh>Pd>Ir.It is noteworthy that the Ni catalyst with a well developed surface is highly active and rather selective in the methanation reaction.In some cases, depending on the mode of preparation, the nature of the support, and other factors, the Ru catalyst proves to be more efficient.75 ANi catalyst with a highly developed surface is applied for the selective CO methanation in the presence of CO2.77 On a number of catalysts, viz., 0.5%Ru/a-Al2O3, Raney nickel of ICI (30%Ni718%Al752%Al2O3), 58%Ni/kieselguhr (Harshaw), 47%Ni/Al2O3 (CCI), and 32%Ni/Al2O3 (Girdeer, G-65), selective hydrogenation of CO in a mixture 0.3%CO+ 80%H2+19.7%CO2 can be carried out at atmospheric pressure and at 125 ± 300 8C.77 In all cases, the maximum conversion occurs in the temperature range 200 ± 250 8C. Some catalytic systems including Ni, Fe, Rh and Ru are more active and selective in the reaction of CO2 methanation than in reaction (1).78, 79 For example, the rate of CO2 methanation on a Ni/MgO catalyst is higher than that of CO methanation, espe- cially after the catalyst has been reduced at 600 8C.77 Examples of higher activity and selectivity of different metal catalysts in the hydrogenation of CO2 are reviewed.80 Inui and Takeguchi 81 described a Ru-containing supported Ni catalyst of methanation with a low content of components (3%Ni, 2.5% La2O, and 0.7%Ru).The catalyst has a meso- macro-bimodal porous structure with finely dispersed compo- nents and displays a very high activity in the reaction of CO2 methanation. The role of Ru in this catalyst is explained by its involvement in the hydrogen spillover to the active Ni centres.Numerous studies and reviews are devoted to the problems of the catalyst formation and the mechanism of CO hydrogena- tion.75, 82 ± 86 Krylov et al.85 consider in detail the results of studies into the formation mechanism of the active centres of Fe catalysts for the Fischer ± Tropsch synthesis and of Ni catalysts of methan- ation and their effect on the catalytic activity and selectivity.Much attention is given to the investigation of the formation of active centres in finely dispersed metal particles obtained by decomposition of carbonyl clusters on oxide supports. The carbonyl clusters interact with the hydroxyl groups of the support forming bridging bonds with the metal (M) atoms in a lower oxidation state: Al7OH+M2+(CO)n Al7OM(CO)n71 +CO +H+.(7) Active and selective contacts can be obtained by deposition of carbonyl complexes [e.g., Co2(CO)8] and a number of complexes with organic ligands on g-Al2O3. In this case, the hydroxyl cover of the support is of critical importance.85 Some complexes are oxidised to form Co2+ particles which aggregate into clusters on the surface. Heating under an inert gas atmosphere favours formation of a cluster containing 3 or 4 Co atoms.Further heating to 220 8C leads to dimerisation of the complex with subsequent formation of metal particles. To increase the activity of a catalyst in the COhydrogenation, it is necessary to have large Co0 clusters. As to selectivity, it is increased when oxidised Co species that are produced on the hydroxylated Al2O3 surface upon deposition of low-nuclear complexes are present on the surface.Carbonaceous ligands present in the cluster inhibit the process of hydrogen spillover and impede the complete hydrogenation of CO to CH4. The use of carbonyl clusters containing two different metals allows the preparation of catalysts with new properties. For example, hydrogenation catalysts prepared from Co7Fe car- bonyl clusters manifest the activity of cobalt and the selectivity of iron.High-resolution electron microscopy and X-ray phase analy- sis in situ were used to study the mechanism underlying the formation of Ni catalysts of methanation.85, 87 In contrast to hydrogenation on Fe or Co, the hydrogenation of CO on Ni yields only methane. After the dissociation of CO on the Ni surface, the consecutive transformations C?CH?CH2?CH3 take place. Active carbon exists on the surface in a rather narrow temperature range: it is not formed below 200 8C since there is no COdissociation, while above 450 8Cit is rapidly transformed into inactive graphite.The methanation rate at the Ni (100) face is higher the larger the carbon concentration on the surface, and the yield of higher hydrocarbons is larger the higher the CO concen- tration.84 Two nickel specimens were shown to behave differently 85, 87 under the conditions of CO methanation.Nickel oxide obtained by a plasmochemical technique forms cubic crystals with the most developed face (100). Under the reaction conditions this face is the site of epitaxial growth of cubic Ni crystals, while the polycrystal- line carbide Ni3C is formed only in later stages.NiO prepared by decomposition of NiCO3 is composed of lamellar crystals with the developed face (111) on which nickel carbide builds up epitaxially, whereas the phase of metallic nickel appears only after heating to 283 8C and decomposition of carbide. No direct correlation between the methanation rate and the carbide content was established.The reaction occurs most likely at the Ni/NiO inter- face (possibly, on NiO defects). The nature of the support influences the catalytic properties of Co and Ni in hydrogenation of CO to methane and its homo- logues.86 Thus, the activity and selectivity of supported metal catalysts depend on the nature and acid-base properties of the support.The specific catalytic activity of supported Co catalysts decreases in the following series of supports: SiO2>- TiO2>ZrO2> BeO>Cr2O3>Al2O3>MgO>WO3, while that of supported Ni catalysts decreases in the order: ZrO2> TiO2>BeO>MgO>Al2O3>SiO2>Cr2O3>WO3. The heat of adsorption of carbon monoxide (qCO) and, consequently, catalytic activity and selectivity strongly depend on the mean size of metal particles.According to quantum- chemical calculations,88 the M7CO bond is weakened on tran- sition from small clusters to large crystallites. Co/SiO2 (mean particle size d=32 nm) and Ni/WO3 catalysts (d=97 nm) are the most active in the reaction ofCOhydrogenation. Co/WO3 and Ni/WO3 catalysts (d=60 and 120 nm, respectively) manifest noticeable activities only at T>300 8C.89 The only reaction product is methane.These data indicate that the optimum sizes of Co and Ni particles are different. Prokhorenko et al.86 explain this by the difference in their qCO values. The heat of CO adsorption on Co must be lower than that on Ni, and this was confirmed experimentally. In order to have similar qCO values for both metals, the nickel particles should be larger than those of cobalt.At small qCO values, the filling of the Co and Ni surfaces is insignificant and the selectivity is low. With the increase in qCO, the extent of filling increases as well as the probability of non- linear stages of reactions of CO leading to the formation of higher hydrocarbons. However, the heat of CO adsorption is not the only factor affecting the selecitivy.86 For example, Co/ZrO2 and Co/SiO2 have virtually the same degree of dispersion (d=97 and 103 nm, respectively), though their selectivities differ by 34%. Transfor- mations of an intermediate compound are supposed to involve the support and to occur at the metal/support interface.The increase in the acidic character of the support on transition from ZrO2 to WO3 results in an increase in the selectivity of methane formation to 100%. 5. Catalysts for hydrogenation of carbonyl compounds Catalysts for the hydrogenation of carbonyl compounds (alde- hydes, ketones, and carboxylic acids) present in alcohols as admixtures are of commercial interest. In particular, this reaction is used for hydrorefining of alcohol-based motor fuels.Hydro- + 592 M D Navalikhina, O V Krylovgenation of aldehydes and ketones is also of an independent significance since it allows the preparation of alcohols free of by- products. For example, 2-ethylhexanol is produced from 2-ethyl- hexanal on an industrial scale.90, 91 The first industrial hydrogenation of aldehydes on copper catalysts dates back to the 1920s.However, after the appearance of Raney nickel (1925) most processes of hydrogenation of aldehydes were modified to use catalysts of this type. The industrial importance of techniques for the alcohol production from aldehydes increased after the discovery of oxosynthesis in 1938.90 In parallel, a strong impetus was given by the production of plastics [specifically, poly(vinyl chloride)], where respective alco- hols are used as plasticisers.Lower alcohols are used as solvents, while higher alcohols are used in the production of detergents. In 1992, the world production of alcohols was (millions of tons): n-butanol, 1.7; isobutanol, 0.4; 2-ethylhexanol, 2.4; higher alco- hols, 0.2.90 Despite the very high activity of Ni-containing catalysts, they prove sometimes to be unsuitable for the hydrogenation of aldehydes, especially when high selectivity of the process is required.Thus, the preparation of n-butanol from crotonalde- hyde proceeds more selectively on copper catalysts. In addition, Ni unlike Cu, often catalyses splitting of CO from saturated aldehydes: (8) When a Ni catalyst is sufficiently active, subsequent hydrogena- tion of alkenes occurs with the predominant formation of alkanes.Modified Ni catalysts on supports are free of these short- comings. Their activity may be reduced using such supports as SiO2 subjected to high-temperature treatment, or employing a special impregnation technique of Na+-containing SiO2, and calcinating the specimens additionally at 600 ± 700 8C.16 To this end, TiO2 may also be used as a support.90 The high activity of Ni and Ni7Cu catalysts allows their use in liquid-phase reactions, which facilitates the removal of heat released during the reaction.In addition to Cu and Ni, supported Co and noble metals (Pt, Pd, Ru, and Rh) are also applied for the hydrogenation of carbonyl compounds. Of all noble metals only palladium is a component of catalysts for the selective hydrogenation of a,b- unsaturated aldehydes to saturated ones.Furthermore, along with nickel, palladium ensures complete hydrogenation of unsa- turated aldehydes, e.g., 2-ethylhexenal.91 III. Nickel hydrogenation catalysts 1. Methods for preparation Industrial production of nickel catalysts relies basically on pre- cipitation of components from aqueous solutions and suspensions and on impregnation of specially prepared supports with aqueous solutions of active metal compounds.The activity of hydrogena- tion catalysts prepared by the impregnation of supports is usually higher than that of catalysts produced by co-precipitation.9, 12, 22 In the late 1980s, the relevant literature (especially the Rus- sian) opposed the impregnation and `sorptional' catalysts.17 ± 20 In our opinion, the technology for the preparation of the latter does not differ substantially from impregnation.It consists of the deposition of an active component on a support from different solutions (including non-aqueous ones) using organometallic compounds. In this case, optimally distributed active centres are formed on the support surface, and their characteristics are best suited for hydrogenation.Generally, the procedure for the preparation of Ni catalysts includes the stages of impregnation, drying, dehydration, activa- tion (calcination), reduction, and passivation. The preparation of precipitated and impregnated Ni catalysts involves the following processes.92 1. Impregnation of supports (Al2O3, SiO2, carbon, zeolite) with a salt solution, e.g., nickel nitrate. Then the catalyst is dried and calcined at high temperature (400 ± 500 8C) in air. 2. Mixing of precipitated active hydroxide Ni(OH)2 with a hydrogel of the support. The mixing lasts for 20 ± 30 h, followed by extrusion, drying, and calcination. 3. Deposition of one precipitate onto another. For instance, nickel hydroxide is precipitated with ammonia from a nickel nitrate solution in the presence of a suspension of hydrated alumina.The precipitate is filtered off, thoroughly washed, dried, and calcined at 500 8C for 2 h in air. 4. Co-precipitation. For example, a boiling solution of sodium silicate is slowly added to a nickel nitrate solution. The precipitate is filtered and washed. Also, ammonia may be added to a mixed solution of nickel and aluminium nitrates at 70 8C.Then the precipitate is filtered off, washed, dried, and calcined. 5. Hydrothermal treatment of mixed precipitates. A mixed paste of Ni(OH)2 and SiO2 is heated with water vapour in an autoclave for 25 h at 250 8C and 50 atm; the precipitate is filtered off, dried, and calcined. 6. Superhomogeneous co-precipitation. Three salt solutions, e.g., Ni(NO3)2, NaNO3, and Na2Si2O5, are mixed to form a supersaturated solution.Under intense stirring, a homogeneous precipitate is formed which is separated, washed, dried, and calcined. Prior to their use in reactions, the catalysts are reduced with hydrogen (sometimes with the reaction mixture) at 400 ± 500 8C. The catalysts prepared by impregnation are the easiest to reduce, while those obtained by co-precipitation and hydrothermal treat- ment undergo reduction with more difficulty.In the supported catalysts the specific surface of the active nickel is usually 100 ± 120 m2 g71, whereas in the precipitated catalysts it varies from 20 to 70 m2 g71 (cf. Ref. 93). Depending on the treatment and reaction conditions, the metal particles on supports have sizes from 2 to 10 nm.94, 95 In recent years, efficient catalysts on supports have been prepared using the so-called CVD (chemical vapour deposition) technique.This consists of fixation of a metal in a layer of finely dispersed particles with the aid of vapour-phase adsorption of intermediate compounds, e.g., metal acetylacetonates, which are subject to subsequent decomposition on the supports.96 In this case, the surface is the first formation site for catalytic precursors, ensuring high metal dispersion upon their subsequent decompo- sition.The Amoco Co. proposed the use of CVD method for binding the silanol groups in zeolite matrices to prepare metal oxide catalysts of the Fe/Mo/support type. German scientists prepared catalysts for hydrogenation and a number of other processes in which Pt, Cr2O3, and V2O5 are fixed on SiO2 or Al2O3.96 Active film structures of the metal ± oxide type on supports, the analogues of metal-ceramic films, in which metal particles are regularly distributed over the surface, were also designed.97, 98 Fixation of metal on supports is ensured by co-precipitation of salts of an active metal (Pd, Pt, etc.) and salts of higher carboxylic acids (resinates, stearates, abietates, and others) from organic solvents.Similar structures can also be prepared using aqueous solutions of inorganic salts and compounds of noble metals in the presence of water-soluble polymers, colloidal oxide solutions, etc. The common feature of all these techniques is the formation of a film on the surface of a support.The preparation by this method of highly selective catalysts for the hydrogenation of acetylene that are suitable for large-scale hydropurification of ethane ± ethylene fractions from acetylene admixtures has been patented,99 as well as of catalysts of oxida- tion, isomerisation, and dehydrogenation.100 ± 104 The preparation of film catalysts is based on a specific combination of several conventional techniques.97 ± 100 This leads not only to regular distribution of active components in the surface layer of the support and high dispersity of metal particles, but also to the thermal stability of the catalysts. R1R2CHCHR3CHO R1R2C CHR3+CO+H2 .H2 Heterogeneous catalysts of hydrogenation 593Nonetheless, impregnation from aqueous solutions still remains an industrially preferable way of preparing the hydro- genation catalysts on supports. This takes advantage of the recent developments in the field of formation of the catalyst active phase.For example, it was established that if conditions favouring retention of a metal in the ionic form in supports (Ni/Al2O3, Ni/ SiO2) are met, this favours the preparation of active and stable Ni hydrogenation catalysts.15, 104 In particular, two-stage deposition of nickel should be applied: the first stage consists of the incorporation of difficultly reducible Ni2+ ions into the support as a result of ion exchange, and the second stage includes physical sorption of a readily reducible salt.The Ni2+ ion appears to be the centre of crystallisation and stabilisation of Ni crystallites.Heat- ing of a Ni/TiO2 system often leads to a decrease in the active surface of nickel due to coating of the crystallites with metal oxide.101, 102 This may be regarded as a manifestation of a strong metal ± support interaction (SMSI) effect. The SMSI effect is enhanced in the presence of hydrogen which affects the activity and selectivity of catalysts. 2. Influence of the preparation procedure on the structure, physicochemical properties, and activity of catalysts Principal factors that affect the catalytic and physicochemical properties of impregnated low-percentage Ni/Al2O3 catalysts have been studied systematically.47, 49, 103 ± 111 The conditions for the realisation of major stages in the preparation of these catalysts were shown to be related to their structure, activity, selectivity, and the degree of nickel reduction and dispersion.For example, the active surface of nickel and the proportion of the reduced metal are increased by the increase in the H2 flow rate during reduction. Passivation of the reduced Ni/Al2O3 catalyst at room temperature in a flow of nitrogen containing 1% of air prevents ageing and bulk oxidation of nickel, which may result from its contact with the atmosphere.The first stage in the preparation of impregnated nickel catalysts is highly important. When a support is impregnated with an aqueous solution of a salt (or salts) precursors appear on the support surface and in its bulk, and their formation is influenced by the electrolyte concentration, pH of the medium, and ionic strength of the solution.105, 112 ± 114 The first stage is supposed to generate several forms of isomorphically substituted nickel in the support lattice.At low pH, aluminium oxide is solubilised, which favours isomorphic substitution in the final forms. The effect of pH on the concentration of nickel precipitated from a Ni(NO3)2 solution onto Al2O3 is illustrated by the data presented in Table 2.The concentration of electrolyte, ionic strength, and pH of impregnation solution influence not only the content of Ni in the catalyst, but also the amount of active nickel precursors and hence the physicochemical properties and the hydrogenation activity of the catalysts formed.103, 105 The procedure for the preparation of Ni/Al2O3 catalysts affects their structure, activity, and selectivity, e.g., in the hydro- genation of CO.47, 49, 103 ± 108 Temperature-programmed desorp- tion (TPD) of H2 and its interaction with CO allowed us to obtain data on the accessibility of surface Ni centres to H2 and CO at various conditions of preparation of low-percentage catalysts.Surprisingly, the variation of Ni content from 0.8 mass% to 8.3 mass% proved to have little effect on the adsorption of H2 and the active nickel surface calculated from the results obtained (Table 3).However, the metal dispersity varies noticeably depending on the concentration of nickel precipitated and precip- itation conditions, i.e., it decreases at higher concentrations. At the same time, the analysis of temperature-programmed CO desorption spectra points to the existence of two types of adsorptional and reactive centres.The specimen with 0.8% of Ni is an exception, since its spectrum has a single peak at 260 8C. In all other cases, low-temperature (at 150 ± 200 8C) and high- temperature peaks were observed. The increase in Ni content to 2.3% increases the portion ofCOdesorbed in the low-temperature range.In the case of specimens containing 2.3% ± 7.3% of Ni, the contribution of both peaks is nearly the same. According to Kester and Falconer,48 the two CO peaks correspond to two types of active centres of different composition on the surface of the Ni catalyst. The low-temperature peak is supposed 48, 105 to corre- spond to NiO, while the high-temperature peak corresponds to NiAl2O4.Stronger metal ± support bonding is predominant in catalysts with low Ni content.105 Low pH values facilitate iso- morphic substitution of Ni in the octahedral lattice of Al2O3, which favours the predominant existence of this metal in the NiAl2O4 form. Only this structure was revealed in the 0.8% Ni/ Al2O3 specimen. The results reported by Huang and Schwarz 103 are at variance with those obtained by Tsai et al.112 who have found that the increase in Ni content results in an increase in its specific surface calculated from the CO adsorption from 3.9 to 15.8 m2 g71.The increase in Ni concentration from 0.8% to 2.1% is accompanied by a decrease in the degree of Ni dispersity, also calculated from the CO adsorption, from 70% to 31%, which then remained nearly constant (27% ± 35%). The catalytic activity and selectivity of the Ni/Al2O3 speci- mens described above in the reaction of CO hydrogenation are closely related to their composition and conditions of preparation.This allows the prediction of their catalytic properties.104 The limiting stage of the methanation is determined by the type of active centres prevalent on the surface.According to Huang and Schwarz,47 the process of CO methanation includes the following stages: 1) COads COgas, 2) COads Cads+Oads, 3) Cads+Oads COgas, 4) COads+Oads CO2 ads, Table 2. Dependence of the Ni content in a catalyst on pH of the impregnation solution of Ni(NO3)2.103 [Ni] in the solution pH [Ni] in the catalyst (mass %) (mass %) a 0.7 1 0.2 3 1.2 5 2.1 2.9 1 3.5 3 4.1 5 3.0 5.2 1 6.6 3 8.3 5 7.3 a With respect to the mass of the Al2O3 support.Table 3. Content of nickel in supported catalysts, its specific surface (SNi), and dispersity (DNi).103 [Ni] (mass %) SNi /m2 g71 DNi(%) 0.8 5.2 94 1.2 5.7 70 2.1 5.6 40 3.0 6.0 29 3.5 6.0 26 4.1 5.7 21 6.6 5.1 12 7.3 5.2 11 8.3 5.3 10 594 M D Navalikhina, O V Krylov5) H2 gas 2Hads, 6) COads+(2+x)Hads CHx ads+H2O, 7) 2COads Cads+CO2 ads, 8) CHx ads+(47x)Hads CH4 gas.Investigation of the surface interaction of CO and H2 on Ni/ Al2O3 catalysts with variable Ni content by X-ray photoelectron spectroscopy (XPES), thermoprogrammed reaction (TPR) 113 and other methods confirmed that in the case of prevalence of NiAl2O4 centres on the surface (catalysts with low Ni content) the methanation rate is determined by the rate of CO dissociation [stage (2)].On catalysts with high Ni content, in which nickel is in the form of particles, the methanation rate is determined by that of hydrogenation of the CHx fragments [stage (8)].47 If the reaction of methanation is regarded as being structure-sensitive, then the changes in its mechanism are well explained by the lower degree of Ni dispersity and the increase in its content from 0.8 mass% to 8.0 mass %.XPES and TPR113 studies have shown that NiAl2O4 centres predominate on an Ni/Al2O3 catalyst with 0.8% Ni, while Ni+NiAl2O4 centres predominate on a catalyst with 6.5% Ni. Contrary to the CO methanation, the hydrogenation of aromatic hydrocarbons cannot occur on active centres of the NiAl2O3 type in which Ni is structurally bound.This requires that the active centres were formed by finely dispersed Ni particles on a support. Kester et al.115 consider the hydrogenation of aromatics as structure-insensitive reactions. There are fundamental difficul- ties that prevent the preparation of long-lived Ni catalysts on g-Al2O3 with the Ni content below 5% for the hydrogenation of aromatic compounds.The point is that local overheating taking place during the exothermic reactions of hydrogenation induce the formation of inactive surface spinel and hence the absorption of the active component.48, 116 ± 118 This explains why the industrial Ni catalysts of hydrogenation usually contain considerable quan- tities of metal (up to 65%) and include oxide promoter additives (Table 4) diminishing the SMSI effect.Nickel catalysts on other types of supports (e.g., SiO2) modified by Ti, Zr, and Ce oxides also manifest low activity in the hydrogenation of benzene provided the SMSI effect is real- ised.116, 119 They are one order of magnitude less active than a catalyst analogous in composition [7%Ni/SiO2 ± TiO2(ZrO2)], prepared by precipitation of Ni(OH)2, which is free from the SMSI effect.The technique of homogeneous precipitation was used to prepare a highly efficient Ni catalyst (Euro-Ni-1) containing 25 mass% of Ni.13 ± 15 In this case, the major part of nickel is in the form of hydrosilicate. Therefore, it was necessary to activate the catalyst at elevated temperatures in a flow of hydrogen (Table 5).The maximum degree of reduction of nickel (96%) is achieved at 630 ± 650 8C, though in this case its specific surface is somewhat decreased, possibly because of sintering. Nonetheless, it exceeds nearly eightfold the surface of Ni in a precipitated Ni7Cr catalyst containing 50 mass%± 60 mass% of Ni. The main part of the active component is bound by the support (SiO2), which results in the formation of new surface phases.This means that this method cannot yield industrial low-percentage, highly active nickel hydro- genation catalysts on the SiO2 support. Romero et al.120 showed that the specific surface of unmodi- fied Ni/SiO2 and Ni/Al2O3 catalysts and their capacity to absorb H2 during thermal treatment or under hydrogenation conditions decrease very rapidly on heating to 650 ± 750 8C.Thus, after treatment at 650 ± 750 8C, the overall specific surface of an Ni/ SiO2 catalyst diminishes from 199 to 159 m2 g71, and the absorp- tion of H2 becomes more than twice as low. Transmission electron microscopy data showed that the size of Ni particles increases from 2.9 to 6.8 nm. After an analogous thermal treatment, the Ni particles in the Ni/Al2O3 catalyst are enlarged from 2.7 to 9.3 nm, though in this case the overall specific surface and H2 absorption do not change as strongly as upon the thermal treatment of Ni/SiO2.In both cases the degradation of porous structures takes place. Thermal treatment of Ni/SiO2 and Ni/Al2O3 decreases the specific surface of Ni by 70% and 25%, respectively. Thermal inactivation of supported Ni catalysts can be caused by the growth of Ni crystallites, degradation of a support, metal ± support interaction, or by combinations of these factors. 3. Effect of the nature of supports on the catalytic properties and specific surface of catalysts The chemical nature of supports and their structure influence the activity and selectivity of hydrogenation catalysts.116 ± 119 Pavlenko et al.117 investigated the adsorption and catalytic properties of Ni catalysts deposited on ZrO2, MgO, TiO2, and SiO2 in the reaction of CO hydrogenation (Table 6).Table 4. Industrial nickel catalysts for hydrogenation. Composition of the ctalyst Specific Pore surface volume Ni (mass%) additive support /m2 g71 /cm3 g71 Catalyst of Harshaw Co. (USA) 60 ± 65 ± clay 155 ± 165 0.22 40 ± Al2O3 160 0.69 68 ± special support 125 0.35 46 ± " 14 WO3 Al2O3 140 0.38 Catalyst of BASF Co.(Germany) 4 ± 10 WO3 active clay ± ± Table 5. Effect of temperature on the extent of Euro-Ni-1 catalyst reduction and the specific surface of nickel.15 Temperature Extent SNi SNi0 of reduction /8C of reduction (%) /m2 (g Ni)71 m2 (g Ni0)71 (see a) (see b) 430 ± 450 85 181 213 500 90 181 201 630 ± 650 96 113 118 a Total specific surface of nickel.b Specific surface of totally reduced nickel. Table 6. Kinetic and thermodynamic parameters of CO hydrogenation on supported nickel catalysts.117 Catalyst Mean size of T /K Maximum E qCO Ni particles /nm pore radius /nm /kJ mol71 Ni/ZrO2 97 495 137.0 43 45 97 508 198.0 43 45 97 533 292.0 43 45 Ni/TiO2 103 526 55.0 59 46 103 545 92.0 59 46 103 561 128.0 59 59 Ni/MgO 1 471 2.3 93 101 1 495 7.2 93 101 1 506 12.0 93 101 Ni/SiO2 6 498 1.5 52 76 6 498 3.4 52 76 6 498 8.1 52 76 Heterogeneous catalysts of hydrogenation 595The heat of adsorption qCO increases as the dispersity of Ni increases.This is in agreement with the results of quantum- chemical calculations 88 according to which the metal7CO bond becomes weaker on transition from the primary clusters to large crystals reaching a constant strength at a mean size of Ni particles 10 ¡À 20 nm.The specific catalytic activity of specimens at 220 8C decreases in the following order of supports: ZrO2 > TiO2> >MgO>SiO2, i.e. there is no direct correlation with the size of crystallites.The selectivity of formation of higher hydrocarbons depends on the nature of the support and varies in the same order, while that of methane formation varies in the opposite direction. Thus, the nature of the support affects mainly the size of the crystallites and the heat of CO adsorption qCO, which in turn influence the reaction parameters. There are other factors, too. Thus, according to Huang and Schwarz,49, 107 with the same metal dispersity, the selectivity of CO hydrogenation depends on the acid-base properties of supports.This dependence can be deter- mined, for example, by the intermediate interaction of the surface complexes CHxOH bound through an oxygen atom to acidic Lewis centres (L) at the metal ¡À support interface According to this scheme, the selectivity of formation of higher hydrocarbons should depend on the strength of CO molecule bonding (qCO) and the energy of acid ¡À base interaction.The influence of the nature and acidity of supports on the properties of Ni-containing catalysts and the metal state in them is clearly manifested in the formation of industrially important bifunctional catalysts containing from 0.5 mass%to 2.0 mass% of Ni.120 ¡À 123 Depending on the differential acidity of supports, i.e., on the ratio of the Br��nsted (B) to Lewis (L) centres, catalysts are formed with different activity and selectivity in reactions of hydroisomerisation, aromatisation, and hydrogenolysis. The acidic properties of zeolite supports influence the dispersity of nickel, the extent of its reduction and the magnitude of the SMSI effect, which in turn affects the activity and selectivity of catalysts.Hoang et al.121 investigated the state of nickel in the Ni-c- ontaining catalysts by IR spectroscopy of adsorbed CO and NH3, thermoprogrammed reduction, chemisorption of H2 and O2, and thermoprogrammed desorption of NH3. When zeolite is impreg- nated with a nickel salt solution at pH 1.8, ion exchange of B- centres in the channels of zeolite occurs.The nickel fixed in these centres is difficult to reduce; it remains mainly in the NiO form and hence cannot be determined by IR spectroscopy of adsorbed CO. Impregnation of zeolite with a nickel salt solution at pH 4.8 results in the formation of metallic Ni localised mostly outside the channels.In these two cases, the Ni-zeolite catalysts manifest different activity and selectivity in reactions of aromatisation, dehydrocyclisation, and hydrogenolysis of n-hexane. The acidity of the support and concentration of B- and L-centres also influence the state of Ni in a reduced catalyst (Table 7). In absence of L-centres on the surface (specimen I), NiO particles are formed that interact strongly with the support, the extent of their reduction being a minimum.The presence of even a small number of L-centres (specimen II) leads to the formation of reduced particles Ni0, which are less dispersed than those in specimen III with the maximum quantity of L-centres. The specimen III contains highly dispersed Ni particles. In all cases, part of the Ni interacts with the B-centres to form presumably Ni(OH)+ ions, which are not reduced to Ni0.121 Similar results were reported by Romero et al.120 who used physicochemical methods to study the metal incorporation into zeolite.The use of a number of additives, e.g., nickel acetylacet- onate, and impregnation of zeolite with a Ni salt solution produces catalysts possessing different acidity, which influences the metal state, activity and selectivity of contacts in the process of hydrocracking of n-paraffins.Analogous conclusions about the effect of the character of distribution of acidic centres over the support with regard to the strength of these centres on the metal state, activity, and selectivity were drawn from a number of studies dealing with the preparation of Pt, Pd, and Ni catalysts on zeolite supports.124 ¡À 127 These catalysts are applied for dearomatisation in installations for hydrorefining of diesel fuels.40 According to Hoang et al.,121 the acidic function of the support controls the formation of the active surface of nickel.Impregnation Ni catalysts are less active than those obtained by ion exchange, since the extent of Ni dispersity is lower in the former case.The use of a set of physicochemical methods (IR spectroscopy of adsorbed CO, ferromagnetic resonance, chemisorption of O2 and H2, thermoprogrammed desorption of NH3) made it possible to reveal the effect of strong interaction of Ni with zeolite supports of the Y type characterised by the different extent of Na+ exchange for NH4 + ions and hence by different total acidity.127 With the increase in the extent of ion exchange, the total acidity increases with simultaneous appearance of a trend for a larger size of Ni particles and a smaller number of d-electrons per nickel atom.The results obtained allowed one to make a conclusion on the formation of nickel clusters on the surface of zeolite supports and a substantial role of electronic factors in this process. By control- ling the formation of clusters, including SMSI regulation, one may prepare highly efficient Ni catalysts with different active centres enabling the occurrence of specific reactions involving H2.Studies of the effect of the surface nature, porous structure and support crystallinity on the formation and properties of Ni catalysts are of considerable interest.49 Table 8 presents the main characteristics of Al2O3 specimens of different origin (Degussa, Engelhardt, and Kaiser products) used to prepare impregnation catalysts of composition 6%Ni/Al2O3.The preparation of catalysts involves several stages. In the first stage, the support is brought into contact with an electrolyte, which results in adsorption of precursors, most frequently by ionic adsorption. There are two ways of cation adsorption on the support surface.If the support is protonated, i.e., it contains no heterometallic ions, the ion adsorption may be realised by treating the support with a solution containing metal ions at a sufficiently high pH. Then the surface H+ ions are exchanged for metal ions, e.g., Ni2+, following the scheme An increase in pH shifts the equilibrium to the right, and the concentration of metal ions may prove to be insufficient for precipitation of the metal in the form of hydroxide.In the first stage, upon contact with an Ni(NO3)2 solution at pH 5 the support is partially dissolved and interacts with the electrolyte, whose pH is thus changed, and a portion of Al3+ ions H C OH + M L H H O L (9) C M MH H C M C OH M L 7H2O MH H3C C OH M L C2H6+H2O. 2Al3+OH¡¦s +Ni2�¢ aq +2OH¡¦aq [Al3+O27]2Ni2�¢ s +2H2O. Table 7. Effect of the acidity of the support on the state of nickel in a Ni/HZSM-5 catalyst after reduction.121 Speci- [Ni] Concentration Extent of Ni dis- SNi , men (mass %) of acidic centres reduction persity /m2 - g71 of Ni (%) (%) Ba Lb I 0.5 1.8 0 3 ¡À ¡À II 0.5 1.6 0.09 75 6.2 0.15 III 2.0 1.1 0.38 75 4.8 0.48 a Determined from the absorption band of adsorbed NH3 (1460 cm71) in the IR spectrum. b Determined from the absorption band of adsorbed NH3 (1) in the IR spectrum. 