摘要:

Russian Chemical Reviews 69 (12) 1001 ± 1019 (2000) Tandem transformations initiated and determined by the Michael reaction E V Gorobets, MS Miftakhov, F A Valeev Contents I. Introduction II. A tandem Michael reaction III. Tandem `Michael reaction ± nucleophilic substitution' sequence IV. Tandem `Michael reaction ± aldol reaction' sequence V. Tandem Mannich and Michael reactions VI. Tandem Michael reaction and heterocyclisation VII. Tandem `anionotropic rearrangement ± Michael reaction' sequence VIII. Tandem Michael and Dieckmann reactions IX. Tandem Mukaiyama and Michael reactions X. Miscellaneous tandem transformations involving Michael reactions Abstract. the by initiated transformations tandem on Data Data on tandem transformations initiated by the Michael reaction published over the last five years are analysed, Michael reaction published over the last five years are analysed, systematised and generalised.The bibliography includes 86 refer- systematised and generalised. The bibliography includes 86 refer- ences. I. Introduction Syntheses of polyfunctional organic compounds with complex structures from simpler compounds by tandem one-pot trans- formations were described in a series of reviews (see, for example, Refs 1 ± 6). Tandem transformations are multistep transforma- tions in which each subsequent step is strictly determined by the preceding one and occurs under the same reaction conditions. The advantage of these transformations is the formation of several bonds and complication of the structure of the compound in one step, which is generally accompanied by the high stereoselectivity of the reaction.Minimisation of waste products, a decrease in the amount of required solvents, reagents and adsorbents, reduction in the energy and a decrease in the number of laboratory operations provide the economic and ecological feasibility of this type of processes. According to the mechanism of the first step, tandem con- versions are divided into cationic, anionic, radical, pericyclic, photochemical and transition-metal-induced processes. The terms `tandem', `cascade' and `domino' reactions are used as synonyms. Most tandem transformations proceed through the formation of an anionic intermediate. This first step governs the subsequent transformations.Most often, deprotonation of the CH group of E V Gorobets,MS Miftakhov, F A Valeev Institute of Organic Chemistry, Ufa Research Centre of the Russian Academy of Sciences, prosp. Oktyabrya 71, 450054 Ufa, Russian Federation. Fax (7-347) 235 60 66. Tel. (7-347) 235 57 11. E-mail: valeev@anrb.ru (E V Gorobets, F A Valeev). Tel. (7-347) 235 52 88. E-mail: bioreg@anrb.ru (M S Miftakhov) Received 17 July 2000 Uspekhi Khimii 69 (12) 1091 ± 1110 (2000); translated by T N Safonova #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n12ABEH000614 1001 1001 1006 1009 1011 1012 1014 1014 1015 1016 the substrate occurs giving rise to a carbanion followed by its reaction with an electrophile to form a new anionic intermediate.This anion can attack another electrophilic centre, etc. The sequence is completed with trapping of an electrophile (for example, H+) or elimination of the leaving group X7. In many cases, anionic transformations are either initiated or governed by the Michael reaction. In the present review, most attention is given to this type of transformation studied in recent years.{ II. A tandem Michael reaction One of the approaches which is most popular for the formation of C7C bonds is based on the Michael reaction involving com- pounds containing an activated methylene group and a,b-unsatu- rated esters or ketones. A combination of two consecutive Michael addition reactions has considerable promise for the efficient construction of cyclic and polycyclic compounds because the second addition reaction often results in ring closure.This type of transformation is named the Michael Ring Closure (MIRC) or the Michael-Michael Ring Closure (MIMIRC).2 Of all tandem transformations, a combination of two Michael reactions has been described in most detail. The examples considered below clearly demonstrate the importance of this approach in synthetic organic chemistry. In studies of intramolecular double Michael addition reac- tions, a series of polycyclic structures were prepared 7 which served as the basis for the total synthesis of some natural compounds. The total synthesis of a sesquiterpenoid fungicide ()-culmorin (1) was developed.8 The tricyclo[6.3.0.03,9]undecan-10-one skeleton of 1 was constructed in one step by intramolecular cyclisation of 4-substituted cyclopent-2-enone 2 according to a scheme of double conjugate addition. { The data on tandem transformations published earlier have been surveyed in a monograph 1 and reviews.2 ±6E V Gorobets,MS Miftakhov, F A Valeev 1002 O LHMDS ±78 to 0 8C CO2Me (CH2)3 2 The synthetic potential of enimines in a two-step process involving the Michael addition followed by cyclisation has been examined.13 Readily accessible 1-aza-1,3-butadienes (enimines) 11 and 12 reacted with methyl acetoacetate and acetylacetone in the presence of catalytic amounts of LiI to form unsymmetrically sub- stituted 1,4-dihydropyridines 13 or cyclohexenone derivatives 14. 3 steps O HO OH CO2Me 1 (94%) LHMDS is lithium hexamethyldisilazide. The reaction pathway depends on the structure of the enimine used.The reaction of methyl acetoacetate with N-benzylenimines of type 11 gives rise to 1,4-dihydropyridine derivative 13. If the benzyl group in the acceptor is replaced by the tert-butyl group (compound 12), the reaction follows another pathway to form the cyclohexenone derivative 14. In these reactions, either two (path a) or three (path b) carbon atoms of the 1,3-dicarbonyl compound are involved in the formation of the six-membered ring. R2 R1 R1 R2 R3 MeO2C R3 The same strategy was used for the preparation of sesquiter- penoids ()-8,14-cedranediol (3) and ()-8,14-cedrane oxide (4) based on 5-substituted cyclopent-2-enone 5.9, 10 a MeO2C Me N NCH2Ph O O 5 steps O 13 CH2Ph 11 CO2Et H R2 R1 R1 ZnCl2, Et3N, TMSCl 1,2-Cl2C6H4, 150 8C R2 5 CO2Et (91%) R3 MeO2C R3 b MeO2C O NBut O 14 12 R1=Alk, Ar; R2, R3=H, Alk.O OH OH 3 4 TMSCl is trimethylsilyl chloride. This divergence was not observed in the reactions of enimines 11 and 12 with dimethyl 3-oxoglutarate. Both enimines formed derivatives of bicyclo[3.3.1]nonan-3-one 15. H R1 R1 E R2 R2 E O H E NR3 O E E The total synthesis of alkaloid cylindricine A (6) of animal origin, which was first isolated from ascidia in 1993, was performed.11 The Michael addition of ammonia to the conjugated diene system of the ketone 7 afforded the key synthon, viz., perhydroquinolinone 8.Its subsequent reaction with N-chloro- succinimide and radical cyclisation gave rise to a 55 : 45 mixture of cylindricineA(6) and its epimer, which could be readily separated. E A 11, 12 O O O E R1 E R1 2 steps R2 R2 E E C6H13-n NH4OH±MeOH, NH4Cl 73 8C NH OH C6H13-n O E E E E OH OH 15 CH2 8 (47% ± 56%) CH2 7 E=CO2Me; R1, R2=H, Alk, Ar; R3=CH2Ph (11), But (12). O C6H13-n N 6 Cl This result indicates that two molecules of the oxo diester are involved in the reaction. When the diester and enimine were taken in a ratio of 2 : 1, the product 15 was obtained in higher yield. The possible mechanism of transformation involves the formation of the intermediate A, which enters into the second Michael reaction followed by cyclisation.The total synthesis of alkaloid cylindricine C (9) related to compound 6 was performed in two steps 12 starting from the azide 10. Selective reduction of the azide in the presence of the diene system was carried out with the use of CrCl2. The resulting amine entered into the highly convergent and efficient double Michael addition reaction. C6H13-n O O The double Michael addition makes it possible to construct the bicyclo[2.2.2]octane skeleton of many natural compounds based on oxophorone (16) and its derivatives.14 The reactions of oxophorone enolates with a,b-unsaturated carbonyl compounds 17 gave rise to the intermediate A, which underwent 6-endo-trig cyclisation following the path a to form the only product 18.The isomer 19, which could be formed in the case of the path b, was not detected. C6H13-n N3 1) CrCl2, H3O+ 2) TBAF, THF OTBS N OH 10 9 (45%) TBS is the tert-butyldimethylsilyl group, TBAF is tetrabutylammonium fluoride.Tandem transformations initiated and determined by the Michael reaction OM O O O16 M a b O ORO A M=Li, SiMe3; R=H, Alk. The double conjugate addition of enones 20 to fulvene (21) offers a convenient approach to tricyclo[5.3.0.n2,5]alkane sys- tems.15 The dienolate anion generated from enone enters into the Michael reaction with fulvene to give an anionic intermediate. The latter undergoes cyclisation as a result of the intramolecular Michael reaction giving rise to tricyclic ketone 22 or 23 with high stereoselectivity. The reaction is particularly efficient with the use of cyclopentenone as the Michael acceptor.In the latter case, tricycloalkanes are obtained in 75% ±96% yields. 1) LDA, THF, 730 8C R1 2) O 21 R2 (CH2)n 20 (CH2)n O R1 R2 22 R1, R2=H, Alk; n=1 ± 3; LDA is lithium diisopropylamide. The new method for the construction of polyfunctional cyclic compounds containing several substituents developed by Gross- man et al. 16 ± 19 is based on the use of the double Michael reaction of carboxylates 24 with but-3-yn-2-one. Y Z HC CCOMe (H2C)n (H2C)n CO2Et NaH, THF 24 Z Y, Z=CN, CO2Et; n=0, 1. Z NaOEt, EtOH, H+ n=1 Z H2, Pd/C, AcOH Y=CO2Et, Z=CN; n=1 Cyclic products 25 are convenient starting compounds for the synthesis of biologically active polycyclic compounds.In partic- ular, the Dieckmann reaction and intramolecular reductive ami- O R 17 O R O a O 18 (6-endo-trig) O O b R O 19 (5-exo-trig) (CH2)n O or R1 R2 23 Y Z CO2Et COMe Z 25 O Y OEt 26 CO2Et CN CO2Et Me NH 27 1003 nation of these compounds afforded trans-decalin 26 and trans- perhydroisoquinoline 27, respectively. Three new C7C or C7N bonds, the ring system of bicyclo[4.4.0]decane and up to four new stereocentres can be constructed starting from two readily acces- sible acyclic compounds. Tandem reactions giving rise to angular polyquinanes were reported.20 The derivative of squaric acids, viz., cis-13-methyltri- cyclo[10.3.0.04,9]pentadeca-4(5),12(13)-diene-3,14-dione (28), was involved in intramolecular cyclisation initiated by PhS7.Enolate 29 formed in the first intermolecular Michael reaction underwent intramolecular transannular cyclisation in the second step to give the angular polyquinane 30. H Me Me H 0.1 equiv. PhSH, 0.1 equiv. PhSNa O O O THF, D, 12 h H H 30 (93%) 28 O ± PhS7 PhS7 H Me H Me O O O SPh H H 29 O7 SPh The Michael reactions of methyl (chloro)cyclopropylidene- acetate (31) with cyclic dienolates formed from enones 32 and 33 were studied.21 These reactions gave rise to tricyclic adducts 34 and 35, respectively. It was proposed that the compound 35 can be used as an intermediate in the synthesis of sea diterpenoids, viz., mediterraneols A and B (36).CO2Me 1) LDA, THF, 778 8C (CH2)n CO2Me O 2) 32a ± c (H2C)n 31 Cl 34a ± c O n = 5 (a), 6 (b), 7 (c). Yield (%) Compound 34 abc 40 34 78 CO2Me O OBn 1) LDA, THF, 778 8C 2) 31 OBn O 33 35OH O O OH O 36 The methodology of tandem 1,4-inter- and intramolecular addition reactions has been used for the preparation of pyra- zolo[1,5-a]quinoline ring systems 38 and 39 inhibiting DNA synthesis in bacteria starting from the quinolone 37.22 The key tricyclic synthons 40a,b were synthesised by the reactions of the quinolone 37 with different acrylates.1004 OH O F F CO2Et CO2Et H 1) NaH, DMF, 0 8C 2) CO2R N N F F CO2R HNMe F F N 37 Me 40a,b (88% ± 97%) R=Me (a), But (b).OH OH F F CO2Et CO2Et N N F F N F N F Me Me 39 38 The double Michael reactions of 1-nitrocyclohexene (41) with ynols 42a,b afforded nitrooctahydrobenzofurans (43a) and -pyrans (43b), respectively.23 The possible reaction mechanism involves the addition of the alcohols 42a,b at the double bond of the cycloalkene 41 and intramolecular cyclisation of the resulting anion A via the transition state B to give the anionic intermediate C containing cis-fused rings. Subsequent protonation of the anion C affords a mixture of the E- and Z-isomers 43a,b. COR3 COR3 O2N NO2 (CH2)n ButOK HO + THF ± ButOH R2 R1 O (CH2)n R2 R1 41 42a,b H 43a,b (97%) E :Z=55 : 45 n = 0 (a), 1 (b); R1, R2=H, Alk; R3=OAlk, NMe2.O7 O7 COR3 + COR3 N NO2+ O H 7O A O7 + N NO2 O7 O7 O 43a C C C H R3 H R3 O O B C The use of terminal alkynes enables one to change the order in which the reagents enter into the reaction. In this case also, the reaction results in a five-membered heterocyclic compound.24 Thus the reaction of acyclic nitroalkene 44 with the ynol 45 NO2 CH2 R1 ButOK, THF R3 R2 R4 H R1 NO2 OHR4 R3 O46 45 NO2 R2 H CH2 R1 44 ButOK, THF NHMe R2 N H 47 48 Me R1=Alk; R2=Alk, Ar; R3, R4=H, Alk. E V Gorobets,MS Miftakhov, F A Valeev afforded tetrahydrofuran derivative 46. The aza-Michael addition of N-methylprop-2-ynylamine (47) to nitroalkenes 44 proceeded regio- and stereoselectively to form 3-methylenepyrrolidines 48.A new procedure for the stereoselective preparation of tri- cyclo[5.3.1.02,6]undecane-4,11-diones (49a) and tricyclo- [5.4.1.02,6]dodecane-4,12-diones (49b) is based on the inter- and intramolecular Michael reactions of 1-pyrrolidinocyclohexene (50a) and -cycloheptene (50b), respectively, with the mesylate 51.25 O N R 1) MeCN, D 2) Et3N, D + 3) AcOH, H2O (CH2)n OMs 51 50a,b R O H 7 H (CH2)n R=Alk, Ar; n = 1 (a), 2 (b); Ms is mesyl. Organolithium compounds initiate Michael cyclisation of a,b,c,o-unsaturated bisphosphonates 52.26 Thus the addition of PhLi to a solution of tetraester 52 afforded cyclic bisphosphonate 53, which is a product of the tandem Michael reaction. PO(OEt)2 PhLi THF, 778 8C PO(OEt)2 52 PO(OEt)2 Li PO(OEt)2 Ph The reaction of cyclohexane-1,3-dione (54) with ethyl acrylate proceeded as an intermolecular double Michael addition.27 O CHCO2Et CH2 NaH, DMF, 80 8C, 4 h O 54 O CO2Et O 55 O CO2H CO2H O H (H2C)n O O O R H 49a,b PO(OEt)2 PO(OEt)2 Ph 53 (50%) cis : trans=3.5 : 1.0 CO2Et O CO2Et + O 56 OMeTandem transformations initiated and determined by the Michael reaction O7 CHCO2Et CH2 56 O 54 54 0 O O O7 CO2Et OEt OEt O O7 O 550 56 0 O7 CO2Et CHCO2Et CH2 O 55 00 H+ 55 In the presence of catalytic amounts of a base, the reaction gave rise to a mixture of two products, viz., 55 and 56, in a ratio of 3 : 1.The reaction in the presence of an equimolar amount of a base afforded only the monoaddition product 55. The observed difference can be explained as follows. An appropriate amount of the enolate anion 54 0 is generated under the action of catalytic amounts of a base. The Michael reaction of this anion with ethyl acrylate yields the anion 55 0 whose intramolecular isomerisation gives rise to the anion 55 00. The latter either undergoes protona- tion to form the product 55 or enters into the second Michael reaction. The enolate anion 56 0 abstracts the labile proton from the diketone 54 yielding the anion 54 0 and the diester 56. As a result, a mixture of the products 55 and 56 is obtained.In the presence of one equivalent of a base, the enolate anion 54 0 is formed in quantitative yield. The Michael reaction affords the anion 55 0, which undergoes intramolecular deprotonation ± pro- tonation to give the anion 55 00 and then exclusively the mono- addition product 55. Esters of a,b,a0,b0-dienedioic acids can serve as potential substrates for anionotropic tandem transformations. The inter- mediate enolate which is generated in the first step of intermolec- ular conjugate addition of the nucleophile attacks the second a,b- unsaturated ester group according to an intramolecular mecha- nism. The use of a chiral nucleophile in these tandem conversions allows the preparation of optically active adducts. Thus the reaction of homochiral lithium [(1R)-methylbenzyl]benzylamide with dimethyl octa-2,6-dienedioate (57) gave rise to chiral cyclo- pentane derivative 58, the configurations of the C(1) and C(2) atoms being completely controlled and the configuration of the C(5) atom being controlled to a large degree [the content of the C(5)-epimer 59 was<5%].28 Ph Ph CO2Me NLi (1 ± 12 equiv.) CO2Me 57 N N Ph Ph Ph Ph CO2Me CO2Me + CO2Me CO2Me 59 58 The reaction of 4,4-dimethoxycyclohexa-2,5-dienone (60) with ethyl acetoacetate and acetylacetone proceeded as a tandem 1005 Michael reaction.29 Compounds 61a,b were formed as a result of successive Michael C- and O-addition.OMe MeO OMe MeO O O O X O electrolysis X O O60 OMe MeO O O C O X 61a (98%), 61b (57%) X=OEt (a), Me (b).The bifunctional reagent, viz., 2-bromomethyl-3-phenylsulfo- nylprop-1-ene (62), reacts with different electrophilic alkenes to give [3+2]-cycloadducts.30 The use of acyclic (E)-enoates in these reactions allows the preparation of stereohomogeneous trans, trans-trisubstituted methylenecyclopentanones 63 in high yields (90%). The tandem reactions of bromosulfone 62 with a,b- unsaturated esters containing the O atom at the g position, in particular, with (4S)-enoate 64, also proceeded with good facial selectivity. Subsequent ozonolysis of the alkene 65a gave rise to enantiomerically pure cyclopentanone 66. CH2 PhO2S CH2 CO2R2 LDA Br + PhO2S R1 CO2R2 62 R163 (>90%) 62, LDA O O CO2Me 64 CH2 CH2 PhO2S PhO2S + CO2Me CO2Me H H O O O O 65b 65a (91%) a : b=95:5 O3 O PhO2S CO2Me H O O 66 (90%) Abundant data on tandem reactions of 1,6-anhydro-3,4- dideoxy-b-D-glycero-hexopyranos-3-en-2-one, the so-called levo- glucosenone (67), are now available in the literature (see, for example, Refs 31 ± 33).Interesting conversions take place upon treatment of levoglucosenone with pyrazole under conditions of electrolysis. The first stage involves the cathode-initiated addition of pyrazole followed by the selective addition of yet another molecule 67 to the adduct to form the dimer 68 in quantitative yield.34 Previously, the dimer 68 was obtained in low yield (8%)1006 along with trimers 33 by heating levoglucosenone in aqueous Et3N.O O O N 7 7 O 67 N O N N 67 O O O O O N O O H ± N N 7 N 7 O O O O O O 68 The introduction of the halogen atom at position 3 of levoglucosenone extends the possibilities of its tandem trans- formations. The effects of the temperature and the reagent ratio on the structure of the product were examined using the reaction of sodium ethyl acetoacetate with 3-iodolevoglucosenone (69) as an example.35 O O O CO2Et NaH, THF O I O 69 O O O O O O O O 20 8C O O HO + O CO2Et H H O O CO2Et O 70 (73%) O 71 (20%) 69 20 8C O O O O EtO2C I ±60 8C OH + EtO2C O O O 73 (40%) 72 (42%) The reactions with an excess of the ketone 69 (2 equiv.) at room temperature afforded two products, viz., stable oxetene derivative 70 and polycyclic compound 71 36 containing the cyclo- butane fragment.The reaction of equimolar amounts of the reagents at 760 8C gave rise to 1 : 1 adducts 72 and 73. It should be noted that the reaction of the compound 72 with 3-iodolevo- glucosenone in the presence of NaH yielded the adduct 70.36 The assumed mechanism of conversions is shown in Scheme 1. The reaction of yet another representative of CH-acids, viz., acetylacetone, with 3-iodolevoglucosenone (69) proceeded analo- gously. O O O O O O O O O O 69 + O NaH, THF, 20 8C H O O O (40%) (20%) HO E V Gorobets,MS Miftakhov, F A Valeev O O O d O I 1,5-shift O I 7 O CO2Et O d O O7 O I OH EtO2C I O O O O H O I O O EtO2C 7 III.Tandem `Michael reaction ± nucleophilic substitution' sequence As mentioned above, the addition of a nucleophile to an acceptor gives rise to an anionic intermediate, which can act as a new nucleophile. If the molecule of the latter contains a good leaving group X, the subsequent intramolecular nucleophilic substitution can occur. The procedure `tandem oxa-Michael addition ± intramolecu- lar SN2 0-substitution' was used 37 for the efficient and stereo- controlled synthesis of poorly studied allenetetrahydrofurans. Nitrocycloalkenes 74a ± c react smoothly with 4-chlorobut-2-yn- 1-ol to give bicyclic adducts 75a ± c. The reaction mechanism is analogous to that of the well-studied reactions of nitroalkenes with O-nucleophiles. Initially, the oxa-Michael addition affords the nitronate A, which undergoes cyclisation of the SN2 0 type to form the allene derivative 75.NO2 ClCH2C CCH2OH (H2C)n ButOK, THF, 0 8C R R 74a ± c NO2 (H2C)n C R R O 75a ± c (70% ± 80%) R=H, n = 1 (a); R =H, n = 2 (b); R±R=OCH2O, n = 2 (c). Aprocedure for cyclopropanation of cyclic and acyclic enones is based on the `intermolecular Michael reaction ± intramolecular alkylation' sequence.38 Thus the reactions of readily accessible Scheme 1 72 73b a O O O O I c c ba 7 7O O O CO2Et O CO2Et 69 O O I O7 EtO2C O O 70 H O I O 7HI 71 O EtO2C H O7 +N (H2C)n O7 R Cl R O A1007 Tandem transformations initiated and determined by the Michael reaction a-haloenones 76 and 77 with mild C-nucleophiles 78 gave rise to cyclopropane derivatives 79 and 80, respectively.O O Z Y 78 (H2C)n (H2C)n H THAB, K2CO3, PhMe Br Z H 79 76 Y Z Y n=1, 2; Y = CO2Alk, CN, H; Z=CO2Alk, NO2; THAB is tetra- hexylammonium bromide. O 78 R2 R1 R1 R2 THAB, K2CO3, PhMe O X 80 77 X=Cl, Br; R1, R2 =Alk, Ar. Cyclopropane derivative 81 was synthesised by the reaction of the malonate anion with 3-iodolevoglucosenone (69) as an electro- philic component.39 O O O O EtO2C CO2Et NaH, THF, 20 8C O O I EtO2C 69 second cyanoacetate anion. This fact was confirmed by the reaction of the individual cyclopropane 82 with sodium ethyl cyanoacetate or the malonate anion at 20 8C giving rise to tetracyclic compounds 83 and 84, respectively.The classical tandem reaction of polyfunctional substrates, viz., of electrophile 85 and nucleophile 86, (Scheme 2) has been described.40 It appeared that the result of transformations is governed by the order in which the reagents are added. The addition of dimethyl 3-oxoglutarate (86) to a suspension of the dibromide 85 (i.e., in the presence of an excess of 85) and Na2CO3 in dry THF afforded substituted benzo[g]isochromene 87. On the contrary, the addition of the dibromide 85 to a suspension of the diester 86 (i.e., in the presence of an excess of 86) and Na2CO3 in THF gave rise to a derivative of cyclopenta[1,3]cyclopropa[1,2-b]- naphthalene 88.The observed selectivity can be explained as follows. Alkylation of the b-oxo ester 86 with the highly reactive benzyl-type bromide 85 leads to the intermediate A. The latter can react in two different ways. In the presence of an excess of the bromide 85, the intermediate A is subjected to O-alkylation to form the intermediate B, which undergoes the intramolecular Michael cyclisation with elimination of HBr to give benzoiso- chromene C. Isomerisation of the double bond in the latter affords the final product 87. If the reagents are added in the reverse order, the intramolecular Michael reaction takes place and the inter- mediate A is converted into the spiro compound D. The latter undergoes smooth cyclisation to form cyclopropane E and sub- sequent alkylation at the activated CH group with the second molecule of the dibromide 85 affords the product 88.CO2Et 81 (90%) A new strategy for the synthesis of carbocyclic b-amino acids was reported.41 The conjugate addition of the chiral analogue of ammonia, viz., (S)-(7)-1-(trimethylsilylamino)-2-(methoxyme- thyl)pyrrolidine (TMS ± SAMP), to o-halogen-substituted enoates 89 gave rise to intermediate ester A, Hal O CO2R 7 Apparently, a 1,3-shift of electron density occurs in the intermediate carbanion A resulting in generation of the more stable carbanion B, which eliminates the iodide ion to give the product 81 containing the cyclopropane ring. O O O I I 81 * NH2 7I7 7 A 7 EtO2C EtO2C O CO2Et B O CO2Et A which underwent intramolecular cyclisation to give compound 90 with high stereoselectivity.Subsequent removal of the protective trimethylsilyl group, reduction of the N7N bond and hydrolysis of the ester group afforded carbocyclic b-amino acid 91. O OBut Hal (CH2)n 1) TMS± SAMP, ButLi, THF, ± 78 8C 2) HMPA, ± 78 8C 3) NaHCO3, H2O Analogously, 3-iodolevoglucosenone (69) reacted with sodium ethyl cyanoacetate at 760 8C to give the corresponding cyclopropane 82.35 However, the reaction of the ketone 69 with ethyl cyanoacetate at 20 8C afforded imidate 83, which is the product of the cascade reaction of the cyclopropane 82 with a 89 O OMe O NH2 EtO2C CN CO2H SiMe3 N NaH, THF 4 steps N O I CO2But O O 69 (CH2)n 7 O O 91 EtO2C CO2Et ±60 8C 90 (CH2)n CO2Et O O CO2Et EtO2C EtO2C n=0 ± 4; HMPA is hexamethylphosphoramide; OMe SiMe3 .N TMS±SAMP= CN 82 (80%) 84 NH 7 NH EtO2C CN O O A procedure for the synthesis of chiral bicyclo[3.3.1]nonanes from homologues of carvone (92) has been developed.42 The reactions of allylic bromides 93 with different CH-acids occur as CO2Et 20 8C O CN EtO2C NH 83 (72%)1008 O O CO2Me MeO2C Br 86 Br O85 O Br CO2Me CO2Me 85 O O O Br O B O BrCO2Me O OD CO2Me the tandem `intermolecular alkylation ± intramolecular Michael reaction' sequence giving rise to bicyclic compounds 94. O O NBS CH2XY K2CO3 R R Br 93 92 R=H, Me; X, Y=CO2Alk; NBS is N-bromosuccinimide. Cl OC SO2 CH2 O2N S 95Cl O 7 C C SO2 O2N S (CH2)2CO2Et AO O2N S (CH2)2CO2Et 96 (74%) DBU is 1,8-diazabicyclo[5.4.0]undec-7-ene. O Br CO2Me CO2Me OH O AO CO2MeO BrCO2Me O O O C O CO2Me 85 O OE CO2Me X R Y The reaction of 2-(2-chloro-4-nitrophenylsulfonyl)acetylthio- phene (95) with ethyl acrylate proceeds in a `Michael reaction ± nucleophilic substitution' sequence.43 The possible mechanism involves generation of the anions A and B under basic conditions.The oxygen atom of the anion B bearing a negative charge can readily attack the C(2) atom of the benzene ring resulting in intramolecular nucleophilic substitution to form cyclic product 96 (Scheme 3). O 94 (68% ± 97%) DBU, DMF O2N CHCO2Et CH2 + 50 ± 55 8C Cl SO2 O2N B S E V Gorobets,MS Miftakhov, F A Valeev OO O CO2Me OO O MeO2C Br O 88 Cl O CH C SO2 (CH2)2CO2Et 7O DBU, DMF C C 50 ± 55 8C S (CH2)2CO2Et Scheme 2 CO2MeO BrCO2Me O O 87 Scheme 3 DBU, DMF 50 ± 55 8C S1009 Tandem transformations initiated and determined by the Michael reaction IV.Tandem `Michael reaction ± aldol reaction' sequence anion. The phenols 101 were obtained in high yields. Only the reaction with the use of propynal (R=H) as an acceptor gave the product in low yield (11%), which is attributable to side anionic polymerisation of propynal. OH CR The basic medium required for the Michael reaction favours a subsequent aldol reaction. This type of transformation has been described in the literature in detail. O OHCC 102 CO2Me MeO2C The short stereospecific synthesis of bicyclo[3.3.1]nonane CO2Me MeO2C 86 derivatives (97) from acyclic diester 86 and enal 98 was performed by Michael addition followed by intramolecular aldolysation with the use of the reagent ratio 86 : 98=2:1.44 R 101 (45% ± 88%) R3 O R=H, Alk.TBAF H R1 CO2Me + MeO2C O 86 R298 OH MeO2C CO2MeR2 CO2Me HO R1 The tandem `conjugate addition ± aldol condensation' sequence has been used in the synthesis of the b-lactam fragment of thienamycin (103).47 The highly stereoselective conjugate addition of lithium (R)-(a-methylbenzyl)allylamide (104) to an unsaturated ester, viz., to (E)-tert-butylpenta-2,4-dienoate (105), followed by aldol condensation with acetaldehyde (106) led to assembly of all three stereocentres of azetidinone 103 with the required absolute stereochemistry.R3 Me Me CO2Me N Ph N Ph 97 (53% ± 99%) 5 steps H R1, R2, R3=H, Alk, Ar. Li (R)-104 CO2But CO2But Me OH 105 H H (82%) O Me An efficient procedure was developed for the construction of five ± seven-membered carbocycles by the tandem `Michael reac- tion ± aldol cyclisation' sequence based on o-oxo-a,b-unsaturated esters 99.45 Treatment of the latter with lithium a-toluenethiolates afforded cyclisation products 100 with a good stereoselectivity and in high yields. 