596 M D Navalikhina, O V Krylovpass into solution. Here, ion exchange can take place with participation of both the OH groups and the surface cations.The hydroxy groups are donors of protons, while the hydrolysed cations are donors of hydroxyls. The departing Al3+ cations are substituted by hydrogen as a result of hydrolysis. Subsequently, the Al3+ ions are readsorbed. A mixed adsorption leads to the formation of nickel aluminate on the surface. As is seen in Table 8, the crystallinity (ordering) of the structure of Al2O3 and its solubility in the electrolyte change in opposite directions: solubilisation of a support with a deformed lattice occurs more easily.Transition of Ni2+ ions from solution to support may be explained by ion exchange with a neutral or negatively charged surface of Al2O3 and incorporation of Ni2+ into the deformed lattice of Al2O3. The ion exchange requires a definite pH at which Al2O3 is dissolved in the electrolyte.The larger the number of Al3+ ions passing into solution, the more deformed the support lattice, which facilitates incorporation of Ni2+ into the support. The most soluble Al2O3 of Kaiser Co. adsorbs the largest quantity of Ni2+, whereas the Al2O3 of Degussa Co. has the lowest solubility and adsorbs the smallest quantity of Ni2+. The order of Ni2+ adsorbability on various supports is the reverse to the order of their crystallinity. The efficiency of a catalyst is substantially dependent not only on the quantity of adsorbed Ni2+, but also on the size of its active surface and the structure of the latter.The catalysts of the Kaiser and Engelhardt companies have similar pore sizes. However, according to Ghen et al.,49 the adsorption of Ni2+ from aqueous solutions and treatment with HNO3 do not affect the structure of the support of Kaiser Co., but modify the structure of Al2O3 of Engelhardt Co.Thermoprogrammed reduction (TPRd) of catalysts with hydrogen reveals differences between the porous (Engelhardt and Kaiser) and non-porous (Degussa) specimens. The TPRd spectrum of the Ni catalyst on the Degussa support has no high- temperature peak, whereas the catalysts on the porous supports are characterised by two peaks (at 141 ± 149 and 223 ± 230 8C), though their ratio for the two above-mentioned specimens is not the same.It may be suggested that the catalyst on the Degussa support possesses only the exchangeable surface Ni and centres of the NiO type. The presence of two peaks in the spectrum of TPRd on this catalyst is indicative of the presence on the surface of reduced nickel (Ni0) and nickel aluminate, i.e.a surface compound carrying incorporated Ni. Al2O3 of Kaiser Co. has very little surface Ni and much incorporated Ni in comparison with Al2O3 of Engelhardt Co. The quantity of surface (exchangeable) nickel decreases in the order Degussa>Engelhardt>Kaiser, which correlates with the data on crystallinity, solubility, and adsorp- tion.The porous structure of the support was shown 128 to influence the extent of metal reduction. On the Al2O3 support with a smaller pore size, the mean size of metal particles and the quantity of H2 sorbed in the low-temperature range are larger. The readily reducible nickel oxides are believed 129 to be localised in small pores, while the oxides that are difficult to reduce occupy large pores.The oxides of the first type are close in their structure to the bulk-distributed NiO and interact very weakly with the support. The reduction-resistant oxides may occur as small crystallites and as surface compounds. The next section is concerned with the effect of the porous structure of supports on physicochemical and catalytic properties of modified nickel catalysts studied by one of the authors of this review.There are two peaks in the TPR spectra of CO for all reduced samples. It is supposed that the low-temperature peak corre- sponds to theCOadsorbed on Ni,49 whereas the high-temperature peak is associated with the CO spillover on Al2O3 or with its adsorption on a surface nickel compound. In the spectrum of the catalyst on the porous support of Kaiser Co.with an abnormally small specific surface of nickel (SNi=2.17 m2 g71), the low- temperature peak is smaller than that in the spectrum of the catalyst on the support of Engelhardt Co. (14.89 m2 g71). The low-temperature peak is larger (14.22 m2 g71) in the case of the sample on the non-porous support of Degussa Co.Thus, by selecting the type and porosity of a support, one can control the formation of precursors of active Ni centres, the extent of metal reduction, and hence the activity and selectivity of a catalyst in specific hydrogenation reactions, and regulate thermo- stability and physicochemical and catalytic properties of impreg- nation catalysts. 4. Effect of modifiers and promoters on the properties of catalysts The mode of support preparation and its nature substantially influence the activity, selectivity and physicochemical properties of supported catalysts; however, the strongest effect is achieved by the use of promoters and modifiers. According to Roginskii,130 modification, unlike promotion, is `a more complicated phenomenon associated with more profound and complex changes in the catalytic properties and specific action of additives' the introduction of which induces `not merely changes in the absolute magnitude of activity, but also a profound modification of the surface quality'.This can result in a change in the reaction route, i.e., not only in the catalyst activity, but also in its selectivity.This concerns, for example, the modified Ni catalysts of hydrogenation, if the additives, as shown below, `can react with the surface layer of the solid'. The application of modern physicochemical methods allows a scientifically sound selection of modifiers and promoters for preparing efficient hydrogenation catalysts for any purpose. At present, Ni-containing catalysts of hydrogenation are prepared mostly on supports, such as Al2O3, SiO2, aluminosili- cates, zeolites, etc., which are modified by oxides.5, 12, 20, 116, 131 Modifiers, which are for the most part inorganic additives (SiO2, zeolites, Mg, Zr, and Ti oxides, boric and phosphoric acids), are introduced into supports by precipitation or impregnation tech- niques.This results in the enhanced thermostability and mechan- ical strength of the support.The activity and selectivity of the catalyst change insignificantly when the additives introduced do not affect the extent of nickel dispersity, the degree of its reduction and the magnitude of SMSI. The fact that thermal treatment influences, as has been shown above, the dispersity and structure of supported Ni catalysts makes topical the enhancement of thermostability by modifica- tion.For example, addition of 3%± 10% of BaO to g-Al2O3 inhibits sintering and improves thermostability of the support.12 Table 8. Effect of the type of support on pH of the impregnation solution and Ni adsorption from solution.49 Type of Ssp X-Ray crystallinity pH at zero Pore volume Al3+ con- Solubility Ni adsorption support /m2 g71 /rel.units charge (from N2) centration /rel. units /cm3 g71 /ppm a (see b) g (g Al2O3)71 /rel. units Degussa 97 1.8 5.2 0.003 0.26 1.00 3.25 . 1074 1.0 Engelhardt 195 1.4 6.8 0.54 0.84 13.51 4.39 . 1073 13.5 Kaiser 200 1.0 9.3 0.5 2.87 35.08 1.14 . 1072 35.0 a Concentration of Al3+ in the electrolyte formed upon Al2O3 solubilisation and ion exchange with Ni2+.b Solubility of Al2O3 in an Ni(NO3)2 solution at pH 5.0. Heterogeneous catalysts of hydrogenation 597Addition of phosphoric acid to aluminium hydroxide favours an increase in the total porosity and the proportion of micropores. Other modifiers are also used. Stabilisation of the existing Ni-containing centres, and crea- tion of newer ones, often requires that considerable quantities of modifying oxide additives (up to 20 mass%± 40 mass %) be incorporated into the catalyst.12, 132 Thus, the addition of 5% of B2O3 to the NiO ±Al2O3 system leads solely to a lower concen- tration of Ni3+ ions that are not reduced to the metallic state.Addition of 10% of B2O3 precludes the formation of Ni3+ ions and stabilises Ni2+ ions (a sample with 3%Ni/Al2O3), though even in this case a large portion of nickel remains in the form of surface oxide associates, whose reduction to the metallic state is problematic.An analogous situation is observed in the modification of g-Al2O3 with oxides, such as WO3 and Cr2O3, and zeolites. With hydrogenolysis of thiophene as an example Surin et al.133 described the promoting effect of zeolite on the hydrotreating catalyst, Ni ± Mo/Al2O3.Studies involving a set of techniques (DTA, TG, EPR, FMR) showed that zeolite forms thermostable compounds with molybdenum and interacts with g-Al2O3. Intro- duction of zeolite in the Ni7Mo catalyst favours more uniform distribution of acidic centres, which exerts a substantial effect on the formation of an active hydrogenation phase: a more dispersed phase is formed upon reduction of nickel molybdates.In recent years there has been a much greater use of nickel catalysts modified by surface hydrides formed in the process of catalyst preparation 13 ± 15, 134 and of the above-mentioned sorp- tional catalysts. The latter catalysts are highly dispersed, thermo- stable, and very active in hydrogenation processes, although the method of their preparation is not well-suited to industrial conditions because it includes decomposition on supports of organometallic compounds of elements belonging to Groups IV ± VII (Mo, W, Re, and Sn) or Group VIII.18 ± 22, 110, 116, 134 The most probable reason for the enhanced thermostability is the interaction of the active component with low-valent ions to form metal ± metal bonds between cations of these modifiers and the metal of organometallic compound:135 where M(p-C4H7)2 is the organometallic compound; M0 is the metal; Ex+ is the modifying element of Group IV ± VII.Certain procedures developed for preparing sorptional cata- lysts, including nickel-containing ones, make it possible to reach a degree of dispersion close to the atomic state.18 ± 22 Slinkin et al.116 described a process for preparing nickel catalysts based on ternary oxide systems Ni ± SiO2(MO2) synthesised by the alkoxide tech- nology.Catalysts prepared by precipitation from alcoholic or aqueous solutions and promoted with Ti, Zr, and Ce oxides were used for comparison. The extent of nickel dispersion in these catalysts and others obtained from the ternary systems using organometallic compounds proved to be much the same. How- ever, due to the absence of a SMSI effect, the activity of the precipitated catalysts 7%Ni/SiO2 ± 10%TiO2(ZrO2) in the hydro- genation of benzene proved to be one order of magnitude higher than that of the systems synthesised by chemical mixing, where the SMSI effect takes place.In their capacity to stabilise high- dispersity crystals of nickel, the modifiers form the following series:116 CeO2>ZrO2>TiO2.A highly dispersed Ni catalyst for benzene hydrogenation was prepared by modifying the support surface with an organolithium compound from methanolic solution:18 In accordance with this scheme the so-called grafted Ni/SiO2 catalysts with 0.8 mass% ± 4.6 mass% of metal were prepared. Physicochemical and catalytic studies of such catalysts showed that they possess a number of advantages compared to the impregnation systems of similar composition. The degree of reduction of the grafted catalysts at 250 ± 400 8C does not reach 100%, and the size of Ni particles in them ranges from 4.8 to 8.5 nm, while it is 12.5 nm in the impregnation catalysts with 100% reduction.The maximum activity of grafted catalysts in the reaction of benzene hydrogenation is observed at 350 ± 360 8C, which is 12 ± 15 times larger than that of impregnation systems.18 The drawbacks of the method of support modification with organometallic compounds include a narrow set of compounds possessing satisfactory solubility even in organic solvents, com- plexity of their synthesis, high explosive tendency and fire hazard of the process, and the need for special pretreatment of supports.13 However, this method makes it possible to modify the support surface at the atomic level and to form active metal centres with a predetermined environment.In our opinion, a method for preparing low-percentage Ni catalysts of hydrogenation by impregnation with mixed inorganic complex salts from aqueous solutions is more promising.136 ± 145 In all cases, the role of modifier or active phase is played by heteropolycompounds (HPC) with anions of the general formula [HwAxByOz]n7, where A=Ni or Co; B=Mo, W, or V.Their decomposition yields supported Co7Mo, Ni7W, and Ni7Mo catalysts;146 they may also be used as modifiers of Ni-containing catalysts.136 ± 145 Heteropolycompounds with anions [NiM6O24H6]47 (M=Mo or W) were synthesised to prepare hydrogenation catalysts of this type aimed at modelling the known Co7Mo, Co7W and Ni7W catalysts of hydrotreating.146 Their decom- position on the surface yielded the active phase.Previously, this type of salt and acid were mostly applied as catalysts of oxidation, hydration ± dehydration and some other processes.147 ± 151 Supported Ni7Moand Ni7Wcatalysts prepared by impreg- nation of their support with heteropolycompounds were tested in reactions of hydrodesulfurisation and hydrogenation.146, 152 They were found to be less active than the industrial Ni7Mo and Ni7W catalysts.This is rationalised by the lower content of Ni which cannot be increased because in this case HPC are used as an ordinary active phase.In the 1980s, one of the authors of this review designed nickel catalysts of a fundamentally new type on different supports (activated carbon, Al2O3, and SiO2) in which HPC of different composition (HPC1, HPC2, HPC3, etc.) play the role of modi- fiers.136 ± 143 They were tested in a broad spectrum of reactions (hydrogenation of alkenes, aromatic compounds, aldehydes, ketones, and CO; hydrogenation of coal; hydrodemethylation of toluene; hydrocracking of petrol fractions; dehydrogenation of cycloalkanes).It was shown that they possess high and control- lable activity and selectivity. Common distinctive features of these catalysts are low metal content (2% ± 6% Ni), modifying salts on the surface of a support and a well-developed surface of metallic nickel with high dispersity.These catalysts exhibit high thermo- stability, and the procedures for their preparation match techno- logical demands. Table 9 presents values of the specific surface of nickel (SNi), its dispersity (d Ni), and the degree of reduction (a) in modified catalysts of the new type on g-Al2O3 with different HPC modifiers and, for comparison, in some industrial Ni catalysts of hydro- genation.The extent of reduction of modified samples reaches high values, despite the fact that the reduction conditions are consid- erably milder (400 8C, 6 h) than, for example, those for the Euro- Ni-1 catalyst.15 In some cases, the a value for modified catalysts with low Ni content (e.g., 2%Ni7HPC1) reaches 100%.As noted above, this index is not reached for non-modified Ni/Al2O3 catalysts, since in these a fraction of the Ni is incorporated into the lattice forming a spinel. (O)x . Ex+M(p-C4H7)2 (O)x . Ex+M(p-C4H7)2, (O)x . Ex++M(p-C4H7)2 (O)x . Ex+M0+2C4H8, OH OH+CH3Li OH O NiCl O O Ni OLi OLi OLi NiCl2 7LiCl 7CH4 598 M D Navalikhina, O V KrylovThe HPC-modified catalysts with 2%± 4%of nickel manifest enhanced activity in the hydrogenation of mono- and bi-cyclic aromatic hydrocarbons, and their selectivity reaches 100%.Hydrogenation of toluene on HPC-modified Ni catalysts with 2%± 4%of Ni occurs considerably more easily than on industrial Ni7Cr catalysts with 50% of Ni (Fig. 1).138 The catalyst Ni7HPC1/Al2O3 with 2% of Ni is more active at 3 MPa and 220 8C than its analogue with 4% of Ni, and complete hydrogenation of benzene on this catalyst occurs within a contact time less than 0.1 h.140 These results confirm direct participation of HPC in the formation of active centres of hydrogenation. This follows in particular from the data on the formation of HPC oxometallates, and the formation of W-containing hydrogen bronze upon reduction of highly dispersed particles in the Pt/WO3 sys- tem.141, 150, 152 Preparation of modified Ni7HPC catalysts on carbon sup- ports is accompanied by the formation of an active phase in which Ni particles are distributed uniformly.Figure 2 presents micro- photographs of modified and non-modified catalysts on a carbon support with a specific surface of 600 m2 g71.In the absence of HPC the Ni particles are nonuniformly distributed, and their size exceeds that of particles in the modified catalyst.138 In the catalyst 4%Ni7HPC/C reduced at 300 ± 420 8C, the size of particles can be as small as 5 ± 7 A, and in the catalyst on g-Al2O3 they can be even smaller. To achieve optimum dispersity of Ni (5 ± 7 A) and maximum reduction, it is necessary to have specific HPC concen- tration and a Ni :HPC ratio on each support.Complete reduction of Ni is possible at a low metal content (2% ± 4%). In this case, nickel virtually does not interact with g-Al2O3 since it is distrib- uted on the modifier surface. Catalysts Ni/g-Al2O3 containing less than 5% of metal are virtually inactive in hydrogenation of benzene because of the difficulty in reducing NiO that is in contact with Al2O3.Modification of the Ni/Al2O3 catalyst containing 4% of Ni with heteropolycompounds shifts the range of CO hydrogenation towards higher temperatures (250 ± 400 8C) and substantially diminishes the CO2 concentration in the products on total CO conversion.139 The transition of centres of type A (Ni/NiO) to centres of type B (Ni/NiO) ensures selective hydrogenation, which occurs at a lower temperature.In the absence of a modifier, the centres of the B type are possibly not formed at all. The assumption of the different structure of Ni centres on modified and non-modified catalysts was confirmed by the analysis of electronic spectra of diffuse reflection. It was shown that Ni catalysts modified by HPC have two types of Ni2+ ions with different crystalline fields, while Ni3+ ions are absent.Spectral parameters do not correspond to the compound NiAl2O4, which shows that the modifier is involved in the distribution of nickel and that the direct interaction between Ni and Al2O3 is absent. The B2O3 modifier added in quantities considerably larger (up to 10 mass%± 20 mass %) than HPC influences the thermo- stability of catalysts since it destroys the Ni7O7Ni clusters, inhibits the formation of spinel with Al2O3, and stabilises the Ni2+ ions on the surface.132 Addition of B2O3 to the Ni/Al2O3 system strongly decreases the concentration of Ni3+ ions, which form the oxide associates Ni7O7Ni and alters the acid-base properties of a catalyst.Heteropolycompounds serve not only as chemical modifiers, they also influence the microporous structure of catalysts.Table 9. Specific surface of nickel, its dispersity, and the extent of reduction in HPC-modified and some industrial catalysts. Catalyst [Ni] SNi dNi a a (%) (%) /m2 (g cat- /A alyst)71 Ni7HPC1/Al2O3 4 220 26 70 Ni7HPC2/Al2O3 4 334 17 87 Ni7HPC3/Al2O3 4 56 100 87 Ni7Cr 48 ± 63 25 ± 27 Euro-Ni-1b 25 ± 52 181 48 ± 56 85 ± 96 Ni/Al2O3 (Exson) 8 ± 10 197 34 a Found from H2 adsorption.b Reduction for 10 ± 26 h at 630 8C. 1 2 3 4 0.4 0.8 1.2 1.6 t /h 100 80 60 40 20 0 x (%) Figure 1. The extent of toluene hydrogenation vs. time on an industrial Ni7Cr catalyst (1) and modified Ni7HPC2/Al2O3 catalysts with 2 mass %, 3 mass %, and 4 mass% of Ni [(2) ± (4), respectively].b Figure 2. Microphotographs of reduced (at 400 8C) catalysts containing 4% of Ni on carbon supports in the absence of modifier (a) and the presence of HPC1 (b). 0.1 mm a Heterogeneous catalysts of hydrogenation 599Figure 3 shows changes in the porous structure of g-Al2O3 after introduction of HPC in the preparation of a Ni hydrogenation catalyst. The presence of micropores 20 ± 30 A in diameter influ- ences rather favourably the activity of catalysts for the hydrogenation of aromatic hydrocarbons.The micropores are the sites where readily reducible nickel oxides are formed, after activation with hydrogen the oxides are transformed into highly dispersed metal particles, which ensure rapid hydrogenation of benzene. By modifying the porous support structure with HPC, one can prepare active and selective catalysts for the hydro- genation of aromatic and carbonyl compounds or CO methana- tion.Table 10 presents some characteristics of HPC-modified cat- alysts with the same content of nickel and supports used for their preparation by impregnation. The supports were g-Al2O3 with a predominant pore size of 9 ± 12 nm (Al2O3-1) or Al2O3 of the X type (X1, X2, X3) with a pore size of 1 ± 3 nm.Figure 4 shows the dependence of changes in the volume of pores (V) as a function of their radius (r) for the supports of the X type and Al2O3-1. The data presented in Table 10 show that the catalysts based on supports with similar total porosity (specimens 1, 4 and 8) and volume of micropores (specimens 1 and 4) exhibit different activities.The catalyst No. 9 prepared on the support with the largest volume of micropores is the most active in hydrogenation of benzene but is little active in the hydrogenation of acetone. On the contrary, specimens 2 and 3 prepared on Al2O3-1 are not very active in the hydrogenation of benzene and highly active in the hydrogenation of acetone. The modified Ni catalysts with a well- developed microstructure (the predominant pore size 1 ± 3 nm) were shown to have active centres of benzene hydrogenation and those with large pores (9 ± 12 nm) � to contain active centres of acetone hydrogenation.Dehydration of modified samples under conditions of ther- moprogrammed heating in the temperature range 50 ± 700 8C yields water as the sole product.Modification with HPC substan- tially alters the properties of the initial surface of Al2O3. Specif- ically, hydrophilicity of the catalysts prepared was found to be higher than that of initial Al2O3 over all the measurement range. After deposition of Ni on modified Al2O3, the liberation of adsorbed water in the thermolysis takes place in the temperature intervals 50 ± 100 8C, 200 ± 300 8C, *400 8C, 500 ± 600 8C, and even near 700 8C.The most actively modified Ni catalysts possess an increased capacity for H2O (12% and more). They are characterised by the presence of water removable only in the high-temperature range.153 IV. Catalysts based on noble metals The Ni-containing catalysts are less expensive than those based on noble metals (Pt, Pd, Ru, Rh, Ir, and Os).However, for a number of industrial hydrogenation processes preference is given to the latter because of their high catalytic activity allowing operations at lower temperature and pressure. The lifetime of these catalysts often extends to several years. Pd, Pt, Ru, and Rh catalysts and 2 1 dV dr 8 6 4 2 0 7.5 15 r /nm Figure 3. Changes in the porous support structure after formation of the active phase of Ni catalyst in the presence of HPC: (1) g-Al2O3 support after thermal treatment; (2) Ni7HPC1/Al2O3 catalyst. Table 10.Effect of the porous structure of supports on the activity of Ni catalysts in the hydrogenation of benzene and acetone. Entry Catalyst or support Ssp /m2 g71 Total porosity Microporosity Rate constant of Degree of acetone /cm3 g71 /cm3 g71 benzene hydrogenation conversion (%) /h71 cm73 1 Al2 O3-1 170 0.59 0.10 ± ± 2 Ni7HPC2/Al2O3-1 ± ± ± 0.40 80 3 Ni7HPC5/Al2O3-1 ± ± ± 0.12 60 4 X1 193 0.42 0.11 ± ± 5 Ni7HPC2/X1 ± ± ± 0.86 ± 6 X2 214 0.45 0.14 ± ± 7 Ni7HPC2/X2 ± ± ± 1.16 20 8 X3 201 0.51 0.20 ± ± 9 Ni7HPC2/X3 ± ± ± 1.44 12 60 120 180 r /A 15 5 7 3 1 10 5 dV dr a b Figure 4.Distribution of the volume of pores in the supports g-Al2O3 of the X type (a) and Al2O3-1 (b). 600 M D Navalikhina, O V Krylovbimetallic systems are of special interest. The use of special supports makes it posble to bring the content of noble metals down to 0.01% ± 0.005% and thus make catalysts less expensive, while preserving their qualities.154 ± 157 Modern physicochemical methods allow the investigation into the mechanism of action of these catalysts, optimisation of their composition, and development of methods for controlling the metal dispersity on the support surface.Some catalysts based on noble metals are resistant to poisoning with S- and N-containing compounds.158, 159 1. Hydrogenation of CO A number of studies (see, e.g., Refs 160�163) showed that supported Pt, Pd, Ir, Ru, and Rh catalysts allow one to bring about the methanation of CO under atmospheric pressure.Hydrogenation of CO under elevated pressure in the presence of Pt/Al2O3, Pd/Al2O3, and especially Ru/Al2O3 and Rh/Al2O3 (in contrast to Ni-containing catalysts) leads to the formation of alcohols and other oxygenated compounds 164 ± 166 CO+2H2=CH3OH, 2CO+4H2=C2H5OH +H2O, 2CO+3H2=CH3CHO+H2O.The Fischer ± Tropsch synthesis may be also carried out on Ru/Al2O3 and Ru catalysts on specially selected supports to produce higher (C5±C20) hydrocarbons (with selectivity up to 81%) as well as methane, C1±C5 alcohols, and C21±C40 hydro- carbons.167 ± 170 nCO+2nH2 [7CH27]n+nH2O, 2nCO+nH2 [7CH27]n+nCO2 . Industrial processes (Sasol-II, SMDS, and Gulf Badger) on Ru-containing catalysts allow selective formation of a diesel fraction of synthetic liquid fuel or higher paraffins the cracking of which affords high-quality diesel and rocket fuels.167, 170 During the period 1987 ± 1990, the catalytic activity of plati- num group metals on supports in the reaction of CO hydro- genation, its mechanism and kinetics were systematically studied.117, 118, 163 The dispersities (mean size of metal particles, dmean) of the catalysts studied, their activities (k), and selectivities are presented in Table 11.As seen in Table 11, the specific catalytic activity decreases and the selectivity increases in the series Ru ? Rh ? Ni ? Pt ? Ir?Pd. The catalytic activity decreases as qCO increases.163 The qCO values determined from kinetic data increase in the following series: Ru<Rh<Ni<Pt Ir<Pd.161, 171 On the whole, the specific catalytic activity of platinum group metals and nickel is lower the larger the size of the metal particles (cf.Table 11). The effect of the phase structure of the support (Al2O3) was studied for Rh catalysts.172 Rhodium was deposited on d-, y-, and a-Al2O3 with specific surfaces from 5 to 90 m2 g71.In these catalysts, rhodium was found to be present in three forms: encapsulated, strongly interacting, and non-interacting with the support. The maximum extent of dispersity (11%) was reached on a-Al2O3 at low Rh content (<1 mmol m72). At higher Rh content (>1 mmol m72), the extent of dispersity was below 3%. The activity of a catalyst is enhanced if a-Al2O3 is used as a support, if Al2O3 is treated with water vapour prior to Rh deposition, and if the content of deposited rhodium is low. 2. Hydrogenation of carbonyl compounds The similarity of the processes underlying hydrogenation of CO and carbonyl compounds on platinum group metals was reported in different studies.161, 167 ± 171 The character of the dependence of specific catalytic activity on the heat of adsorption of the reactant proved to be nearly the same in both cases.The same holds for the influence of dispersity: the heat of adsorption of reactants increases with a decrease in the size of crystals leading to a lower catalytic activity. The mechanisms underlying the hydrogenation of CO and acetone were supposed to be similar (the Langmuir ± Hinshelwood mechanism): 1) CO+Z ZCO, 2) H2+2Z 2ZH, 3) ZCO+ZH ZCOH+Z, 4) ZCOH+ZH hydrogenation products and 10) (CH3)2CO+Za Za .OC(CH3)2 , 20) H2+2ZH 2ZHH, 30) Za . OC(CH3)2+ZHH Za .HOC(CH3)2+ZH , 40) Za .HOC(CH3)2+ZHH hydrogenation products (Z, Za, and ZH are the catalyst active centres). The limiting stages are the addition of the first hydrogen atom to an adsorbed CO molecule or to the CO group of acetone, respectively, with the formation of a `semi-hydrogenated' reactant form.However, certain differences also exist: in the CO hydro- genation, the most active catalyst is Ru, while the hydrogenation of carbonyl compounds is best catalysed by Pt. Some investigators believe 167 ± 171 that the reason for this difference consists in the stronger binding of C andOatoms (q CO) in a complex of the type compared to the carbonyl compound binding (qCO) in the com- plex and thus q CO, opt<qCO, opt.According to approximate esti- mates, q CO, opt'20 kJ mol71 and qCO, opt'100 kJ mol71 . These differences are reflected in the series of catalytic activity and selectivity. Table 12 presents data on the dispersity of metals of the platinum group and nickel deposited on Al2O3 and their specific activities in the acetone hydrogenation. The specific activities of catalysts with similar metal dispersion decreases in the order: Pt>Ir>Rh>Ni>Ru>Pd.The position of indi- vidual metals in this series depends on both their chemical nature and the dispersity. Hydrogenation of the carbonyl group of a ketone is a structure-sensitive reaction. This means that the probability of multipoint adsorption of complex molecules and the rate of their hydrogenation depend on the size, shape, and arrangement of microcrystals on the support.Possibly, the adjacent sites of the support surface take part in the substrate activation due to hydrogen spillover.165 C O C O Table 11. Certain characteristics of different metal catalysts for the CO methanation (pCO=20 kPa, pH2=50 kPa, T=275 8C).Catalyst dmean 107 k Selectivity of /nm /mol m72 s71 methane formation (%) Ru/Al2O3 4.4 6.7 87 Rh/Al2O3 2.5 5.5 94 Ni/Al2O3 2.4 4.7 96 Pt/Al2O3 3.9 0.4 99 Ir/Al2O3 6.8 0.2 100 Pd/Al2O3 5.0 0.1 100 Heterogeneous catalysts of hydrogenation 6013. Hydrogenation of unsaturated hydrocarbons In some cases, noble metals rather than nickel are preferred as catalysts for the hydrorefining of motor fuels (dearomatisation, hydrogenation of alkenes).29, 40 Catalysts based on noble metals are more active and more resistant to poisoning by sulfur compounds.Platinum and palladium deposited on Al2O3 are easily deactivated at a sulfur concentration below 1 ppm;173 however, their resistance to sulfur poisoning increases on acidic supports, such as zeolites.158, 159 For instance, 98% of the products of dodecane hydrocracking are alkanes on Pd/LaY under H2S pressure of 63 kPa, i.e. when all palladium would be expected to pass to PdS; the catalyst activity and selectivity remain virtually unaltered.159 Such an unusual behaviour of this noble metal is explained by the SMSI effect or changes in the electronic state of palladium upon its interaction with the acidic centres of zeolite.158 Platinum catalysts are prepared by ion exchange or impreg- nation of the support with a solution of Pt(NH3)4Cl2; its cation binds to an acidic centre of zeolite:Al7O7. The next important stage consists of the formation of centres 2(:Al7O7)Pt2+ upon calcination in air.On subsequent reduction with hydrogen, platinum becomes involved in clusters of the type m(:Al7O7H) .. . Ptn. The clusters are in close contact with strong acidic centres, which makes possible the transfer of electrons from this noble metal and leads to the formation of electron-deficient metal particles (Ptd+), which are responsible for the specific properties of this catalyst.174 In most cases, hydrogenation is conducted on supported palladium catalysts.Table 13 presents the data from studies of the hydrogenation reactions of CO, aldehydes, and unsaturated hydrocarbons on Pd/TiO2 catalysts prepared by different methods and differing in the size of particles and the active surface of the metal.176 The data in Table 13 indicate that the rates of CO hydro- genation and low-temperature hydrogenation of propionaldehyde to propanol increase with the Pd dispersity.However, in the hydrogenation of the C=C double bonds in ethylene, butadiene, and acrolein the reverse order of activity is observed, being paicularly noticeable in the hydrogenation of the conjugated bonds of butadiene. The relationship between the catalytic properties of supported palladium and its dispersity is widely discussed in the litera- ture.174 ± 184 In some reports, this phenomenon is explained by changes in the electronic state of Pd resulting from the SMSI effect,175, 179 while in others, by the participation of both the catalyst and the support in the reaction.180 Table 13 shows three Pd catalysts on the same support (TiO2), which possess different activities and selectivities.Therefore, according to Barbier et al.,173 the reason for the differences should be sought in the properties of the particles of palladium itself: in variations of their size, shape, electronic structure, and the nature of their interaction with the molecules subject to hydrogenation. For example, in the adsorption of CO an electron is transferred from Pd to CO, while in the adsorption of ethylene it is transferred in the opposite direction.This results in inverse dependences of the catalytic activity of Pd/TiO2 on the degree of metal dispersity in different hydrogenation reactions. Highly dispersed palladium is more active in the hydrogena- tion of electron-acceptor compounds, whereas the low dispersed palladium is more active in the hydrogenation of electron-donor compounds.As was shown above (see Table 11), Pd/Al2O3 is the least active catalyst of CO methanation. The metal in bulk is also weakly active in this reaction.181 Consequently, in Pd/TiO2, palladium forms clusters or small particles which are endowed with special properties different from those of bulk Pd or Pd on Al2O3.Of all the catalysts listed in Table 13, the strongest Pd ± support interaction may be expected for the first specimen, which is characterised by a well-developed active surface and small particle size. In this case, an effect of the SMSI type is possible. 4. Bimetallic and modified catalysts Considerable interest arose in bimetallic catalysts on supports in connection with the development of sulfur-resistant hydrogena- tion catalysts based on noble metals.For example, combination of Pt and Pd deposited on zeolite is known 183, 184 to enhance sulfur tolerance, the optimum Pt : Pd ratio being equal to 1. The support used was zeolite of the Y type with a high Si/Al ratio (Shell), and mordenites, borosilicates, and Y- and b-zeolites (Amoco). The Haldor Topsùe company has developed two types of catalysts for the two-stage hydrotreating of petroleum fractions: Ni7Mo/Al2O3 for the first stage and TK908 containing a supported noble metal for the second stage.185 The latter was in operation for 5600 h and the extent of its deactivation was less than 1% after 1000 h, despite the fact that the content of sulfur in raw materials reached 1000 ppm and that of nitrogen reached 50 ppm.Ru7Pd catalysts prepared by impregnation of a support with salts of both metals are highly active. These catalysts were used to study the hydrogenation of benzene.186 Bastiaux et al.187 described in detail a new approach to the preparation of highly selective bimetallic catalysts for the hydro- genation of alkynes and dienes in the hydrorefining of pyrolysis gases and petroleum fractions. Industrial Pd catalysts were sub- stantially upgraded by promoting them with a second metal.The choice of the second metal promoter was made on the basis of modern theories of homogeneous catalysis and of current con- cepts of changes in ligand mobility in organometallic com- pounds.188 Alkanes, alkenes, and alkadienes localised on the surface of Pd catalysts may be regarded as ligands which form strong bonds with palladium in the absence of promoters, thus Table 12.Specific activity of some catalysts in the acetone hydrogenation reaction and the deposited metal dispersity. Catalyst dmean 107k qCO /nm /mol m72 s71 /kJ mol71 Pt/Al2O3 3.9 44 12 Ir/Al2O3 6.8 11.7 20 Rh/Al2O3 2.5 7.7 15 Ni/Al2O3 2.4 7.5 85 ± 25 Ru/Al2O3 4.4 700 30 Pd/Al2O3 5.0 0 ± Table 13.Mean activity (catalyst turnover) of different Pd/TiO2 catalysts in the hydrogenation of CO, aldehydes, and alkenes. 176 Catalyst characteristics Mean catalyst activity in hydrogenation [Pd] Ssp d H: Pd CO CO in C=O bond C2H4 C=C C4H6 (%) m2 g71 /nm (methana- CH3OH in propion- (80 8C) bond in (80 8C) tion) (230 8C) aldehyde acrolein (80 8C) (80 8C) 5.5 207 4.5 ± 6.5 0.32 ± 0.67 7.361074 4.361074 0.16 380 0.09 150 5.2 9 186 0.17 2.861074 4.061075 4.561073 500 ± 680 4.4 9 295 0.067 5.561074 0 4.561073 580 0.4 1.056103 602 M D Navalikhina, O V Krylovdecreasing the catalyst's activity.The strength of the alkyl ± metal bond may be reduced by the addition of an electron-donor compound or a metal promoter which favours higher selectivity of acetylene hydrogenation.187 New procedures for the prepara- tion of highly selective bimetallic hydrogenation catalysts with controlled dispersities were developed using recent progress in the field of inorganic synthesis and synthesis of organometallic compounds.189, 190 Higher yield of the target product and higher stability of catalysts are achieved by modification of the electronic properties of palladium by the second metal.This approach was employed to prepare the industrial catalyst LT-279 for the selective hydrogenation of acetylene in an ethane ± ethylene frac- tion.187 As noted above, the selectivity of acetylene hydrogenation may be considerably increased by preparing Pd catalysts on supports in the form of active film structures of the cermet type.99 In this case, the electronic properties of palladium are modified in a different way, namely, due to the formation of a surface catalyst layer with incorporated modifying oxide additives (Al2O3).To this end, a simple technique may be used, such as impregnation, which is inapplicable for the production of bimet- allic catalysts. The latter are synthesised by making recourse to a more complex technology using, e.g., organometallic compounds or vapour condensation of two metals on a support.187 Film-type Pd catalysts modified with Al2O3 in their surface layer contain 0.1% ± 0.3% of Pd. They are thermostable and may be subjected to oxidative regeneration.Metal particles are formed in them on the surface of g-Al2O3 supplemented with a thermo- stable alumina oxide in the ratio Pd :Al2O3 = 1 : 1.Alteration of Pd properties leads to higher catalytic activity in the reaction of acetylene hydrogenation, which allows one to conduct it at low temperature (30 8C), low H2 :C2H2 ratio (2 : 1), and high flow rates (up to 10 000 h71), with a substantial increase in the selectivity of C2H2 hydrogenation, i.e., the same effects as with bimetallic catalysts.Modification of film-type palladium catalysts is brought about by adding very small quantities of oxide (Pd :Al2O3=1 : 1), and the procedure for their preparation is much simpler than that of bimetallic catalysts. As for the selectivity of acetylene hydrogenation, the Al2O3- modified Pd (0.1%Pd) is more than twice as effective as its industrial analogue MA-15 (0.5%Pd).99 On a bimetallic LT-279 catalyst, transformation of acetylene increases by no more than 10%±20% compared to that on a common Pd hydrorefining catalyst.187 Selective hydrogenation of unsaturated hydrocarbons can be performed on new highly dispersed Pt/MoO3 and Pt/WO3 cata- lysts in which the metal particles interact on the surface with the modified support.152 The ultradispersed MoO3 andWO3 particles obtained in a plasma reactor are reduced in a stream of hydrogen at 300 8C.Subsequent interaction with Pt or Pd leads to hydrogen bronzes, which serve as hydrogen reservoirs for hydrogena- tion.191, 192 The hydrogen bronzes are formed upon deposition of 0.2% Pt on MoO3 or 0.1% Pt on WO3 owing to spillover of H2 from the metal centres to the support under the hydrogenation conditions.This results in changes in the specific surface and the size of particles that form the centres active in the hydrogenation of alkenes. The molybdenum and tungsten bronzes themselves (without hydrogen) can also catalyse ethylene hydrogenation.193 V. The mechanism of hydrogenation on metal catalysts The mechanism of catalytic hydrogenation, in particular the sequential formation of different surface compounds, has been reviewed recently.194 Hydrogenation is assumed to involve atomic hydrogen formed upon dissociation of H2 on metals, such as Ni, Pt, Pd, and others.This assumption is confirmed by an intense isotope exchange H2+D2 2HD and formation of monodeuterioethane C2H5D upon deuteriation of ethylene.On oxides, in particular on ZnO, deuterioethane C2H4D2 is formed from ethylene, which indicates the participa- tion of molecular hydrogen in this reaction. However, other data 195 point to the possibility of participation of both atomic and molecular hydrogen in the hydrogenation on metals. Numerous forms of adsorbed alkenes were detected on the surface of metals that were found to compete for the active centres.Strongly bound forms of the type CH37CH=M, CH37C:M, etc. are hydrogenated at high temperatures, whereas at low temperatures only weakly bound forms of the p-type and p-allyl complexes undergo hydrogenation. The hydrogenation rate may be varied by regulating the strength of alkene adsorption. For example, modification with alkylsilanes increases the activity of ZnO 20- and 150-times in the hydrogenation of butadiene and pentadiene, respectively.195 This is explained by a steric effect: bulky alkylsilanes prevent strong adsorption of dienes, but do not hinder hydrogen adsorption.It is suggested that the centres of hydrogen activation involved in hydrogenation are localised on the metal surface. Therefore, the larger the surface of a metal the higher should be its catalytic activity.If the size of metal crystallites has no effect on the specific activity of a catalyst, the catalysed reaction is called structure- insensitive. The hydrogenation of the C=C bond in alkenes is an example of such reactions. On the contrary, the hydrogenation rates of an aromatic ring, the C=O and C:C bonds, and conjugated compounds of the butadiene type depend on the metal dispersity and hence such hydrogenations are regarded as structure-sensitive reactions.Investigation of hydrogenation of cinnamaldehyde on Ru/ Al2O3 showed a decrease in the specific rate constant for theC=C bond hydrogenation with an increase in the metal dispersity.196 This indicates that this reaction requires a planar adsorption of substrate, e.g., in the form of a p-complex.At the same time, the rate constant for the C=O bond hydrogenation increased with increase in the metal dispersity. Evidently, this reaction requires point-type adsorption (Ru . . .O=C ). The adsorption coefficient of the C=O bond of cinnamaldehyde is independent of disper- sion, while the adsorption coefficient of the C=C bond decreases with an increase in the metal dispersity.Apparently, intermediate semi-hydrogenated forms can man- ifest properties of free radicals (C2H5 . from C2H4, C3H7 . from C3H6 , etc.).197 Small palladium particles with a size <2.5 nm (Pd/MgO, Pd/Al2O3, Pd/SiO2, Pd films and Pt black) are less active than larger particles,95, 198 ± 200 and they become totally inactive when their size is <1 nm.This is explained by the fact that very small particles cannot dissolve hydrogen. Sintering of particles leads to enhancement of the catalyst's activity and selectivity in the hydro- genation of unsaturated compounds. The hydrogen necessary for the hydrogenation can be accu- mulated not only on the surface but also in the bulk of the metal.The properties of the surface- and bulk-absorbed hydrogen differ considerably. For example, in the hydrogenation of CO and CO2 on Pd/Al2O3, Pd black, and Pd foil the bulk hydrogen favours formation of oxygenated compounds, CH3OH and CH3CHO, while the surface hydrogen favours formation of CH4 and other hydrocarbons.201 Other phases of a complex catalyst can also absorb hydrogen and deliver it rapidly to the compound subject to hydrogenation. Hydrogenation reactions often require the activation of both hydrogen and the compound to be hydrogenated.The latter process can be brought about on another phase. The unsaturated compound can also be localised on the metal, although covering of the surface of the metal prevents adsorption of H2 and hampers the delivery of hydrogen in quantities necessary for hydrogena- tion. Hydrogenation can occur as a result of alternating inter- actions of different components of a catalyst with hydrogen and the compound to be hydrogenated and hydrogen spillover from Heterogeneous catalysts of hydrogenation 603one component to another.The latest data on hydrogen spillover have been reviewed recently.202 Hydrogen exchange between the metal phase and the support occurs with participation of acidic centres of the support.This was proved by experiments on the poisoning of Pt/Al2O3 catalysts with ammonia under the conditions of butadiene hydrogenation.203 The results are interpreted by the authors as a redox process Pt0+(Al7OH) (Al7O)n7+Ptn++1 2H2 . The increase in the catalyst crystalline lattice parameter in the course of hydrogenation may be regarded as an indirect evidence of hydrogen dissolution.The use of neutron diffraction revealed the intercalation of hydrogen into the ZnO phase in CO hydro- genation on a Cu7ZnO catalyst.204 Such an intercalation leads to the formation of a solid solution of Cu in ZnO in which half of the copper ions are replaced by protons.Carbon monoxide is adsorbed on copper, while hydrogen, though also being activated on copper, is involved in the hydrogenation from the subsurface ZnO layer. Neutron diffraction also revealed the dissolution of hydrogen in the CuCrO2 catalyst.205 The defects &Cu2+ stabilise the dis- solved hydrogen. The catalyst has in its bulk two types of hydro- gen: atomic hydrogen H in the lattice interpoint space and hydrogen in the form of the OH groups resulting from Cu2+ exchange for H+.The intercalation of hydrogen into the lattice of CeMxO catalyst (M=Cu and Ni; 0<x<5) was proved by X-ray, XPES, EPR, and DTG methods.205 The intercalation of H atoms and reduction of Ce4+ to Ce3+ results in the extension of the CeO2 lattice: 2Ce4++2O27+H2 2Ce3++2OH7, 2OH7 H2O+O27+&, H2+O27+& OH7+H7, Ni2++O27+H2 Ni0+H2O+&, 2Ce4++Ni0 2Ce3++Ni2+. In the above consideration of Ni supported catalysts, we repeatedly mentioned the existence of two types of active centres on the surface: metallic (Ni) and oxidised (Ni2+) centres.Different centres are already formed in the preparation of a catalyst, e.g., upon incorporation of NiO into the lattice of g-Al2O3.Non- modified catalysts with a Ni content of 3.2% ± 4.4% were studied by the method of TPR of carbon monoxide.206 The degree of Ni reduction at 500 8C did not exceed 25%. Two types of CO adsorption were observed, which followed from the two peaks of methane formed upon hydrogenation. The methanation proceeds according to the following scheme: (values of E are given in kJ mol71). The conversion of CO (B) into CO (A) at temperatures below 250 8C proceeds very slowly, but upon completion of this tran- sition, CO (A) is rapidly hydrogenated at 142 ± 247 8C.Thus, the rate of methanation on low-percentage Ni/Al2O3 catalysts is determined by the rate of B?A conversion, i.e. of Ni atoms interacting with NiO, into the Ni centres on the metal surface.Hydrogenation of CO (B) requires a considerably higher activa- tion energy than that of CO (A). The presence of centres with Ni in different oxidation states is also characteristic of HPC-modified Ni catalysts on supports. We consider here a possible hydrogenation mechanism on such catalysts. The hydroxylic cover formed on an HPC-modified surface differs substantially from that of the initial support and a binary Ni catalyst.It is homogeneous and depends on the type and concentration of the modifier and metal concentration. Heteropolycompounds (HPC) are directly involved in the formation of metal-containing active centres. Judging from elec- tron microscopy (see Fig. 2), at the optimum HPC concentration the active phase is regularly distributed thus creating seeds of Ni microcrystal formation and stabilising the metal formed upon reduction.Nickel particles are distributed in the defect layer of the modifier where Ni2+ can also be present. They have no direct contact with the support and are exceptionally thermostable: the specific surface of the metal is not changed up to 600 8C. Hydro- gen is adsorbed on the Ni surface and diffuses to the modifier (M) surface following the spillover mechanism with the possible participation of the hydroxy groups of the modifier, as shown below (the arrows indicate the direction of spillover): An unsaturated organic compound seems to be adsorbed on the modifier phase.Thus, the functions of the two catalyst's phases viz., Ni and HPC, are different. Presumably, Ni2+ ions formed from the metal are incorporated into the HPC structure.The initial activation of hydrogen takes place on the Ni particles: 2Ni+H2 2NiH (or 2Ni+2H). Then a reversible reaction occurs. 2H+Ni2+ Ni+2H+. This results in the formation of very small Ni particles, which in some cases cannot be detected even by electron microscopy. HPC can interact with the organic compound with transfer of the accumulated hydrogen 2H++R+Ni RH2+Ni2+. Probably, the active hydrogenation centres of unsaturated compounds and aromatics differ from those of CO and carbonyl compounds.VI. Carbide catalysts Boudart and Levy 207 were among the first investigators to study the catalytic properties of carbides. They showed that Mo2C possessed an unusually high catalytic activity close to that of platinum in reactions of hydrogenation and isomerisation. Nearly the same activity was found in tungsten carbide.These com- pounds are formed upon reactions ofMoorWsalts or oxides with hydrocarbons. For example, a powder of tungsten carbide with a specific surface of 3.5 m2 g71 can be obtained in the reaction of WCl5 vapour with butane at 1.3 kPa and 1340 8C.208 On the surface of tugsten carbide there is polymeric carbon, which decreases its hydrogenation activity.Adsorption of oxygen facil- itates adsorption of H2 by inducing the formation of islets of tungsten bronze, HxWO3, on the surface which is possibly a hydrogen reservoir. Under the conditions of catalysis the WO3 oxide is not reduced. Thus, the catalytic activity is apparently manifested not by the carbide itself, but rather by a biphasic structure.The interaction of MoO3 with a CH4+H2 mixture at 630 8C produces a powder ofMo2C of hexagonal structure with a specific surface ranging from 50 to 100 m2 g71.209 This also has a polymeric carbon on its surface which is removed under the hydrogenation conditions. A metastable carbide a-Mo2C1-x (x=0 ± 0.5) with a specific surface of 200 m2 g71 and particle size of 3 nm was prepared in the TPR mode by the reaction of MoO3 with ammonia and methane.210 First, the reaction Mo3+NH3 yields nitride Mo2N, which is then transformed into CO (A)+3H2 CH4 (A)+H2O E=51 CO (B)+3H2 CH4 (B) +H2O E=145 E<51 O Al O Al O Al O Al O Al O Al O Al O Al O Al O NiNiNi NiNiNi H H H H H NiNiNiNi H H RC CR H NiNiNiNi ONi2+O MNi2+OMO M O M M O M O Ni2+O H H 604 M D Navalikhina, O V Krylovcarbide upon reaction with a CH4+H2 mixture.When some platinum is added to MoO3, the temperature of the synthesis is reduced by 100 ± 150 8C and a thermodynamically stable but less active Mo2C is formed. Carbides can also be synthesised by carbidisation of the supported oxide MoO3/Al2O3.211 It is also known 210, 211 that in the reactions involving hydro- carbons the carbides are precursors of the `coke', i.e., graphite-like formations, which cover the surface of the catalyst and poison it.The coke results from realisation of the so-called carbide cycle.212, 213 For example, the interaction of Fe with C4H6 involves the following transformations: Depending on the conditions, either the metal phase or the carbide phase may be stable.Pathway (1) is the catalytic forma- tion of carbonaceous deposits on the metal phase; pathway (2) is the transformation of a carbide-like phase (Fe7C) into the genuine carbide phase via an intermediate carbide-like compound; pathway (3) is the formation of carbon on the stable carbide phase. The carbide phase is stable at low temperatures, while at high temperatures the metal phase is stable, i.e., the pathway (1) is predominant. Molybdenum carbide was used to prepare catalysts for the hydrodesulfurisation.214 The surface layer was sulfidised, while the carbide remains the support.Numerous hydrogenation reactions occur at low temper- atures, where the carbide phase is stable. This raises the question whether one should seek to prepare the carbide phase with a large specific surface in the synthesis of catalysts for hydrogenation.This problem has been investigated using iron and nickel catalysts for the CO hydrogenation (the Fischer ± Tropsch synthe- sis).85, 87, 215 ± 218 The most detailed studies were concerned with the catalysts based on the oxides a-Fe2O3 or g-Fe2O3 as the initial phase.87, 216, 217 The use of a-Fe2O3 allowed the preparation of an active catalyst yielding basically methane and a small propo- tion of alkenes in the CO+H2 reaction at 250 ± 300 8C.The analysis of the phase composition of this catalyst showed that after the reaction it contained equimolar quantities of K-Fe5C2 and Fe3O4. On the contrary, the use of g-Fe2O3 led to a system totally inactive as a catalyst composed of Fe3O4, Fe3C, and a-Fe.During catalyst preparation according to the Bell ± Boudart reaction, a considerable quantity of CO2 is released; however, under stationary catalysis conditions hardly any release occurred, which was indicative of the lack of carbon formation. An investigation into the catalytic activity together with the application of X-ray phase analysis in situ and MoÈ ssbauer spectroscopy made it possible to trace the sequence of catalyst formation based on a-Fe2O3 under the reaction conditions used.216 First, a-Fe2O3 is transformed into a defective magnetite Fe3O4 possessing low activity in the hydrogenation reactions.A sharp increase in the activity and selectivity (for C2±C4 hydro- carbons) occurs at the moment the K-Fe5C2 carbide phase appears.With an increase in the K-Fe5C2 : Fe3O4, ratio the catalytic activity of Fe3O4 steadily increases, while the selectivity of alkene formation passes through a maximum. Electron microscopy studies 87, 216 showed that the general morphology of particles remained unaltered under the reaction conditions. The particles are overgrown with amorphous carbon; however, the latter does not seem to block the active centres.The size of the Fe3O4 and K-Fe5C2 particles is*60 nm and*25 nm, respectively. The most interesting finding is evidence of intercala- tion of the carbide formed in the Fe3O4 lattice. In addition to electron microscopy, the electronographic analysis of selected sites and the anisotropy of the MoÈ ssbauer spectra confirm the formation of such a structure.The active centres of CO hydro- genation were supposed to be localised at the carbide ± oxide interface. The higher the proportion of intercalated carbide the more extended the interface and the higher the catalytic activity. The maximum yield of alkenes is formed at an equal percentage of Fe3O4 and K-Fe5C2 phases.Jung and Thomson 219 undertook an analogous study of the formation of CO hydrogenation catalysts based on a-Fe2O3 using the X-ray phase analysis in situ. The results obtained are also in favour of the initial formation of non-stoichiometric magnetite, the establishment of steady-state activity after the appearance of the Fe5C2 carbide phase, and the formation of magnetite and carbide phases in approximately equal proportions under steady- state conditions.These authors attribute the permanent activity of the catalyst produced to the presence of a single stable carbide phase. In another study,220 where the initial phase was metallic iron, they detected the formation of a mixture of e0-Fe2.2C and Fe2.5C carbides. Investigations into CO hydrogenation by MoÈ ssbauer spectro- scopy using the system 64% Fe2O3 + 5% Cu + 1% K2O + 30% kaolin revealed that under the reaction conditions the oxide Fe3O4 and carbide K-Fe5C2 are formed, which were further trans- formed into e0-Fe2.2C.Rao et al.221 presume that the carbide formation is a necessary but insufficient condition for the mani- festation of catalytic activity. A catalyst with the maximum carbide content was not the most active.An investigation of an Fe ± ZSM-5 system by MoÈ ssbauer spectroscopy and X-ray phase analysis 217 revealed that the initial catalyst comprised Fe3+-containing particles 5 ± 15 nm in size. Treatment with hydrogen reduces 90% of iron to a-Fe. Under the conditions of CO hydrogenation, small spinel clusters, goethite a-FeOOH and K-carbide were formed.Under the same condi- tions, in an Fe7Mn7zeolite system small clusters of inverted spinel, K-carbide, and MnO were formed. In analogous studies of a nickel catalyst for the CO hydro- genation (methanation),85, 87, 218 two procedures for preparing the initial NiO phase were considered: (1) a plasmochemical method yielding cubic NiO crystals possessing numerous point defects with the most developed plane (100); (2) decomposition of NiCO3 yielding lamellar crystals with the most developed plane (111). Reduction by hydrogen of the former specimen leads to the epitaxial formation of metallic nickel.Under the conditions of the H2+CO reaction, polycrystalline carbide Ni3C is formed in a later stage. This is followed by the formation of a three-phase system metal+carbide+oxide, which is active in the CO metha- nation.In the specimen prepared by the latter method, the epitaxy on the NiO plane (111) first induces the formation of hexagonal Ni3C single crystals which correspond in size to NiO crystallites. Upon reaching its critical size, the carbide crystal liberates the NiO surface for the build-up of the next particle. In this case, the metal nickel phase appears only after a temperature increase to 283 8C, which may be due to carbide decomposition.The rate of CO hydrogenation is higher in the latter than in the former specimen; nonetheless, no direct correlation between the reaction rate and the carbide content was found. The formation of carbides was also observed on a cobalt catalyst in the Fischer ± Tropsch synthesis.222 It occurs under the conditions of CO hydrogenation according to the Boudart reac- tion.Subsequently, the participation of carbides in the synthesis is possible due to direct interaction with hydrogen or intermediate interaction with water. In the presence of carbides, the selectivity of the process is shifted towards the synthesis of higher hydro- carbons. Thus, it is systems in which the oxide ± carbide phases can be formed under the reaction conditions that are the active hydro- genation catalysts.Most recent studies have shown that the formation of oxycarbide structures favours the reactions with the participation of unsaturated compounds on tungsten and molybdenum carbides. Fe [Fe ±C] C4H6 1 2 3 Fe+C C4H6 FexC [Fe3C17x]n+nxC Heterogeneous catalysts