106 H H ButMe2SiO THF (CH2)nCHO (CH2)nCHO + PhCH2SLi OR 0 8C CO2R Me N OLi SCH2Ph 99a ± c O H OH 103 (CH2)n CO2R SCH2Ph 100a ± c n Yield of 100 (%) R Compounds 99 and 100 abc 96 95 33 012 Et Me Et The tandem `Michael addition ± aldol reaction' and `Michael addition ± aldol reaction ± Michael reaction' sequences have been described.48 Dimethyl fumarate 107 reacted with heterocyclic ketone 108 in the presence of Et3N to give the expected Michael adduct 109 (the yield was 38%) along with pyrrolo[1,2-a]indoles 110 (16%) and 111 (3%) (Scheme 4).Heating of the individual compound 109 with Et3N afforded the product 110 the formation of which is attributable to spontaneous intramolecular cyclisation of the diester 109 followed by elimination of MeOH. The Michael addition of the pyrrole 110 to the dimethyl fumarate 107 afforded the tetraester 111.An efficient one-pot procedure was proposed for the con- struction of trisubstituted phenols 101 starting from prop-2-ynals 102 and the diester 86.46 This tandem process involves the Michael addition followed by aldol cyclisation of the intermediate enolate The `Michael reaction ± intramolecular aldol cyclisation' sequence was used for the synthesis of cycloalkenols 112 from oxo aldehydes 113 under the action of piperidine.49 The resulting Scheme 4 O O H CO2Me O O MeO2C CO2Me CO2Me CO2Me 107 N OMe N + + Et3N N N CO2Me Ac 108 CO2Me OMe 109 Ac CO2Me 110 Me CO2Me Et3N MeO2C 111 CO2Me O O 7 OMeCO2Me H CO2Me OMe N N CO2Me Me O Me1010 O CO2Et CO2Et O N NH + S S 114 115 CO2Et CO2Et O O NH NH 114 O S S b CO2Et NHS cyclic intermediate A (for R=Ph, this intermediate is stable for several days) eliminated piperidine (slowly for R=Ph) to give the enol 112.NH O OH O NH, R S 117 OCDCl3, 20 8C R (CH2)n N 2 ± 5 days 7 NH (CH2)n 113 A O OH R (CH2)n 112 (20% ± 25%) The adduct 122 was obtained in good yield and high syn- selectivity by the reaction of a mixture of lithium thiophenoxide, an a,b-unsaturated ester and an aldehyde.51 The reaction began with the nucleophilic attack of the thiolate at the b-carbon atom of the acceptor giving rise to intermediate enolate which entered into the aldol reaction with the aldehyde to form the polyfunctional- ised adduct 122 as a mixture of the syn- and anti-isomers.An analogous mixture of isomers was obtained in the reaction with the use of lithium selenolate generated in situ under the action of methyllithium on diphenyl selenide. R=Alk, Ph; n=1, 2. PhSLi R1, R2=H, Alk. PhSeLi The reaction of enamine 114 with methyl vinyl ketone at room temperature afforded only the product 115 in 95% yield (Scheme 5).50 Refluxing of the reactants gave rise to the tetracycle 116 (the yield was 72%) along with the compound 115 (10%). Apparently, this is attributable to the initial addition of one molecule of methyl vinyl ketone at the a-carbon atom of the enamine followed by isomerisation of the double bond of the adduct and addition of the second molecule of methyl vinyl ketone.The process is completed with intramolecular aldol condensation (path a) giving rise to the tricycle 115. Upon heating, aldol condensation (path b) occurs faster than the reaction following the path a and product 116 is formed. R=H, Alk. The reactions of heterocyclic enamines 114 and 117 with dimethyl acetylenedicarboxylate (DMAD) afforded a mixture of the corresponding Michael addition (118 and 119) and cyclisation products (120 and 121). The individual compounds 118 and 119 were quantitatively converted upon heating into the cyclic ketones 120 and 121, respectively. O CO2Me CO2Me EtO2C NH N DMAD CO2Et + 114 CO2Me Previously, it had been demonstrated that the reaction of levoglucosenone (67) with an excess of nitromethane afforded the addition product 123 (the yield was 89%).The reaction in the presence of an excess of levoglucosenone gave rise to the adduct 124 (95%).31 The reaction with the use of an equimolar mixture of the reagents gave both products 123 and 124 in 18% and 61% yields, respectively.32 The electrochemical method proved to be more selective.34 Electrolysis of an equimolar mixture of nitro- methane and levoglucosenone afforded exclusively the ketone 124 (the yield was 92%). S 120 (2%) S 118 (16%) E V Gorobets,MS Miftakhov, F A Valeev CO2Et NS O 116O NH a O S O NHS CO2Et EtO2C NH DMAD S119 (62%) OH CH2 R2 CHCO2R1, R2CHO PhS syn-122 CH2 CHCO2But, RCHO RPhSe syn : anti=8.5 : 1.0 Scheme 5 O CO2Et 115 OH CO2Et 116 O7 CO2Me CO2Me CO2Et + CO2Me N O S 121 (16%) OH CO2R1 CO2R1 R2 + PhS anti-122 syn : anti=9: 1 OH CO2ButTandem transformations initiated and determined by the Michael reaction O OH O O H 1) MsCl, DMAP H (CH2)5CO2Me ALB (CH2)5CO2Me 2) Al2O3 127 CMe(CO2Bn)2 7 128 (86%) CMe(CO2Bn)2 126 MsCl is mesyl chloride; DMAP is dimethylaminopyridine.O O O O MeNO2 OH O O and (or) OH O 67 O O2N NO2 123 O2N O O 124 The assumed mechanism of formation of the adduct 124 involves the Michael addition of the anion A, which is generated by the addition of the nitromethane molecule to the ketone 67, to the second levoglucosenone molecules followed by intramolecular aldol condensation of the intermediate diketone B.O O electrolysis 67 MeNO2 O O 67 7 O O O2N O2N A OO H+ O 124 7O O2N O O B The `Michael reaction ± aldol reaction' sequence was used as the key stage in the synthesis of the prostaglandin 11-deoxy- PGF1a (125).52 The reaction of cyclopent-2-enone with dibenzyl methylmalonate (126) and the aldehyde 127 in the presence of aluminium lithium bis[(S)-1,10-bi-2-naphthoxide] (ALB) as a catalyst produced the expected three-component coupling prod- uct 128 as a mixture of two diastereomers in a ratio varying from 6 : 1 to 17 : 1. The use of this catalyst not only initiates tandem conversions, but also provides high stereoselectivity of addition.Mesylation of the adduct 128 followed by elimination of MsOH afforded optically pure compound 129, which can be converted into the target product 125 (Scheme 6). V. Tandem Mannich and Michael reactions It is known that Mannich bases are very convenient intermediates in a number of syntheses. These bases can readily be converted into vinyl ketones and used in the Michael reaction as electro- philes. Due to their proneness to polymerisation, acceptors are generated from Mannich bases in situ. Since decomposition of Mannich bases produces secondary amines, which can catalyse the Michael reaction, an additional amount of base need not be added. The Mannich and Michael reactions of imines with the Danishefsky diene proceed in tandem.53 Electron-excessive 1-methoxy-3-trimethylsilyloxybutadiene (130) reacted with the imine 131 in the presence of a Lewis acid to give the addition product A.Acid treatment of the latter led to rapid intramolecular 1011 Scheme 6 O O (CH2)5CO2H (CH2)5CO2Me C5H11-n CMe(CO2Bn)2 OH 125 129 (92% ee) Michael addition and elimination of MeOH to form cyclic enaminone 132. OSiMe3 Me H ZnI2, MeCN OBn N Ph + 720 8C, 4 h OBn 131 130 MeO O O H H H+ OBn N OBn MeO HN OBn OBn Me Me Ph Ph A 132 (75%) The tandem Mannich and Michael reactions were used for the enantioselective synthesis of chiral piperidine derivatives.54 Aldi- mines derived from tetra-O-pivaloyl-b-D-galactopyranosylamine (133a,b) and the diene 130 gave 2-substituted dehydropiperidine derivatives 134a,b, respectively. The latter were transformed into alkaloids coniine (135) and anabasine (136).OPiv OPiv PivO O 1) 130, ZnCl2, THF, 720 8C PivO 2) HCl, H2O O O N R N PivO PivO OPiv OPiv R H 134a,b 133a,b Piv=C(O)CMe3. Yield of 134 (%) Compounds 133 and 134 R ab 96 90 Prn 3-Pyridyl 4 steps 134a NH 135 4 steps 134b NH136 N An efficient and short procedure was developed for the syn- thesis of tetra- and hexahydrobenzoquinolizinones, which are precursors of psychoactive alkaloids of the berberine series.55 The Mannich and Michael reactions of 2-bromohomoveratryl- amine (137) and silyloxydiene 130 gave rise to N-substituted enaminone 138, which underwent Heck cyclisation to form the heterocycle 139.MeO 1) RCHO 2) 130, ZnCl2 ±THF or EtAlCl2±CH2Cl2 Br NH2 MeO 1371012 MeO N MeO Br O 138 MeO N MeO 139 O R=H, Alk or Ar; dba is dibenzylideneacetone. VI. Tandem Michael reaction and heterocyclisation The Michael addition of cyclohexaneimines (which react as enamines) to b-substituted nitroalkenes is accompanied by cycli- sation and elimination of the nitro group to form substituted tetrahydroindoles.56 The imine 140 reacted with trans-b-nitro alkenes 141a,b to give the expected tetrahydroindoles 142a,b, respectively. R O2N141a,b N PhMe, 0 to 20 8C 140 Ph R+ O7 N N OH Ph R Compounds 141 and 142 ab Ph Et Allene-1,3-dicarboxylate (143) acts as an extremely reactive acceptor in the Michael reaction.This compound reacts with secondary amines to give the ion A as a result of conjugate addition. The ion A can also exist as the dianion B. The latter contains a good leaving group and undergoes smooth cyclisation. The final product, viz., azabicycle 144, represents a skeleton of pyrrolizidine alkaloids.57 DMSO, Et3N MeO2C CO2Me O 86 7 CO2Me MeO2C N Cl Cl A CO2Me MeO2C N 144 (83%) Pd2(dba)3 .CHCl3, PPh3, K2CO3, DMF R R R O7 N N O Ph R N Ph 142a,b Yield of 142 (%) 70 76 H MeO2C (C2H4Cl)2NH C Et3N H CO2Me 143 MeO2C CO2Me 7 7 NaH, DMF + N Cl Cl B E V Gorobets,MS Miftakhov, F A Valeev The procedure for the synthesis of various different function- alised 3-methylidenetetrahydrofurans 145 58 is based on the con- jugate addition of anions of propargylic alcohols 146 to alkylidene- or arylmethylidenemalonate initiated by catalytic amounts of a base and subsequent in situ Pd-catalysed exo-dig cyclisation.Z1 Z2 BunLi + THF R1 O7 R3 R2 146 Z1Z2 R1 R3 O R2145 (47% ± 94%) R1, R2=H, Alk, Ar; R3=Alk, Ar; Z1, Z2=CO2Et, CN. This approach was used in the synthesis of 3-methylidenepyr- rolidines 148.59 N-Substituted propargylamine 147 is used as the nucleophilic substrate, which reacts with different acceptors to give pyrrolidines 148 in good yields. Z1 Z2 BunLi, Pd(OAc)2(PPh3)2 + THF, 20 8C R2 NH R1 147 R1, R2=Alk, Ar; Z1, Z2=CO2Et, CN.The tandem `Michael addition ± intramolecular condensation' sequence was used in the synthesis of functionalised pyrroloisoin- doles.60 Treatment of methyl ester of N-phthaloylalanine (149) with lithium hexamethyldisilazide afforded carbanion A, which reacted with a,b-unsaturated compounds to give intermediates B and C and then pyrrolo[2,1-a]isoindol-5-ones (150) according to a standard scheme. O Me LHMDS N THF, ± 78 8C CO2Me O 149 O Li+ Me R 7 N CO2Me A O R Z O7 CO2Me N Me O C R=H, CO2Et; Z=CO2Et, CN. The reactions of levoglucosenone (67) with binucleophiles proceed with the participation of the C=C and C=O bonds and can be used for the synthesis of fused heterocyclic systems containing the carbohydrate fragment.Thus the addition of nitroacetamide (151a) to levoglucosenone was accompanied by cyclisation to yield exclusively the stereoisomer 152a.34 The reactions of N-hexylnitroacetamide (151b) and a-nitropropion- Z2 Z1 7 Pd R1 O R3 R2 Z1Z2R2 NR1 148 Z Li+ O 7 R Z NMe CO2Me O B Z 1 2 R HO 9 3 H2O CO2Me N Me 5 6 O 1501013 Tandem transformations initiated and determined by the Michael reaction amide (151c) with levoglucosenone produced compounds 152b,c, respectively.61 O O Et3N, MeCN O + O2NCHCNHR2 OH O 20 ± 40 8C, 0.2 ± 24 h the intermediate A and the latter undergoes cyclisation to give compound 156. On the contrary, if the C=C bond is subjected to attack by the nitrogen atom remote from the amino group (the intermediate B), cyclisation does not take place because it should lead to the closure of the seven-membered ring, which is much less probable than the closure of the six-membered ring.R1 N R1 O 151a ± c 67 O O R2 O2N O O 152a ± c O N Yield of 152 (%) R2 Compounds 151 and 152 R1 N XN NH2 B abc 6249 69 H H HMe (CH2)5Me H The reactions of levoglucosenone with urea (153a), thiourea (153b), N-cyanoguanidine (153c) and N-nitroguanidine (153d) in the