of hydrogenation 605Many recent studies 223 ± 226 were devoted to the studies of hydrogenolysis of 2-methylpentane, n-hexane, and methylcyclo- pentene on tungsten carbide WC.This catalyst was synthesised by heating WO3 in a mixture of 20% CH4+80% H2 in the TPR mode. In the process of reduction, intermediate species, viz., W20O58, W18O49, and WO2, were produced and a carbon layer was formed.After heating in this mixture at 800 8C for 10 h, the carbon layer disappeared to give way to WC with specific surface 12 m2 g71 possessing a biporous structure (the pores 2 nm and 6 nm in diameter predominate). On WC, which is similar in its catalytic properties to Ru, Ir and Ni, cracking and, to a much lesser extent, isomerisation take place in the presence of hydrogen.According to Leclercq et al.,227 it is advantageous to prepare the tungsten carbide in two stages: first, carbidisation of metallic W or WO3 with a mixture of 20%H2+80%CH4 at 650 8C producesW2C and a small amount of free C 2W+CH4=W2C+2H2, CH4 CH2+H2, CH2 C+H2 , and then WC is formed at 800 8C in the reaction W2C+CH4=2WC+2H2 . In the presence of O2 at 100 ± 300 8C a complex mixture containing oxycarbide phases, WC, W2C, WO2, and metallic tungsten is formed.228 The properties of the catalyst are strongly changed, and it becomes similar to a bifunctional platinum catalyst and selectively accelerates skeletal isomerisation.The development of procedures for preparing carbides with large specific surface (up to 200 m2 g71) should be regarded as an important achievement.229 ± 233 The hexagonal phase of Mo2C is formed upon interaction of activated carbon with MoO3 vapour.The unconsumed carbon is burned in the bulk of the carbide and can enhance the mechanical strength of the catalyst grains. According to the XPES data, the surface is covered by carbide alone, though the contact with oxygen leads to the formation of an oxide film.The oxide oxygen cannot be removed from the Mo2C surface by heating in hydrogen at 850 8C without partial decar- bidisation. To avoid this, the catalyst is treated with a mixture of H2 plus a hydrocarbon. Alternatively Mo2C or WC is impreg- nated with small quantities of Group VIII metal salts which accelerate decarbidisation. However, oxygen is not always a poison for carbide catalysts.A discovery was made owing to the accidental aspiration of air into a catalytic installation:229 ± 233 in the presence of hydrogen, the oxidised Mo2C catalyses isomerisation of n-hexane into methyl- pentanes and dimethylbutanes without formation of aromatic compounds. This makes it possible to reform hydrocarbons into high-octane and non-carcinogenic fuel.The most efficient catalyst was obtained by treating the carbide with air for 14 h at 350 8C. This resulted in the formation of an oxycarbide layer on the catalyst surface. According to the results of XPES, the major part of the molybdenum exists in the Mo(VI) state on the surface of this catalyst (in contrast to Mo2C). X-Ray studies showed that its structure corresponds to that of MoO3, but the oxygen vacancies &O on the surface are filled with carbon atoms that block the formation of MoO2.195 Such an oxycarbide possesses not only good selectivity, but also high activity and resistance to coking.It is 25 times as active as the Pt/Al2O3 catalyst but, unlike the latter, is little poisoned with the coke and retains its catalytic activity for a long time.The addition of methylcyclopentane to n-heptane results in slow deactivation of the catalyst due to blocking of active centres by the alkenes formed, which are readily hydrogenated. Moderate heating of the catalyst in an oxygen atmosphere is sufficient for its regeneration.233 High catalytic activity in the hydrogenation of isopropylben- zene is exhibited by molybdenum oxynitride.In this reaction its activity is increased due to carbidisation:234 MoN0.7O0.7 is trans- formed into MoC0.27N0.14O0.76. VII. Mechanochemical treatment of hydrogenation catalysts In most cases, mechanochemical treatment of solids leads to their fragmentation and the appearance of various defects. Many investigators have used mechanochemical techniques for the treatment of catalysts, since the catalytic activity usually (though not always) increases with the surface area and the number of defects.It is important that the changes obtained are preserved under catalysis conditions. Moroz et al.235 studied the formation of defects in the mechanochemical treatment of a typical ZnO catalyst in a planetary centrifuge mill using stainless steel balls 0.5 cm in diameter. Electron microscopy and X-ray data showed that short mechanochemical impacts (t<180 s) produce insignificant changes in the structure and size of the particles.At 180 s<t<600 s, one observes cracking of crystalline blocks, their mutual displacement, and rotation, relaxation of micro- distortions in the direction (001) and their growth in other directions.At 600 s<t<1200 s the crystals begin to undergo disintegration into primary crystalline blocks 5 ± 10 nm in size with the formation of secondary aggregates of globular shape 1000 nm in size along with relaxation of microdistortions arising as a result of crystal breakdown. Thus, the general picture is not reduced to a mere increase in the catalyst surface and defectiveness of its lattice.Streletskii et al.236 investigated the mechanism of changes induced by mechanical treatment of complex oxides with the perovskite structure. Powders of BaTiO3 and SrTiO3 were sub- jected to mechanical treatment in an eccentric mill with calori- metric measurement of the absorbed energy dose.237 The increase in the absorbed energy dose led to an increase in the specific surface from 7 ± 8 to 25 m2 g71, a decrease in the coherent scattering range (CSR) from 150 to 20 nm, and to the appearance of microdistortions and point defects (paramagnetic centres with a concentration up to 1018 g71).Each particle of the powder was shown to consist of several CSR blocks, the overall area of CSR being *60 m2 g71. Only a small fraction of the mechanochem- ical energy is spent in generating the defects preserved after the completion of mechanical treatment.The temperature of sinter- ing of particles after mechanical treatment decreases by 300 8C. When such a treatment is performed in water, the CSR size corresponds to the particle size, while microdistortions are insig- nificant. The mechanism underlying the mechanochemical transforma- tions depends on the chemical and electronic properties of solids.238 In covalent dielectrics, e.g.in SiO2, the defects after mechanical treatment are better preserved than in semiconducting metals. The application of mechanochemical techniques in the synthesis of catalysts can result not only in the formation of defects and an increased surface, but also in various chemical effects.In such a way one can prepare, for example, FeAl or FeW alloys, which cannot be formed in the usual processes because of the low mobility of the atoms, and even amorphous alloys. Recourse to mechanochemical treatment is often made in the synthesis of complex oxide systems from oxide mixtures, where the conventional thermal synthesis requires very high temper- atures, which are difficult to achieve.Mechanochemical decom- position of salts and hydroxides may be used to prepare oxides with the preservation of the disordered state of their atomic structures. There are some studies 85, 215 of the effect of mechanochemical treatment on the properties of catalysts in CO hydrogenation. We have already mentioned the formation of the two-phase oxide ± - carbide system in the catalyst based on a-Fe2O3 under the conditions of CO hydrogenation.A catalyst prepared from a- Fe2O3 after its treatment at high pressure with shearing on the Bridgeman anvil was also studied. Such a mechanochemical treatment is known to form a great number of defects. The impact of the reaction medium on the `defective' specimen is manifested 606 M D Navalikhina, O V Krylovearlier than on the `defect-free' one.In the first stage, the defective a-Fe2O3 is rapidly transformed into magnetite, however its catalytic activity is lower than that of the defect-free specimen. Electron microscopy of the defective specimen reveals individual oxide and carbide particles and a small number of intercalated particles. The particle size in this specimen is smaller than that in the defect-free specimen, but the extension of the carbide ± oxide interface, which determines the catalytic activity, is also smaller.The CO hydrogenation reaction was also used for an inves- tigation of the activity of amorphous, highly dispersed Fe7Ni7O alloys prepared by friction between two rapidly rotating discs made of iron and nickel.By modifying the friction conditions, it was possible to prepare an alloy of Fe20Ni20O60 composition on the surface of which clusters are formed with a predominantly ionic type of Fe7O bonds, and an Fe30Ni40O30 alloy in which, according to XPES and Auger spectroscopy, clusters with covalent-type of Fe7O bonds are prevalent. The former alloy is catalytically active in the CO hydrogenation to CH4 and C2±C4 hydrocarbons, while the latter proved to be inactive.Under catalysis conditions a non-stoichiometric oxide Fe3O47x, carbide FeCy and metallic nickel are formed on the former alloy. The surface of the latter alloy is covered by a nonreactive oxide FeOx, while in its bulk the spinel NiFe2O4 and a solid solution of iron in nickel are formed. The formation of active centres of CO hydrogenation is crucially influenced by the type of chemical bonding.Iron carbide and oxide phases exist on the surface of active catalysts, while metallic iron does not take part in the active centre formation. In a binary alloy catalyst, iron is mostly in the oxidised state, while nickel is predominantly in the metallic state. The activation of H2 seems to take place on Ni, whereas the CO activation occurs on Fen+.The Fischer ± Tropsch synthesis was brought about directly in a mechanochemical reactor by grinding the metal mixtures.239 Grinding allows formation of amorphous alloys Ni40Zr60, Ni60Zr40, Pd30Zr70, M30Ti70 (M=Fe, Co, Cu), M30Ti70 (M=Ni, Pd), which catalyse the CO hydrogenation. The alloy Ni30Ti70 manifests 100% selectivity.Detailed studies of structural changes in Ni7Zr alloys in the course of CO hydrogenation showed that prolonged grinding results in the formation of ZrO, ZrO2, Ni, and possibly ZrH2 and ZrC phases. Zirconium oxides are formed due to CO dissociation on the segregated nickel. Hydrogenation of CO was also studied under the conditions of mechanochemical treatment of Ti, V and Zr hydrides.In contrast with the reaction on amorphous titanium and zirconium alloys, the formation of hydrocarbons on the above-mentioned hydrides proceeds without any induction period. The CO conversion reaches a stationary level and begins to decrease, when the hydrogen content in the metal becomes insufficient for the maintenance of the reaction stoichiometry. Analogous work was done by Morozova et al.240, 241 Powders of an amorphous alloy ZrNi1-x were treated in a vibrational mechanochemical reactor consisting of a cylindrical steel chamber filled with steel balls 6 ± 7 mm in diameter.242 The amorphous alloy was prepared in the same reactor by mechanochemical fusion of Ni and Zr.Experiments were carried out in the static and flow modes; in the first case, the products were analysed by mass spectrometry, while in the second case, by gas chromatog- raphy.Under conditions of mechanochemical treatment, sorption of hydrogen rapidly increased with the dose of absorbed energy and reached a maximum. The quantity of sorbed hydrogen was approximately proportional to the zirconium content in the alloy. X-Ray phase analysis indicated that the reaction of hydro- gen with zirconium and ZrNi yields ZrH2 and a-ZrNiH2, respec- tively.After the termination of mechanochemical treatment, sorption of H2 continued at a decreasing rate until a stoichiomet- ric hydride was formed. Under these conditions, chemisorption of CO increased proportionally to the absorbed energy and Zr content. It follows a dissociative mechanism and is accompanied by oxidation of Zr CO+Zr ZrOxCy (x=y, x+y<1).In addition, reversible sorption of CO occurs, which corre- sponds to the nickel content. At large doses of absorbed energy, zirconium oxycarbides are decomposed ZrOxCy+CO ZrC+ZrO2+C. After saturation of the alloy with hydrogen, i.e., after for- mation of ZrH2 and ZrNiH2, methane, some ethane and hydrogen are formed under the conditions of mechanochemical treatment in a CO flow: ZrH2+CO ZrOxCy+H2 ZrHC +ZrO2+CH4 , a-NiZrH2+CO ZrHC +ZrO2+H2+CH4 .After termination of mechnochemical treatment, the forma- tion of methane continues at a decreasing rate. Such treatment may be conducted in the flow of a CO+H2 mixture. In this case, the rate of methane formation is 5 ± 10 times lower than that in the stepwise process (in the hydrogenation of CO with dissolved hydrogen) under the same conditions.In the mechanochemical treatment of a NiZr alloy, graphite also undergoes hydrogenation with formation of methane (and ethane at a high dose) a-NiZrH2+C ZrHC +Ni+CH4 . Thus, in the process of mechanochemical treatment the dissolved hydrogen acquires mobility indispensable for the real- isation of reactions both on the surface and in the bulk of a powder.Interestingly, nickel, which is a typical hydrogenation catalyst, is a ballast in this case, while hydrogen is activated with participation of zirconium. In addition to the inertial process of CO hydrogenation, an inertia-free mechanochemical process leading to the formation of short-lived intermediate compounds takes place on the surface.VIII. Sulfide catalysts for hydrotreating 1. Preparation of oxide precursors In the first sections of this review mention was made of the use of Ni7Mo, Ni7W, Co7Mo and Co7W oxide catalysts for the hydrogenation of unsaturated compounds, especially for hydro- treating. However, industrial processes rely most widely on the application of sulfide hydrotreating catalysts prepared from oxide precursors.The major components of sulfide catalysts for hydrotreating (hydrodesulfurisation, hydrodenitration, hydrodemetalation, hydrogenation of oxygenated and aromatic compounds in petro- leum) are as a rule Co (or Ni) and Mo (or W). Nickel is cheaper than cobalt and ensures a higher hydrogenation rate but it favours rehydrogenation, which results in an increased hydrogen uptake and a lower octane number of the products.Tungsten is similar to molybdenum in its catalytic properties, but it is more expensive. High activity is displayed by ruthenium, but this metal is very costly. The NiRu/Al2O3 catalyst is 30 times more active than RuS/ Al2O3 in hydrogenation processes.243 A great number of studies (see, e.g., reviews 5, 244, 245 ) are devoted to the preparation of hydrotreating catalysts, their for- mation under conditions of catalysis and the mechanisms under- lying the relevant reactions.The common practice is to use catalysts deposited on g-Al2O3 or carbon. Catalysts on Z-Al2O3 are more active than those on g-Al2O3, which is explained by the stronger bonding of MoO3 with Z-Al2O3 and easier subsequent sulfidisation.246 Ammonium molybdate (NH4)2MoO4 is a precursor of MoO3.In acidic media, polymolybdates are formed Heterogeneous catalysts of hydrogenation 6077MoO2¡ 4 +8Há Mo7O6¡ 24 +4H2O. At high levels of ammonium molybdate, further condensation takes place Mo7O6¡ 24 +MoO2¡ 4 +2Há [H2Mo8O28]67. At pH>7, polymolybdates dissociate to recover MoO27 4 . These acid-base processes involve the surface of Al2O3, which is amphoteric in its chemical nature AlOH+2Há AlOHá2 , AlOH AlO¡ +Há .In order to form polymolybdates in the preparation of catalysts by the impregnation procedure, it is necessary that the pH be kept below the isoelectric point of a given support (3 ± 6 for Al2O3 and below 2 for SiO2). After impregnation, the catalyst is dried at 250 ± 300 8C and calcined in air at 500 ± 550 8C; the dispersity of MoO3 changes insignificantly.The use of various physical methods (Raman and IR spectroscopy, EXAFS, and others) showed that at typical MoO3 concentrations (10% ± 12%) the initial molybdates form Al7O7Mo bonds with Al2O3 and represent two-monolayer-thick islets containing seven molybde- num atoms.244 Supports may be impregnated with HPC solutions, e.g., with nickel or cobalt heteropolymolybdates.247 In this case, however, because of the strong interaction of HPC with the OH groups of Al2O3 the precursor does not penetrate into the pores, and a `crusty', rather than regular distribution, of the active component over the support takes place.Thioheteroanions are also used for the preparation of such catalysts.248 Another method for the preparation of the Mo7Al7O catalysts is the mechanical mixing of MoO3 and Al2O3 at 400 ± 500 8C in the presence of water.The system thus formed possesses virtually the same catalytic properties as an MoO3/ Al2O3 contact obtained by impregnation.244 A very large specific surface (174 ± 184 m2 g71) is character- istic of a catalyst prepared by precipitation of MoO3 on SiO2 in a solution of nitric acid with a mixture of urea and thioacetamide and with subsequent drying at 120 8C and sulfidisation at 400 8C.249 The second component (`promoter') is introduced by impreg- nating the support with cobalt or nickel nitrates; the sequence of impregnation steps may be different.It is assumed 244 that impregnation of the support first with molybdenum and then with cobalt salts favours higher activity. There is an opposite viewpoint however, according to which it is more advantageous to impregnate the support first with cobalt salts or to bring about simultaneous impregnation with Co- and Mo-containing com- pounds.When Co2+ or Ni2+ are introduced by ion exchange, it is recommended to do it at low pH after deposition of molybdenum.A two-stage method for the preparation of a catalyst includes impregnation of g-Al2O3 in the fluidised bed with (NH4)2Mo7O24 solution at constant pH=7, calcination in air at 500 8C, then impregnation with nickel or cobalt nitrates until the pores are filled, and further calcination at 500 8C.250 Active hydrogenation catalysts are prepared using thiosalts, e.g., by impregnating a support with a mixture of (NH4)2MoS4 and Ni(NO3)2 or Co(NO3)2 and subsequent heating in a mixture of H2S and H2 at 400 8C.251 Unsupported Co7Mo7S catalysts are obtained by treating MoO3 and Co3O4 powders with a hot solution of (NH4)2S or by mixing solutions of Co(NO3)2 and (NH4)2Mo7O24 followed by addition of (NH4)2S, drying and calcination.In the latter case, the catalyst contains larger quantity of labile sulfur and is character- ised by a higher dispersity of the Co9S8 phase and activity, though it is deactivated more rapidly.252, 253 Lewis et al.254 assume that the catalyst activity depends on the acidity of the support.This is rationalised by the presence of surface OH groups capable of interaction with molybdenum and cobalt salts.Catalysts are calcined at 400 ± 600 8C, at which temperature the salts are completely decomposed but undesired solid-phase reactions have not yet occurred. According to results obtained by different physical methods,244 an amorphous MoO3 phase and anions of the (Mo7O24)67 type appear on the surface of Al2O3 after calcination of Mo7Al catalysts. Calcination of mixed catalysts leads to the formation of polymolybdates, MoO3, CoMoO4 or NiMoO4, Co3O4, or NiO, Al2(MoO4)3, Ni(AlxMoO4+1.5x) (x=0.1 ± 1.0) and Co or Ni aluminates.244, 255 ± 258 At low calcination temperatures, the cobalt is mostly in the octahedral coordination.It is with the content of this Co form that the catalyst activity after its sulfidisation correlates. At higher temperatures, intercalation into the Al2O3 lattice results in the formation of tetrahedral cobalt, which under- goes sulfidisation with more difficulty.Therefore, calcination at temperatures above 600 8C is not reasonable. Upon calcination, the nickel acquires octahedral coordination and interacts more strongly than cobalt with the Al2O3 lattice. Chukin et al.259 believe that the formation of nickel or cobalt molybdates or polymolybdates, which can be removed by extrac- tion with water, is beneficial for the subsequent sulfidisation or formation of a mixed NiMo or CoMo phase.The introduction of NiMoO4 prepared beforehand into the Ni7Mo7Al matrix makes it possible to enhance the catalyst activity. The content of the active component depends on the specific surface of the support.Usually, from 8 mass% to 15 mass% of molybdenum can be deposited on Al2O3, which corresponds to a specific Al2O3 surface of*250 m2 g71. The Co :Mo ratio varies from 0.1 to 1.0. The ratio of molybdenum to cobalt or molybde- num to nickel depends on the sequence of impregnation steps and is equal to 2 in the best catalysts. Active catalysts may be produced by decomposition of cobalt and molybdenum carbonyls on supports.260, 261 The maximum effect of promotion of CoMo/g-Al2O3 catalyst with WO3 is attained at W:(W+Mo)=0.25.Impregnation of the support first with a tungsten salt solution and then with a cobalt salt solution yields more active catalysts than those result- ing from the reverse impregnation order. According to the results of EPR and XPES studies, tungsten favours the alteration of the molybdenum coordination from octahedral to tetrahedral and facilitates its reduction.262 2.Sulfidisation of oxides Sulfidisation of Mo/Al2O3 catalysts at low temperatures leads to the substitution of the oxygen atoms by sulfur Mo6+7O+H2S Mo6+7S+H2O and formation of the bridging bonds Al7O7Mo7SH or (Ref. 263). Each of the species thus formed contains seven Mo atoms.Prolonged sulfidisation at 300 8C leads to the reduction of Mo6+ to Mo4+ and formation, as follows from electron micro- scopy data,244 264, 265 of planar hexagonal MoS2 particles below 3 nm in size that are retained on the surface by the Mo7O7Al bonds and mostly situated normally to the Al2O3 surface. Sulfur- containing Co7Ni heteropolycompounds are formed in the incomplete sulfidisation.266 At the optimum sulfidisation temperature (400 8C), the Mo7O7Al bonds are cleaved, and the distribution of the MoS2 particles occurs parallel to the support surface; they are enlarged, the ratio of the basal to lateral planes of particles remaining constant.However, there is no oriented growth (epitaxy) of the MoS2 particles on Al2O3: they are linked to each other by physical forces.The smaller the mean sizes of MoS2 and WS2 particles the more active the catalyst in the hydrodesulfurisation processes.257 Higher catalyst activity appears to be reached with the incomplete sulfidisation of MoO3.267 Complete transformation of the oxy- gen-containing surface phase into MoS2 occurs only at 970 8C, Al Mo O S S 608 M D Navalikhina, O V Krylovthough at such a high temperature the MoS2 phase is sin- tered.268, 269 Apparently the nature of the support does not strongly affect the final form of molybdenum and tungsten oxides and sulfides.In most cases, Mo/C proves to be more active than Mo/Al2O3.270 Isolated molybdenum ions or highly dispersed MoO3 particles were detected on a molybdenum catalyst deposited on activated carbon.These species are sulfidised more rapidly on carbon than on Al2O3: at 400 8C the sulfidisation is already completed. After the sulfidisation, MoS2 forms plates on Al2O3 and three-dimen- sional particles on carbon. This is explained by a stronger interaction of its oxide precursor with the Al2O3 surface than with carbon. The catalytic activity of MoS2 is higher on carbon than on alumina.271, 272 Pratt et al.273 observed a somewhat different orientation of MoS2 particles on supports: plates were found on Al2O3 and SiO2 oriented normally to their surface, while on TiO2 and ZrO2, particles were coplanar. The latter displayed higher catalytic activity.Treatment of a MoO3 ± SiO2 system with a H2S+H2 mixture at room temperature leads to the formation ofMoOxSy.At 150 8C MoS3 particles appear, while at 250 ± 270 8C the MoS2 phase is formed.274 Electron microscopy data showed that in a Co7Mo7S/Al2O3 system, MoS2 particles are formed, and their size is increased from 1 to 1.5 nm at the sulfidisation temper- ature.275 These particles are linked to form chains.At high cobalt content, the Co9S8 phase is formed.276 In Ni7Mo7S and Ni7W7S catalysts, the phase NiS was also observed along with mixed phases.277 According to NMR and EXAFS data, the major active phase of a molybdenum sulfide catalyst consists of MoS2 plates with Mo coordination number 4 or 5 and with the Co atoms at their edges.244 The formation of a continuous MoS2 phase, e.g., after long operation of the catalyst, leads to its deactivation.Analogous phases formed on SiO2 or carbon are less strongly bound to these supports than to Al2O3. To model sulfide hydrotreating catalysts, several layers of sulfur were applied onto the Mo(110) plane with subsequent spraying of Zn or Co.272 Zinc and cobalt were found to accelerate the sulfidisation of molybdenum. The hydrodesulfurisation activ- ity of catalysts was observed to correlate with the ease of molybdenum sulfidisation.The Mo atoms become coordinatively unsaturated in the presence of zinc or cobalt. 3. The role of different phases in sulfide catalysts To elucidate the role of different phases in the active surface formation, the following models of sulfide catalysts may be considered.5 1. A structural model of a Co7Mo/Al2O3 catalyst implies the formation of a molybdenum monolayer on the surface of alu- mina.278 According to this model, molybdenum is bound to the Al2O3 surface through the OH groups, whereas Co2+ ions replace Al3+ ions in the surface layer.Due to its large size, the S27 ion replaces two O27 ions. 2. In the initial version of the intercalation model the cobalt (nickel) ions were supposed to intercalate between the layers of the MoS2 (WS2) structure.279 Later, it was found that intercalation inside the lattice is impossible, and the model has been modified.Now it is assumed that Co ions are intercalated near the sulfide surface: The promoter atoms interact with MoS2.280 Co0+2Mo4+ 2Mo3++Co2+. 3. According to the model of contact synergism,257, 281 hydro- gen is activated on the Co9S8 or NiS phases, while the compound subject to hydrogenation is activated, on the MoS2 phase.Hydro- gen spillover takes place between these phases. A later modifica- tion of the model assumes hydrogen spillover between the CoS2 and CoMoS phases.282 In recent times, the prevalent view is that the sulfide phases Co9S8 and NiS apparently play no special role in the hydrotreating processes.5, 283 ± 296 The application of different physical methods (EXAFS, NMR, XPD, and others) revealed a kind of a `Co7Mo7S phase' in which the Co atoms are localised on lateral facets or edges of MoS2 microplates.According to EXAFS data, the coordination number of Co is 4 or 5. The catalyst MoS2/Al2O3 undergoes no deactivation after its sulfidisation.It is the distribu- tion of Co atoms at the edges of theMoS2 particles that determines the optimum ratio between Co and Mo. At high Co contents all its atoms cannot be placed at the edges of the MoS2 particles. MoÈ ssbauer spectroscopy showed 245 that the sulfidisation of CoMo/Al2O3 and CoMo/C leads to the formation of the same particles as in the Co/Al2O3 and Co/C systems, respectively.This casts doubts on the concept of the `Co7Mo7S phase' and makes one suspect that the MoS2 phase is merely a diluent. However, it is believed that the systems CoMoS/C and CoMoS/Al2O3 differ. The Co/C catalyst is only three times less active than CoMoS/C, whereas Co/Al2O3 and CoMoS/Al2O3 exhibit much larger differ- ences. According to Chianelli et al.,245 in the Co7Mo7S system the Co atoms are situated at the edges of the MoS2 crystallites.The Co9S8 phase is also present; however, its function is reduced not to hydrogen spillover, but rather to the role of an additional support for the Co7Mo7S phase. By analogy with the `Co7Mo7S phase', the `Ni7Mo7S phase' is also supposed to exist.255, 256 The nickel-modified sulfide (MoS2) adsorbs three times as much hydrogen as the pure MoS2.297 The excess phase Ni3S2 can be removed from the Ni7Mo catalyst by dissolution, and this does not affect its catalytic activity.298 Ekman et al.299 synthesised catalysts based on the so-called Chevrel phases (MxMo6Z8, where Z=S, Se or Te and M is a transition metal).In these phases, the transition metal cations may be localised not at the edges but inside the sulfide phase.Such systems differ from the usual mixed sulfide catalysts. The results of 59Co NMR suggested 287, 300 that the high activity of a Co7Mo7S system is associated with the presence of Co atoms with perturbed tetrahedral environments. In this system, octahedral cobalt is also present but the catalytic activity correlates only with the concentration of the tetrahedral cobalt.The active centre is assumed to have the following structure: Electron exchange stabilises the structure Cotetr7 S7Cooct7S7Mo7S. The appearance of the NiWO4 phase in the Ni7W7Al catalysts upon calcination of an NiO+WO3 mixture decreases the efficiency of subsequent reduction and sulfidisation. This results in the formation of an unordered WS2 phase and of Ni2+ and Ni3+ ions rather than an NiS phase.301, 302 The higher activity of an Ni7W/C catalyst than that of Ni7W/Al2O3 is supposed 303 O Al O Al O Al O Al O O Al O Al O Al O Co O Al O Co O Al O Al O Co O S S O Mo O S S O Mo Co Co S S S S Mo Bulk intercalation Surface intercalation Mo S S Co Mo .S S Heterogeneous catalysts of hydrogenation 609to be due to the fact that the active nickel in the former is linked to six sulfur atoms.The major advantage of the use of carbon as the support apparently lies in the fact that, in contrast to Al2O3, it has no acidic properties and does not undergo coking.304 The role of cobalt (nickel) in a mixed catalyst may be dual. Presumably, cobalt (or nickel) favours the reduction of MoO3 305 or WO3.306 According to the opposite viewpoint,307 Ni3+ incor- porated into the MoO3 layer decreases its susceptibility to reduction.It is also believed 268, 285, 299, 308, 309 that cobalt stabilises the dispersed MoS2 microparticles preventing them from sinter- ing. Some investigators 283, 286, 300, 301 presume that Co does not increase the MoS2 dispersity, but participates instead in the hydrogen activation.There are also reports 310 ¡À 312 that cobalt accelerates the sulfidisation of MoO3 and suppresses deactivation of catalysts after sulfidisation.287, 313 In the CoMoS/Al2O3 system, sulfur is hydrogenated at a lower temperature than in the CoS/Al2O3 or MoS2/Al2O3 systems. The ease of sulfur hydrogenation correlates with the catalytic activ- ity.314Ahigher rate of sulfur hydrogenation within MoS2 was also observed when Co, Ni or Fe were deposited on a MoS2 plate.315 The presence of Ni in aNi7Mo hydrodesulfurisation catalyst promotes the reduction of MoO3 to MoO2 and partly to Mo.308 Thus, nickel can act as a catalyst for MoO3 reduction.Cobalt also accelerates the reduction of Mo(VI) to Mo(IV) and even to Mo(III).316, 317 In this case, some vacancies appear in the sulfide layer. 4.The hydrodesufurisation mechanism The most popular hydrodesulfurisation mechanism includes the formation of anion vacancies in the sulfide layer and their subsequent filling with sulfur from a sulfur-containing compound, e.g., thiophene.244, 278, 318, 319 The thiophene molecules are adsorbed normal to the surface. The vacancies may either be initially present in the MoS2 phase, most probably at the edges of the microplates, or be formed in the course of sulfide reduction with hydrogen 2Mo4++S27+H2 2Mo3++&s+H2S, Mo4++S27+H2 Mo2++&s+H2S. Low-valent forms of molybdenum were in fact revealed by IR spectroscopy and EPR.The use of these techniques also provided evidence of S7H group formation under catalysis conditions, while XPES revealed low-valent sulfur in the form of S7 or (S7S)27.316, 320, 321 The disulfide groups, possibly localised at the MoS2 edges, may take part in the hydrogen activation.S2¡¦ 2 +H2 2SH¡¦. Hydrogen forms in sulfide catalysts were studied by IR spectroscopy and neutron diffraction.322 No bands corresponding to vibrations of the Co7H and Mo7H bonds were found after the adsorption of H2 in the systems CoMo1.3S1.6, MoS2 and Co9S8, but S7H vibrations bands were observed.This led to the conclusion that in the course of H2 adsorption processes of the type Mo7S+H2 H7Mo7S7H do not occur, giving way to the reaction 2Mo7S+H2 2Mo7S7H or Mo7S7S7Mo+H2 2Mo7S7H. The Co (Ni) atoms may be directly involved in the formation of active centres or neighbours with the vacancy at the peripheral molybdenum atom.The number of cobalt atoms in the catalyst was observed to correlate with the number of peripheral molyb- denum atoms. According to EXAFS,323, 324 the vacancies are formed near the cobalt atoms. This was confirmed in particular by experiments on the adsorption of selenophene (a thiophene analogue).325 The vacancies available in the CoxMo17xS2d phase are involved in the following reactions: Mo4+ +Co2++&+S2¡¦ +HSR Mo3+ +Co3++SH¡¦ +SR¡¦, Mo3+ +Co3++SH¡¦ +SR¡¦ +H2 Mo3+ +Co3++2SH¡¦ +HR, Mo3+ +Co3++2SH¡¦ Mo4+ +Co2++&+S2¡¦ +H2S.According to the results of TPR,326 sulfur is more weakly bound in the Co7Mo7S system than in Co9S8 or MoS2: the H2S peak is observed at a lower temperature. However, an investiga- tion of the isotope exchange of labelled hydrogen sulfide H2 35S with the NiS/Al2O3, MoS2/Al2O3, and NiMoS/Al2O3 catalysts unexpectedly showed that the exchange rate on NiMoS/Al2O3 was lower than on MoS2/Al2O3, which in turn was lower than on NiS/ Al2O3.327 This surprising finding was explained by the participa- tion of sulfur atoms bound to both molybdenum and nickel in the reaction. The results reported by Tops��e et al.326 may also be explained by the associative mechanism of this reaction without participation of vacancies.Experiments with 35S showed that the loss of sulfur from the catalyst as a result of its interaction with hydrogen can be compensated for at the expense of thiophene sulfur.318 The quantity of mobile sulfur and the catalytic activity increase in the presence of promoters (Co, Ni).According to Markel et al.,328, 329 dihydrothiophene is formed as an intermediate in the reaction of thiophene hydrodesulfurisa- tion The idea of an associative mechanism underlying the hydro- genolysis of thiophene without participation of the sulfide layer vacancies was also put forward by Startsev,323 who assumed that thiophene is adsorbed on the Co(Ni)7Mo7S catalyst in the coordination sphere of cobalt, while dissociative adsorption ofH2 takes place at the terminal sulfur atoms flanking the electroneutral macromolecule MoS2.This viewpoint stands in need of more conclusive evidence. Different opinions exist as to the nature of the active centres in the hydrogenation of various organic compounds. According to the most popular concepts, the role of the active centres in the hydrodesulfurisation is played by the molybdenum or cobalt atoms on the lateral facets of MoS2.5 The activity of hydro- desulfurisation catalysts was found to be proportional to the number of centres on the lateral facets of MoS2 determined by the EXAFS method in situ.326 There is a relationship between the activity of catalysts and the concentration of Mo5+ in them (according to EPR data).330 The activity of WS2 in the hydro- genation of benzene correlates with the concentration of &S vacancies on the lateral sulfide facets measured by EPR.275 The mechanism of thiophene hydrodesulfurisation on the Co7Mo/Al2O3 catalyst was investigated using the TPD and TPR techniques and D7H isotope exchange.331 ¡À 335 The TPR and TPD spectra of thiophene, dihydrothiophene, and butane- thiol were found to be very similar on Mo/Al2O3 and Co7Mo/ Al2O3, but somewhat shifted towards lower temperatures on the Co-containing catalyst.At the same time, the shape of the spectra on Co/Al2O3 was completely different and most of the products formed in the first two systems under TPR conditions were not H+ M S H H M S M H H H H + M S H2 M+H2S + S M M H M S 610 M D Navalikhina, O V Krylovproduced in the latter case, whereas hydrogen was desorbed at a lower temperature.These results confirm the traditional view- point, according to which thiophene is adsorbed on the anionic vacancies of MoS2, while hydrogen is activated on cobalt. We cannot say whether the Co andMocentres are situated in different phases or in different crystallographic positions of the same phase, though they do not appear to be spatially separated.In any event, no results have been reported to disprove this viewpoint. Desorption of thiophene at low temperatures characteristic of all catalysts is most likely to occur from the weakly bound p- adsorbed state. The high-temperature desorption of thiophene characteristic of Mo/Al2O3 and Co/Al2O3 catalysts seems to be related to the dehydrogenated forms.In such cases, strong retention of molecules by the surface is ensured by both the bonding of the sulfur atom and the s-bonds (possibly several) of carbon atoms in the ring. When there is a surface hydrogen shortage the limiting stage of the thiophene desorption may be the interaction of such surface complexes with the hydrogen atoms.To bring about desulfurisation, it is necessary to activate the adsorbed molecule of the S-containinmpound. This first requires the formation of sufficiently strong bonding of the sulfur atom with the surface and, second, surrounding of the reacting sulfur-containing molecule with hydrogen atoms situated in the immediate proximity.Under TPR conditions, such a complex may be formed with the participation of migrating hydrogen according to the mechanism of reverse spillover from the support. The coincidence of kinetic parameters for the formation of C4 hydrocarbons from thiophene and tetrahydrothiophene indicates that the hydrodesulfurisation of these compounds on Mo and Co7Mo catalysts proceeds via the same intermediates, probably dihydrothiophenes. Markedly easier desulfurisation of thiols than of thiophene and tetrahydrothiophene may suggest that the cleavage of the second C7S bond in the surface intermediates is not the limiting stage of the entire hydrodesulfurisation process.In contrast with the desulfurisation of tetrahydrothiophene and thiophene, butadiene is absent from the products of butane- thiol desulfuration, which indicates that the surface butanethio- late cannot be an intermediate compound in the hydrodesulfurisation of thiophene and tetrahydrothiophene, at least under low hydrogen pressure in the gaseous phase.Rozanov et al.334 proposed the following generalised scheme for the desulfurisation mechanism: According to this scheme, thiophene and tetrahydrothiophene can be interconverted via dihydrothiophene as the intermediate compound.Depending on the hydrogen pressure and the reaction temperature, equilibrium may be shifted towards thiophene or tetrahydrothiophene. Butene may be formed from the intermedi- ate dihydrothiophene as a result of two-stage hydrogenolysis of a-C with concomitant cleavage of the C7S bond, while butadiene may result from b-elimination in the first stage and a-hydro- genolysis in the second stage.Presumably, hydrogenation ± dehy- drogenation reactions occur on one type of centres, while hydrodesulfurisation occurs on another type of centre. The limit- ing stage of the whole process is the cleavage of the C7S bond. This conclusion was drawn from studies on the hydrodesul- furisation of thiophene on Rh2S3, Co/C, and Mo/C catalysts.336 It can be mentioned that the limiting stage of the reaction on Co7Mo/C is hydrogenation, which determines the low selectivity of formation of hydrogenation products in this case.Some investigators believe that hydrodesulfurisation and hydrogenation take place on the same active centres,337 while others think that these centres are different.338 Hydrogenation of propylene requires molybdenum(IV), and hydrogenolysis of pro- pane occurs on the more reduced Mo forms, up to its metallic state.On different supports, one may observe different orienta- tions of MoS2 and Co7Mo7S particles.339 When these particles are deposited on Al2O3, their activity in the hydrogenation is higher and that in the hydrocracking is lower than in their deposition on Al2O3 .SiO2, MgO. SiO2 or MgO. The molybde- num- and cobalt-containing phases have no synergistic effect in the hydrodesulfurisation of thiophene, but they do manifest this effect in the hydrogenation of butenes and butadiene.318 In contrast to Co7Mo7S, the Ni7W7S system accelerates hydro- genation of biphenyl more strongly than hydrodesulfurisation of dibenzothiophene.340 According to Topsùe et al.,5 the stronger deactivating effect of H2S on Co7Mo7S catalysts in the hydro- genation of aromatic compounds than in the hydrogenation of alkenes is indicative of the difference between the active centres in these reactions.In conclusion, it is noteworthy that the reactions of hydro- genation and hydrodesulfurisation on Co(Ni)7Mo(W) catalysts are the best studied catalytic processes, judging from the number of publications devoted to them.Nonetheless, the mechanism of these reactions has not been totally elucidated and remains the subject of numerous discussions. This also holds for the prepara- tion of hydrotreating catalysts. Studies in which the reaction mechanism was investigated by physical methods in situ are virtually absent.Without an adequate understanding of the mechanism of the process, it is impossible to develop scientific grounds for the preparation of catalysts. To this end it is necessary to elucidate what component of the catalysts activates hydrogen and what component activates an organic compound, and how to distribute these components most effectively on supports.IX. Conclusion The aim of this review was to consider hydrogenation catalysts, including modified supported Ni catalysts, carbides, sulfides, and others that have been most actively studied in recent years. Investigation into the mechanism of action of hydrogenation catalysts by physicochemical methods made it possible to reveal some of their common features.It was found that hydrogen and compounds subject to hydrogenation are, as a rule, activated on different active centres: the former, on metallic (or reduced) centres, the latter, on oxidised centres which are often formed by ions of the basal metal intercalated into supports or modifiers. Hydrogen transport between centres may be brought about by the spillover mechanism.Therefore, the role of support or modifier is often played by systems capable of accumulating hydrogen. The secondary structure of the support and catalyst (specific surface, dispersity, porosity) has a significant effect on the supply of hydrogen to the compound under hydrogenation, as well as on the thermostability and selectivity of catalytic systems.Co-precipitated metal ± oxide catalysts, complex oxides, skel- etal catalysts, and hydrides of intermetallides, which are also actively studied nowadays, remained almost totally outside the scope of this review. References 1. A K Rhodes Oil Gas J. 89 43 (1991) 2. Oil Gas J. 88 2552 (1990) 3. P Sabatier Die Katalyse in der Organischen Chemie (Leipzig: Akademische Verlagsgesellschaft, 1927) 4.Hydrogenation Catalysts (New Jersey: Noyes Data Corp., Park Ridge, 1977) 5. H Topsùe, B S Clausen, F E Massoth Hydrotreating Catalysis. Structure and Technology (Eds J R Anderson, MBoudart) (Berlin: Springer, 1996) S S S a-H a-H a-H 7b-H S S Heterogeneous catalysts of hydrogenation 6116. A A Krichko,M D Navalikhina Tekhnologiya Organicheskikh Veshchestv (Itogi Nauki i Tekhniki) [The Technology of Organic Substances (Advances in Science and Engineering Series)] (Moscow: Izd.VINITI, 1977) 7. Chem. Week 125 5 (1992) 8. US P. 3 796 764; Chem. Abstr. 80 120 395 (1974) 9. D V Sokol'skii, A M Sokol'skaya MetallyìKatalizatory Gidroge- nizatsii (Metals as Catalysts of Hydrogenation) (Alma-Ata: Nauka, 1970) 10. A B Fasman, D V Sokol'skii Struktura i Fiziko-Khimicheskie Svoistva Skeletnykh Katalizatorov (Structure and Physicochemical Properties of Skeleton Catalysts) (Alma-Ata: Nauka, 1968) 11.Promyshlennye Katalizatory i Nositeli (Industrial Catalysts and Sup- ports) (Novosibirsk: Institute of Catalysis, Siberian Branch of Acad- emy of Sciences of the USSR, 1977) 12. E D Radchenko, B K Nefedov, R R Aliev Promyshlennye Kataliza- tory Gidrogenizatsionnykh Protsessov Neftepererabotki (Industrial Catalysts of Hydrogenation Processes of Oil Refining) (Moscow: Khimiya, 1987) 13.J W E Coenen Appl. Catal. 54 59 (1989) 14. J W E Coenen Appl. Catal. 54 65 (1989) 15. J W E Coenen Appl. Catal. 75 193 (1991) 16. L M Gandia, A Diaz,M Montes J. Catal. 157 461 (1995) 17. N N Lebedeva, A Yu Stakheev, E S Shpiro, A Ya Yuffa Kinet.Katal. 30 212 (1989) a 18. A Ya Yuffa, A Yu Stakheev, E S Shpiro, A A Slinkin, A V Kucherov, E A Fedorovskaya, G V Antoshin, Kh M Minachev Appl. Catal. 39 153 (1988) 19. Yu I Ermakov, Yu A Ryndin, O S Alekseev,M N Vasil'eva, K G Myakishev, V V Volkov Kinet. Katal. 25 1017 (1984) a 20. B N Kuznetsov, in Nanesennye Metallicheskie Katalizatory Pre- vrashcheniya Uglevodorodov (Supported Metallic Catalysts of Conversion of Hydrocarbons) (Novosibirsk: Institute of Catalysis, Siberian Branch of Academy of Sciences of the USSR, 1978) Pt. 1, p. 74 21. V A Likholobov J. Mol. Catal. 55 105 (1989) 22. K I Zamaraev, in Perspectives in Catalysis. Chemistry for the 21st Century (Eds JMThomas, K I Zamoraev) (Oxford: Blackwell, 1992) p. 35 23. L McQuillan Pet.Rev. 49 25 (1995) 24. R C Scherr, G A Smelley, Y M E Normann Oil Gas J. 89 (21) 68 (1991) 25. Oil Gas J. 88 (30) 15 (1990) 26. O V Krylov, Kh M Minachev, V I Panchishnyi Neftekhimiya 29 579 (1989) 27. W J Piel Oil Gas J. 87 (49) 40 (1989) 28. R Corbett Oil Gas J. 85 (13) 56 (1987) 29. G A Unzelman Oil Gas J. 89 (15) 43 (1990) 30. J S Cannon, S L Azimi Int. J. Hydrog. Energy 21 987 (1995) 31.B Kh Kuper, A Stanislaus, P N Khannerup Neftegaz. Tekhnol. (3) 42 (1994) 32. Oil Gas J. 89 (26) 55 (1991) 33. J Goes,M G Asim Oil Gas J. 85 (19) 54 (1987) 34. G Parkinson, E Johnson Chem Eng. 96 (9) 30 (1989) 35. Oil Gas J. 91 (8) 45 (1993) 36. Oil Gas J. 91 (30) 88 (1993) 37. A J Suchanek Oil Gas J. 88 (19) 109 (1990) 38. G Heinrich, S Kaszelan, L Kerdraon Rev.Inst. France du Petrole 49 475 (1994) 39. R M Nash Oil Gas J. 85 (22) 47 (1989) 40. B H Cooper, B B L Donnis Appl. Catal. A137 203 (1996) 41. G A Mills, T W Steffen Catal. Rev. 3 159 (1973) 42. J A Dalmon, G A Martin J. Catal. 84 45 (1983) 43. P W Lesgar, M Sheintuch J. Catal. 127 576 (1991) 44. J A Lacey Energy World 156 2 (1988) 45. J T Richardson, S A Paripatyardar Appl. Catal. 61 293 (1988) 46. K D Froning, Kh Khammer, in Khimicheskie Veshchestva iz Uglya (Chemical Substances from Carbon) (Moscow: Khimiya, 1980) p. 231 47. L J Huang, J A Schwarz Appl. Catal. 36 177 (1988) 48. K B Kester, J L Falconer J. Catal. 89 380 (1984) 49. S L Ghen, H L Zhang, J Hu, C Contescu, J A Schwarz Appl. Catal. 73 289 (1991) 50. I E Maxwell, J E Naber, K P de Jong Appl. Catal. 113A l53 (1994) 51. M F Wilson, J P Fischer, Y P Kriz Ind. Eng. Chem., Prod. Res. Dev. 25 505 (1986) 52. A Nishijima, T Sato, Y Yoshimura, in Proceedings of the 1st World Congress on Environmental Catalysis, Pisa, Italy, 1995 (Roma: Societa Chim., 1995) p. 89 53. E E Shpil'rain,M D Navalikhina Khim. Tv. Topliva (2) 45 (1990) 54. A Elzir Pet. Rev. 41 37 (1987) 55. E D Radchenko Zh. Vses.Khim. O-va im. D I Mendeleeva 34 607 (1989) b 56. E F Kaminskii, V A Khavkin Khim. Tekhnol. Topliv Masel 2 14 (1996) 57. S A Taback, F J Krembeck,WE Garwood AIChE J. 32 1526 (1986) 58. C T O'Connor, M Kojima Catal. Today 6 329 (1990) 59. J Topp-Jorgensen Pet. Techn. 333 11 (1987) 60. J Haggin Chem. Eng. News 65 (25) 22 (1987) 61. J J Dell Amico Pet. Techn. 333 18 (1987) 62. R H Crabtree Chem.Rev. 95 987 (1995) 63. M Dry Catal. Today 6 183 (1990) 64. S T J Van Rensburg Chem. SA 16 (2) 41 (1990) 65. C A Jones, J J Leonard, J A Sofranko Energy Fuels 1 12 (1987) 66. A H Singleton, S Rejier Hydrocarbon Rross. 62 (5) 71 (1983) 67. N Y Chen,W E Garwood Ind. Eng. Chem., Prod. Res. Dev. 25 641 (1986) 68. A A Prokhorova, E M Buslova, E M Bushueva Khim. Tekhnol. Topliv Masel 1 33 (1988) 69.A N Startsev, V N Rodin React. Kinet. Catal. Lett. 48 119 (1989) 70. F Bernasconi ErdoÈ l Kohl Erdgas Petrochem. 40 480 (1987) 71. P A Borril, K R Wild Gas Eng. Manag. 27 6 (1986) 72. Yu A Vyskubenko Izv. Akad. Nauk SSSR, Ser. Energ. 5 11 (1996) 73. A Mills, T W Steffen Catal. Rev. 8 159 (1973) 74. R Levitan, H Rosin, M Levy Sol. Energy 42 267 (1989) 75.Yu B Yan, B K Nefedov, in Sintezy na Osnove Oksidov Ugleroda (The Syntheses Based on Carbon Oxides) (Moscow: Khimiya, 1987) p. 264 76. J T Richardson, J L Pratt J. Catal. 95 467 (1985) 77. A Rehmat, S S Randhava Ind. Eng. Chem., Prod. Res. Dev. 9 512 (1978) 78. N Takezawa, H Teruma, M Shimokawa Appl. Catal. 23 291 (1986) 79. T Takahama, H Oshima, A Veno, Y Kotera Appl. Catal. 5 59 (1983) 80. O V Krylov, A Kh Mamedov Usp.Khim. 64 935 (1995) [Russ. Chem. Rev. 64 877 (1995)] 81. T Inui, T Takeguchi Catal. Today 10 95 (1991) 82. R B Anderson The Fischer ë Tropsch Synthesis (New York: Academic Press, 1984) 83. P Forzatti, E Tronconi, I Pasquon Catal. Rev. 33 104 (1991) 84. G A Somorjai Catal. Lett. 7 169 (1990) 85. O V Krylov, O S Morozova, T I Khomenko Kinet.Katal. 35 805 (1994) a 86. E V Prokhorenko, N V Pavlenko, G I Golodets Kinet. Katal. 29 820 (1988) a 87. O S Morozova, G N Kryukovs, L M Plyasova, O V Krylov J. Catal. 158 13 (1996) 88. H H Dunken, V I Lygin Quantenchemie der Adsorption an Fes- tkoÈ rperoberflaÈchen (Leipzig: VEB Deuscher Verlag fuÈ r Drundstoff- industrie, 1978), 89. R D Kelley, S Semanchik J. Catal. 84 248 (1983) 90.G H Ludwig Hydrocarbon Process. 12 (3) 67 (1993) 91. G S Gurevich Khim. Promst 10 581 (1987) 92. K Morikawa, T Shirasaki, M Okada Adv. Catal. 20 97 (1969) 93. K T Kim, S K Kim J. Catal. 96 12 (1985) 94. G D Zakumbaeva, N A Zakarina, V A Naidin, A M Dostiyarov, N F Toktabaeva, E N Litvyakova Kinet. Katal. 24 449 (1983) a 95. N A Zakarina, G D Zakumbaeva, N F Toktabaeva, B N Dyussenbaev, E N Litvyakova, A Sh Kuanyshev Kinet.Katal. 23 86 (1983) a 96. J Haggin Chem. Eng. News 73 37 (1995) 97. G A Branitskii Vestn. Belorus. Univ., Ser. 2 (3) 8 (1987) 98. G A Branitskii, D I Michko, S N Mal'chenko Dokl. Akad. Nauk Belorus. SSR 28 914 (1984) 612 M D Navalikhina, O V Krylov99. M D Navalikhina, G A Branitskii, M S Kharson, Z T Beisembaeva, D I Mychko, in Nauchnye Osnovy Prigotovle- niya i Tekhnologii Katalizatorov (Tez.Dokl. II Vsesoyuz. Soveshch.), Minsk, 1989 [Scientific Foundations of Preparation and Technology of Catalysts (Abstracts of Reports at the Second All-Union Meeting), Minsk, 1989] p. 184 100. E F Ivan'kovich, L F Buzanova, V P Polyakova, D I Michko, T S Petkevich Vestsi Akad. Navuk Bel. SSR, Ser. Khim. Navuk (3) 36 (1986) 101.R Burch, A R Flambard J. Catal. 85 16 (1984) 102. A R Gonzales-Elipe, A Fernandez, J P Espinoza, G Munuera J. Catal. 131 51 (1991) 103. L-J Huang, J A Schwarz Appl. Catal. 30 239 (1987) 104. L-J Huang, J A Schwarz Appl. Catal. 30 255 (1987) 105. L-J Huang, B T Barrett, J A Schwarz Appl. Catal. 24 241 (1986) 106. L-J Huang, J A Schwarz Appl. Catal. 32 45 (1987) 107.L-J Huang, J A Schwarz Appl. Catal. 32 59 (1987) 108. L-J Huang, J A Schwarz, J R Diehl, J P Baltrus Appl. Catal. 37 229 (1988) 109. C H Bartholomew, R J Farranto J. Catal. 45 41 (1976) 110. C H Bartholomew, R B Pannel, J L Butler J. Catal. 65 335 (1980) 111. C H Bartholomew, R B Pannel J. Catal. 65 390 (1980) 112. W Tsai, J A Schwarz, C T Driscoll J. Catal. 78 88 (1982) 113. M S Heise, J A Schwarz J.Colloid Interface Sci. 107 237 (1985) 114. M S Heise, J A Schwarz J. Colloid Interface Sci. 113 55 (1986) 115. K B Kester, E Zagli, J L Falconer Appl. Catal. 22 311 (1986) 116. A A Slinkin, L V Ermolov, T N Kucherova Kinet. Katal. 34 305 (1993) a 117. N V Pavlenko, E V Prokhorenko, A I Tripol'skii, G I Golodets Kinet. Katal. 30 1364 (1989) a 118. N V Pavlenko, A I Tripol'skii, G I Golodets, Yu N Shevchenko, V A Nazarenko Kinet. Katal. 26 108 (1985) a 119. C H Bartholomew, W L Sorensen J. Catal. 81 131 (1983) 120. M D Romero, H De Lucas, J A Calles, A Rodriguez Appl. Catal. A146 425 (1996) 121. D L Hoang, H Berndt, H Miessner, E Schrier Appl. Catal. A114 295 (1994) 122. J T Richardson, B Turk,M Lei Appl. Catal. 83 87 (1992) 123. B Coughan,M A Keane Zeolites 11 2 (1991) 124.R J Taylor, R H Petit Appl. Catal. A119 121 (1994) 125. A Gil, A Diaz, L M Gandia, M Montez Appl. Catal. A109 167 (1994) 126. W J J Welters, O H van der Waerden, V H J de Beer, R A van Santen Ind. Eng. Chem. Res. 34 1166 (1995) 127. K P Wendlandt, H Bremer, F Vogt Appl. Catal. 31 65 (1987) 128. B Shen, J L Falconer J. Catal. 117 404 (1989) 129. B Mile, D Sirling, A L Zammit, M Webb J.Catal. 114 217 (1988) 130. S Z Roginskii Geterogennyi Kataliz (Heterogeneous Catalysis) (Moscow: Nauka, 1979) 131. V N Vorob'ev, D R Agzamkhodzhaeva, V P Mikita, M F Abidova Kinet. Katal. 25 190 (1984) 132. G C Bond, in Metal-Support and Metal-Additive Effects in Catalysis (Ed. B Imelik) (Amsterdam: Elsevier, 1982) p. 1 133. S A Surin, R R Aliev, B K Nefedov Neftekhimiya 21 385 (1981) 134.A S Lysitsyn, A V Golovin, A L Chuvilin Appl. Catal. 55 235 (1989) 135. Yu I Ermakov, B N Kuznetsov Kinet. Katal. 18 1167 (1979) a 136. A A Krichko,M D Navalikhina, Yu I Petrov, in Khimiya i Tekh- nologiya Molibdena i Vol'frama (Tez. Dokl. IV Vsesoyuz. Soveshch.), Tashkent, 1980 [The Chemistry and Technology of Molybdenum and Tungsten (Abstracts of Reports at the Fourth All-Union Meeting), Tashkent, 1980] Vol. 2, p. 121 137. A M Proskurin, M D Navalikhina, V V Ermakova, in Materialy IV Konf. po Kinetike Geterogenno-Kataliticheskikh Reaktsii, Yaroslavl', 1988 (Proceedings of the Fourth Conference on Kinetics of Heterogeneous Catalytic Reactions, Yaroslavl', 1988) p. 281 138. M D Navalikhina, I L Malkina, V I Garanin Neftekhimiya 30 26 (1990) 139.M D Navalikhina, G S Golovina, Yu I Petrov Khim. Tv. Topliva (6) 89 (1984) 140. Kh M Minachev, V I Garanin,M D Navalikhina, I L Mironova, G S Golovina Neftekhimiya 25 739 (1985) 141. E I Evko,M D Navalikhina, A E Chalykh,M P Glazunov Izv. Akad. Nauk SSSR, Ser. Khim. 2705 (1990) c 142. M D Navalikhina,M V Tsapkina, in Tez. Dokl. III Vsesoyuz. Konf. po Khimicheskim Sintezam na Osnove Odnouglerodnykh Molekul, Moskva, 1991 (Abstracts of Reports at the Third All- Union Conference on Chemical Syntheses Based on One-Carbon Molecules, Moscow, 1991) p. 8 143. MD Navalikhina, V I Spitsyn, I D Kolli, G A Gordeeva, A A Slinkin, G S Golovina Izv. Akad. Nauk SSSR, Ser. Khim. 18 (1985) c 144. O V Krylov, M D Navalikhina Zh. Prikl. Khim. 70 1300 (1997) d 145.O V Krylov, M D Navalikhina, in Plenarnye Doklady na III Ross. Konf. `Nauchnye Osnovy Prigotovleniya i Tekhnologii Katalizatorov', Yaroslavl', 1996 (Plenary Reports at the Russian Conference `Scientific Foundation of Preparation and Technology of Catalysts' (Novosibirsk: Institute of Catalysis, Siberian Branch of the Russian Academy of Sciences, 1996) p. 147 146. A M Maitra, N W Cant, D L Trimm Appl.Catal. 48 187 (1989) 147. H Nishi, K Nowinska J. Catal. 116 480 (1989) 148. S Kasztelan, J B Moffat J. Catal. 116 82 (1989) 149. S Kasztelan, J B Moffat J. Catal. 112 54 (1988) 150. J B Moffat Appl. Catal. A146 65 (1996) 151. M T Pope Heteropoly- and Isopolyoxymetallates (Berlin: Springer, 1983) 152. C Hoang-Van, O Zegaoui Appl. Catal. A130 89 (1995) 153.M D Navalikhina, G V Grebenshchikova, M V Glushchenko Khim. Tv. Topliva (2) 94 (1986) 154. K I Zamaraev Catal. Today 35 3 (1997) 155. P Kluson, L Cerveny Appl. Catal. A128 13 (1995) 156. V D Stytsenko Appl. Catal. A126 1 (1995) 157. I Palinko Appl. Catal. A126 39 (1995) 158. P Gallezot Catal. Rev. 20 121 (1979) 159. H Dauns, S Ernst, J Weitkamp Stud. Surf. Sci. 28 787 (1987) 160.R A Sheldon Chemicals from Synthesis Gas (Dordrecht: Reidel, 1983) 161. M A Vannice J. Catal. 37 449 (1975) 162. G I Golodets, N V Pavlenko, A I Tripol'skii Kinet. Katal. 28 888 (1987) a 163. M A Vannice J. Catal. 50 228 (1977) 164. R Szimansky, H Charkosset, V Perrichon, in Proceedings of the 8th International Congress on Catalysis, West Berlin, 1984 Vol. 2 (Basel: Weinheim, 1984) p. 151 165. S D Jackson, B J Brandrecth, D Winstahley Appl. Catal. 27 325 (1986) 166. Yu-Wen Chen Can. J. Chem. Eng. 64 875 (1986) 167. F King, E Shutt, A J Tompson Plat. Metals Rev. 29 146 (1985) 168. N Takahashi, T Mori, A Miyamoto Appl. Catal. 38 61 (1988) 169. T Okuhara, T Kimura, K Kobayashi Bull. Chem. Soc. Jpn. 57 939 (1984) 170. M Dry J. Organomet. Chem. 372 117 (1989) 171.N M Popova, L V Babenkova, G A Savel'eva Adsorbtsiya i Vzai- modeistvie Prosteishikh Gazov s Metallami VIII Gruppy (Adsorption and Interaction of the Simplest Gases with Group VIII Metals) (Alma-Ata: Nauka, 1972) 172. R W McCabe, R K Usmen,K Ober,H S Gandhi J. Catal. 151 385 (1995) 173. I Barbier, E Lami-Pitara, P Marecot, J P Boitiax Adv. Catal. 37 279 (1995) 174. N S Figoli, P C Argentiere, A Arcoya, K Scoane J.Catal. 155 95 (1995) 175. T N Fleisch, R F Hicks, A T Bell J. Catal. 87 398 (1984) 176. Y Wu,M Funakoshi, H Nagamoto Bull. Chem. Soc. Jpn. 62 493 (1989) 177. M Ishikawa, K Shikakura Shokubai (Catalyst) 21 253 (1979); Chem. Abstr. 95 203 250 (1981) 178. Yu A Ryndin, R F Hicks, A T Bell J. Catal. 70 287 (1981) 179. J M Drissen, E K Poels, J P Hindermann, V Ponec J.Catal. 82 26 (1983) 180. M A Vannice, C Sudhahar, M Freeman J. Catal. 108 97 (1987) 181. C Sudhahar,M A Vannice J. Catal. 95 227 (1995) 182. L Czerveny Chem. Eng. Commun. 83 31 (1989) 183. E C Hsiao, J L Falconer J. Catal. 132 195 (1991) 184. R F Hicks, H Qi, A B Kooh, L B Fisher J. Catal. 124 488 (1990) Heterogeneous catalysts of hydrogenation 613185.B H Cooper, P Sogaard-Andersen, P A Hannerup Catalyst Hydroprocessing of Petroleum and Distillates (New York: Marcel Decker, 1994) p. 279 186. M Viniegna, V Arroyo, R Gomez Appl. Catal. 44 1 (1988) 187. J P Bastiaux, J Cosyns, M Derrien, G Leger Hydrocarbon Process. 4 (3) 51 (1985) 188. R Ugo Catal. Rev. 11 225 (1975) 189. K De Jonc, G P Meima, J W Geus Appl. Surf. Sci. 14 73 (1982) 190.R G Nuzzo, L H Duboix, N E Bowles,M A Trecoske J. Catal. 85 2667 (1984) 191. P A Sermon, G C Bond Trans. Faraday Soc. 76 889 (1980) 192. P A Sermon, G C Bond Catal. Rev. 8 211 (1973) 193. C Hoang-Van, O Zegaoui, I Arand, in Proceedings of the 10th International Congress on Catalysis, Budapest, 1992 (Budapest: Akademiai Kiado, 1993) Vol. C, p. 2099 194. O V Krylov, V A Matyshak Usp.Khim. 64 66 (1995) [Russ. Chem. Rev. 64 61 (1995)] 195. Y Imizo, Y Fujito, Y Yamabe, N Toyoshita, in Posters of the 11th International Congress on Catalysis, Baltimore, 1996 (Baltimore: North Am. Catal. Soc., 1996) Po-016 196. C Neru, A Donato, L Mercadante, L Bonaccorsi, S Galvagno, in Posters of the 11th International Congress on Catalysis, Baltimore, 1996 (Baltimore: North Am. Catal.Soc., 1996) Po-018 197. F King, S D Jackson, F E Hancock, in Posters of the 11th Interna- tional Congress on Catalysis, Baltimore, 1996 (Baltimore: North Am. Catal. Soc., 1996) Po-054 198. N A Zakarina, G D Zakumbaeva, N V Toktabaeva, A Sh Kuyanyshev, E N Litvyakova React. Kinet. Catal. Lett. 26 441 (1994) 199. B Tardy, C Noupr, C Leclercq, J L Berolini, A Hoarez, N Treilleux, J P Faurt, C Nihoul J.Catal. 129 1 (1991) 200. J P Battiaux, J Cosins, S Vasudevan Appl. Catal. 6 41 (1983) 201. S Naito, N Amemiya, in Posters of the 11th International Congress on Catalysis, Baltimore, 1996 (Baltimore: North Am. Catal. Soc., 1996) Po-072 202. V V Rozanov, O V Krylov Usp. Khim. 66 117 (1997) [Russ. Chem. Rev. 66 107 (1997)] 203. A Sarkany, G Steflar, J W Hightower Appl. Catal.A127 77 (1995) 204. T M Yur'eva, L M Plyasova, T A Kriger, O V Makarova Kinet. Katal. 36 769 (1995) a 205. L M Plyasova, L P Solovyeva, T A Krieger, O V Makarova, T M Yurieva J. Mol. Catal. 105 61 (1995) 206. G Wrobel, C Lamonier, A Bennani, A D Huysser, A Aboukakis J. Chem. Soc., Faraday Trans. 92 2001 (1996) 207. M Boudart, R Levy Science 181 547 (1973) 208.M Boudart, J S Lee, K Imura, S Yoshida J. Catal. 103 30 (1987) 209. J S Lee, S T Oyama, M Boudart J. Catal. 106 125 (1987) 210. J S Lee, L Volpe, F H Ribeiro, M Boudart J. Catal. 112 44 (1988) 211. J S Lee, S Locatelli, S T Oyama, M Boudart J. Catal. 125 157 (1990) 212. R A Buyanov,M A Tanatarov, in Nauchnye Osnovy Tekhnologii Katalizatorov (Scientific Foundations of Technology of Catalysts) (Novosibirsk: Coordination Centre, 1981) No. 13, p. 67 213. V V Chesnokov, R A Buyanov Kinet. Katal. 28 403 (1987) a 214. J S Lee,M Boudart Appl. Catal. 19 207 (1985) 215. D P Shashkin, P A Shiryaev, A V Chichagov, O S Morozova, O V Krylov Kinet. Katal. 33 923 (1992) a 216. O S Morozova, Yu V Maksimov, D P Shashkin, P A Shiryaev, V V Matveev, V A Zhorin, O V Krylov, G N Kryukova Appl.Catal. 78 227 (1991) 217. Yu V Maksimov, V V Matveev,D P Shashkin Kinet. Katal. 27 920 (1986) a 218. O S Morozova, A V Ziborov, G N Kryukova, L M Plyasova J. Catal. 144 50 (1993) 219. H Jung,W J Thomson J. Catal. 139 375 (1993) 220. H Jung,W J Thomson J. Catal. 134 654 (1992) 221. K R P M Rao, F E Huggins, G P Huggins, G P Huffman, R J O Brien, B H Davis, in Posters of the 11th International Con- gress on Catalysis, Baltimore, 1996 (Baltimore: North Am.Catal. Soc., 1996) Po-075 222. A Eckstrom, J A Lapszevich, in Proceedings of the 9th International Congress on Catalysis, Calgary, 1988 Vol. 2 (Ottawa: Chem. Inst. Canada, 1988) p. 714 223. V Keller, P Wehrer, F Garin, R Ducros, G Maire J. Catal. 153 9 (1995) 224. A Frennet, G Leclercq, G Maire, R Docros, M Jardinier- Offergeld, F Buillon, J M Bastin, A Loffler, P Blehen, M Dufour, M Kamal, P Wehrer, M Cheval, F Garin, P Kons, P Delkambe, L Binst, in Proceedings of the 10th International Congress on Catalysis, Budapest, 1992 (Budapest: Akademiai Kiado, 1993) Vol.C, p. 927 225. A Miller, V Keller, R Ducros, G Maire Catal. Lett. 35 65 (1995) 226. A Loffler, L Seyfried, P Blehen, S Decker, J M Bastin, A Frennet Catal.Lett. 33 165 (1995) 227. G Leclercq, M Kamal, J M Girandon, P Devassine, L Feigen- baum, L Leclercq, A Frennet, J M Bastin, A Lolberg, S Decker, M Dufour J. Catal. 158 142 (1996) 228. A Katrib, F Hemming, P Welher, L Hilaire, G Maire Catal. Lett. 29 397 (1994) 229. M J Ledoux, C Pham-Huu, H Dunlop, J Guille, in Proceedings of the 10th International Congress on Catalysis, Budapest, 1992 (Buda- pest: Akademiai Kiado, 1993) Vol.C, p. 954 230. E A Blekkan, C Pham-Huu,M J Ledoux, J Guille Ind. Eng. Chem. Res. 33 1657 (1994) 231. P Delporte, F Meunier, C Pham-Huu, P Vennegues,M J Ledoux, J Guille Catal. Today 23 251 (1995) 232. M J Ledoux, P Del Gallo, C Pham-Huu, A P E York Catal. Today 27 143 (1996) 233. C Pham-Huu, A P E York,M Benaissa, P Del Gallo,M J Ledoux Ind.Eng. Chem Res. 34 1107 (1995) 234. G Djega-Mariadassou,M Boudart, G Bugli, C Sayad Catal. Lett. 31 411 (1995) 235. E M Moroz, S V Bogdanov, V I Zaikovskii, V V Molchanov, R A Buyanov Kinet. Katal. 30 993 (1989) a 236. A N Streletskii, V I Lapshin, I V Berestetskaya, E N Gindin, G E Ivanchikhina, I V Kolbanev, P Yu Butyagin Kinet.Katal. 