presence of bases were accompanied by pyrimidine-ring closure to form compounds 154a ± d, respectively.62 In spite of drastic reaction conditions, water was not eliminated from the molecule 154, as in the above-considered reaction, which is in agreement with the Bredt rule.O O The reaction of levoglucosenone with 2-aminopyridine (158) was examined.63 If the reagents were taken in an equimolar ratio, the reaction gave rise to a mixture of the addition product 159 and polycyclic compound 160. The reaction with a twofold molar excess of levoglucosenone afforded virtually exclusively the prod- uct 160. The compound 160 is structurally very similar to the 2 : 1 adduct of levoglucosenone with nitromethane, and, consequently, the probable mechanism of its formation involves the addition of two levoglucosenone molecules at the nucleophilic centre followed by intramolecular aldol condensation. NH O O OH O MeCN + (H2N)2C X 153a ± d + O MeCN±EtOH, D, 30 8C NH HN O 20 8C 67 NH2 O N158 67 O 154a ± d X O Yield of 154 (%) X Compounds 153 and 154 O OH O + N O O NH N O 58 S 5331 48 NCN NNO2 abcd N O 160 O 159 The reactions of levoglucosenone with a-aminoazoles 155a ± d gave rise to azolo[1,5-a]pyrimidine systems 156a ± d, respec- tively.63 In the reaction with the triazole 155b, the Michael adduct 157 was also formed.O O O N NH2 O + HN N X N The reaction of levoglucosenone with dimedone (161a) yielded only the product 162a. The reaction of levoglucosenone with barbituric acid (161b) produced the tetraketone 163 along with the heterocyclisation product 162b. The formation of pyrans of the type 162 in the reactions of dimedone or barbituric acid with compound 67 may result from the enhanced reactivity of the carbonyl group of levoglucosenone as well as from steric factors favouring cyclisation.O NH2 O N O 155a ± d 67 A X N O O O O OH MeCN O O O + X 20 ± 70 8C O O O O OH O Y X 161a,b 67 X X O N NH N Y 162a,b N N X=CH2, Y=CMe2 (a); X=NH, Y =CO (b). N O X N 156a ± d H2N 157 O Yield of 156 (%) X Compounds 155 and 156 O O O HN NH abcd N 7539 76 62 CH CNO2 CCF3 O 163 Six-membered hemiaminals 164a,b smoothly formed in the reactions of the enamine derived from ethyl acetoacetate with levoglucosenone (67)63 and 3-iodolevoglucosenone (69), respec- tively.36 Apparently, deprotonation of the NH group of the hetero- cycle under the action of the base results in the anion, which attacks the double bond of levoglucosenone.If the C=C bond is subjected to the attack of the nitrogen atom adjacent to the amino group, the reaction is not terminated at the stage of formation of1014 O NH2 O CO2Et Et2O, 20 8C X O 67, 69 X = H (67, 164a), I (69, 164b). VII. Tandem `anionotropic rearrangement ± Michael reaction' sequence Heating of diketones 165 with N,N-diethylaniline (DEA) to 200 8C afforded pyranopyrans 166.64 Apparently, the compound 165 undergoes a sigmatropic rearrangement upon heating to give intermediate A, which can undergo intramolecular Michael cycli- sation giving rise to the final product 166 due to the suitable arrangement of the phenolic OH group.O O MeN O NMe O 165O O MeN O NMe HO H+ A O O MeN O NMe O 166 R=H, Alk, Cl, NO2. The tandem `Brook rearrangement ± intramolecular Michael reaction' sequence is suitable for the construction of functional- ised carbocycles.65 Acylsilanes containing the double bond at a suitable position (separated by 2, 3 or 4 carbon atoms) serve as substrates for cyclisation. The addition of phenyllithium to a solution of the acylsilane 167 brings about the Brook rearrange- ment and the intramolecular Michael reaction, which proceed smoothly to form diastereomeric cycloalkanes 168 and 169 in a total yield of 70%± 82%. O SiMe2ButPhLi, THF (H2C)n ± 80 to 0 8C CO2Me 167 Ph (H2C)n H 168 n=1±3.O O X OH EtO2C NH 164a,b O O DEA MeN R R D, 3 h O NMe O O7 O MeN R R O NMe O H+ R O7 SiMe2But Ph (H2C)n CO2Me Ph OSiMe2But OSiMe2But + H (H2C)n CH2CO2Me CH2CO2Me 169 E V Gorobets,MS Miftakhov, F A Valeev VIII. Tandem Michael and Dieckmann reactions The cascade processes involving more than three different reactions remain poorly studied. Of these processes, the `Michael reaction ± aldol condensation ± retro-Dieckmann reaction' sequences, which are also called MARD cascades, have received the most study. Previously, it has been demonstrated 66 that the reactions of a-oxocyclopentanecarboxylates with acroleins under mild condi- tions were accompanied by two-carbon-atom ring expansion. The five-step domino reactions of simple cyclic b-oxo esters 170 and 171 with 2-substituted acroleins 172 were used for the synthesis of valuable highly substituted cycloheptenones 173 and 174, respec- tively, containing two stereocentres, the double bond and two chemically different carbonyl groups.67 The transformation cas- cade involves the Michael addition, intramolecular aldol conden- sation and retro-Dieckmann reaction followed by dehydration and chemoselective saponification of one ester group.The config- urations of the asymmetric centres are determined by the stereo- differentiating stage of the aldol cyclisation. O HO2C CO2Me 170 R R 173 DBU MeO2C H MeOH HO2C O O 172 171 CO2Me R 174 MeO2C R=Alk, Ar.The synthesis of non-symmetrical and symmetrical resorcinols 175 and 176 from aliphatic compounds by the tandem `Michael reaction ± Dieckmann cyclisation' sequence was investigated.68 The reactions of dimethyl 3-oxoglutarate (86) with alkyl alky- noates 177a ± c at 25 8C produced non-symmetrical resorcinols 175a ± c, respectively. The reactions with the use of methyl alkynoates 177d,e at low temperatures afforded symmetrical resorcinols 176a,b, respectively. Apparently, conjugate addition of the enolate 86 to methyl alkynoates 177a ± e affords vinyl O THF, NaH CR MeO2C CO2Me + MeO2CC 25 8C 177a ± e 86 OH CO2Me MeO2C for 177a ± c R HO 175a ± c OH CO2Me for 177d,e R HO 176a,b CO2Me Yield (%) Product R Compound 177 175a 175b 175c 176a 176b 59 84 67 18 34 Me Ph CH2OBz HCO2Me abcdeTandem transformations initiated and determined by the Michael reaction carbanion A.The latter is rapidly protonated with the most acidic methylene proton to give carbanion B, which is a precursor of symmetrical products 176a,b. Contrary to the expected stability, the anion B underwent isomerisation (at 25 8C) to form enolate C, which is a precursor of the non-symmetrical products 175a ± c. It is believed that the delocalised anion B is more rapidly generated from the vinyl carbanion A and is stable only at low temperatures (up to 0 8C). On the contrary, this anion undergoes isomerisation to give the more stable enolate at room temperature.Conse- quently, the enolate A undergoes cyclisation to form symmetrical resorcinols 176a,b when low temperature is required for preven- tion of polymerisation. O O O MeO2C MeO2C CO2Me OMe 7 176a,b 7 R R A MeO2C OMe B O O O 7 175a ± c CO2Me MeO RC MeO2C The tandem `Michael addition ± intramolecular Dieckmann cyclisation' sequence was used as the key stage in the total synthesis of mycophenolic acid exhibiting fungicidal, antibacte- rial, antiviral and immunosuppressive properties.69 The reaction of dimethyl 2-geranyl-3-oxoglutarate (178) with 4-pivaloyloxy- but-2-ynal (179) gave rise to the totally substituted resorcinol 180 with the required arrangement of all substituents. The resulting adduct was converted into the target mycophenolic acid (181) in six steps.O CO2Me HOCC179CCH2OPiv NaH, THF, 25 8C CO2Me 178 OH CO2Me 6 steps HO CH2OPiv CHO 180 (33%)OH OO MeO Me HO2C 181 The use of readily accessible optically pure a,b-unsaturated ketones, such as carvone (94a) or methylcarvone (94b), in the tandem sequence of three Michael reactions and Dieckmann cyclisation allows the preparation of optically pure tricy- clo[5.3.1.03,8]undecanes 182a,b, 70 which are structural cores of seychellane. 1015 O O Me MeO2C MeO2C R Me R Me Me CO2Me LDA±HMPA, THF H 94a,b O Me CO2Me R Me OH 182a,b Yield of 182 (%) R Compounds 94 and 182 ab H 4044 Me IX. Tandem Mukaiyama and Michael reactions The tandem `Mukaiyama reaction ± Michael reaction ± aldol cyc- lisation' sequence was used as the key stage in the synthesis of biologically active compound 183.71 Under conditions of the Mukaiyama reaction, enol silyl ether 184 and spiro compound 185 gave enolate A whose conjugate addition to the chalcone 186 afforded enolate B.Intramolecular aldol condensation of the latter gave rise to compound 187 from which the target product 183 was obtained (Scheme 7). The tandem Michael and Mukaiyama reactions were used in the total synthesis of vitamin D3 (188).72 H OH H D C H A OH 188 HO The reaction of ketene hemithioacetal 189 with 2-methylcy- clopent-2-enone (190) catalysed by a Lewis acid afforded enol silyl ether.The latter entered into the second conjugate addition reaction with methyl vinyl ketone acetal 191 as a reactive intermediate. The key precursor of blocks of the C/D type (192) was prepared in one step starting from three components. 1) TrSbCl6, ButS +O Me3SiO TiCl4 ± Ti(OPri)4 O 2) Amberlyst-15 + OBn O 189 191 190 H ButS(O)C OBn H O 192 (64%) O The tandem `Mukaiyama ± Michael reaction ± Mukaiyama reaction ± aldol reaction' sequence was used for the introduction of two chains at the a and b positions of cyclic a,b-unsaturated ketones.73 The low temperature at which this reaction was performed (760 8C) hindered the formation of by-products of the crotonic type and enabled the achievement of high dia-1016 O Me3Si O + 184 185 O 184+185 Pri Mis a group stabilising enolate.stereoselectivity (cis : trans=98 : 2) and good yields of the adduct 193. This reaction is the first example of the lanthanide-catalysed tandem formation of C7C intermolecular bonds. O OSiMe3 10% SmI2(THF)2 + OMe OSiMe3 X. Miscellaneous tandem transformations involving Michael reactions Recently, a large number of new tandem reactions have been developed. Thus a three-step ring expansion was used for the conversion of methyl dialkylcyclopent-1-en-1-ylacetate 194 into methyl cyclohexylacetate 195.74 The process involved ozonolytic cleavage of the double bond in 194, the Wittig olefination of the product and the tandem sequence of dealkoxycarbonylation of the oxo ester 196 and the Michael addition.Intramolecular cyclisation gave rise to the ester 195. Alk Alk CO2Me 194 An analogous sequence was used for the preparation of lactones and lactams.75 Precursors of lactones 197a ± c were synthesised from the corresponding salicylaldehyde derivatives 198a ± c by introducing the acrylic-acid fragment using the Wittig F TiCl3(OPri) O + CH2Cl2,75 8C F3C 186 F3C F OM 186 MO A O F3C Pri CH2Cl2 O OSiMe3 H PhCHO Ph CO2Me CO2Me 193 (75%) O O CO2Me Alk Alk Alk LiCl, HMPA 120 8C AlkCO2Me CO2Et 196 195 (46% ± 74%) E V Gorobets,MS Miftakhov, F A Valeev F 8 steps O O Pri F3C OH rac-187 (42%) O 187 Bolefination and esterification of the 2-OH group for the subse- quent ring closure.The precursor of the lactam 197d was synthes- ised analogously from 2-aminobenzaldehyde (198d). CHO X X R R 198a ± d CO2Et X O R 197a ± d DMEU is 1,3-dimethyl-2-imidazolidinone. Compounds 197 and 198 R abcd H O 62 5-OMe O 50 3,4-benzo O 52 H NH 76 The reaction of levoglucosenone with malononitrile in the presence of piperidine afforded polycyclic compound 199 (Scheme 8).76 The tandem Michael and Wittig reactions involving five- membered phosphonium ylides 200 and a,b-unsaturated esters gave rise to seven-membered vinyl ethers 201.77 The process occurred via the structurally rigid phosphorus-containing bicyclic intermediate A O7 Ph EtO P+ Ph R A which provided high stereoselectivity of the formation of vinyl ethers.Analogous reactions were observed in the case of a,b- unsaturated thioesters 78 yielding cycloheptene derivatives 202. Scheme 7 F OH F Pri 183 CO2Et LiI O CO2Me DMEU, 100 8C Yield of 197 (%) XTandem transformations initiated and determined by the Michael reaction 67 + CH2 (CN)2 NC O CN O NC 7 NC O O ButOK + THF P P Ph Ph Ph Ph 200 R=Alk, Ar. A new procedure was developed for the solid-phase synthesis of imidazolidin-2-ones with the use of the tandem `aminoacyla- tion ± Michael addition' sequence.79 The reaction of the allylic amine 203 bound to an ion-exchange resin with phenyl isocyanate proceeded regioselectively to form imidazolidin-2-one 204.O 1) PhNCO, Et3N, DMF 2) CF3CO2H±H2O NHBui NH 203 The tandem Michael and Corey ± Chaykovsky reactions of five-membered oxosulfonium ylides 205 with b-acetoxy-a-methyl- idene ketones 206 produced cycloheptene oxide derivatives 207.80 Initially, the Michael addition of the ylide 205 to the enone 206 occurred. Then the acetoxy group was eliminated to form the cation A followed by generation of the ylide and the intramolec- ular Corey ± Chaykovsky reaction, which proceeded via the stable intermediate B. The latter was converted into epoxide 207. O PF¡6ButOLi + THF S S Ph O O Ph 205 S O Ph O O O O O 67 NC O O NC O CN O O O CN O 7 CNCN CNCN NC CN CN NC O O O O OEt R PPh2 D, 16 h EtO R 201 (55% ± 73%) O O SR Ph PPh2 RS Ph 202 (29% ± 58%) O NBui PhN O H2N 204 (70%) OAc 206 ButOLi +S O Ph O AO O 7 S+ Ph B O CN O CNCN 2 CH2(CN)2 NC 72 H2O CN NC O O O CN O CNN7 H+ NC CN NC O O H O S O Ph Me 207 (56% ± 74%) The tandem `Michael addition ± Meerwein ± Ponndorf ± Ver- ley reduction' sequence was used for the asymmetric reduction of a,b-unsaturated ketones giving rise to secondary alcohols.81 The reaction of the enone 208 with (7)-10-mercaptoisoborneol (209) in the presence of Me2AlCl afforded the expected product 210.In this tandem process, the chiral mercaptoalcohol reacted with enone to give an adduct, which was converted into the alcohol 210 as a result of intramolecular proton transfer.Reductive desulfurisation of the latter produced the optically active alcohol 211. Me2AlCl O R2 OH CH2Cl2 + R1 R3 SH 209 208 OH R2 R1 R3 211 R1, R2, R3=H, Alk, Ar. The three-step `reductive amination ± intramolecular Michael reaction ± lactonisation' sequence was used for the construction of the ring skeleton of homoaza-sugars in the synthesis of higher homologues of 1-deoxy-L-ido-nojirimycin (212) exhibiting a broad spectrum of biological activities.82 Apparently, reductive amination of the anomeric centre was succeeded by intramolecu- O 1) BnNH2, AcOH±MeOH 2) Na(CN)BH3 OH EtO O OBn H OH OBn Bn 3 steps N O OH O (79%) 1017 Scheme 8 7H+ O CN O CNNH2 NC CN NC O O 199 Raney Ni, NaPH2O2 pH 5.2 O S R2 HOH R3 R1 210 (85%) NH HOHO OH 212 OH1018 lar ring closure at the activated double bond and lactonisation.Reduction of the ester group and deprotection gave rise to the target product 212. The tandem Michael addition and elimination were used as the key stages in the total synthesis of the S-containing yellow pigment of Australian sponges, viz., benzylthiocrellidone (213),83 which was prepared by the reaction of ketene dithioacetal 214 with dimedone. SBn O O THF SBn + 25 8C O O 214 O 213 (60%) The transformation of cis-1-bromo-2-(1-naphthyl)ethylene containing the dimethylamino group at the peri position (215) into N-methylbenzo[de]quinoline 216 was studied.84 The tandem addition ± elimination reactions can follow two alternative path- ways, viz., 6-endo-trig addition ± elimination (path a) or tandem 5-exo-trig Michael addition followed by the carbene rearrange- ment (path b) (Scheme 9).a Br b b Me 7 H Br N a Me a 215b Me H 7 + Br N H Me 7Br7 Me +N Me Br7 7MeBr A new approach to the synthesis of 9 ± 15-membered cyclic compound was devised.85 Very dilute solutions of substrates 217 ± 219 were treated with cesium fluoride, which induced a sequence of ring opening and ring closure reactions to yield cyclopentadecanone, cyclotetradecanone and cyclononanone derivatives (220, 221 and 222), respectively (Scheme 10). The efficiency of the intramolecular Michael addition, which is the crucial step and consists in the transformation of the carbanion A into the final product, depends essentially on the `bridging' X group.O CO2Me X Z 7Z A X=(CH2)2, CH=CH. The tandem Peterson and Michael reactions have been described.86 The reactions of a-silyl-a-formyl carbanions with carbonyl compounds containing a nucleophilic group gave rise to SBn OH O O Scheme 9 Me +N Me 7Br7 Me +N Me Me N 216 E V Gorobets,MS Miftakhov, F A Valeev Scheme 10 Z Z ButMe2SiO (CH2)3 Z CsF, BnEt3N+Cl7 (CH2)3 DMF, 90 8C, 64 h Z 217 ZZ Z Z Z Z O 220 (10%) Z Z ButMe2SiO CsF, BnEt3N+Cl7 DMF, 90 8C, 64 h Z 218 Z Z Z Z Z Z Z O 221 (57%) Z ButMe2SiO Z CsF, BnEt3N+Cl7 Z Z O DMF, 90 8C, 5 days Z 219 Z 222 (14%) Z=CO2Me.cyclic compounds with a functionalised side chain. Tandem reactions of a-trimethylsilylalkylphosphine oxide (223) or a-tri- methylsilylalkylphosphine sulfide (224) with n-butyllithium and then with lithium tetrahydropyran-2-olate gave rise to products 225 and 226, respectively, in high yields. 1) BunLi 2) X O O R1 X P OLi R2 P R1 CHR3SiMe3 R3 R2 THF, 778 to 20 8C 223, 224 225, 226 X = O (223, 225), S (224, 226); R1, R2, R3=H, Ar. * * * To summarise, tandem transformations, which represent an advantageous combination of the thermodynamic and kinetic parameters of a single process, play an important role in proce- dures for the preparation of complex organic compounds.There is no doubt that the development of one-pot tandem reactions based on compounds containing various functional groups resulting in substantial complication of the starting structures is an important problem of modern organic synthesis. References 1. T-L Ho Tandem Organic Reaction (New York: Wiley, 1992) 2. G H Posner Chem. Rev. 86 831 (1986) 3. L Tietze, U Beifuss Angew. Chem., Int. Ed. Engl. 32 131 (1993) 4. M Ihara, K Fukumoto Angew. Chem., Int. Ed. Engl. 32 1010 (1993) 5. R A Bunce Tetrahedron 51 13103 (1995)Tandem transformations initiated and determined by the Michael reaction 6. L F Tietze Chem. Rev. 96 115 (1996) 7. M Ihara, K Makita, Y Fujiwara, Y Tokunaga, K Fukumoto J. Org. Chem. 61 6416 (1996) 8.K Takasu, S Mizutani,M Noguchi, K Makita,M Ihara Org. Lett. 1 391 (1999) 9. M Ihara, K Makita, K Takasu J. Org. Chem. 64 1259 (1999) 10. K Makita, K Fukumoto,M Ihara Tetrahedron Lett. 38 5197 (1997) 11. B B Snider, T Liu J. Org. Chem. 62 5630 (1997) 12. G A Molander, M Ronn J. Org. Chem. 64 5183 (1999) 13. J K F Geirsson, J F Johannesdottir J. Org. Chem. 61 7320 (1996) 14. H Hagiwara, Y Yamada, H Sakai, T Suzuki,M Ando Tetrahedron 54 10 999 (1998) 15. B-C Hong, J-H Hong Tetrahedron Lett. 38 255 (1997) 16. R B Grossman, M A Varner, A J Skaggs J. Org. Chem. 64 340 (1999) 17. R B Grossman, R M Rasne, B O Patrick J. Org. Chem. 64 7173 (1999) 18. R B Grossman, D S Pendharkar, B O Patrick J. Org. Chem. 64 7178 (1999) 19. R B Grossman, A J Skaggs, A E Kray, B O Patrick Org.Lett. 1 1583 (1999) 20. J-K Erguden, H W Moore Org. Lett. 1 375 (1999) 21. L Hadjiarapoglou, I Klein, D Spitzner, A de Meijere Synthesis 525 (1996) 22. D Barrett, H Sasaki, T Kinoshita, A Fujikawa, K Sakane Tetrahedron 52 8471 (1996) 23. T Yakura, T Tsuda, Y Matsumura, S Yamada,M Ikeda Synlett 985 (1996) 24. E Dumez, J Rodriguez, J-P Dulcere J. Chem. Soc., Chem. Commun. 1831 (1997) 25. G U Gunawardena, A M Arif, F G West J. Am. Chem. Soc. 119 2066 (1997) 26. Y Nagaoka, K Tomioka Org. Lett. 1 1467 (1999) 27. M Konno, T Nakae, S Sakuyama, K Imaki, H Nakai, N Hamanaka Synlett 1472 (1997) 28. J G Urones, N M Garrido, D Diez, S H Dominguez, S G Davies Tetrahedron Asymmetry 10 1637 (1999) 29.S Torii, N Hayashi,M Kuroboshi Synlett 599 (1998) 30. T Yechezkel, E Ghera, N G Ramesh, A Hassner Tetrahedron Asymmetry 7 2423 (1996) 31. A C Forsyth, R O Gould, R M Paton, I H Sadler, I Watt J. Chem. Soc., Perkin Trans. 1 2737 (1993) 32. A C Forsyth, R M Paton, I Watt Tetrahedron Lett. 30 993 (1989) 33. F Shafizadeh, D D Ward, D Pang Carbohydr. Res. 102 217 (1982) 34. A V Samet, M E Niyazymbetov, V V Semenov, A L Laikhter, D H Evans J. Org. Chem. 61 8786 (1996) 35. F A Valeev, E V Gorobets, M S Miftakhov Izv. Akad. Nauk, Ser. Khim. 152 (1999) a 36. E V Gorobets, Candidate Thesis in Chemical Sciences, Institute of Organic Chemistry, Ufa Scientific Centre, Russian Academy of Sciences, Ufa, 2000 37. J-P Dulcere, E Dumez J. Chem. Soc., Chem.Commun. 971 (1997) 38. S Arai, K Nakayama, K Hatano, T Shioiri J. Org. Chem. 63 9572 (1998) 39. F A Valeev, E V Gorobets, M S Miftakhov Izv. Acad. Nauk, Ser. Khim. 1242 (1997) a 40. K Krohn, C Freund, U FloÈ rke Eur. J. Org. Chem. 2713 (1998) 41. D Enders, J Wiedemann Liebigs Ann. Chem. 699 (1997) 42. A Srikrishna, T J Reddy, P P Kumar Synlett 663 (1997) 43. S Cao, J-H Xu, Z Zhang, A-L Fan J. Heterocycl. Chem. 35 477 (1998) 44. K Aoyagi,H Nakamura, Y Yamamoto J. Org. Chem. 64 4148 (1999) 45. M Ono,K Nishimura, Y Nagaoka,K Tomioka Tetrahedron Lett. 40 6979 (1999) 46. A Covarrubias-Zuniga, E Rios-Barrios J. Org. Chem. 62 5688 (1997) 47. S G Davies,D R Fenwick J. Chem. Soc., Chem. Commun. 565 (1997) 48. T Kawasaki, C-Y Tang, H Nakanishi, S Hirai, T Ohshita, M Tanizawa,M Himori, H Satoh,M Sakamoto, K Miura, F Nakano J.Chem. Soc., Perkin Trans 1 327 (1999) 49. F Dinon, E Richards, P J Murphy, D E Hibbs, M B Hursthouse, K M Abdul Malik Tetrahedron Lett. 40 3279 (1999) 50. P Puebla, Z Honores,M Medarde, L Moran, E Caballero, A San Feliciano Tetrahedron 55 7915 (1999) 1019 51. A Kamimura, H Mitsudera, S Asano, S Kidera, A Kakehi J. Org. Chem. 64 6353 (1999) 52. K Yamada, T Arai, H Sasai,M Shibasaki J. Org. Chem. 63 3666 (1998) 53. R Badorrey, C Cativiela, M D Diaz-de-Villegas, J A Galvez Tetrahedron 55 7601 (1999) 54. H Kunz, M Weymann,M Follmann, P Allef, K Oertel, M Schultz-Kukula, A Hofmeister Pol. J. Chem. 73 15 (1999) 55. S Kirschbaum, H Waldmann Tetrahedron Lett. 38 2829 (1997) 56. S Lim, I Jabin, G Revial Tetrahedron Lett. 40 4177 (1999) 57. M Node, T Fujiwara, S Ichihashi, K Nishide Tetrahedron Lett. 39 6331 (1998) 58. X Marat, N Monteiro, G Balme Synlett 845 (1997) 59. B Clique, N Monteiro, G Balme Tetrahedron Lett. 40 1301 (1999) 60. A Reyes, I Regla, M C Fragoso, L A Vallejo, P Demare, H A Jimenez-Vazquez, Y Ramirez, E Juaristi, J Tamariz Tetrahedron 55 11187 (1999) 61. A V Samet, V P Kislyi, N B Chernysheva, D N Reznikov, B I Ugrak, V V Semenov Izv. Akad. Nauk, Ser. Khim. 409 (1996) a 62. A V Samet, A N Yamskov, B I Ugrak, L G Vorontsova, M G Kurella, V V Semenov Izv. Akad. Nauk, Ser. Khim. 415 (1996) a 63. A V Samet, A N Yamskov, B I Ugrak, V V Semenov Izv. Akad. Nauk, Ser. Khim. 553 (1997) a 64. K C Majumdar, U Das J. Org. Chem. 63 9997 (1998) 65. K Takeda, T Tanaka Synlett 705 (1999) 66. M-H Filippini, J Rodriguez,M Santelli J. Chem. Soc., Chem. Commun. 1647 (1993) 67. M N Filippini, J Rodriguez J. Org. Chem. 62 3034 (1997) 68. A Covarrubias-Zuniga, L A Maldonado, E Rios-Barrios, A Gonzalez-Lucas Synth. Commun. 28 3461 (1998) 69. A Covarrubias-Zuniga,A Gonzalez-Lucas Tetrahedron Lett. 39 2881 (1998) 70. S Maiti, S Bhaduri, B Achari, A K Banerjee, N P Nayak, A K Mukherjee Tetrahedron Lett. 37 8061 (1996) 71. H Paulsen, S Antons, A Brandes, M LoÈ gers, S N MuÈ ller, P Naab, C Schmeck, S Schneider, J Stoltefuss Angew. Chem., Int. Ed. Engl. 38 3373 (1999) 72. S Marczak, K Michalak, Z Urbanczyk-Lipkowska, J Wicha J. Org. Chem. 63 2218 (1998) 73. N Giuseppone, Y Courtaux, I Collin Tetrahedron Lett. 39 7845 (1998) 74. R A Bunce, C L Schilling III J. Org. Chem. 60 2748 (1995) 75. R A Bunce, C L Schilling III Tetrahedron 53 9477 (1997) 76. A V Samet,N B Chernysheva,A M Shestopalov, V V Semenov Izv. Akad. Nauk, Ser. Khim. 211 (1999) a 77. T Fujimoto, Y Kodama, I Yamamoto, A Kakehi J. Org. Chem. 62 6627 (1997) 78. N Kishimoto, T Fujimoto, I Yamamoto J. Org. Chem. 64 5988 (1999) 79. D Goff Tetrahedron Lett. 39 1477 (1998) 80. H Akiyama, T Fujimoto, K Ohshima, K Hoshino, I Yamamoto, R Iriye Org. Lett. 1 427 (1999) 81. K Nishide, Y Shigeta, K Obata,M Node J. Am. Chem. Soc. 118 13103 (1996) 82. V N Desai, N N Saha, D D Dhavale J. Chem. Soc., Chem. Commun. 1719 (1999) 83. H W Lam, P A Cooke, G Pattenden,W M Bandaranayake, W A Wickramasinghe J. Chem. Soc., Perkin Trans. 1 847 (1999) 84. D R W Hodgson, A J Kirby, N Feeder J. Chem. Soc., Perkin Trans. 1 949 (1999) 85. A Ullmann, H-U Reissig, O Rademacher Eur. J. Org. Chem. 2541 (1998) 86. T Kawashima,M Nakamura,N Inamoto Heterocycles 44 487 (1997) a�Russ. Chem. Bull. (Engl. Transl

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

Russian Chemical Reviews 69 (12) 1021 ± 1036 (2000) Methods of synthesis of conjugated o-amino ketones Yu V Smirnova, Zh A Krasnaya Contents I. Introduction II. Methods of synthesis of b-aminovinyl ketones, d-amino dienones and a,a0-bis(3-aminopropenylidene)alkanones III. Conjugated o-amino polyenones IV. Conjugated o-amino ketones containing a heterocyclic fragment Abstract. synthesis the on data published The The published data on the synthesis of of b-enamino ketones, polyenones conjugated and dienones d-amino -amino dienones and conjugated o-amino -amino polyenones of bibliography The generalised. are structures diverse of diverse structures are generalised. The bibliography includes includes 129 references 129 references. I. Introduction Conjugated o-amino ketones represent an ample class of organic compounds which attract the attention of a broad circle of organic chemists, biochemists, physicochemists, etc.Owing to their high reactivities, these compounds find application in the synthesis of various biologically active heterocyclic systems, natural com- pounds, dyes and photosensitive materials. The presence in o-amino ketones of an electron-donating and an electron-withdrawing group separated by conjugated double bonds imparts to them specific properties and makes them perfect models for the study of fundamental problems, such as the nature of chemical bonds, electron transfer along the conjugation chain, chromaticity, sensitivity to the action of different kinds of energy, tautomerism, cis ± trans-isomerism, etc.The methods employed in the synthesis of b-aminovinyl ketones (AVK) have been generalised in reviews 1, 2 and the monograph by Freimanis.3 Several novel versions of conventional reactions which make it possible to increase the yields of the target products and to simplify the synthetic procedures have been proposed recently. The approaches to regio- and stereospecific synthesis of enamino ketones containing additional functional groups are being developed intensively. The most practicable methods for the synthesis of AVK and those based on unusual chemical conversions are considered in this review. Special atten- tion is given to conjugated o-amino polyenones prepared with the help of procedures developed in recent years.These compounds include a,a0-bis(o-aminopolyenyl)alkanones, which contain two polymethine chains linked through a carbonyl group and manifest unusual spectral and luminescent properties. Yu V Smirnova, Zh A Krasnaya N D Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Leninsky prosp. 