30 1064 (1989) a 237. P Yu Butyagin, I K Pavlichev React. Solids 1 361 (1986) 238. P Yu Butyagin Kinet. Katal. 28 5 (1987) a 239. G Cocco, G Mulas, L Schiffini Mater. Sci. Eng. A181/182 1065 (1994) 240. P Yu Butyagin, A N Streletskii, O S Morozova, I V Berestetskaya, A B Borunova Dokl. Akad. Nauk 336 771 (1994) e 241. O S Morozova, A N Streletskii, I V Berestetskaya, A B Borunova Catal. Today 38 109 (1997) 242. A R Kuznetsov, P Yu Butyagin, I K Pavlychev Prib. Tekhn. Eksp. 6 201 (1986) 243. J A De Los Reyes, M Vrinat,M Breysse, F Mauge, J C Lavalley Catal. Lett. 13 213 (1992) 244. H Knozinger, in Proceedings of the 9th International Congress on Catalysis, Calgary, 1988 Vol. 5 (Ottawa: Chem. Inst. Canada, 1988) p. 20 245. R R Chianelli,M Daage,M J Ledoux Adv. Catal. 40 177 (1994) 246. F E Massoth J. Catal. 36 164 (1975) 247. N R Galimzyanov, V N Mikhailov, I I Zadko Kinet. Katal. 33 915 (1992) a 248. A J A Konings, A Valster, V H J de Beer, R Prins J. Catal. 76 466 (1982) 249. V Meeyoo, A A Adesina, G Foulds React. Kinet. Catal. Lett. 56 231 (1995) 250. S Kasztelan, J Grimblot, J P Bonnelle, E Payen, H Toulhoat, Y Jacquin Appl. Catal. 7 91 (1983) 251. S Fuentes, C Diaz, F Pedraza, H Rojas, N Roses J. Catal. 113 535 (1988) 252. M Vrinat,M Breysse, R Frety Appl. Catal. 12 151 (1984) 253. M Breysse, R Frety,M Vrinat, P Grange,M Genet Appl. Catal. 13 363 (1985) 254. J M Lewis, R A Kidd, P M Boormann J. Catal. 120 413 (1989) 255. M V Landau, L N Alekseenko, B K Nefedov,M A Kipnis, D A Agievskii, G D Chukin, S A Sergienko, V I Kvashonkin Kinet. Katal. 25 934 (1984) a 256. D A Agievskii, V I Kvashonkin, A I Pavlova, G D Chukin, V I Mikhailov, S A Surin, M V Landau, B K Nefedov Kinet. Katal. 27 178 (1986) a 257. B Delmon React. Kinet. Catal. Lett. 13 203 (1980) 614 M D Navalikhina, O V Krylov258. L V Manuilova, V P Fedin, Yu V Mironov, E M Moroz, G V Pasynkova, V I Zaikovskii, T S Sukhareva React. Kinet. Catal. Lett. 54 281 (1985) 259. G D Chukin, V G Sidel'kovskaya, S A Surin, L I Nikulina, L D Konoval'chikov, B K Nefedov Kinet. Katal. 32 1261 (1991) a 260. F Mauge, A Vallet, J Bachelier, J C Duchet, J C Lavalley Catal. Lett. 2 57 (1989) 261. Y Okamoto, M Odawara, H Onimatsu, T Imanaka Ind. Eng. Chem. Res. 34 3703 (1995) 262. D K Lee, I C Lee, S K Park, S Y Bac, S I Woo J. Catal. 159 219 (1996) 263. F Delanney Appl. Catal. 16 135 (1985) 264. A I Startsev, V I Zaikovskii Kinet. Katal. 35 288 (1994) a 265. E Diemann, Th Weber, A Muller J. Catal. 148 288 (1994) 266. A A Spozhakina, N G Kustova, I N Yuchnovski, D M Shopov, T M Yurieva, T Kh Shokireva Appl. Catal. 39 331 (1988) 267. Z Paal, T I Koran'i, M A Dobrovol'skii Kinet. Katal. 28 127 (1987) a 268. D A Agievskii, L I Pavlova, V I Kvashonkin,M A Kipnis, M A Landau, G D Chukin, B K Nefedov Kinet. Katal. 29 1452 (1988) a 269. S M A M Bouwens, D C Koningsberger, V H J de Beer, R Prins Catal. Lett. 1 55 (1988) 270. Z B Wei, Q Xin, S S Sheng, H P Chen Cuihua Xuebao 16 359 (1995); Chem. Abstr. 123 29 1291 (1995) 271. J P R Vissers, B Scheffer, V H J de Beer, J A Moulijn, R Prins J. Catal. 105 277 (1987) 272. B Scheffer, P Arnoldy, J A Moulijn J. Catal. 112 516 (1988) 273. H C Pratt, J V Sandri, V Christov J. Catal. 124 416 (1990) 274. M De Boer, A J van Dillen, D C Koningsberger, J W Geus J. Phys. Chem. 98 7862 (1994) 275. R Candia, J Villadsen, N-Y Topsùe, B S Clausen, H Topsùe Bull. Soc. Chim. Belg. 93 763 (1984) 276. Yu I Yermakov, A N Startsev, V A Burmistrov, O N Shumilo, N N Bulgakov Appl. Catal. 18 33 (1985) 277. J A Rodriguez, S Y Lee, J Hrbek, H H Huang, C-Q Lee J. Phys. Chem. 100 14476 (1996) 278. J M J G Lipsch, G C A Schuit J. Catal. 15 163 (1969) 279. R J N Voorhoeve J. Catal. 23 236 (1971) 280. A L Farragher, P Cossee, in Proceedings of the 5th International Congress on Catalysis, Miami Beach, 1972 Vol. 2 (Amsterdam: North Holland, 1973) p. 1301 281. B Delmon React. Kinet. Catal. Lett. 22 1 (1983) 282. M Karroua, H Matralis, P Grange J. Catal. 133 371 (1993) 283. S I Kim, S I Woo Appl. Catal. 74 109 (1991) 284. S I Kondrat'ev, S B Nikishenko, G V Antoshin, A A Slinkin, E A Fedorovskaya, Kh M Minachev Kinet. Katal. 25 1169 (1984) a 285. M W J Crajg, V H J de Beer, A M van der Kraan Appl. Catal. 70 L7 (1991) 286. R Candia, B S Clausen, J Bartholdy, N-Y Topsùe, B Rengeler, H Topsùe, in Proceedings of the 8th International Congress on Catalysis, West Berlin, 1984 Vol. 2 (Basel: Weinheim, 1984) p. 375 287. M J Ledoux, G Maire, S Hantzer, O Michaux, in Proceedings of the 9th International Congress on Catalysis, Calgary, 1988 (Ottawa: Chem. Inst. Canada, 1988) Vol. 1, p. 74 288. H Topsùe, B S Clausen, R Candia, S Moing J. Catal. 68 435 (1981) 289. R Candia, B S Clausen, H Topsùe J. Catal. 77 364 (1982) 290. S M A M Bouwens, F B M van Zon,M P van Dijk J. Catal. 146 375 (1993) 291. O Sorensen, B S Clausen, R Candia, H Topsùe Appl. Catal. l3 363 (1985) 292. B Scheffer, E M van Oeri, P Arnoldy, V H J de Beer, J A Moulijn Appl. Catal. 25 303 (1986) 293. M W J Craje, V H J de Beer, J A R van Veen, A M van der Kraan J. Catal. 143 601 (1993) 294. A M van der Kraan,M W J Craje, E Gerkema, W L T M Ramselaar Appl. Catal. 39 L7 (1988) 295. J P R Vissers, V H J de Beer, R Prins J. Chem. Soc., Faraday Trans. 1 83 2145 (1987) 296. K Imamura, R Prins J. Catal. 147 515 (1994) 297. T Komatsu,W K Hall J. Phys. Chem. 96 8131 (1992) 298. E D Radchenko, M V Landau, B K Nefedov, L I Nikulina, G D Chukin, P A Zhdan, A A Slinkin, in Proceedings of the 8th International Congress on Catalysis, West Berlin, 1984 Vol. 2 (Basel: Weinheim, 1984) p. 399 299. M E Ekman, J W Anderegg,G L Schrader J. Catal. 117 246 (1989) 300. M J Ledoux, G Michaux, G Agostini, P Panissod J. Catal. 96 189 (1985) 301. D A Agievskii, L I Pavlova, M A Landau, V I Kvashonkin, G D Chukin Kinet. Katal. 26 1207 (1985) a 302. S N Kondrat'ev, S B Nikishenko, G V Antoshin, A A Slinkin, E A Fedorovskaya, Kh M Minachev Kinet. Katal. 26 699 (1985) a 303. S P A Louwers, R Prins J. Catal. 139 525 (1993) 304. A W Scaroni, R G Jenkins, P L Walker Appl. Catal. 14 173 (1985) 305. J Abart, E Delgado, G Ertl, H Jeziorowski, H Knozinger, N Thilde, X Zh Wang Appl. Catal. 2 155 (1982) 306. B L Meyers, R L Mieville Appl. Catal. 14 207 (1985) 307. J L Garcia Fierro, J Soria, A Lopez Agudo Appl. Catal. 3 117 (1982) 308. A A Slinkin, T N Kucherova, G A Ashavskaya, T K Lavrov- skaya Kinet. Katal. 25 1213 (1984) a 309. M W J Craje, S P A Louwers, V H J de Beer, R Prins, A M van der Kraan J. Phys. Chem. 96 5445 (1992) 310. A Lopez Agudo, F J Gil Leambias, P Reyes, J L Garcia Fierro Appl. Catal. 1 59 (1981) 311. N S McIntire, T C Chan, P A Spevack, J R Brown Appl. Catal. 63 391 (1990) 312. P Gajardo, A Mathieux, P Grange, B Delmon Appl. Catal. 3 347 (1982) 313. M L Vrinat, L De Mourgues Appl. Catal. 5 43 (1983) 314. B Scheffer, J Mangnus, J A Moulijn J. Catal. 121 31 (1990) 315. B H Upton, C C Chen, N M Rodruguez, R T K Baker J. Catal. 141 171 (1993) 316. J C Duchet J. Catal. 81 386 (1987) 317. R R Chianelli, A F Ruppert,M Jose-Yacaman, A Vazquez- Zavala Catal. Today 21 269 (1995) 318. G V Isagulyants, A A Greish, V M Kogan Kinet. Katal. 28 243 (1987) a 319. J Jayamurthy, S Vasudevan J. Phys. Chem. 98 6777 (1994) 320. L Portela, P Grange, B Delmon, in Proceedings of the 10th Inter- national Congress on Catalysis, Budapest, 1992 (Budapest: Akade- miai Kiado, 1993) Vol. A, p. 559 321. L Portela, P Grange, B Delmon J. Catal. 156 243 (1995) 322. F C H Mitchel, D A Green, E Payen, A C Evans J. Chem. Soc., Faraday Trans. 91 4467 (1995) 323. A N Startsev Kinet. Katal. 36 515 (1995) a 324. A N Startsev Catal. Rev. 37 353 (1995) 325. A N Startsev, S A Shkuropat, V V Kriventsov, D I Kochubei, K I Zamaraev Mendeleev Commun. 6 (1991) 326. H Topsùe, B S Clausen, N-Y Topsùe, J K Norskov, C V Ovesen, C J H Jacobson Bull. Soc. Chim. Belg. 104 283 (1995) 327. T Kabe, W Qian, A Ishihara J. Catal. 148 171 (1994) 328. E J Markel, G I Schrader, N N Sauer, R J Angelici J. Catal. 116 11 (1989) 329. N N Sauer, E J Markel, G I Schrader, R J Angelici J. Catal. 117 295 (1989) 330. E J M Hansen,M J Vissenberg, V N J de Beer, J A R van Veen, R A van Santen, in Posters of the 11th International Congress on Catalysis, Baltimore, 1996 (Baltimore: North Am. Catal. Soc., 1996) Po-363 331. V V Rozanov, Tsao Yamin, O V Krylov Kinet. Katal. 34 131 (1993) a 332. V V Rozanov, Tsao Yamin, O V Krylov Kinet. Katal. 37 598 (1996) a 333. V V Rozanov, Tsao Yamin, O V Krylov Kinet. Katal. 37 603 (1996) a 334. V V Rozanov, Tsao Yamin, O V Krylov Kinet. Katal. 37 871 (1996) a 335. V V Rozanov, Tsao Yamin, O V Krylov Kinet. Katal. 37 877 (1996) a 336. B G Supernagel, T A Pecoraro, R R Chianelli J. Catal. 78 380 (1982) a 337. D A Agievskii, L I Pavlova, V N Kvashonkin Kinet. Katal. 30 1268 (1990) a 338. J Yasumori, M Yamada, M Houalla, D M Hercules, in Proceed- ings of the 10th International Congress on Catalysis, Budapest, 1992 (Budapest: Akademiai Kiado, 1993) Vol. B, p. 1867 Heterogeneous catalysts of hydrogenation 615339. C Muralidher, F E Massoth, J Shabtai J. Catal. 85 44 (1984) 340. M Lacroix, M Vrinat,M Breysse Appl. Catal. 21 73 (1986) a�Kinet. Catal. (Engl. Transl.) b�Mendeleev J. (Engl. Transl.) c�Russ. Chem. Bull. (Engl. Transl.) d�Russ. J. Appl. Chem. (Engl. Transl.) e�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) 616 M D Navalikhina,

ISSN:0036-021X    

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Abstract. Data on the synthesis of reactive derivatives of nucleic acid analogues and their application for the study of structure and functions of biopolymers are generalised. The main types of such analogues including photoactivated reagents containing azido- aryl, halogeno, and thiol groups, psoralen and its derivatives, platinum-based reagents, and nucleic acid analogues containing substituted pyrophosphate or acyl phosphate internucleotide groups are presented. The mechanisms of interaction of these compounds with proteins and nucleic acids are considered.The prospects for the in vivo application of reactive nucleic acids in various systems are discussed. The bibliography includes 76 refer- ences. I. Introduction Synthetic fragments of nucleic acids containing modified units are widely employed in solving practical problems of molecular biology, biotechnology, and medicine.Derivatives of nucleic acids containing reactive groups are of special interest. These compounds can form covalent bonds with the reactive groups of biopolymers and are promising tools for the study of the top- ography of active centres of proteins, their functional features, and molecular mechanisms of DNA± protein recognition.The interest in these compounds has been considerably increased in recent years owing to the development of sense biotechnology. The main attention is concentrated on the search for new active groups and elaboration of methods for their directed introduction into nucleic acids. This review describes the main types of reactive analogues of nucleic acids, viz., photoactivated reagents, platinum-based reagents, and new promising analogues of nucleic acids containing substituted pyrophosphate or acyl phosphate internucleotide groups.The use of reactive nucleic acids in the analysis of proteins and nucleic acids and prospects for their further application in various systems in vivo are surveyed.A large class of alkylating reagents first proposed by Grineva and Knorre in the late 60's is beyond the scope of this review because this material has been considered rather comprehensively in recent reviews (see, e.g., Refs 1, 2). For the same reason, reagents with redox mechanisms of action 1, 2 have been omitted from consideration. II. Photoactivated nucleic acid analogues In recent years, reactive nucleic acids containing photoactivated groups have gained wide acceptance as helpful tools for directed modification of proteins and nucleic acids and for the investiga- tion of the structure and functions of these biopolymers.Nor- mally, groups that are activated upon irradiation with light of wavelength >300 nm are employed. This excludes the possibility of other photochemical transformations of the biopolymers themselves.Among photoactivated compounds, nucleic acid analogues carrying azidoaryl, halogeno, or mercapto groups and various psoralen derivatives are the most popular. 1. Nucleic acid derivatives containing the azido group Reactive azido derivatives of oligonucleotides include synthetic oligonucleotides that are modified at the sugar-phosphate back- bone or at heterocyclic bases by reagents containing the azido group.Irradiation converts them into reactive intermediates (nitrenes), which can further react with reactive groups of proteins and nucleic acids. a. Introduction of azidoaryl groups into nucleic acids Azidoaryl groups are introduced into synthetic oligonucleotides either in the course of their synthesis with the use of modified nucleotide components 3 or by postsynthetic modification of oligonucleotides by various aryl azide derivatives.The most popular reagents contain p-azidophenyl, 2-nitro-5-azidophenyl, and azidoperfluorophenyl groups,4,5 or azidoproflavin.6 Synthesis of azido derivatives of oligonucleotides makes use of the reactive groups inherent or incorporated into the oligonucleo- tides.The latter kind of groups involves aliphatic amino and thiol groups. Thus, an oligodeoxyribonucleotide azidoperfluorophenyl derivative 1 was obtained by acylation of the terminal amino group that had been introduced into the oligonucleotide using p-azidotetrafluorobenzoic acid N-hydroxysuccinimide ester 2.5 I E Kanevskii, SAKuznetsova Department of Chemistry,MVLomonosov Moscow State University, Leninskie Gory, 119899 Moscow, Russian Federation.Fax (7-095) 939 31 81. Tel. (7-095) 939 31 53 Received 9 September 1997 Uspekhi Khimii 67 (7) 688 ± 704 (1998); translated by R L Birnova UDC 547.863.32; 577.113.4 Synthesis of reactive nucleic acid analogues and their application for the study of structure and functions of biopolymers I E Kanevskii, S A Kuznetsova Contents I.Introduction 617 II. Photoactivated nucleic acid analogues 617 III. Platinum-containing nucleic acid analogues 626 IV. Reactive nucleic acids containing substituted pyrophosphate and acyl phosphate groups 628 V. Prospects for the application of reactive oligonucleotides for directed in vivo modification of biopolymers 630 Russian Chemical Reviews 67 (7) 617 ± 631 (1998) #1998 Russian Academy of Sciences and Turpion LtdR is the oligodeoxyribonucleotide residue, n=2 or 6.The azido derivative of E. coli tRNAPhe 3 was obtained by acylation of the amino group of the phenylalanine residue that had been introduced into tRNA with N-(p-azidobenzoyl)glycine N-hydroxysuccinimide ester.7 A similar approach was used to introduce the azido group at the exocyclic amino group of cytidine.p-Azidobenzoic acid N-hydroxysuccinimide ester was used as the acylating reagent. This reaction gave the pentanucleotide azido derivative 4.8 R�TAG Various azido derivatives of the oligonucleotide d(UCCACTT) 5 were obtained by acylation of the amino group attached through a spacer to C(5) of the terminal deoxyuridine residue with p-azidotetrafluoro-, 5-azido-2-nitro-, or p-azidoben- zoic acids N-hydroxysuccinimide esters.4,9 5-[3-(p-Azidobenzoyl)aminoprop-1-enyl]deoxyuridine 50-tri- phosphate (6) was used to incorporate the azido group into the tRNATyr SUP4 gene.This compound was obtained by the reaction of 5-(3-aminoprop-1-enyl)deoxyuridine 50-triphosphate with p-azidobenzoic acid N-hydroxysuccinimide ester (DMF, pH 8.5, 20 8C, 3 h in the dark).10 This was followed by incorpo- ration of compound 6 into DNA using DNA polymerase T4.The DNAthus modified was used for probingDNA± protein contacts in a specific transcription complex comprising the modified DNA, RNA polymerase III, and several transcription factors. Recently, a novel azido-containing reagent 4-[2-(p-azidoben- zoyl)aminoethyl]deoxycytidine 50-triphosphate (7) has been pro- posed for this purpose.11 This compound was also obtained using azidobenzoic acid N-hydroxysuccinimide ester.The first step of this synthesis included the hydrogen sulfite- catalysed transamination of the amino group of deoxycytidine 50-triphosphate with ethylenediamine and isolation of the reaction product by column chromatography.The reaction of 4-(2-amino- ethyl)deoxycytidine 50-triphosphate with a 10-fold excess of p-azidobenzoic acid N-hydroxysuccinimide ester was conducted at reduced illumination. The target product 7 was also isolated by column chromatography and analysed by TLC. Compound 7 was introduced into a specific region of the tRNATyr SUP4 gene. The modified DNA was used for the probing of DNA± protein contacts in the transcription complex.The reaction of 5-fluoro-2,4-dinitrophenylazide with the amino group of the ethylenediamine spacer attached at the terminal phosphate group of oligoribonucleotides gave photo- activated analogues of mRNA 8, which represent oligoribonu- cleotide derivatives of various lengths having an arylazido group at the 50-end of the oligoribonucleotide fragment.12 pN=pA, n=5; pN =pU, n=4, 7, and 8.The reagents 8 were used for affinity modification of the 50-region of the mRNA-binding centre of the E. coli ribosome. To study the localisation of the protein environment of DNA in chromatin, Kobets et al.13 have used a photoactivated deriva- tive of hexadecathymidylate 9, in which the p-azidoaniline residue was linked to the 50-terminal phosphate group through a 1,4- butanediamine spacer.RO P O O7 NH(CH2)nNHN O C O O O N3 F F F F + 2 RO P O O7 NH(CH2)nNH F F F F N3 O C 1 tRNAOCCHNH2+ N O O OCCH2NHCC6H4N3-p O O O CH2Ph pH 8.2 20 8C, 1.5 h p-N3C6H4CNHCH2CNHCH O O CH2Ph 3 (95%) C O O tRNA OpdR O N pdTpO N O NHCO N3 4 R=d(CCACTT); X=CH2, CH2OCH2CH2, CH2NHCOCH2CH2; R0 = F F N3 F F N3 NO2 N3 , , .O HN N O R X NH2 + N O O OR0 HN N O O R X NHR0 5 C O C O C O HN N O O dRib ± ppp CH 6 CHCH2NHC C6H4N3-p O N N O dRib ± ppp NH2 a HN N O dRib ± ppp NH(CH2)2NH2 b N N O dRib ± ppp NH(CH2)2NH C O 7 C6H4N3-p (a) H2N(CH2)2NH2, NaHSO3; (b) , pH 8.5, 3 h, 20 8C. NOCOC6H4N3-p O O O2N NO2 F+H2N(CH2)2NH(pN)n NaHCO3 (pH 9.8) 20 8C, 40 min N3 O2N N3 NO2 NH(CH2)2NH(pN)n 8 618 I E Kanevskii, S A KuznetsovaCompound 9 was obtained by phosphorylation of 4-(p- azidoanilino)butylamine with the oligonucleotide N-methylimi- dazolide.Earlier, Godovikova et al.14 have studied the ability of amino acid residues to undergo modification by oligonucleotide derivatives in which the p-azidoaniline fragment was linked to the terminal phosphate group of the oligonucleotide directly rather than through a spacer.It was found that these reagents yield p-benzoquinonimide derivatives on irradiation and subsequent reaction with amino acids results in splitting of the reagent from the oligonucleotide. On the contrary, modification of proteins with reagents of the type 9 occurs specifically to give covalently bound nucleic acid ± protein complexes.13 In addition to arylation and aroylation of aliphatic amino groups, alkylation or acylation of a thiol group that had been introduced into oligonucleotides is also employed for the prepa- ration of azido derivatives of oligonucleotides using the corre- sponding halide derivatives of aryl azides.Thus acylation of the 50-terminal thiolphosphate of oligothymidylate with p-azidoben- zoyl bromide gives the p-azidobenzoyl derivative 10.15 R is the oligothymidylate residue.Analogously, the reaction of 3-azido-6-(3-bromopropylami- no)acridine 6 or p-azidobenzoyl bromide 16 with the terminal thiolphosphate group of oligothymidylate yielded azidoproflavin (11) or p-azidobenzoyl (12) derivatives of oligonucleotides, respec- tively. R is the oligodeoxyribonucleotide residue.p-Azidophenacyl bromide 17 was also used for modification of the oligonucleotide 13 containing an internucleotide thiolphos- phate group that had been introduced in the oligonucleotide synthesis.17 R and R0 are fragments of the oligonucleotide sequence. The azido group can be incorporated through a thiol group into a heterocyclic base as well. Thus 5-[(4-azidophenacyl)thio]- cytidine 50-triphosphate (15) has been synthesized according to Scheme 1.18 All the reactions were carried out in the dark or under reduced illumination.Scheme 1 Treatment of cytidine 50-triphosphate 16 with methyl hypo- bromite gave a bromomethoxy derivative 17. The reaction prod- uct of compound 17 with sodium disulfide was isolated by extraction and column chromatography.The yield of 5,5-dithio- bis(cytidine 50-triphosphate) 18 was 10%. Such a low yield was attributed to the negative charge of both reacting molecules.18 An alternative route to the disulfide from a nucleoside was also attempted.18 In this case, the yield of the reaction products was much higher. Then, the 20- and 30-OH groups of ribose were protected and the 50-OH group of the disulfide derivative was phosphorylated.The final stage of this synthesis was more complicated and labourious due to a greater number of purifica- tion steps and more complex isolation procedure of intermediate compounds, which resulted in nearly the same yield of 5,500-dithio- bis(cytidine 50-triphosphate) 18. 5-Mercaptocytidine 50-triphos- phate 19 was obtained from compound 18 by reduction with sodium borohydride. The reaction mixture was treated with p-azidophenacyl bromide in DMSO. The target product 15 was isolated by extraction and column chromatography (80% yield).The azido derivative of the nucleoside triphosphate 15 was introduced into various positions of RNA in the course of polymerase synthesis. Thus, the methods available allow one to incorporate the azidoaryl groups both into heterocyclic bases of nucleic acids and the sugar-phosphate backbone.They involve acylation of ali- phatic amino groups of nucleic acids with N-hydroxysuccinimide esters of azidobenzoic acid and its derivatives or alkylation (acylation) of a thiol group that had been introduced into nucleic acids using the corresponding azidoaryl halides.As a rule, these reactions occur with high efficiency (the product yield is 80%± 99%). However, whereas the former route makes use of the exocyclic amino groups of heterocyclic bases of nucleic acids for the incorporation of the azido groups, in the case of azidoaryl halides an additional step is required , viz., the incorporation of a thiol group into nucleic acids.This may complicate the procedure and reduce the yields of the target products. b. Cross-coupling of nucleic acid derivatives containing the azido groups with biopolymers On irradiation with light (photolysis), compounds containing azidoaryl groups generate highly reactive short-lived radicals, viz., nitrenes. In the singlet state, the lifetime of nitrenes depends on the nature of substituents in the aromatic nucleus and varies from nano- to microseconds.19 This state of nitrene radicals is very unstable: they either undergo intramolecular rearrangements to yield various reactive intermediates or react with neighbouring reactive groups.In the latter case, intermolecular covalent bonds may be formed. Using azidophenyl derivatives as examples, it has been shown that introduction of fluorine atoms or nitro groups into the benzene ring decreases the probability of intramolecular N3 NH(CH2)4NH (pdT)16 9 N3 CBr O +HS P OR O O7 N3 C O S P OR O7 O 10 N N3 NH(CH2)3S P (d(a)T)8 O O7 11 N3 C S O O7 O OR P 12 ROT O P O S7 O CGAOR0 + BrCH2 C O N3 pH 7.0 3 h, 20 8C 13 ROT O P O S O CGAOR0 CH2 C O N3 14 (99%) N N O NH2 Rib ± ppp O 16 a N N O NH2 Rib ± ppp Br OMe 17 b N N NH2 Rib ± ppp S S Rib ± ppp NH2 O N N 18 (10%) c d N N O NH2 Rib ± ppp SH N N O NH2 Rib ± ppp S CH2COC6H4N3-p 19 15 (a) MeOBr, 4 8C, 12724 h; (b) Na2S2, 4 8C, 5 h; 20 8C, 48 h; (c) NaBH4, H2O, 0 8C, 3 h; (d) BrCH2COC6H4N3-p, DMSO, 4 8C, 12 h.Synthesis of reactive nucleic acid analogues and their application for the study of structure and functions of biopolymers 619rearrangements and simultaneously increases the probability of nitrene involvement in intermolecular reactions.20 For the first time, azido derivatives of oligonucleotides have been used for the study of the formation of cross-coupled DNA duplexes 20 from a 27-membered matrix oligonucleotide and a complementary octa(a-thymidylate) carrying an azidoproflavin residue linked with the 50-thiolphosphate group of the a-anomer through a spacer arm.6 UV irradiation of oligonucleotides constituting the system 20 (l>300 nm) in 0.1 M NaCl resulted in cross-linking.The photo- addition product was treated with an alkali, which resulted in selective cleavage at modification sites. The residues T9 and T18 (1 : 3) of the DNA template were found to be the main sites of this cleavage. This ratio was rationalised by postulating parallel and antiparallel orientations of the oligonucleotide probe during the formation of a complementary complex or the existence of a triplex structure.Further experiments showed that the second hypothesis was more probable, since octa(a-thymidylate) in 0.1 M NaCl can form ternary complexes with the template.Photomodification of the pentadecanucleotide 21 and the hexadecanucleotide 22 with the oligonucleotide derivatives 23 ± 28 carrying the azido group at various positions of the nucleotide probe has been carried out by Levina et al.21 The oligonucleotide derivatives employed contained the azido group attached through a spacer to the 30(50)-terminal phosphates (reagents 23, 25, and 26) or to the C(5) atom of a terminal deoxyuridine (reagent 24) or deoxyuridine inside the chain (reagents 27 and 28).21 The reagents complementary to the adjacent sites of the target were also used in pairs: tandems of reagents 23 + 25 and 23 + 26.UV irradiation at l = 303 ± 365 nm yielded products of covalent addition of one or two (in the case of tandems) reagents to the target.The efficiency of photo- addition of reagents 23 and 26 containing the azido group (at the 50- or 30-ends, respectively) to the pentadecanucleotide was approximately the same, and the yields of the photoaddition products were 70%. However, the use of a tandem of these reagents increases the efficiency of photomodification up to 80%. The modification predominantly affects the G9 residue of the oligonucleotide 21, which is not involved in the formation of the complementary complex.The adduct yield decreases to 55% and 51%, respectively, when reagents 27 and 28 containing the uridine azido derivative inside the chain are used. The decrease in the photoreaction efficiency is apparently due to the involvement of the modified base in a complementary reaction.This is confirmed by the fact that the efficiency of photoaddition in the case of reagent 24 having an azidonucleotide at its 50-end reaches 69% and modification affects predominantly theG7 residue of the oligonucleotide 22, whereas the G8 residue is modified in a lesser degree. Highly reactive nitrenes generated from the azido derivatives of oligonucleotides upon photolysis can instantaneously react with the neighbouring reactive groups as well as with amino acid residues of a DNA± protein complex.Investigation of the structure of the transcription factor TFIIIC of RNA polymerase III was carried out using a genomic DNA sequence of yeast tRNATyr SUP4 containing one or several azido derivatives of deoxyuridine 6.10 A complex of modified DNA with the protein was irradiated with UV light (l=254 nm) for 5 min.Analysis of photoaddition products allowed one to identify four subunits of TFIIIC and to establish their position relative to SUP4 in the DNA± protein complex. This study was extended by using a genomic DNA sequence SUP4 containing azido derivatives of cytidine 7.11 In contrast with the previous example, a non-purified protein fraction was used for modifica- tion.Analysis of photoaddition products revealed a protein with molecular mass of 40 kDa, which could not be detected when purified TFIIIC has been used. It was assumed that this protein is a component of the transcription complex. Affinity modification of the 50-region of the mRNA-binding centre of the E. coli ribosome was studied using N-[2-(2,4-dinitro- 5-azidophenyl)aminoethyl]-50-phosphoramidites of oligouridy- lates 8.12 It was found that compounds 8 (N = U, n = 4, 7, and 8) stimulate binding of code-specific aminoacyl-tRNA with the ribosome.Upon UV irradiation of the ribosome ± reagent 8 ±tRNA ternary complex (l>350 nm), covalent binding of the reagent 8 with the ribosome occurred. Up to 10% of the reagent 8 in the ternary complex was cross-linked with the ribosomal proteins of the 30S and 50S subunits. It was found that in the 30S subunit modification affected proteins S3, S4, S9, S11, S12, S14, S17, S19, and S20.In the 50S subunit, it was proteins L2, L13, L16, L27, L32, and L33 that were modified. The array of the modified proteins strongly depended on the length of the oligo- nucleotide fragment of the reagent and on the binding of the tRNA molecule to the site A of the ribosome. By virtue of their ability to form stable covalent bonds with the reactive groups of proteins, azido derivatives of oligonucleotides are successfully employed for identifying amino acid residues in the active centres of DNA-recognising proteins.Thus identifica- tion of the amino acid residues of the restriction endonuclease TaqI that reacts with the recognition site was made using the photoreagent 13 containing the azido group just at the TaqI cleavage site.17 The complementary complex `oligonucleotide probe 13 ± template DNA' was UV-irradiated (l = 306 nm, 5 min) following 5-min incubation with TaqI.The efficiency of formation of the covalently bound DNA± protein complex was 15%± 20%.An analysis of photoaddition products allowed one to identify the amino acid residue involved in the formation of a covalent bond with the oligonucleotide probe as Tyr161. Modified fragments of RNA containing 5-(4-azidophen- acyl)thiocytidine residues at the 30-end or in the middle of the chain were used for photoaffinity modification of T7 and E.coli RNA polymerases.18 The complexes of modified RNA with T7 and E. coli RNA polymerases were UV-irradiated (l=302 nm) for 2 ± 4 min at 20 8C. An analysis of photoaddition products revealed RNA-binding domains in the corresponding RNA polymerases. In recent years, azido derivatives of oligonucleotides have been employed for modification of the proteins of chromatin. Thus reagent 9 was used to study structural rearrangements in chromatin during the cell cycle.13 To this end, the spectra of proteins modified in the nuclei isolated from HeLa cells in two different phases of the cell cycle, viz., the S-phase and in the region X= ; Sp is the 50-thiolphosphate residue. 50d(....5TGAGTAAAAAAAATGAGT22....) X7Sp[a]d(TT T TTT TT) 20 N3 + N NH(CH2)3 H 27 50 TAAGTGGAGTTTGGC 50 AGAAAGTGAGTGTATC 21 22 30 TUCACCT 30 X7pCAAACCG 30 X7pAAACCG 25 26 30 TTCACCTp7X 30 CTTTCACU7X1 23 24 X=HN(CH2)3NHCOC6H4N3-p, X1=CH2O(CH2)2NHCOC6H4N3-p.X1 28 30 CACUCAC X1 620 I E Kanevskii, S A Kuznetsovaof transition from the G1 to the S-phase (G1/S), were compared. Modification of proteins in the nuclei isolated in the S-phase of the cell cycle gave five nuclear proteins with molecular masses 160, 87, 72, 59, and 46 kDa and was specific.The spectrum of the modified proteins changed on going from the S-phase to the G1/S interface. In particular, the 160 kDa protein was not modified. The results obtained by other authors 13 confirm the idea that photoactivated oligonucleotides are promising for studies of the structure and functions of chromatin.Wide application of azido derivatives of nucleic acids in the studies of nucleic acid ± nucleic acid and nucleic acid ± protein interactions is indicated by the existence of reproducible protocols for the synthesis and their high reactivity. However, in some cases the high reactivity of nitrenes photogenerated from the azido derivatives may decrease the specificity of the reactions and thus imposes certain limitations on the application of azidoaryl deriv- atives of nucleic acids in structure ± function studies of biopol- ymers. 