47, 119992 Moscow, Russian Federation. Fax (7-095) 135 53 28. Tel. (7-095) 135 89 61. E-mail: yusmir@cacr.ioc.ac.ru (Yu V Smirnova), kra@cacr.ioc.ac.ru (Zh A Krasnaya) Received 26 September 2000 Uspekhi Khimii 69 (12) 1111 ± 1127 (2000); translated by R L Birnova #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n12ABEH000625 1021 1021 1032 1034 II. Methods of synthesis of b-aminovinyl ketones, d-amino dienones and a,a0-bis(3-aminopropenylidene)alkanones 1.Reactions of b-diketones with amines Condensation of readily accessible b-diketones with ammonia and amines is a traditional approach to the synthesis of b-aminovinyl ketones.1± 8 The starting diketone reacts in both diketo- and keto- enol forms. This reaction is usually carried out by boiling b-diketone ± amine mixtures in solvents (e.g., benzene, toluene) which allow removal of the water formed.4±7 In the reaction of non-symmetrical b-diketones 1 (R1=R3), one or both regioiso- mericAVK2, b-aminovinylimine 3 or their mixture can be formed due to the non-equivalence of the electrophilic centres C(1) and C(2).1, 3, 6 ± 8 R2 R2 R2 R3 R1 R3 R1 R3 R1 N N N O O O 3 2 H1 The difference in the reactivities of the electrophilic centres C(1) and C(3) is especially pronounced in the case of non- symmetrical polyfluorinated b-diketones 4 due to the difference in the inductive effects of the substituents RF and R1 (Refs 6 ± 8).Polyfluorinated b-diketones 4 react with ammonia and amines at only one of the electrophilic centres to give fluorine-containing AVK 5.6± 8 R1 RF R1 R1 RF RF R2R3NH + O O O R2R3N O NR2R3 6 5 H4 RF=CF3, H(CF2)2, H(CF2)4; R1=Me, Bun, But, Ph, C6H4Cl, C6H4Br, C6H4NO2, C5H4Me, C6H4OMe; R2R3N=NH2, NHPh, NH , O, N NHMe, N . It is of note that the condensation of arylfluoroalkyl diketones 4 with NH3, NH2Me and NH2Ph affords a mixture of regioiso- meric AVK 5 and 6; the proportion of the isomer 6 with a geminal amino group and the fluoroalkyl residue depends on the length of the fluoroalkyl substituent and the nature of substituents in the amine.7, 8 This reaction is carried out with azeotropic distillation of the water eliminated.In the case of low-boiling amines, this reaction is difficult to perform, since this requires repeated1022 bubbling of the amine into the reaction mixture for the reaction to complete.4, 6, 7, 9 This reaction is also accompanied by scission of the b-diketones at the C(1)7C(2) bond and by secondary con- densations.1, 8 Therefore, modifications of the process 10 ± 14 which exclude the formation of by-products, increase the selectivity and simplify the procedure are of special interest.Thus, a method for the synthesis of AVK 2 was proposed which includes boiling 1,3-diketones with ammonium acetate or acetates of low-boiling amines in benzene in the presence of AcOH; the water formed is removed by azeotropic distillation.10 O O NHR4 R3 R1 R4NH3OAc, PhH, 80 8C R2 R1 R3 OH R2 1 2 (70% ± 98%) R1=Me, Ph; R2=Me; R3=H, CH2CH=CH2; R27R3=(CH2)3, CH2CH(Me)CH2; R4=H, Et. Satisfactory results were obtained in the condensation of methyl diketones with primary amines on inorganic supports.11, 12 Under heterogenous conditions, the reactions proceed with higher selectivities and under milder conditions than the corresponding homogenous reactions. O O HNR2 O a or b +NH2R2 R1 Me R1 Me 2 (77% ± 99%) 1 R1=Me, Ph; R2=H, Me, Et, But, Pri, Ph, Bn, CH2CN; (a) Al2O3; (b) montmorillonite K10.Neutral alumina or montmorillonite K10 can be used as solid supports. The reagents must be completely adsorbed on the support. This procedure is extremely simple as regards its imple- mentation. To isolate enamino ketones 2, the reaction mixture is diluted with CH2Cl2, the solid phase is filtered off, the solvent is removed in vacuo and the residue is subjected to distillation. In reactions with acetyl acetone (1, R1=Me), primary amines can be used as aqueous solutions, which is especially important in the case of low-boiling amines.12 If heterogenous reactions are carried out in a microwave (MW) oven,13, 14 the reaction time decreases from 24 h to 1 ± 20 min.Both primary and secondary amines can be introduced into the reaction.13 O O R3NR4 O a or b +HNR3R4 R1 R2 R1 R2 2 (87% ± 99%) 1 R1=R2=Ph, R37R4=(CH2)4, (CH2)5; R1=Me, R2=Ph, OMe; R37R4=(CH2)4; (a) SiO2 , MW, 1 ± 20 min; (b) montmorillonite K10, MW, 1 ± 20 min. It was shown 13 that primary amines yield acyclic AVK 2 with Z-configuration of the a,b-double bond. The use of symmetrical disubstituted ureas 7 14 instead of amines also results in AVK 2 having a Z-configuration of the double bond and stabilised by an intramolecular hydrogen bond. O O O a or b + NHR3 R3HN CH2R2 R1CH2 7 R3NH O R1CH2 CH2R2 2 (87% ± 99%) R1=R2 =Me, Ph; R3=H, Me, Et, Ph; (a) SiO2 , MW, 1 ± 20 min; (b) montmorillonite K10, MW, 1 ± 20 min.Yu V Smirnova, Zh A Krasnaya 2. Reactions of unsaturated alkoxy ketones with amines The alkoxy group in readily accessible b-alkoxyvinyl ketones 815 ± 18 is exchanged for an amino group upon reactions with ammonia and primary and secondary amines.3, 18 ± 22 This reac- tion is widely employed in the synthesis of halogen-containing AVK 9.18 ± 22 As a rule, the reaction of amines with b-alkoxyvinyl ketones proceeds at room temperature or with gentle heating and results in high yields of AVK 9 even in the case of sterically hindered amines.15, 19 ± 22 O O R1R2NH, THF EtO CCl3 CCl3 R1R2N9 (84% ± 95%) 8 , R1R2N= PhCH(OH)CHMe , HNCH(Me)Ph, CON CO NH PhCH(OH)CHPh, MeNCH(Me)Ph. NMe The reactions of alkoxy-substituted 2H- and 4H-pyrones with amines also belong to this type of transformations.3, 23 ± 25 Thus prolonged boiling of 6-methyl-4-methoxy-2-pyrone 10 with amines in the absence of solvents resulted in the formation of AVK2.24 The details of these experiments have not been reported.OMe OMe O R1R2N R1R2NH OH Me O Me O 10 O R1R2N Me Me 2 R1=H;R2=C7H15; R17R2=(CH2)4. It is believed 24 that AVK 2 are formed as a result of an attack by the amine at position 6 of the pyrone ring with subsequent decarboxylation and hydrolysis of the diacetoacetic acid deriva- tive.Treatment of dihydropyrones 11 with aqueous or methanolic ammonia gives amino dienones 12.25 O O NH2 Me 20 8C, 2 days Me +NH3 R R Me Me 12 (32% ± 46%) O 11 R=CF3, CCl3. d-Alkoxy hexadienone 13 obtained by the acylation of 1-ethoxybuta-1,3-diene (14) with trifluoroacetic anhydride reacts with amines much like b-alkoxyvinyl ketones.26 The alkoxy group in the dienone 13 is readily exchanged for an amino group to give d-amino dienones 15.O R1R2NH (CF3CO)2O, Py EtO EtO CF3 14 13 (38%) O R1R2N CF3 15 (78% ± 98%) R1=R2=H, Et; R1=Me, R2=Ph. d-Amino dienones 15 were also prepared 27 from 3,4-dihydro- 2H-pyran derivative 16. Compound 16 reacts with various ali- phatic and aromatic amines to give substituted diketones 17. Presumably,27 the putative mechanism of this reaction involvesMethods of synthesis of conjugated o-amino ketones ethoxy-substituted dienedione 18 as an intermediate. Acid hydro- lysis of diketones 17 with aliphatic substituents at the nitrogen atom results in d-amino dienones 15.Diketones 17 with aromatic substituents at the nitrogen atom do not undergo deacetylation under conditions of acid hydrolysis. O BuiO CF3 7BuiOH EtO CF3 O 16 O O EtO CF3 CF3 R1R2NH 18 EtO O O CF3 CF3 O R1R2N CF3 6 M HCl, THF CF3 O 17 (77% ± 100%) O R1R2N CF3 15 (58% ± 78%) R1=R2=Me, Et, Bn; R1=Me, R2=Ph; R17R2=(CH2)4, (CH2)5. Non-fluorinated derivatives of 3,4-dihydro-2H-pyran do not react with amines. An attempt to apply this approach to the synthesis of 6-dimethylaminohexa-3,5-dien-2-one (19) was unsuc- cessful.27BuiO O O Me Me2NH, MeCN Me Me2N 19 Me EtO O 3. Reductive cleavage of isoxazoles Reductive cleavage of the isoxazole ring under conditions of catalytic hydrogenation is one of the most well-studied and widely employed procedures for the preparation of enamino ketones.28 ± 31 Diaminodivinyl diketones 20 were obtained by the reduction of the diisoxazolylmethane derivatives 21 in the pres- ence of Raney nickel (Ni/Ra).R2 R1 O O NH2 H2N H2, Ni/Ra N R2 R1 O O N 20 21 R1=R2=Me, Ph; R1=Ph, R2=Me. Catalytic hydrogenation of isoxazoles affords enamino ketones selectively, other functional groups of the starting com- pounds are not affected.31 O O (CH2)nCO2Me (CH2)nCO2Me H2, Pd/C or Ni/Ra Me Me N R COR H2N O n=3, 5; R = C5H11, CH2OC6H4F-2. The enamino ketones 2 were obtained in good yields by reductive cleavage of 3,4,5-substituted isoxazolium salts 22 with lithium dialkyl cuprates generated in situ from methyl- or butyl- lithium and copper iodide.32 This reaction proceeds very quickly 1023 and under very mild conditions (720 8C, 15 min) and affords enamino ketones 2 with different (including functional) substitu- ents in the a-position.32 HNR4 O R3 R2 Me2CuLi or Bu2CuLi + R3 R1 N O R1 R4 22 BF¡ R2 2 (66% ± 90%) 4 R1=R3=R4=Me, Ph; R2=H, Br, Cl, COPh, COMe, CO2Et, CN, NO2.More readily accessible 5-substituted isoxazolium salts 23 do not react with lithium cuprates. In this case, preliminary treatment with triethylamine resulting in oxo ketene imines 24 is required. The reactions of the latter with organocuprates 25 afford AVK 2, simultaneously a substituent is introduced into the b-position of the AVK formed.33 R32 CuLi (25) + Et3N CH2Cl2 N R1C(O)CH C NR2 24 O R1 R2 X7 23 O HNR2 R1 R3 2 (51% ± 93%) X=ClO4, BF4; R1=Me, Ph; R2=But, Et; R3=Me, Bun, But, Ph.The reaction of oxo ketene imine 24 (R1=Me,R2=But) with lithium diallyl cuprate (26) leads to the isomerisation of the allyl group into the propenyl group to yield a mixture of enamino ketones 27 and 28 containing conjugated and isolated double bonds. CHCH2)2CuLi (26) (CH2 MeC(O)CH C NBut 24 O HNBut O HNBut + Me Me Me 28 (31%) 27 (40%) 4. Condensation of acid nitriles with methyl ketones Until recently, information concerning the synthesis of AVK by condensation of nitriles with ketones was scarce.3, 34 A method for the synthesis of enamino ketones containing a primary amino group from nitriles and methyl ketones has been developed in the last few years where (phenyl)ethylaminomagnesium bromide is used as a condensing reagent. This procedure affords enamino ketones of various structures, which are difficult to prepare by other methods.35, 36 The yields of enamino ketones depend on the nature of substituents in the starting compounds.The best results were obtained in the reaction of trichloroacetonitrile with aryl(hetaryl) methyl ketones and tert-alkyl methyl ketones with acid nitriles containing electron-withdrawing aromatic substituents.36 O O H2N Ph(Et)NMgBr, 20 8C R1CN+Me R1 R2 R2 2 (26% ± 73%) Yield (%) R2 Yield (%) R1 R2 R1 PhOCH2 CCl3 Ph 4-MeOC6H4 4-BrC6H4 But But But 48 26 80 48 58 34 73 But But Pri Ph 60 But N 71 S But Bn 541024The presence of a double bond 37 or a functional group 38 ± 46 in the molecule of a methyl ketone does not prevent the reaction with nitriles, the cyano group of which is activated by a polyhaloge- noalkyl substituent, which makes it possible to synthesise enam- ino ketones containing several functional groups. Thus the reaction of trichloro- and trifluoroacetonitriles with a,b-unsaturated ketones 29a ± d results in b-aminodivinyl ketones 30a ± d.37 O R2 O R2 H2N Ph(Et)NMgBr R1CN+Me R1 R3 R1 30a ± d (27% ± 58%) 29a ± d R1=CCl3, CF3; R2=R3=Me (a); R2=H: R3=Ph (b), (d).4-MeOC6H4 (c), O Ph(Et)NMgBr The condensation of b-acylaminoalkyl methyl ketones 31 with halogenoacetonitriles gives AVK 32.38 This is the first example of the involvement of b-acylaminoalkyl methyl ketones as a methyl- ene component in aldol-type reactions. Me R1CN+ MeCCH2CNHCOR2 O Me 31 O H2N Me R1 CH2CNHCOR2 32 (19% ± 25%) Me R1=CCl3, CF3; R2=Me, Ph. The reaction of trichloro- and trifluoroacetonitrile with the so-called diacetone alcohol 33 proceeds as nucleophilic addition to the nitrile group and yields b-aminovinyl b0-hydroxyalkyl ketones 34; the latter are stable compounds, which occur exclu- sively in the acyclic form.39 OH O NH2 Ph(Et)NMgBr RCN+ Me2CCH2CMe Me2C R 34 33 OH O R=CCl3, CF3.The condensation of trichloro- or trifluoroacetonitrile with 2-hydroxyacetophenone 39 and 2-acetyl-1-naphthol 40 gives the corresponding hydroxy enamino ketones 35 and 36, which occur as Z-isomers with a coplanar s-cis-configuration stabilised by intramolecular hydrogen bonds. H H H H NH O O NH O O R R 36 35 R=CCl3, CF3. On going to sterically hindered b-hydroxyalkyl methyl ketones, the structures of the reaction products depend on the nature of the halogen in the halogenoacetonitrile.41, 42 Thus the reaction of trichloroacetonitrile with 4-hydroxy-3,3-dimethylbu- tan-2-one (37) is not stopped in the formation of an open form of Ph(Et)NMgBr HOCH2CMe2CMe+CCl3CN 37 O O O NH2 Me HOCH2CMe2C Me NH2 CCl3 38 CCl3 O39 (45%) Yu V Smirnova, Zh A Krasnaya the hydroxy enamino ketone 38; its spontaneous cyclisation yields tetrahydropyranone 39.41 The condensation of trichloroacetonitrile with 1-acetyl-1- hydroxymethylcyclohexane 41 and 1-acetyl-2-naphthol 40 pro- ceeds similarly to give the cyclic products 40 and 41.Acyclic forms were absent in both cases. O O NH2 NH2 O O CCl3 CCl3 41 40 The reactions of polyfluorocarboxylic acid nitriles with steri- cally hindered b-hydroxyalkyl ketones 42 under the same con- ditions afford hydroxy enamino ketones 43 as the final products; the latter are stable on storage and do not cyclise spontaneously into tetrahydropyrones.42 R1 Ph(Et)NMgBr HOCH2CCOMe +RFCN 42 R2 O O NH2 R1 HOCH2 R2 NH2 RF R1 R2 RF O 43 (55% ± 63%) R1=R2=Me, R1±R2=(CH2)5; RF=CF3, (CF2)2H. The reaction of trichloro- and trifluoroacetonitrile with acy- clic hydroxyalkyl ketones 44 in the presence of (phenyl)ethylami- nomagnesium bromide results in the corresponding hydroxy enamino ketones 45.43, 44 O NH2 O Ph(Et)NMgBr Me Me +R2CN R2 Me R1 R1 HO 44 HO 45 (38% ± 93%) R1=Me, Et, But; R2=CCl3, CF3.As in the case of sterically hindered b-hydroxyalkyl methyl ketones, the structures of the products formed in the reactions of nitriles with sterically hindered a-hydroxyalkyl methyl ketones depend on the nature of the halogen which activates the cyano group.45, 46 Thus the condensation of 1-acetylcyclohexan-1-ol (46) with trichloroacetonitrile does not afford a-hydroxy enamino ketone 47 as might be expected, but rather gives its cyclic isomer 48.45 Ph(Et)NMgBr +CCl3CN Me HO46 O O CCl3 CCl3 HO O NH2 O NH2 47 48 (60%) The condensation of a-hydroxy ketone 46 with trifluoroace- tonitrile results in a-hydroxy enamino ketone 49 with a geminal amino group and the trifluoromethyl substituent.46 CF3 +CF3CN Ph(Et)NMgBr Me HO HO NH2 46 O O 49 (36%)Methods of synthesis of conjugated o-amino ketones 5.Modification of b-aminovinyl ketones based on the use of their lithium salts Enamino ketones containing other functional groups in addition to the aminoenone group cannot be prepared in the majority of cases using traditional methods.2, 3, 19, 21, 31 ± 33 Therefore, proce- dures based on directed introduction of functional groups into more readily accessible AVK acquire special importance.The use of N,a 0- and N,g-dilithium salts of AVK in reactions with electro- philes is an example of a successful application of synthetic approaches to regio-controlled introduction of functional groups into AVK. Studies of reactivities of N,a 0- and N,g-dilithium derivatives with respect to various electrophiles have shown that these are convenient starting compounds for the synthesis of heterocyclic compounds and for the introduction of functional groups into the a 0- and g-positions of AVK.47 ± 58 The alkylation of cyclic AVK50 is the first example of such an approach.47 It was shown that depending on conditions of preparation of lithium salts (viz., the reagent ratios, the nature of a base, temperature), the alkylation predominantly occurs either at position 6 or at position 4 to give the AVK derivatives 51 or 52, respectively.O R2 1) LDA 2) R2X O N(Me)R1 51 (70% ± 78%) O N(Me)R1 50 1) LTMSA 2) R2X N(Me)R1 R2 52 (80% ± 83%) R1=Me, C6H11; R2=Bn, Me; X=I, Br; LDA is lithium N,N-diiso- propylamide; LTMSA is lithium bis(trimethylsilyl)amide. Treatment of the enamino ketone 50 with lithium N,N-diiso- propylamide (LDA) in ether or THF at 778 8C results in deprotonation of the a 0-position to afford a kinetically more favourable dianion (termed an a 0-dianion); its alkylation gives rise to the cyclic AVK 51. Deprotonation of the g-position which gives a thermodynamically more favourable dianion (termed a g-dianion) occurs upon treatment of compound 50 with lithium bis(trimethylsilyl)amide (LTMSA). Subsequent alkylation results in a practically exclusive formation of AVK 52.47 However, this approach could not be used for the alkylation of acyclic AVK, since it gave a mixture of alkylation products irrespective of the base used.47 The conditions favouring the preparation of a 0- or g-dilithium derivatives from acyclic b-(N-monoalkylamino)vinyl ketones and the factors influencing the regioselectivity of their formation have been studied in detail.48, 49 Treatment of 4-(N-alkylamino)pent-3-enones 2a ± e with an excess of a strong base (2.5 equiv., THF) gives both a 0- (53) and g-dilithium derivatives 54 which exist in a dynamic equilibrium (see Ref.48).{ Li+ NR O HNR Li+ Li+ O Li+NR O B7 7 7 7 7 Me Me Me Me CH2 CH2 54a ± e 2a ± e 53a ± e R=Me (a), Pri (b), cyclo-C6H11 (c), PhCHMe (d), But (e). Depending on the strength of the base and the substituentRin the starting compound (AVK 2), the equilibrium can be practi- { Hereinafter, for convenience sake we shall refer to the dilithium derivatives 53 and 54 as dianions and omit Li cations in the structural formulae. 1025 cally completely shifted towards one of the dianions.48 ± 51 The a 0-dianion 53 which is predominantly formed upon treatment of AVK 2 with lithium 2,2,6,6-tetramethylpiperidide (LTMP) is kinetically more favourable. Thermodynamic calculations dem- onstrate that under these conditions the conformation of the a 0-dianion 53 prevents its isomerisation into a thermodynamically more stable g-dianion 54.48 It should be noted that the regiospe- cificity of formation of the a 0-dianion 53 depends on the size of the substituent at the nitrogen atom.The predominant formation of the a 0-dianion 53 is observed only in the case of bulky substitu- ents, such as But and Ph.49, 51 In the presence of less bulky substituents (R=Me, Pri) the g-dianion 54 is formed together with the a 0-dianion 53.51 The predominant formation of a thermodynamically more favourable g-dianion 54 takes place under the action of a stronger base, viz., methyllithium, in the presence of N,N,N 0,N 0-tetramethylethylenediamine (TMEDA).Under these conditions, the a 0-dianion which is formed first adopts a conformation which allows its rapid isomerisation into a more stable g-dianion 54.48, 51 The size of a substituent at the nitrogen atom does not influence the regiospecificity of formation of the g-dianion 54.48 The dianions 53 and 54 can also be prepared from enamino ketones with secondary amino groups.48 ± 50, 54, 58 The reactions of the dianions 53 and 54 with electrophiles proceed at low temperatures (7100 to720 8C) and take as a rule from 5 min to 2 h. The alkylation of dianions of the type 53 and 54 afforded the corresponding a 0- and g-alkyl-substituted AVK in good yields. It is noteworthy that this reaction proceeds as C-alkylation and does not affect the primary amino group.48, 52 NBut NHBut O O RX 7 7 LTMP THF, 20 8C Me Me Me CH2 53e NHBut O RCH2 Me (84% ± 85%) R=Me, Bn; X=Br, I.NR2 O NHR2 O R3X 7 7 MeLi, TMEDA THF, 20 8C R1 Me R1 CH2 54 NHR2 O R1 CH2R3 (56% ± 97%) R1=Me, Ph, 4-MeC6H4, 4-MeOC6H4; R2=Me, Pri, But; R3=Me, Et, Pri, Bn, n-C10H23, CH2CH=CH2; CH2C:CH; X=Cl, Br, I. In addition to alkyl halides, allyl and propargyl bromides can be used as alkylating agents.52 a 0-Dianions of the type 53 react smoothly with nitriles under mild conditions to give bis(aminovinyl) ketones 55. During chromatography on SiO2, compounds 55 are partly converted into 4-pyridones 56. This reaction is quantitative when solutions of enamino ketones are kept over SiO2 for 2 days.50 NR1 O O HNR1 LTMP 7 7 R2 R3CN, THF 750 8C, 15 min Me CH2 CH2R2 53 O O HNR1 H2N SiO2 R2 R3 R2 R3 56 55 (61%± 88%) NH R1 =Pri, Ph; R2=Et, Bn, H; R3 =Ph, Bun, 2-ClC6H4.1026According to 1H and 13C NMR and X-ray diffraction data, compounds 55 exist in both solutions and crystalline state as tautomers A containing two enamine fragments, but not as oxo- (B) or hydroxy imines (C).50 NH O HNR1 O HNR1 NH2 R2 R2 R3 R3 B A HO HNR1 HN R2 R3 C Hydroxy derivatives of enamino ketones were obtained in the reaction of a 0- (53) or g-dianions (54) with oxiranes,49 aldehydes, ketones 51, 53 and esters.54 The reaction of dianions of the type 53 and 54 with oxiranes 57 is accompanied by the opening of the epoxide ring to give the corresponding g 0- or e-hydroxy derivatives 58 and 59.This reaction proceeds as bimolecular SN2 nucleophilic substitution; non-symmetrical epoxides react at the least substituted carbon atom. The opening of the epoxide ring occurs in a stereo- and regioselective manner resulting in only one of the possible iso- mers.49 O HNR1 R2 O O NR1 g0 HO 7 7 a0 + b0 Me Me R3 R2 CH2 R3 57 58 (71% ± 83%) 53e,f R1=But (53e), Ph (53f); R2=H: R3=H, Me, Et, Bu, Ph; R27R3=(CH2)4. O O NR1 O HNR1 R2 b d OH g 7 7 a e + Ph R3 Ph R2 CH2 57 54 R3 59 (74% ± 98%) R1=Me, Bu; R2=H, Ph, Me3Si; R3=H, Me, Et, CH2Cl; R27R3=(CH2)4. The condensation of the a 0-dianion 53f with aldehydes and ketones was used to obtain b-aminovinyl b 0-hydroxy alkyl ketones 60.51 Like aminovinyl b 0-hydroxyalkyl ketones 34, these are stable in the acyclic form.NPh O HNPh O OH O 7 7 + R1 R2 b0 Me Me R1 R2 CH2 a0 60 (58% ± 97%) 53f R1=H, Me, Ph; R2=Me, Pr, But, n-C6H13, Ph, (CH2)2Ph. Compounds 60 are produced in high yields even if sterically hindered aldehydes and ketones (trimethylacetaldehyde or benzo- phenone) and enolisable ketones (e.g., acetophenone) are intro- duced into the reaction. In contrast to the a 0-dianion 53f, the condensation of the g-dianion 54a with aldehydes and ketones is non-regiospecific and results in the formation of both d- (60) and b 0-hydroxyenamino ketones (61).51 O CH2 Me Me CH2 g 7 7 + a 7 7 NMe NMe O O R2 R1 b 53a 54a O OH O HNMe R1 HNMe OHR1 + g b0 Me Me d R2 R2 a b61 60 R1=H, Me, Ph; R2=Me, Prn, But, n-C6H13, Ph, (CH2)2Ph, CH=CHPh.Yu V Smirnova, Zh A Krasnaya Presumably,48, 51 the attack of bulky electrophiles, such as carbonyl compounds, at the sterically hindered g-position of the g-dianion 54a which exists in the syn-conformation proceeds with great difficulty. However, owing to the presence in the reaction mixture of a small amount of a more reactive a 0-dianion 53a the latter reacts with carbonyl compounds much faster to give a b 0-hydroxy derivative, which is accumulated in the mixture due to rapid isomerisation of the anion 54a. This suggestion is corroborated by the fact that the yield of the undesirable b 0-hydroxy ketone 60 increases in the case of sterically more hindered ketones, e.g., benzophenone.51 Its formation could be completely excluded using an enamino ketone with a phenyl substituent as a starting compound.O HNMe 1) MeLi, TMEDA 2) R1COR2 Ph Me Ph O HNMe OHR1 R2 (61% ± 96%) R1=H, Me, Ph; R2=Me, Et, Ph, cyclo-C6H11. Direct aldol condensation of g-dianions of the type 54 with aldehydes and ketones gave chiral anti-b-amino-d-hydroxyalk-1- enyl ketones 62.53 The stereoselectivity of this reaction depends critically on steric factors, being increased in the case of sterically more hindered starting enamino ketones and carbonyl com- pounds.53 Me H Ph N OHR4 R1 R3 R2 Ph R1=H;R2=H, Me, Bn; R17R2=(CH2)2; R3=H, Me; R4=Me, (CH2)2Ph, But. O62 Under conditions of acid hydrolysis, enamino ketones 62 undergo cyclisation into 4-imino-2,3-dihydropyrans 63.Com- pounds 63 are stable in acid media, apparently due to the formation of a more stable protonated form 64. However, in neutral media iminopyrans 63 are unstable and isomerise into enamino ketones 65.53 Me Me N Ph Ph NH OHR2 R1 3M HCl, THF R3 R2 R1 Ph O Ph R3 63 (68% ± 81%) O62 Me Me NH Ph Ph R2 O HN R1 R3 Ph R2 R1 R3 Ph O 64 65 (84% ± 91%) R1=H, Bn, Me; R2=Me, But, Ph; R3=H, Me. The reaction of the dilithium derivative 53f with esters results in enamino enols 66.54 NPh O 7 7 +R1CO2R2 Me CH2 53fMethods of synthesis of conjugated o-amino ketones Li Li OH O HNPh NPh O O NH4Cl R1 Me R1 Me R2O 67 66 (32% ± 88%) R1=Ph, 4-ClC6H4, EtCHBr; R2=Me, Et, C(Me)=CH2.It was proposed 54 that the adduct 67 stabilised by bridging bonds, is the intermediate in this reaction. The yields of com- pounds 66 are very high in the case of aromatic acid esters, but decrease dramatically if aliphatic acid esters are used. Moreover, the reaction of dilithium derivatives with aliphatic acid esters is accompanied by the formation of ester self-condensation prod- ucts; their proportion increases with time. It is suggested 54 that a transmetallation reaction between the dilithium derivative 53f and the ester favours self-condensation of esters. According to 1H NMR spectroscopic data, compounds 66 exist in solutions as two tautomers, viz., ketoenols 66a and diketones 66b, the equi- librium being markedly shifted towards the enol form 66a.54 O HNPh O OH O HNPh R1 R1 66b 66a It should be noted that the reaction of dilithium derivatives of the type 54 with esters yields unsaturated amino alcohols 68 as the main product; these are formed as a result of condensation of two molecules of the enamino ketone with one molecule of the ester.Obviously, the intermediate 69 which is formed first is far less stable than the adduct 67 and is decomposed with elimination of the alcoholate anion to give the monolithium derivative 70. The carbonyl group of the latter is not coordinated with the lithium ion and behaves as the free carbonyl group. Its reaction with the second molecule of the g-dilithium derivative 54 affords the polyfunctional compounds 68.54 The expected products of this reaction, viz., enamino diketones 71, are formed in insignificant amounts.NR1 O R2CO2R3 (92) 7 7 Ph CH2 54 Li O NR1 OLiR2 7OR3 Ph OR3 69 Li O O NR1 1) 54 2) NH4Cl Ph R2 70R2 O HNR1 OH R1NH O O HNR1 O + Ph Ph R2 Ph 71 (6% ± 20%) 68 (57% ± 86%) R1=Pri, Me; R2=Me, Et, Ph, Pr; R3=Me, Et, C(Me)=CH2. The AVK derivatives 