2.Derivatives of nucleic acids containing halogen or mercapto groups In recent years, photoactivated oligonucleotide derivatives with bromo, iodo, or mercapto groups have been widely employed for the study of nucleic acid ± protein or nucleic acid ± nucleic acid interactions.The interest in these compounds is due to their ability to generate free-radical intermediates that can form covalent bonds with electron-donor groups of proteins and nucleic acids on irradiation with long-wave UV light. Reactions according to radical mechanisms are also possible, but only in the case where the reacting groups are at a distance of several angstroÈ m.22 This circumstance makes possible the use of oligonucleotide halogeno and thiol derivatives for identifying direct nucleic acid ± nucleic acid or nucleic acid ± protein contacts.a. Synthesis of halogeno and mercapto derivatives of oligonucleotides Oligonucleotide derivatives carrying halogeno or mercapto groups are usually obtained by chemical or enzymic syntheses using modified nucleotide components, such as 4-mercapto or 5-bromo(iodo) uridine or cytidine.DNA derivatives containing halogeno or mercapto groups are obtained by solid-phase phos- phoramidite synthesis with bromo-, iodo-, or mercapto-deoxy- ribonucleoside phosphoramidites as monomeric units. Thus, 5-bromo-20-deoxyuridine (5-BrdU) derivatives of DNA that included a recognition site for the transcription factor NF-kB were obtained by Liu et al.23 The modified uridine was introduced into various positions of DNA by solid-phase synthesis with 5-BrdU phosphoramidites.The 5-iodo-20-deoxyuridine derivative of the deoxyribo- nucleotide GCTCACCGAA± IdU ±GC 29 was synthesised in a similar way 24 from compound 30 obtained in quantitative yield according to the following scheme: It should be noted that the high sensitivity of aromatic iodides to nucleophilic substitution prevents deprotection of internucleo- tide phosphates and heterocyclic bases by conventional treatment with NH4OH at 55 8C.Removal of protective groups leading to the oligomer 29 could be carried out by treatment with NH4OH for 48 h at room temperature.Under these conditions, no nucleophilic substitution of iodine in the aromatic nucleus occurred, while phosphate groups and heterocyclic bases were completely deprotected. Halogeno or mercapto derivatives of oligoribonucleotides are also obtained using modified nucleoside triphosphates as sub- strates for T7 RNA polymerase in the polymerase-catalysed synthesis of RNA.Thus a 19-membered oligoribonucleotide 31 containing 5-bromouridine (5-BrU) at position (75) was used to identify amino acid residues in the site of binding of the bacter- iophage R17 membrane protein with genomic RNA.25 The syn- thesis of compound 31 was carried out using 5-bromouridine 50- triphosphate by a T7 RNA polymerase-catalysed synthesis on a synthetic DNA template.The same approach was used to obtain oligoribonucleotides with a nucleotide sequence analogous to that of compound 31 but with U or C 5-iodo derivatives at position (75).26 A method for the synthesis of 4-(2-cyanoethyl)thiouridine 32 phosphoramidite has been proposed (Scheme 2).27 This com- pound was further used in the chemical synthesis of modified RNA. The 20-OH group of ribose was protected by a 1-(2- fluorophenyl)-4-methoxypiperidin-4-yl group. The mercapto group of uridine was protected as a 2-cyanoethyl derivative.The reaction of 20-O-[1-(2-fluorophenyl)-4-methoxypiperidin- 4-yl]-50-O-dimethoxytrityluridine with methoxyacetic anhydride in dry pyridine gave the derivative 33 in quantitative yield. To obtain the uridine derivative 35 with a cyanoethyl group at position 4, compound 33 was treated with 2,4,6-triisopropylben- zenesulfonyl chloride, after which the sulfonate 34 was introduced into reaction with 3-mercaptopropionitrile.The resulting com- pound was treated with ammonia in methanol to deprotect the 30-hydroxy group of ribose. Phosphitylation of the 30-OH group of the protected ribonucleoside 35 was carried out with 2-cyanoethyl bis(diisopropylphosphoramidite).After purification, compound 32 was obtained in 44% yield and further used in a solid-phase phosphoramidite synthesis to obtain photoactivated derivatives OH O N HO HN O O I OH O N DMTrO HN O O I a b (a) DMTrCl=(p-MeOC6H4)2C(Ph)Cl, Py; (b) (Pri 2N)2P(Cl)OMe, Pri 2NEt, CH2Cl2, 20 8C, 10 min. Pri 2N O O N DMTrO HN O O I P OMe 30 C G G C C G G C G C A A A A G G G ABrU75 ppp 31 Synthesis of reactive nucleic acid analogues and their application for the study of structure and functions of biopolymers 621ofRNAcontaining 4-mercaptouridine in different positions of the recognition sites of the HIV Rev protein and the bacteriophage MS2 membrane protein.b. The use of oligonucleotide halogeno and mercapto derivatives for cross-linking to proteins and nucleic acids As has been mentioned above, UV irradiation of bromo(iodo) or mercapto derivatives of oligonucleotides generates reactive inter- mediates, which interact with reactive groups of proteins and nucleic acids. Using an oligonucleotide derivative containing 5-BrU as an example, it was shown that the mechanism of photoactivated reaction of halogeno derivatives with the protein amino groups depends on the wavelength.28 Upon UV irradiation with l=308 nm, the 5-BrU molecule passes from the low-energy singlet state to the reactive triplet state n,p* responsible for the binding of BrU ±DNA with the associated protein, presumably through the initial electron transfer.The intermolecular covalent binding of BrU with amino acid residues, such as cysteine, glutathione, or tryptophan, is accompanied by elimination of HBr.Upon UV irradiation with l=254 nm, the 5-BrU molecule passes into a high-energy singlet state p,p*, which results in the cleavage of the C± Br bond to yield free radicalsU . and a Br atom. The highly reactive radical U . either reacts with the reactive groups of the nearest environment (in particular, with the for- mation of a covalent bond with the amino groups of a protein) or passes into the triplet state, 3p,p*.29 Synthetic analogues of the promoter region of DNA contain- ing several 5-BrdU residues were used to study covalently bound complexes of a series of transcription factors of the albumin gene (HNF1, APF, vAPF, a1TEB, HPI) with the promoter regions of this gene.30 Yet another aim of this study was to identify the polypeptide chains involved in the formation of a vAPF ±APF complex and the nucleotide residues involved in the interaction with the protein.Complexes ofDNA derivatives with the proteins were irradiated at l=321 nm. An analysis of cross-linking products revealed that the DNA± protein contact sites were distributed asymmetrically on a pseudopalyndromic recognition site: the nucleotides crucial for the recognition were located in the middle of the pseudopalyndromic sequence but the recognition site had other contacts of the protein with several nucleotides.In order to investigate the formation of cross-linked com- plexes of a membrane protein with an operator of bacteriophage R17 transcription (R17 RNA) and to identify the amino acid residue responsible for the formation of the cross-linkage, Willis et al.25 proposed to use a synthetic analogue of R17RNAcontaining a 5-BrU residue at position (75) (compound 30).Its complex with the membrane protein was irradiated at l=308 nm for 5 min after which the protein concentration was increased and the complex was irradiated for an additional 5 min.The efficiency of formation of the cross-linked complex was 22%. The amino acid residue involved in this process was identified as Tyr85. It was found that cross-linking did not occur if Tyr85 was substituted by Ser.25 These results are in good agreement with the above- described mechanism of the photoinduced reaction of oligonu- cleotides containing 5-BrU with the amino acid residues of the protein according to which 5-BrU derivatives cross-link only with aromatic amino acids upon irradiation at l=308 nm.29 Earlier, it was found 31 that oligonucleotides containing 5-IU residues are more efficient reagents for cross-linking to proteins than those containing 5-BrU.The efficiency of covalent photo- addition of iodine-containing reagents is usually 3 to 5 times higher than that of their bromo analogues.Therefore, the 5-BrU residue at position (75) of compound 30 was substituted by the 5-IC residue.26 Five additional 5-IC residues were introduced into various positions of RNA 36 in order to demonstrate the specificity of formation of the cross-linked complex. R is the oligonucleotide residue. HN N R O O Br H N CO2Me NHAc + HN NR O O NHAc CO2Me H N hn 7HBr 36 ppp U IC75 G IC IC G U A IC G G IC A A A A G G G U IC U A U 37 ppp A U75 IC G G IC IC G G IC IC G A A A A G G G OFfmp OH O N DMTrO HN O O OFfmp O O N DMTrO HN O O CH2OMe O a b 33 OFfmp O O N DMTrO O O CH2OMe O S O7 O 34 d NCCH2CH2S OFfmp OH O N DMTrO N O 35 (59%) c N OFfmp O O N DMTrO N NCCH2CH2S O P Pri 2N OCH2CH2CN 32 (44%) (a) (MeOCH2CO)2, Py, 20 8C, 24 h; (b) 2,4,6-Pri 3C6H2SO2Cl, Pri 2NEt, 20 8C, 12 h; 2.NH3, MeOH, 20 8C, 9 h; (d) (Pri 2N)2PO(CH2)2CN, CH2Cl2, ,208C, 48 h. Me2N N, Pri 2NEt, 20 8C, 12 h; (c) 1. HS(CH2)2CN, MeCN, N NH N N Scheme 2 622 I E Kanevskii, S A KuznetsovaA complex of the membrane protein with compound 36 was irradiated with UV light (l=325 nm) for 2 h. The efficiency of cross-linking was 75%± 95%.It is noteworthy that the formation of the cross-linked complex was accompanied by partial cleavage of the RNA chain. When an analogous RNA devoid of the 5-IC residue at position (75) (compound 37) was used, the cross-linked RNA± protein complex was not formed. On irradiation at l>300 nm, oligonucleotides containing 4-mercaptouridine reacted by a free radical mechanism and cross-linked through reactive groups with proteins and nucleic acids.32 In the case of nucleic acids, the cross-linking was the most efficient between the adjacent uridine or adenine residues.This makes it possible to use mercapto derivatives of oligonucleotides for the studies of both nucleic acid ± protein and nucleic acid ± - nucleic acid contacts. Thus a series of model mRNAs containing 4-mercaptouridine residues in various positions were synthesised to study cross-linking between mRNA and 18S RNA within a complex with the 80S ribosome (Fig. 1).33 The synthetic mRNAs contained encoding sequences for glycine and/or tryptophan (GGG and ACC, respectively).The mRNA± 80S ribosome complex was irradiated with UV light (l=320 ± 365 nm) in the presence and in the absence of tRNAGly and/or tRNAThr.The efficiency of formation of a cross- linked mRNA± 18S RNA complex depended on the presence of tRNA only for the mRNA 90 derivative and was 20% in the presence of tRNA and 40% in its absence. The efficiency of cross- linking was8%and 14% formRNA80 andmRNA10 derivatives, respectively. It was shown that a covalent bond is formed with U630 and U1111/A1112 of the 80S ribosome 18S rRNA.Synthetic nucleic acid analogues having reactive thiol groups are widely employed in the studies of the structure and function of ribosomes. Thus a series of model RNAs have been synthesised to study the sites of binding of 5S RNA with the 50S subunit of the ribosome 34 in which 18% of uridine residues were statistically substituted by 4-mercaptouridine residues.The 5S RNA± ribo- some 50S subunit complex was irradiated at l=300 nm for 10 min. The yield of the product of cross-linking of 5S RNA to the 23S ribosomalRNAwas more than 75%. The structures of the products formed were not established, although it was shown that U89 of 5S RNA and U2477 of 23S RNA of 50S ribosomal subunits were cross-linked.Thus, the indisputable advantage of halogeno and mercapto derivatives of nucleic acids is their high reactivity and the ability to operate at a zero distance, which makes possible the probing of direct nucleic acid ± protein contacts. At the same time, the high reactivity of these intermediates may be the cause of nonspecific modification of proteins and other cell components and their partial destruction.26 Moreover, the need for additional activation with light of definite wavelengths restricts the application of halogeno and mercapto derivatives of nucleic acids in vivo. 3. Psoralen derivatives of nucleic acids Nucleic acids that contain psoralen residues or its derivatives belong to the most efficient, selective photoreagents that are widely employed for affinity modification and structural analysis of nucleic acids. a.Introduction of psoralen into oligonucleotides Psoralens (7H-furo[3,2-g][1]benzopyran-7-ones) { are aromatic compounds with three linearly fused rings of a-pyrone, benzene, and furan. They contain two photoactivated double bonds, C(3)=C(4) in the pyrone ring and C(40)=C(50) in the furan ring. Irradiation with long-wave UV light (l=300 ± 365 nm) can result in cycloaddition of both these bonds to the C(5)=C(6) bond of thymidine or uridine to form a cyclobutane ring.Thus psoralen derivatives 38a and 38b of a dodecanucleotide containing only one thymidine residue were obtained.35,36 R=d(GAAGC); R0=d(ACGAGC). Photochemical addition of psoralen gives reactive derivatives only if the C(40)=C(50) bond of the furan ring is involved in the reaction with the thymidine residue. In the case of thymidine cycloaddition to the C(3)=C(4) bond of the pyrone ring, the psoralen derivative formed is incapable of further photochemical 40 4 O O O 3 50 OR0 O N RO H N O O O O O Me Me Me Me CH2OH 38a OR0 O RO OH O Me Me Me O O Me O O HN N 38b CH2 B (50) ...CACUAUCUGCACAGGAGCGCAACGGGACCGCACAGCCGAGAGC 720 716 CAGA...C (50) ...AGAGCGGCAAGGAGCGCUAUGGGUGUGCACAGCCGAGAGC 71 CAGACGA +6 +4 73 (50)...AGAGCGCACAGGAGCGCAACGGGACCGCACAGCCGAGAGU A CUGUCUA +20 +26 Figure 1. Model mRNA containing 4-mercaptouridine residues: A � mRNA 80; B � mRNA 90; C � mRNA 100. The encoding sequences are bold- faced, and modified bases are underlined. { Hereinafter the names used in the original publications are given.Synthesis of reactive nucleic acid analogues and their application for the study of structure and functions of biopolymers 623addition.37 Therefore, the introduction of psoralen into terminal phosphate groups of oligonucleotides has now become an espe- cially widespread procedure, for it allows modifications that involve the reactive centres of the psoralen molecule.Thus a psoralen derivative of a 16-membered oligonucleotide 39 has been synthesised by alkylation of its 50-terminal thiolphosphate group with 5-(o-iodohexyloxy)psoralen.38 R=TTTTCTTTTCCCCCCT. 40-Aminoalkylamino-4,50,8-trimethylpsoralen 40 was attached to the 50-phosphate group of a methylphosphonate analogue of the oligonucleotide using 1-ethyl-3-(3-dimethylami- nopropyl)carbodiimide (CDI) as the activating reagent.39 X=NH±CH2; NH± (CH2)n ±NHCH2, n=2, 4; R is a methylphosphonate analogue of the oligodeoxyribonucleotide; EDC is EtN=C=N(CH2)2CH2NMe2. The reaction was performed in a lutidine hydrochloride buffer (pH 7.5) for 14 h at 20 8C.The yield of compound 41 was 85%. The attachment of psoralen to the 50-end of an oligonucleotide can also be carried out in the solid-phase oligonucleotide syn- thesis.Thus a 4,50,8-triethylpsoralen residue 42 was incorporated into the oligonucleotide by phosphoramidite synthesis using psoralen b-cyanoethyl-N,N-diisopropylphosphoramidite 45 at the final stage.40 (a) 1. tetrazole; 2. I2; 3. NH3, 55 8C, 6 h; R=d(TAGCCGCTATCGGTTAGT). The first step of this synthesis consisted in chloromethylation of 4,50,8-trimethylpsoralen 42 with chloromethyl methyl ether.Then, a linker group was introduced into the resulting compound 43. Phosphitylation of 40-(2-hydroxyethoxymethyl)-4,50,8-trime- thylpsoralen 44 with b-cyanoethyl-N,N-diisopropylamidophos- phochloridite was carried out in the presence of ethyl- diisopropylamine for 30 min at 20 8C. The oligonucleotide psoralen derivative 46 was obtained on an automated DNA synthesiser using a standard phosphoramidite protocol.The psoralen phosphoramidite 45 was attached to the 50-end of the oligonucleotide at the last stage of this synthesis. Oxidation and deprotection with concentrated ammonia gave 50-psoralen deriv- ative 46 in 99% yield. A similar strategy has been used to obtain derivatives of oligonucleotides 47 and 48.41 Compound 48 with a disulfide bond, was synthesised using a mercaptoalkyl phosphoramidite that was introduced into an automated oligonucleotide synthesis prior to the attachment of the psoralen derivative.Thus, psoralen derivatives of nucleic acids can be obtained both by photochemical cycloaddition of psoralen to the hetero- cyclic bases of nucleic acids and by chemical attachment of psoralen derivatives to terminal phosphate groups of nucleic acids.The latter approach has become more popular because the ability of psoralen for subsequent photochemical addition is fully preserved in this case. Moreover, photochemical modification of heterocyclic bases of nucleic acids with psoralen makes the N-glycosidic bond more labile and can sometimes result in the destruction of the polynucleotide chain. b.Cross-linking of psoralen-containing oligonucleotides to biopolymers Psoralen-containing oligonucleotides are mostly used for directed modification of nucleic acids. A detailed study of their cross- linking to complementary sites of nucleic acids has been carried out by Shi.35 To this end, the dodecanucleotide GAAGCT*AC- GAGC derivatives containing a psoralen residue fused to thymi- dine both by its pyrone and furan rings (compounds 38a and 38b, O O O O (CH2)6 S P OR O O7 39 O O O XNH2 Me Me Me + HO P OR O O7 EDC 40 O O O XNH Me Me Me 41 O7 O OR P O O O O Me Me Me 42 ClCH2OMe O O O Me Me Me CH2O(CH2)2OH 44 (78%) Pri 2N P O(CH2)2CN Cl O O O Me Me Me CH2Cl 43 (63%) HO(CH2)2OH 100 8C, 10 min O O O Me Me Me CH2O(CH2)2O 45 (85%) ROH, a O O O Me Me Me CH2O(CH OR O O7 46 (99%) P NPri 2 O(CH2)2CN O O O Me Me Me CH2O(CH2)16OPOAGGAAGGGGG 47 O O O Me Me Me CH2O(CH2)8SS(CH2)8OPOAGGAAGGGGG 48 O O7 O O7 624 I E Kanevskii, S A Kuznetsovarespectively) were used.Modified oligonucleotides 38a and 38b were hybridised with a complementary template and irradiated with monochromatic light of various wavelengths.It was shown that irradiation of the nucleic acid ± oligonucleotide 38a complex with light in a broad range of wavelengths resulted in the predominant splitting of the 40-hydroxymethyl-4,50,8-trimethyl- psoralen residue. In this case, the efficiency of cross-linking between complementary oligonucleotides was extremely low. With irradiation of the complementary nucleic acid ± oligonucleo- tide 38b complex at l<313 nm, both splitting of the psoralen residue and simultaneous formation of a covalent oligonucleoti- de ± nucleic acid bond occur.Irradiation of the latter complex with l>313 nm results in highly efficient formation of a cross-linked complex, and psoralen is not split off. These results are in good agreement with other data on the stability of the psoralen monoadduct with thymidine.It was shown that thymidine fused with psoralen at the furan ring 38b was more resistant to irradiation at l<313 nm than the compound of the type 38a with the pyrone ring fusion. Introduction of a psoralen derivative of thymidine into an oligonucleotide only slightly increases the monoadduct resistance to photocleavage.42 Psoralen-containing oligonucleotides have been used in stud- ies of DNA transcription with the RNA polymerase of E.coli.36 40-Hydroxymethyl-4,50,8-trimethylpsoralen was introduced into one of the chains of a 140-membered DNA duplex containing a promoter of E. coli RNA polymerase. Upon irradiation at l=320 ± 380 nm, psoralen fused with a particular thymidine residue in one of the duplex chains reacted with the closely situated thymidine residue of the complementary chain.The yield of the cross-linked DNA duplex was 80%. It was shown that tran- scription of this duplex was blocked at a distance of one nucleotide from the site of cross-linking. Methylphosphonate analogues of oligonucleotides having psoralen derivatives at their 50-ends and linked to the phosphate group of the oligonucleotide through various spacer groups were used for attachment to two 35-membered deoxyribonucleotides.39 X=NHCH2, NH(CH2)nNHCH2, n=2, 4; Ps is the 4,50,8-trimethylpsoralen residue.The fragments of oligonucleotides containing internucleotide methylphosphonate groups are underlined. The cross-linked complex was obtained by irradiation of the duplexes 49 and 50 with light of l=365 nm at 4 8C for 20 min.The yield of the photoaddition products strongly depended on the length of the spacer arm and reached maximum values (75% and 80% for systems 49 and 50, respectively) at n = 4. Studies of the temperature dependence of the photoaddition reaction revealed that the efficiency of cross-linkage of DNA duplexes was deter- mined by the stability of the complementary complex formed.Irradiation of photoaddition products at l = 254 nm led to the cleavage of the cross-linked DNA duplexes and recovery of the original compounds. It was thus concluded that cross-linking of modified oligonucleotides is a result of formation of a cyclobutane ring between the C(5) ± C(6) bond of thymine and the C(3) ± C(4) bond of the pyrone ring of psoralen.The adduct with the furan ring of psoralen is not formed due to steric hindrances. Psoralen attached to the terminal phosphate group of an oligonucleotide through C(5) retains its capacity to form a bisadduct. Therefore, this type of psoralen-containing oligonuc- leotides within a triplex can form a ternary cross-linked oligonuc- leotide complex.The structure of the triplex is stabilised by intercalation of psoralen residues into a double helix of DNA. Thus psoralen-containing, 16-membered homopyrimidine oligo- nucleotide 39 was used for cross-linking to a double-helical region of DNA of the human immunodeficiency provirus 51 38 (the site where the formation of a triplex structure takes place is bold- faced). The sequence 51 contains a recognition site for the restriction endonuclease DraI (TTTAAA) overlapping with two or three nucleotides in the formation of a triplex with the oligonucleotide 39.Plasmids pLTR and pBT1 containing one or two copies of sequence 51, respectively, were incubated with the oligonucleotide 39 and irradiated with UV light (l>310 nm). The photomodified plasmids were used as substrates for the restriction enzyme DraI.It was found that photomodification with the psoralen derivative 39 selectively inhibited the enzymic splitting at the site overlapping with the triplex structure. The ability of psoralen-containing oligonucleotides to form cross-linked triplex complexes has been used for site-specific insertion of mutations into animal cells in order to investigate their repair mechanism.41 To this end, the frequency and spectrum of mutations induced by various psoralen-containing oligonu- cleotides in several cell lines with various repair disturbances, viz., xeroderma pigmentosum variant (XPV) and xeroderma pigmento- sum group A (XPA), were compared with the corresponding parameters of normal and Fanconi animia (FA) cells.In these studies, compounds 47 and 48 were used, which form triplexes with the 167 ± 176 region of the supF gene. It is noteworthy that the psoralen residue of compound 48 is linked with the oligonu- cleotide through a disulfide bond, which can be cleaved by dithiothreitol; this allows one to study mutagenesis in the absence of the site-directed oligonucleotide part of the probe.The triplex structure of compound 47 (or 48) with the supF gene in the vector pSP189 was irradiated with long-wave UV or visible light. UV irradiation gave a cross-linked triplex in 65% yield, while upon irradiation with visible light the yield did not exceed 1%. The cross-linkage was formed between T166 and T167 residues. The DNA vector thus modified was transformed into the appropriate cultures of human cells.Analysis of mutagenesis in the line XPV revealed an unusually high frequency of mutations (more than 30%) with a unique spectrum that differed from those of XPA, FA, and normal cells. These data led Raha et al.41 to conclude that this approach is useful both for the description of repair defects in XPV cells and other genetic analyses of human cells.The reactive oligonucleotide 46, modified at the 50-end with 40-hydroxyethoxymethyl-4,50,8-trimethylpsoralen, was used to obtain a cross-linked DNA duplex within the complementary complex 52. This DNA duplex was obtained in 75% yield upon UV irradiation of the complex 52 at l=350 nm at 20 8C.40 X is 40-hydroxyethoxymethyl-4,50,8-trimethylpsoralen. The glycosidic bond of the nucleotide involved in cross-linking of psoralen-containing oligonucleotides to nucleic acids becomes more labile. This property has been exploited in site-specific cleavage of DNA.35 The cleavage site was mostly determined by the length of the spacer arm between the psoralen residue and the oligonucleotide end.Thus, owing to their ability for selective cycloaddition to thymidine or uridine residues, psoralen-containing nucleic acids are widely employed for the study of structure and function of nucleic acids.The possibility of formation of bisadducts makes them helpful tools for the study of mechanisms of formation of triplex structures. 30 d(...GTAT TTATT TT TT TTAATCA...) Ps7X7pd (AATAAAAAAAAT) 49 30 d(...GATTAAAAAAAATAAATACGGT...) Ps7X7pd (ATTT TTT T TAT) 50 50 CCACTTTTTAAAAGAAAAGGGGGGACTGG AATTTTCTTT TCCCCCCT 51 X7pTAGCCGCTATCGGTTAGT GCCATCGGCGATAGCCAATCA 52 Synthesis of reactive nucleic acid analogues and their application for the study of structure and functions of biopolymers 625III.Platinum-containing nucleic acid analogues Such compounds as cis- (cisplatin, cis-53) and trans-diamminedi- chloroplatinum(II) (trans-53) and its analogues are widely used for the synthesis of Pt-containing oligonucleotides.As is known, cisplatin-derived reagents, unlike trans-isomers, have wide therapeutic applications as efficient antitumour drugs.43 A great number of investigations have been recently devoted to the study of peculiarities of interaction of cis- and trans-isomers of diamminedichloroplatinum(II) with two-helical sequences of DNA.When bifunctional platinum derivatives, such as cis- and trans-53 interact with DNA duplexes, they can form covalent bonds both between the neighbouring bases of the same chain and between spatially close bases in the complementary chains of a duplex. When cis-53 is used, cross-linking occurs between two neighbouring guanine residues or an adenine and guanine of the same chain and, to a lesser degree, between two guanine residues of the same chain separated by one nucleotide or between closely arranged guanine residues of the complementary chains.Trans-53 induces cross-linking mostly between the closely arranged guanine and cytidine residues of the complementary chain and, to a lesser degree, between two guanine residues of the same chain separated by one nucleotide.The difference in the mode of interaction of cis- and trans-stereoisomers of diamminedichloroplatinum(II) with DNA results in the change in the conformation of the modified DNA duplexes. This, in turn, affects the binding of these duplexes with DNA-recognising proteins.44 Recently, it has been demon- strated that some proteins of nucleic acid metabolism can effi- ciently recognise DNA with cross-links induced by cis-53.44, 45 On the other hand, the DNA cross-linked by both stereoisomers is recognised by these proteins far less effectively.44, 45 These results suggest that the differences in the conformations of native cellular DNA resulting from the interactions of cis- and trans-isomers of diamminedichloroplatinum(II) play a crucial role in the therapeutic efficiency of these compounds as antitumour drugs. 1. Methods for introduction of platinum reagents into oligonucleotides Derivatives of diaminodichloroplatinum(II) are bifunctional reagents, which can interact with nucleophiles. Both nucleophilic centres of residues A, G, and C and the thio groups incorporated into oligonucleotides can be used for modification of oligonucleo- tides by these reagents.Prior to the introduction of an aliphatic thio group into a 16-membered oligonucleotide, the latter was treated with cystamine in the presence of N-methylimidazole and water-soluble carbodiimide and the resulting compound was treated with dithiothreitol.46 Reaction of compound 54 with K2PtCl4 was carried out in an aqueous medium at pH 7 ± 8 and gave a reactive nucleotide derivative with a spacer 2-thioethyl- amino group at the 50-end.The synthesis was performed according to the following scheme: The binuclear Pt complex 55 was used for modification of oligonucleotides 56 ± 59 carrying a 50-terminal or an internucleo- tide thiophosphate group.47 The reaction between modified oligonucleotides 56 ± 59 and compound 55 was carried out in an aqueous medium in the presence of KBH4 for 90 or 60 min at 37 8C.In the former case, the Pt-containing oligonucleotides were used for cross-linking to the complementary fragments of human immunodeficiency virus DNA without preliminary isolation from the reaction mixture. In the latter case, the products were isolated by dialysis.An analogous binuclear complex 60 containing two Pt atoms with different ligands and differing strongly in their reactivity has been proposed for direct modification of heterocyclic bases of oligonucleotides.48 Compound 60 binds with the greatest efficiency with the N(7) atom of the guanine base through its Pt2+ moiety; this reaction gives adducts of the type 61 in virtually quantitative yields (H2O, pH 5.0, 4 h, 20 8C). The nucleophilic centres of adenine and cytosine residues can also undergo modification. In some cases, where it was necessary to obtain oligonucleo- tide monoadducts with trans-diamminedichloroplatinum, the latter were treated with AgNO3 in the presence of NaClO4 (prior to the synthesis or directly in the reaction mixture containing oligonucleotides) (see, for example, Refs 49, 50).This reaction is conducted in an acetate buffer (pH 3.6) at 37 8C for 15 ± 30 min. As a result, diamminedichloroplatinum is converted into aqua- diamminechloroplatinum [Pt(NH3)2Cl(H2O)]+. The latter selec- tively modifies the guanine residues of an oligonucleotide sequence. Platinum-containing oligonucleotides are isolated by H3N Pt Cl NH3 Cl Cl Pt Cl NH3 NH3 cis-53 trans-53 H2NCH2CH2SSCH2CH2HN P OR O7 O RO P OH O7 O a b (a) [H2N(CH2)2S]2, , EDC; (b) [HSCH2CH(OH)]2; (c) K2PtCl4, pH 778, 20 8C, 2 h.R is oligonucleotide residue; NMe N HSCH2CH2HN P OR O7 O 54 c SCH2CH2HN P OR O7 O Cl3Pt (50%) n=4, 5, 6. Cl Pt NH2(CH2)nNH2 Pt Cl 2Cl7 NH3 NH3 NH3 NH3 55 50 SpCCTTTTCCTTCCCTTTT 56 50 CACAATTCSCACACAAAC 57 50 SpCACAATTCCCACAAC 58 50 SpGCTCACAATTCCACAC 59 Sp= P O ; HO O7 S S= P O .O O7 S 2+ Pt N H2N H2N H2O (CH2)6 N Br (CH2)2 (CH2)2 NH2 (H2C)2 NH2 2+ + 60 Pt (H2C)2 2+ + HN N O N Pt N H2N H2N N (CH2)6 N Pt Br (H2C)2 (CH2)2 (CH2)2 NH2 (H2C)2 NH2 R H2N 61 626 I E Kanevskii, S A Kuznetsovaion-exchange chromatography. However, when cis-aquahydr- oxydiammineplatinum (cis[(NH3)2Pt(OH)(H2O)]NO3 is used, modification affects all of the G, C, and A residues of the oligonucleotide.In this case, the covalent bond can also be formed between the nucleotides of the same chain. Reagents with one platinum residue can be prepared from oligothymidylates con- taining only one G, C, or A residue. Thus treatment of oligothy- midylate having a single cytidine residue with a 10-fold excess of cis-aquahydroxydiammineplatinum nitrate 51 gave compound 62.The degree of conversion of the original oligonucleotide was 97%± 99%: R=d(TT), R0=TTTTTTT. The reaction was conducted in an aqueous medium (pH 7.0) for 24 h at 30 8C; the degree of conversion of the original compound and the homogeneity of the final product were estimated by microcolumn chromatography. Compound 62 was used for modification of DNA polymerase a of human placenta.Thus, modern methods allow one to incorporate platinum both into heterocyclic bases and into a predetermined position of the sugar-phosphate backbone of nucleic acids. In the latter case, the synthesis of platinum-containing derivatives must be preceded by introduction of mercapto groups into the sugar-phosphate backbone. 2. Cross-linking of Pt-containing oligonucleotides with biopolymers Bifunctional Pt-containing compounds are often used to obtain cross-linked nucleic acid ± nucleic acid or nucleic acid ± protein complexes.52, 53 The synthesis of platinum-containing oligonu- cleotides makes it possible to increase considerably the efficiency and selectivity of modification of biopolymers by these reagents. Cross-linking of Pt-containing oligonucleotides with biopol- ymers can be carried out in two ways: (1) by direct reaction of Pt- containing oligonucleotides with complementary sites of nucleic acids or DNA-recognising proteins and (2) by treatment of a preformed complex of a biopolymer with an oligonucleotide probe containing nucleophilic SH or NH2 groups at predeter- mined positions with bifunctional Pt reagents.Thus treatment of a complementary complex 63 formed by a 322-membered DNA target and a 16-membered oligonucleotide probe carrying a cysteamine residue at its 50-end with K2PtCl4 resulted in cross- linking. The efficiency of cross-linking varied from 10% to 50% depending on K2PtCl4 concentration.46 X=HSCH2CH2NH The reaction involved the attachment of Pt coordinated with three chloride ions to the 50-SH group of the oligonucleotide probe and subsequent modification of the DNA target.A similar approach was used by Chu and Orgel 54 who carried out affinity modification of the transcription factors CREB and JUN using platinum-containing oligonucleotides. To this end, six internucleotide thiophosphate groups have been introduced into each CRE and TRE recognition site for the transcription factors CREB and JUN in covalently closed DNA duplexes.The com- plexes of oligonucleotide probes with the corresponding proteins were treated either withK2PtCl4 or with trans- or cis-diamminedi- chloroplatinum(II). The efficiency of cross-linking between oligo- nucleotides and proteins with K2PtCl4 or trans-53 was from 30% to 50% depending on the reagent concentration.Treatment of a DNA± protein complex with cis-53 did not result in cross-linking. The use of analogous probes devoid of internucleotide thiophos- phate groups resulted in several-fold lower efficiency of cross- linking. Presumably, this method 54 can be employed for the isolation of DNA-recognising proteins.A Pt-containing nonanucleotide 61 (R=TCCGCCTTT) was used for modification of a complementary hexanucleotide GGCGGA within a two-helical complex 64.48 The yield of the cross-linked complex was 80%. Unlike psoralen derivatives, treatment of guanine residues of template oligonucleotides with platinum-containing reagents favours stabilisation of the glycosidic bond. An analysis of products of partial scission of the template at purine bases allows one to identify particular guanine residues that have been modi- fied.It was found that one of the two 5-terminal guanine residues of the template was involved in cross-linking.48 A detailed study of cross-linking of oligonucleotides modified with the reagent 55 at the 50-terminal or at internucleotide thiophosphate groups to complementary fragments of DNA was carried out by Gruff and Orgel.47 The reaction was studied both within DNA duplexes and within ternary complexes using oligo- nucleotides 56 ± 59 and 65 ± 68.Using duplexes 65 ± 56 as an example, it was shown that cross- linking was the most efficient (64% yield) when the reagent 55 (n=6) was used. This reagent was chosen for subsequent experi- ments.The reaction was carried out for 60 min at 37 8C in the presence of NaClO4. Modification affected residues G19 and G20 of the oligonucleotide 65 (marked bold). Under the same con- ditions, the efficiency of cross-linking in duplexes 58 ± 68 and 59 ± 68 was much lower (46% and 41%, respectively), which was attributed to the formation of intrachain cross-links in oligonu- cleotides 58 and 59 containing purine bases.47 In the case of the oligonucleotide 57, the nucleotide composition of which is iden- tical with that of the oligonucleotide 58 but which contains a platinum reagent inside the chain (because the thiophosphate group is present in the middle of the sugar-phosphate backbone), the efficiency of cross-linking decreased to 37%.This low effi- ciency may be due to the involvement of the modified base in complementary interaction. A ternary complex was formed by oligonucleotides 66 and 67 and an analogue of a Pt-containing oligonucleotide 56 in which all cytidine residues had been sub- stituted by 5-methylcytidine (oligonucleotide 69).Such a substi- tution significantly stabilised the triplex structure. The reaction carried out at 25 8C for 60 min resulted in cross-linking of the oligonucleotide 69 with the chain 66 enriched with purine nucleo- tides and with 53%± 68% efficiency. It was shown that the efficiency of this reaction strongly depended on the stability of the triplex structure.Thus a rise in temperature up to 37 8C sharply decreased the yield of cross-linked products. Replacement of the oligonucleotide probe 69 by 56 also decreased the yield down to several per cent.OR0 O N RO N+ O NH2 Pt Cl NH3 NH3 62 (50) Xpd(CACAATTCCACACAAC) (30) d(....ACGAGTGTTAAGGTGTGTTGTA....) 63 (50) TCCGCCTTT (30) AGGCGG XBr 64 TTTAAAAGGGAAGGAAAAGGTAAAGAC 19 20 65 CAGAAATGGAAAAGGAAGGGAAAATTT 66 AAATTTTCCCTTCCTTTTCCATTTCTG 67 TCGTATGTTGTGGGAATTGTGAGCGGATAACAATTT 68 Synthesis of reactive nucleic acid analogues and their application for the study of structure and functions of biopolymers 627As can be seen from the above data, modification of DNA with derivatives of cis-diamminedichloroplatinum (cis-53) is of particular interest. Yet, attempts were made to obtain protein ± - nucleic acid complexes in whichDNAmolecules were modified by trans-diamminedichloroplatinum (trans-53).The main problem remains to establish the reason for poor recognition of DNA modified with trans-53 by DNA-binding proteins. As can be evidenced from a detailed study carried out by Brabec and Leng,49 modification with trans-53 decreases the thermal stability of modifiedDNAduplexes and significantly distorts the geometry of the double helix. Apparently, these conformational effects account for low yields of proteins cross-linked with DNA modi- fied by trans-53.Thus, affinity modification of biopolymers by platinum-con- taining reagents can be achieved both by direct treatment of biopolymers with Pt-containing nucleic acids and by treatment of preformed nucleic acid ± nucleic acid or nucleic acid ± protein complexes with Pt-containing reagents.In the latter case, nucleic acid derivatives containing SH or NH2 groups are mainly used. The attention of investigators is mainly focussed on the elucida- tion of mechanisms of interaction of Pt-containing nucleic acid derivatives with biopolymers and of the effect of this interaction on the functional properties of the latter. We believe that further developments in this field will enlarge the scope of applications of Pt-containing nucleic acids and will open up new possibilities for their therapeutic applications in various systems in vivo.IV. Reactive nucleic acids containing substituted pyrophosphate and acyl phosphate groups A new type of oligonucleotide derivatives containing reactive groups in the sugar-phosphate backbone was developed and used for affinity modification of proteins at the Laboratory of Nucleic Acid Chemistry of the Moscow State University in the late 80's.55 ± 57 This type of compounds contains substituted internu- cleotide pyrophosphate (70) or acyl phosphate (71) groups.R,R0 are the oligonucleotide residues; X is the residue of an aliphatic alcohol or an amine. The distinctive features of this class of compounds are high stability in aqueous solutions in the absence of strong nucleo- philes, the ability to react with nucleophilic groups of a protein at a zero distance under conditions close to physiological ones, and the fact that they can be used without additional activation.Thus, these compounds are promising tools for the study of structure and functions of nucleic acid-recognising proteins and for their affinity modification.In addition, they can find application in sense biotechnology and medicine as therapeutic preparations, particularly in the treatment of genetic and oncological diseases and AIDS. 1. Directed introduction of substituted pyrophosphate and acyl phosphate groups into oligonucleotides Synthesis of nucleic acid analogues containing substituted reactive pyrophosphate groups at predetermined positions of the sugar- phosphate backbone is based on chemical ligation, which consists in condensation of two oligonucleotides the ends of which are converged on a complementary template.One of these oligonu- cleotides bears an alkoxy or an alkylamino group in the 30(50)- terminal phosphate, while the 50(30)-end of the other oligonucleo- tide is phosphorylated.55, 58 The reaction proceeds according to the following scheme: Oligonucleotide alkyl esters used in this reaction are obtained by postsynthetic modification of 30-phosphorylated oligonucleo- tide with alcohols mediated by coupling reagents 59 or in the course of automated oligonucleotide synthesis.60 Oligonucleotide amides are usually obtained by postsynthetic modification of 30-phosphorylated oligonucleotides by amines in the presence of coupling reagents.59 Chemical ligation was used to obtain oligodeoxyribonucleo- tides containing O-methyl- or N-alkyl-substituted pyrophosphate groups.55 The experiments were carried out with system 72 having a recognition site for the restriction-modification enzymes EcoRII, SsoII, and MvaI (boldfaced).Activation of the phosphate group for the synthesis of com- pounds with a substituted pyrophosphate group was carried out with CDI or N-hydroxybenzotriazole esters. The reaction pro- ceeds with the greatest efficiency when CDI is used as the coupling reagent. The yields of chemical ligation products are 35% ±80% depending on the nature of the non-nucleotide substituent. With triazole esters, the yields decreased down to 10%± 15%.Reactive DNA duplexes 73 containing an O-methyl-substi- tuted pyrophosphate group in the recognition site of the restric- tion-modification enzymes EcoRI and RsrI 56 (boldfaced) were obtained in a similar way. In this case, a methoxy group was used as a non-nucleotide substituent, because it induces minimum distortions in the geom- etry of the double helix.This is important for the subsequent investigation of the interaction of modified DNA duplexes with proteins. The modified oligonucleotide containing an O-methyl- substituted pyrophosphate group was obtained in 25% ±30% yield. A series of DNA duplexes 74 containing an O-methyl-sub- stituted pyrophosphate group in various positions of the HNF-1 recognition site have been synthesised to probe DNA contacts with the transcription factor HNF-1.61 (The recognition site HNF-1 is marked with a straight line; the positions of modified groups are indicated by arrows).RO P O P OR0 O7 X O O RO P O7 O O P OR0 O7 NH CH2 70 71 O O C O P O7 X O O P Y O7 O 50 30 7Y7 O P O X O 50 30 X is the alkoxy or alkylamino group; Y is the residue of the coupling reagent. O O7 P O 50 ACCTACCp pTGGTGGT 30 TGGATGG 72 X=OEt, NH(CH2)3Me, NHCH2COOEt, X ACCACCA O .N 50 CATGCAAGp pAATTCAAGAC 30 GTACGTTC TTAAGTTCTG 73 MeO 50 GTT AGT GTG GTT AAT GAT CTA CAG TTA TTG 30 30 CAA TCA CAC CAA TTA CTA GAT GTC AAT AAC 50 74 13 12 11 10 9 8 7 6 5 4 3 2 1 14 15 16 628 I E Kanevskii, S A KuznetsovaThe reactive group was introduced into the upper and the lower chains of the duplexes by chemical ligation.Each modified duplex had only one substituted pyrophosphate group. The average yield of chemical ligation products was 15%. Chemical ligation was also used to introduce a substituted pyrophosphate group into DNA duplexes of different lengths and compositions.58, 62 Introduction of substituted pyrophosphate groups into RNA was first described by Naryshkin et al.who studied the interaction of TAR RNA with the Tat-protein of HIV-1.63 These authors synthesised a series of RNA duplexes 75 containing a reactive group in various positions of the recognition site of the boldfaced Tat-protein (the position of the reactive group is marked by arrows). Since the efficiency of chemical ligation of oligoribonucleo- tides is extremely low (1% ± 2%), oligoribonucleotides containing 20-deoxyuridine O-alkylphosphate at their 30-ends were used to introduce a substituted pyrophosphate group intoRNAduplexes. The ligation was performed on an oligodeoxyribonucleotide template.Then the modified chain was isolated, and an equimolar amount of a complementary oligoribonucleotide was added to obtain RNA duplexes.Chemical ligation products were obtained in 35% ±55% yields. Analogously, an internucleotide acyl phosphate group was introduced into the duplex structure by chemical ligation. In this case a modified oligonucleotide with a glycine residue at the 30-end was used.57 The modified oligonucleotide containing a glycine residue was obtained by direct condensation of the heptanucleotide ACCTACCp with glycine ethyl ester in the presence of CDI followed by saponification of the ester bond.Condensation of oligonucleotides within the DNA duplex was also carried out in the presence of CDI. The yield of the chemical ligation product was 24%. This method was also used to obtain the DNA duplex 76 containing a reactive acyl phosphate group in the recognition site of NF-kB (the kB site is boldfaced).64 The starting oligonucleotide modified at the 30-terminal phosphate group with a glycine residue was obtained using N-hydroxybenzotriazole esters.59 2.Affinity modification of nucleic acid-specific proteins by nucleic acid analogues containing substituted internucleotide pyrophosphate or acyl phosphate groups Investigation of the properties of reactive nucleic acids containing substituted pyrophosphate or acyl phosphate groups revealed that the modified group of these compounds is stable in aqueous buffers in the absence of strong nucleophiles at pH 6.00 ± 8.75.At the same time, it is easily and quantitatively cleaved by primary and secondary amines, including nucleophilic amino acids of a protein. The cleavage of the modified group is accompanied by a transfer of the nucleic acid fragment onto the nucleophile.55, 57 X is an alkoxy or alkylamino group; R1,R2 are the oligonucleotide residues.By virtue of their ability to form covalent bonds with nucleophilic amino acid residues of nucleic acid-recognising proteins, these compounds are successfully used for the study of enzymes of nucleic acid metabolism as well as for affinity modification and probing of binding sites of transcription factors.Thus the reactive DNA duplex 73 was used for affinity modifica- tion of the restriction-modification enzymes R.EcoRI, R.RsrI, M.EcoRI, and M.RsrI.56 The protein complex with a modified substrate was incubated for 20 ± 24 h at 25 ± 37 8C.The maximum yields of cross-linked complexes (40% and 30%, respectively) were observed for methyl transferases M.EcoRI and M.RsrI. It was found that it is the phosphoramide covalent bond that is formed between the oligonucleotide residue and the protein. Affinity modification of the restriction endonuclease EcoRI (R.EcoRI) by the reactive DNA duplex 72 has been carried out by Shabarova et al.65 The duplex 72 containing an O-ethyl-substi- tuted pyrophosphate group in the recognition site of R.EcoRI was incubated with the enzyme for 18 h at 37 8C.The reaction was carried out in an aqueous buffer containing N-methylimidazole because this reagent catalyses the reactions of nucleophilic sub- stitution in aqueous media.66 The efficiency of cross-linking of the modified duplex 72 with R.EcoRI in the presence of N-methyl- imidazole was 15%.At the same time, the yield of the cross- linking product in a buffer devoid of N-methylimidazole did not exceed 1%. This confirms the catalytic role of N-methylimidazole in nucleophilic substitution reactions of oligonucleotides in aque- ous media.56 The active centre of the transcription factor HNF-1 was probed 61 using a series of DNA duplexes containing an O-methyl-substituted pyrophosphate group in various positions (1 ± 16) of the HNF-1 recognition site.These studies were carried out both with a gene-engineered protein, which represented a DNA-binding domain of HNF-1, and with proteins isolated from nuclear extracts of rat liver. Modified substrates were incubated with the protein or a nuclear extract at 4 8Cor 20 8C for 10 min or 12 h.This reaction yielded cross-linked complexes both for the DNA-binding domain of HNF-1 and the HNF-1 of the nuclear extract. Cross-linking occurred only if a DNA duplex contained a reactive group in position 10. The efficiency of cross-linking was 50 AGCCAGA GAGCAGC 30 UCGGUCU CUCGUCG 75 U U U 40-41 39-40 38-39 42-43 EDC 30 TGGATGG ACCACCA 50 ACCTACC 30 TGGATGG ACCACCA 50 ACCTACC O P O7 O NHCH2 C O7 O HO P O O O7 TGGTGGT O P NHCH2C O7 O O TGGTGGT O7 O O P O GAGCCTTTCA GGGGAGA 76 ACCTCGGAAAGT O P NHCH2C X O O CCCCTCT O7 O O P O R1O P O P OR2 O X O O7 NuH 7H+ R1O P O7 O X + Nu P OR2 O O7 R1O P NHCH2C O O7 O O7 O OR2 P O NuH R1O P NHCH2C O O7 Nu O + HO O7 O OR2 P Synthesis of reactive nucleic acid analogues and their application for the study of structure and functions of biopolymers 62940% of the non-covalent binding.It has been shown for the first time that cross-linking of a modified substrate with HNF-1 isolated from a nuclear extract occurred in a specific manner and was not accompanied by modification of other components of the nuclear extract. This finding holds much promise for the applica- tion of oligonucleotides with substituted pyrophosphate groups in various in vivo systems and for creation on their basis of new generations of therapeutic agents.Compounds with a substituted pyrophosphate group were also used for establishing the contact site for the transcription factor NF-kB and DNA.58 RNA duplexes containing substituted pyrophosphate groups can also be used as reactive derivatives for affinity modification of biopolymers.The interaction of RNA duplexes TAR 75 with the Tat-protein and the Tat-peptide (37 ± 72) of HIV-1 has been studied.63 The RNA± protein complex was incubated for 18 ± 24 h at 10 8C. Cross-linking with the Tat-peptide (37 ± 72) was especially efficient when an RNA duplex containing a modified group in position 38 ± 39 was used.The product yield was 50%. Analysis of the complex revealed that Lys51 of the main region of the Tat-peptide is involved in the formation of a covalent bond. A first report concerning the successful application of a DNA duplex containing an acyl phosphate group for affinity modifica- tion of the p50 subunit of NF-kB has been published recently.64 TheDNAduplex 76 containing a reactive acyl phosphate group in position 6 of the kB region was used for cross-linking with the protein.The efficiency of cross-linking was 10%. Acyl phosphates are acylating reagents.57 The group formed as a result of their interaction with nucleophilic amino acids of a protein is carbox- amide, which unlike the phosphoramidate is relatively stable in acidic media.The application of nucleic acid analogues with an acyl phosphate group for affinity modification of DNA-recognis- ing proteins may significantly facilitate both the isolation of protein ± nucleic acid complexes and subsequent identification of amino acids that react with the reactive group. Thus, reactive nucleic acids containing substituted pyrophos- phate or acyl phosphate groups are successfully employed for the study of nucleic acid ± protein interactions.It should be noted, however, that introduction of these groups into the sugar-phos- phate backbone of nucleic acids increases the length of the internucleotide bond, which, in turn, may decrease the efficiency of its binding with the protein.61 This problem can be overcome through the use of nucleic acid duplexes containing reactive groups at other positions of the protein binding site.61 V. Prospects for the application of reactive oligonucleotides for directed in vivo modification of biopolymers One of the most interesting applications of reactive oligonucleo- tides is the design on their basis of antiviral and antitumour drugs the effects of which are due to inhibition of gene expression in viral or oncogenic cells.The ability of reactive oligonucleotides to covalently bind to proteins of nucleic acid metabolism may suppress their activity and thus control the expression of the genetic material. It was found 67 that human tRNA3 Lys containing 4-mercaptouridine residues efficiently interacts with HIV reverse transcriptase to form a cross-linked complex.This interaction results in the block of a region in reverse transcriptase that is responsible for the initiation of reverse transcription of HIV. Of particular interest is the application of reactive oligonu- cleotide derivatives in various in vivo systems. The first studies in this area revealed a number of limitations related to weak penetrability of these reagents through the cellular membrane and their enhanced sensitivity with respect to cellular exonu- cleases.Besides, nonspecific modification of functional compo- nents of various cell structures by these reagents also presents some problem.68 The permeability problem can be circumvented through the use of hydrophobic oligonucleotide derivatives (see Ref. 69). Several approaches are now used to stabilise DNA fragments in the cell, such as construction of covalently closed dumbbell-like DNA duplexes,70 cyclic DNAs,71 and DNA duplexes with covalently linked chains.72 A serious limitation for the in vivo application of a vast number of reactive oligonucleotides is that an additional treat- ment (UV irradiation, heating, chemical reagents) are required for their activation.Therefore, the most promising reagents for in vivo application are apparently platinum derivatives and reactive nucleic acids containing substituted pyrophosphate or acyl phos- phate groups. These compounds do not require additional activa- tion; the mechanisms of their interaction with nucleic acid- recognising proteins have been studied in sufficiently great detail.Besides, they have already been used in nuclear cell extracts.54, 61 It should be noted, however, that modification of proteins with platinum derivatives of oligonucleotides in nuclear extracts is not always specific.54 Besides, this procedure is inefficient unless mercapto derivatives of nucleic acids are used. At the same time, it has been shown 61 that modification of the transcription factor HNF-1 in nuclear extracts of rat liver is highly specific and efficient.This circumstance gives every reason to believe that such compounds are promising for in vivo applications. In order to increase their resistance to cell exonucleases, it may be necessary to change their structure, e.g., to design on their basis covalently closed DNA duplexes, or to synthesise DNA duplexes with covalently linked chains.The methods of their synthesis have already been developed 61, 73 ± 75 and the first encouraging results of the application of structures with substituted pyrophopsphate groups for affinity modification of the proteins of nucleic acid metabolism have been obtained.76 This study was accomplished within the framework of the proghramme `Leading Scientific Schools' (Grant 96 ± 15 ± 97707).The authors are grateful to the Doctor of Chemistry V K Potapov (Shemyakin and Ovchinnikov Institute of Bioor- ganic Chemistry) for assistance in writing the review and for valuable discussion. References 1. D G Knorre, V F Zarytova, A G Badashkeeva, O S Fedorova, in Biotekhnologiya (Itogi Nauki i Tekhniki) [Biotechnology (Advances in Science and Engineering Series) (Moscow: Izd.VINITI, 1991) Vol. 37, p. 1 2. D G Knorre, V V Vlasov, V F Zarytova, A V Lebedev, O S Fedorova, in Design and Targeted Reactions of Oligonucleotide Derivatives (Boca Raton, FL: CRC Press, 1994) p. 366 3. L A Sylvers, J Wower Bioconjug. Chem. 4 411 (1993) 4. D R Tabatadze,M I Dobrikov, A S Levina, T A Prikhodko, G V Shishkin, V F Zarytova Nucl.Acids Symp. Ser. 24 247 (1991) 5. M I Dobrikov, V F Zarytova, N I Komarova, A S Levina, S G Lokhov, T A Prikhod'ko, G V Shishkin, D R Tabatadze, M M Zaalishvili Bioorg. Khim. 18 540 (1992) a 6. T Le Doan, L Perrouault, D Praseuth, N Habhoub, J-L Decout, N T Thuong, J Lhomme, C Helene Nucl. Acids Res. 15 7749 (1987) 7. P Mitchell, K Stade,M Osswald, R Brimacombe Nucl.Acids Res. 21 887 (1993) 8. M G Ivanovskaya, N I Sokolova, Z A Shabarova Bioorg. Khim. 5 35 (1979) a 9. V F Zarytova, A S Levina, L M Khalimskaya, Z A Sergeeva Bioorg. Khim. 18 640 (1992) a 10. B Bartholomew, G A Kassavetis, B R Braun, E P Geiduschek Embo J. 9 2197 (1990) 11. B J Lannutti, J Persinger, B Bartholomew Biochemistry 35 9821 (1996) 12. O I Gimautdinova,M A Zenkova, G G Karpova, L M Podust Mol.Biol. 18 907 (1984) b 13. N D Kobets, A V Gorozhankin, T S Godovikova, V N Sil'nikov, D G Knorre Dokl. Akad. Nauk 349 822 (1996) c 14. T S Godovikova, M V Berezovskii, D G Knorre Bioorg. Khim. 21 858 (1995) a 630 I E Kanevskii, S A Kuznetsova15. D Praseuth, M Chassignol,M Takasugi, T Le Doan, N T Thuong, C Helene J. Mol. Biol. 196 939 (1987) 16. A S Boutorin, T Le Doan, J P Battioni, D Mansuy, D Dupre, C Helene Bioconjug. Chem. 1 350 (1990) 17. A N Mayer, F Barany Gene 153 1 (1995) 18. M M Hanna, Y Zhang, J C Reidling, M J Thomas, J Jou Nucl. Acids Res. 21 2073 (1993) 19. A Reizer, H M Wagner, in The Chemistry of the Azido Group (Ed. S Patai) (New York: Interscience, 1971) p. 441 20. T-Y Liang, G B Schuster J.Am. Chem. Soc. 109 7803 (1987) 21. A S Levina, D R Tabatadze, V F Zarytova, M I Dobrikov, V N Gorn, L M Khalimskaya Bioorg. Khim. 20 21 (1994) a 22. Y L Dubreuil,A Expert-Bezancon,A Favre Nucl. Acids Res. 19 3653 (1991) 23. J Liu, M Sodeoka,W S Lane, G L Verdine Proc. Natl. Acad. Sci. USA 91 908 (1994) 24. R D Sheardy, N C Seeman J. Org. Chem. 51 4301 (1986) 25. M C Willis, K A LeCuyer, K M Meisenheimer, O C Uhlenbeck, T H Koch Nucl.Acids Res. 22 4947 (1994) 26. K M Meisenheimer, P L Meisenheimer, M C Willis, T H Koch Nucl. Acids Res. 24 981 (1996) 27. A McGregor,M V Rao, G Duckworth, P G Stockley, B A Connolly Nucl. Acids Res. 24 3173 (1996) 28. T M Dietz, R J von Trebra, B J Swanson, T H Koch J. Am. Chem. Soc. 109 1793 (1987) 29. T M Dietz, T H Koch Photochem. Photobiol. 46 971 (1987) 30. S Cereghini,M Blumenfeld, M Yaniv Genes Develop. 2 957 (1988) 31. M C Willis, B J Hicke, O C Uhlenbeck, T R Cech, T H Koch Sci- ence 262 1255 (1993) 32. O Dontsova, A Kopylov, R Brimacombe Embo J. 10 2613 (1991) 33. D M Graifer, D I Juzumiene, P Wollenzien, G G Karpova Biochemistry 33 3878 (1994) 34. O Dontsova, V Tshkov, S Dokudovskaya, A Bogdanov, T Doring, J Rinke-Appel, S Thamm, B Greuer, R Brimacombe Proc. Natl. Acad. Sci. USA 91 4125 (1994) 35. Y B Shi, J E Hearst Biochemistry 26 3792 (1987) 36. Y B Shi, V Gamper, J E Hearst Nucl. Acids Res. 15 6843 (1987) 37. D Kanne, K Straub, J E Hearst, H Rapoport J. Am. Chem. Soc. 104 6754 (1982) 38. C Giovannangeli, N T Thuong, C Helene Nucl. Acids Res. 20 4275 (1992) 39. B L Lee,A Murakami,K R Blake, S B Lin, P S Miller Biochemistry 27 3197 (1988) 40. U Pieles, U Englisch Nucl. Acids Res. 17 285 (1989) 41. M Raha, G Wang,M M Seidman, P M Glazer Proc. Natl. Acad. Sci. USA 93 2941 (1996) 42. Y B Shi, J E Hearst Biochemistry 26 3786 (1987) 43. A L Pinto, S J Lippard Biochim. Biophys. Acta. 780 167 (1985) 44. J Kasparkova, V Brabec Biochemistry 34 12379 (1995) 45. B A Donahue, M Augot, S F Bellon, D K Treiber, J H Toney, S J Lippard, J M Essigmann Biochemistry 29 5872 (1990) 46. B C Chu, L E Orgel Nucl. Acids Res. 17 4783 (1989) 47. E S Gruff, L E Orgel Nucl. Acids Res. 19 6849 (1991) 48. V V Vlassov, V V Gorn, E M Ivanova, S A Kazakov, S V Mamaev FEBS Lett. 162 286 (1983) 49. V Brabec, M Leng Proc. Natl. Acad. Sci. USA 90 5345 (1993) 50. M Boudvillain, R Dalbies, C Aussourd,M Leng Nucl. Acids Res. 23 2381 (1995) 51. G A Nevinskii, V N Podust, A S Levina, O V Khalabuda, O I Lavrik Bioorg. Khim. 12 357 (1986) a 52. F Oshita, A Eastman Oncol. Res. 5 111 (1993) 53. V Brabec, V Kleinwachter, J L Butour, N P Johnson Biophys. Chem. 35 129 (1990) 54. B C Chu, L E Orgel Nucl. Acids Res. 20 2497 (1992) 55. S A Kuznetsova, M G Ivanovskaya, Z A Shabarova Bioorg. Khim. 16 219 (1990) a 56. A A Purmal, Z A Shabarova, R I Gumport Nucl. Acids Res. 20 3713 (1992) 57. S A Kuznetsova, M G Ivanovskaya, Z A Shabarova Bioorg. Khim. 17 1633 (1991) a 58. I A Kozlov, E A Kubareva, M G Ivanovskaya, Z A Shabarova Mol. Biol. 31 63 (1997) b 59. M G Ivanovskaya,M B Gottikh, Z A Shabarova Nucleosides Nucleotides 6 913 (1987) 60. N A Naryshkin,M G Ivanovskaya, T S Oretskaya, E M Volkov, M Dzh Geit, Z A Shabarova Bioorg. Khim. 22 691 (1996) a 61. S A Kuznetsova, C Catherine, U Edgardo, I Elias,M Vasseur, M Blumenfeld, Z A Shabarova Nucl. Acids Res. 24 4783 (1996) 62. E A Kubareva, O A Fedorova,M B Gottikh, H Tanaka, C Malvy, Z A Shabarova FEBS Lett. 381 35 (1996) 63. N A Naryshkin,M A Farrow,M G Ivanovskaya, T S Oretskaya, Z A Shabarova, M J Gait Biochemistry 36 3496 (1997) 64. M G Ivanovskaya, I A Kozlov, E A Kubareva, E A Taran, T V Sarazev, Z A Shabarova Mol. Biol. 31 435 (1997) b 65. Z A Shabarova, G Ya Sheflyan, S A Kuznetsova, E A Kubareva, O H Sysoev,M G Ivanovskaya, E S Gromova Bioorg. Khim. 20 413 (1994) a 66. Z A Shabarova, M G Ivanovskaya,M G Isaguliants FEBS Lett. 154 288 (1983) 67. Y Mishima, J A Steitz Embo J. 14 2679 (1995) 68. D E Wilman, T A Connors in Molecular Aspects of Anti-Cancer Drug Action (Weinheim: Verlag Chemie, 1983) 69. A S Boutorin, L V Gus'kova, E M Ivanova, N D Kobetz, V F Zarytova, A S Ryte, L V Yurchenko, V V Vlassov FEBS Lett. 254 129 (1989) 70. B C F Chu, L E Orgel Nucl. Acids Res. 20 5857 (1992) 71. N G Dolinnaya,M Blumenfeld, I N Merenkova, T S Oretskaya, N F Krynetskaya,M G Ivanovskaya,M Vasseur, Z A Shabarova Nucl. Acids Res. 21 5403 (1993) 72. S I Antsypovich, T S Oretskaya Usp. Khim. 67 274 (1998) [Russ. Chem. Rev. 67 247 (1998) 73. S A Kuznetsova, I E Kanevskii, T S Oretskaya, Z A Shabarova Bioorg. Khim. 22 532 (1996) a 74. I E Kanevskii, S A Kuznetsova, E M Volkov, Z A Shabarova Mol. Biol. 30 1144 (1996) b 75. S A Kuznetsova, I E Kanevskii, T S Oretskaya, Z A Shabarova Nucleosides Nucleotides 22 1231 (1996) 76. I E Kanevskii, Candidate Thesis in Chemical Sciences, Moscow State University, Moscow, 1996 a�Russ. J. Bioorg. Chem. (Engl. Transl.) b�Mol. Biol. (Engl. Transl.) c�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) Synthesis of reactive nucleic acid analogues and their application for the study of structure and functions of biopolyme

ISSN:0036-021X    

出版商:RSC    

年代:1998

数据来源: RSC