72 with a carboxy group in the a 0-position were obtained by the reaction of the a 0-dianions 53 with carbon dioxide.55 For R1=H, the acids 72 are unstable and undergo decarboxylation within several hours to give the starting AVK. For R1=H, the acids 72 retain their stability at lowered temperatures over a period of several months.The dianions of the type 54 do not react with carbon dioxide.55 1027 O NR2 O HNR2 LTMP 7 7 R1 R1 1) CO2 2) NaHCO3 3) HCl R3 R3 53 O HNR2 R1 R3 CO2H 72 (59% ± 90%) R1=H, Me, Et, Bn; R2=Me, Pri, Ph, cyclo-C7H12; R3=Me, Ph(CH2)2. The a 0-dianions of the type 53 react with diethyl carbonate to give enamino oxo acids 73.55 The maximum yields of compounds 72 and 73 were obtained for compounds with bulky substituents at the nitrogen atom.55 HNR1 O O NR1 1) (EtO)2CO 2) NH4Cl 7 7 R2 R2 CH2 CO2Et 73 (56% ± 95%) R1=Me, Pri, Ph; R2=Me, Et. The g-dianions of the type 54 react with diethyl carbonate to produce a-amino-b-benzoylacrylates 74.55 O HNR NR O 1) (EtO)2CO 2) NH4Cl 7 7 Ph CH2 CO2Et Ph74 (45% ± 94%) 54 R=Me, Pri, Ph.It is noteworthy that the condensation of (EtO)2CO with the dianions 54 containing an aliphatic substituent instead of a phenyl substituent yields a complex mixture of products.55 The dianions 53 and 54 add to the double bond of nitroalkenes to give g 0- (75) or e-nitro-b-enamino ketones (76). As a rule, this reaction yields only one of the possible diastereomers.56 O HNR1 R3 R2 R3 53b,f g0 O2N Me b0 a0 NO2 R2 75 (60% ± 75%) R1=Ph (53f), Pri (53b); R2=H, Me; R3=Ph, Et, Ph(CH2)2, 4-MeOC6H4. R4 R3 O NR2 7 7 O HNR2 R4 b + e NO2 d R1 R1 CH2 O2N g a 54 R3 76 (53% ± 83%) R1=Ph, Me; R2=Me, Pri, Ph; R3=H, Me; R4=Et, Ph, Ph(CH2)2, 4-MeOC6H4. The cross-conjugated enamino ketones 77 were obtained by the reaction of aldehydes and ketones with dianions formed upon treatment of a 0-(trimethylsilyl)enamino ketones 78 with bases.57, 58 NR1 O NR1 O R2COR3 B7 7 7 Me Me Me3SiCH2 Me3SiCH2 78 O HNR1 OH R2 TMEDA Me R3 79 SiMe3Yu V Smirnova, Zh A Krasnaya 1028 O HNR1 R3 O HNR1 R2 NMe2 + Me R2 Me R3 O (5E )-77 (5Z)-77 NRR1=Pri, But; R2=Ph, PhCH=CH, Et; R3=H, Me.88 (R=H, Et). Dialkyl acetals of N,N-dialkylformamides 8077, 78 react with 2-acetylcycloalkane-1,3-diones 89 exclusively at the acetyl group to give enamino triketones 90.77 O O OMe 20 8C, 0.5 ± 1.5 h Me OMe +R2N X 80 89 O O O The structures of the products formed depend on the reaction conditions.In tetrahydrofuran, only silyl-substituted hydroxy enamino ketones 79 or their mixture with compounds 77 are isolated. In the presence of tetramethylethylenediamine which binds lithium cations and thus facilitates the elimination of trimethylsilanol only a 0,b 0-unsaturated amino ketones 77 in a total yield of 53% to 82% are obtained. It is noteworthy that the ratios of the (5E)- to (5Z)-isomers of compounds 77 depend on the reaction time, viz., the (5Z)-isomer is mainly formed upon elimination after 2 h, whereas the (5Z)-isomer is isomerised into the (5E)-isomer with an increase in the reaction time to 20 h.58 NR2 X 90 (85% ± 95%) O 6. Reactions of amide and lactam acetals and ester aminals with compounds containing an activated methylene group R=Me, Et; X=CH2, (CH2)2, CH2CHMe.The dioxo derivatives of tetrahydrobenzopyrans 92 formed upon cyclodeamination of enamino triketones 93 are the reaction products of 2-propionylcyclohexane-1,3-diones 91 with N,N- dialkylformamide acetals 80.78 O O OR2 Reactions of amide and lactam acetals and ester aminals with compounds containing an activated methyl or methylene group is one of the most popular approaches to the synthesis of enamines including enamino ketones.1, 59 ± 62 Among amide and lactam acetals and ester aminals synthesised to date,59, 61 ± 66 dialkyl acetals of N,N-dialkylformamides 80 andN,N-dimethylacetamide 81 as well as diethylacetals of N-methylbutyro- (82a) and -valero- lactams (82b) are the most well-studied and popular synthetic reagents.Et + (H2C)n X OR3 OR2 R12 N OEt O MeOR4 80 91 R1R2N OR3 Me2N OEt OR4 81a,b 80 Me N O O O O NR1282a,b Me 7R2NH Me X X R1=R2=Et, Me; R17R2=(CH2)5; R3=Me, Et; R4=Me (81a), Et (81b); n = 1 (82a), 2 (82b). O 92 (56% ± 90%) O 93 In solutions of amide acetals, an equilibrium exists between acetals 83a, ambidental cations 83b and a-alkoxy enamines ; R2=H, Me; X=CH2, CMe2. R12 N=Me2N, Et2N, N 83c.66 ± 71 CH2 CH2OR OR +ROH OR N N Diaminodienediones 94 were synthesised by heating cyclic enamino b-diketones 95 with an excess of acetals 80;79 in this case, the reaction occurs at the methyl group which is separated from the carbonyl function by a double bond. Only few examples of this kind of reactions are documented.79 ± 81 N + OR OR7 83b 83c 83a HNR1 O OMe Me 80 ± 100 8C, 15 ± 20 h Me + OMe R22 N O Me 80 95 O HNR1 Reactions of nucleophilic substitution of amide acetals pro- ceed in the absence of catalysts, since the abstraction of a proton from an activated methyl or methylene group occurs by virtue of the basic properties of the acetal amide itself.In some cases, azeotropic distillation may be required to remove the alcohol formed. NR22 The condensation of dimethylformamide acetal and alkylfor- Me Me O 94 (59% ± 65%) mate aminal with various ketones (e.g., aromatic, heterocyclic) and diketones gave enamino ketones 84,72 85,73, 86,74 87 75 and 88.76 R1=Me, Et; R2=H, Me. O NMe2 N Ph NMe2 The condensation of crotonamide 96 with a large excess of tert-butylformate bis(N,N-dimethylaminal) 97 at 140 8C yields NMe2 84 O 140 8C O 85 + O O O O OBut Me NH2 Me2N 96 N 97 O Me R NMe2 Me2N NMe2 N Me2N 98 (65%) NMe2 86 Bun 87 (R=Bun, Ph, Me)Methods of synthesis of conjugated o-amino ketones enamidine 98 as a result of reactions at both methyl and the primary amino group.81 An unusual reaction was observed in the study of the reactions of N,N-dimethylacetamide diethyl acetal 81b and acetals of N-methylbutyro- (82b) and -valerolactams (82b) with enamino diketones.1, 62, 69 ± 71, 82 Quite unexpectedly, the reaction of the enamino diketone 99 with the acetal 81b occurs not at the primary amino group to give enamidine 100,83 but at the double bond and yields a mixture of diaminodiene diketone 101 and its dimethyl- aminoethylidene derivative 102.1, 62, 69, 82 CH2 MeOEt OEt OEt Me2N Me2N 81c 81b O O Me N NH2 Me Me +81b NMe2 O O Me Me 100 99 + O NH2 NMe2 PhMe, 5 min OEt 99+81c Me Me O7 103O O NMe2 NMe2 Me N Me NMe2 NMe2+Me Me O O Me 102 101 It is assumed that the reaction with enamino diketone 99 involves ketene acetal-aminal 81c rather than dimethylacetamide acetal 81b which is in equilibrium with the former.High electron density in the b-position of ketene acetal-aminal 81c, which is provided by two strong electron-donor substituents, favours its easy addition to the double bond of the enamino ketone 99. This results in the zwitterion 103 the alkoxy group of which reacts with dimethylamine formed in the course of the reaction to give the diaminodiene 101 or with ammonia and an excess of acetal 81b to give amidine 102.69 It is clear from the scheme presented why the dimethylformamide acetals 80 for which the equilibrium of the type 81b>81c is impossible do not enter into such reactions.Obviously, this reaction is ambiguous because of the presence of the primary amino group in the enamino diketone 99. The introduction of N-substituted enamino diketones 104 and 105 into the reaction with acetal 81b allows selective synthesis of the diamino dienediketone 101.69 O NHR Me O NMe2 O Me 104 NMe2 81b Me O O Me 101 NMe2 Me O Me 105 R=Ph, Bn.The reaction of N,N-dimethylacetamide acetal (81b) with other cyclic and acyclic enamino dicarbonyl compounds also occurs at the double bond and results in diamino dienediketones 106.69, 71 O O b g d a R1 R1 81b NMe2 R2 R2 O O 106 (40% ± 71%) R1=R2=Me, Ph; R17R2=OCH2O, (CH2)3. According to 1H NMR spectroscopic data, d,d-diamino di- enediones 101 and 106 exist predominantly as s-trans-conformers with respect to the C(b)7C(g) bond (J=14.2 ± 14.7 Hz). N-Methylbutyro- and -valerolactam acetals (82a) and (82b) 70 react with the enamino diketones 99 and 105 in a similar way. Thus the enamine 99 gives cyclic derivatives of dienediamines 107a,b in high yields.70 (H2C)n NH2 Me N (H2C)n O O OEt O + OEt Me N Me 82a,b Me99 n=1 (a), 2 (b).The structure of the reaction product of the lactam acetal 82a with the enamino diketone 105 depends on the reagent ratio, viz., the dihydropyrrole diene 108 is formed in the case of an equimolar ratio, whereas triene 109 is formed in the presence of an excess of the lactam acetal.70 OEt OEt OEt Me N Me N82a NMe2 O O Me Me10582a Me N Me N Me2N O O O Me Me108 Me109 Dimethylformamide acetal (80) can also be used in the syn- thesis of di(b-aminovinyl) ketones.84 ± 87 Thus the synthesis of 1,5-bis(dimethylamino)penta-1,4-dien-3-one (111) is based on the reaction of dimethylformamide acetal (80) with 3-oxoglutaric acid (110).84, 85 O OMe 70 8C, 1 h + CO2H HO2C OMe Me2N 80 110 O NMe2 Me2N 111 (100%) 1029 NMe2 NMe2 N Me N(CH2)n O Me Me 107a,b (70% ± 80%) NMe2 Me NO Me1030Di(aminomethylidene)cycloalkanones 112 are formed in the reaction of dimethylformamide acetal 80 with cyclopentanone or cyclohexanone which proceeds under very drastic conditions; the yields of the target products are usually low.86 O OMe 115 ± 160 8C, 10 h + OMe Me2N (H2C)n 80 O Me2N NMe2 (H2C)n 112 (11% ± 14%) n=1, 2. In the presence of 1,5-diazabicyclo[4.3.0]non-5-ene as a cata- lyst, the yields of compounds 112 increase up to 36%± 50%.87 7.Synthesis of d-aminodienones and a,a 0-bis(3-aminopropenylidene)alkanones Some examples of the synthesis of d-amino dienones were described earlier.88 ± 90 In most cases, the procedures employed for the synthesis of enamino ketones are inapplicable to the synthesis of conjugated d-amino dienones and, particularly, of polyunsaturated o-amino ketones where the number of conju- gated double bonds between the NMe2 and CO groups is more than two.Thus the reaction of the enynone 113 with diethylamine was used for the synthesis of d-aminodienone 19. The ketone 113 was synthesised from 4-dimethylaminobut-3-en-2-one (114) and ethy- nylmagnesium bromide.88 O +CH CMgBr Me Me2N 114 O O Et2NH C Me Me Me2N 19 113 HC Polyunsaturated amino ketones can, in principle, be synthes- ised by reiteration of this procedure. However, it has been shown 89 that this approach is ineffective even in the case of trienones.The d-amino dienones 115 possessing aryl substituents were obtained by the reaction of aryl b-chlorovinyl ketones 116 with enamines 117 in the presence of triethylamine.90 R1 O Et3N + N R3 Cl R2 116 117 R1 O N R3 115 R2 R1=H,R2=Me, Et, Ph; R17R2=(CH2)3; R3=Ph, 4-MeC6H4, 4-ClC6H4, 4-BrC6H4. The condensation of 3-dimethylamino-1,1,3-trimethoxypro- pane (118) (see Ref. 91) with ketones is successfully employed in the synthesis of d-dimethylamino dienones.86, 92 ± 95 Depending on the conditions, the reaction can be directed towards the formation of ketones containing one (119a ± j) or two (120a ± j) 3-dimethyl- aminopropenylidene substituents in the a-position.86, 92, 93, 95 OMe MeO +R1CH2COCH2R2 OMe Me2N 118 Yu V Smirnova, Zh A Krasnaya O CH2R2 Me2N 119a ± j R1 O NMe2 Me2N 120a ± j R2 R1 R1=R2=H(a); R1=H,R2=Me (b); R17R2=(CH2)2 (c), (CH2)3 (d), (CH2)4 (e), (CH2)2CHMe (f), CH2CH(Me)CH2 (g), CH2CH(Me)CHMe (h), CH2OCHMe (i), CH2OCMe2 (j). b-Dimethylaminoacrolein acetal-aminals or aminals 121 can be used for the synthesis of compounds 119 and 120 instead of 3-dimethylaminotrimethoxypropane 118.96 In both cases, the condensation occurs in the presence of a catalyst and virtually does not result in the formation of by-products.An important advantage of this method is the possibility to synthesise com- pounds 120 containing two aminopolyenoic chromophores. NMe2 119a ± j + R1CH2COCH2R2 Y Me2N 120a ± j 121 Y=OMe, NMe2.Owing to the interaction of two chromophores through the carbonyl group, the ketones 120 possess specific properties, such as pronounced solvatochromism 97 and thermochromism,98 fluo- rescence and the ability to generate laser radiation with high efficiency.99 The exceptionally easy alkylation and protonation of these compounds at the carbonyl group is accompanied by drastic changes in their colour.100, 101 The reaction of b-dimethylaminoacrolein acetals or acetal- aminals 121 with cyclic ketones 122a,b gave the intermediate compounds (diamines 123a,b) which are quantitatively converted into a-(3-dimethylaminopropenylidene)cycloalkanones 124a,b in CHCl3 in the presence of SiO2.94 O NMe2 85 ± 100 8C (CH2)n Y +Me2N 121 122a,b O O NMe2 NMe2 NMe2 SiO2 (CH2)n (CH2)n 124a,b 123a,b (70% ± 75%) Y=OMe, NMe2; n = 9 (a), 12 (b).The reaction of acetals or acetal-aminals 125 102 substituted in position 2 with ketones often gives only ketones 126a ± h with one 3-dimethylaminopropenylidene substituent.103 ± 105 O NMe2 80 8C, 0.5 ± 3 h Y R1CH2 CH2R2 + Me2N125 R3 O b d g a NMe2 R1CH2 R3 R2 126a ± h (10% ± 55%) Y=OMe, NMe2.Methods of synthesis of conjugated o-amino ketones R3 R2 R1 Compo- und 126 Compo- und 126 OEt H HH H Me ef H H Cl g H H Br h abcd According to 1H NMR spectroscopic data, ketone 126a has a trans-configuration of the a,b-double bond (J=15 Hz).105 The cyclic ketones 126e ± h represent E-isomers with a trans-arrange- ment of the carbonyl group and a substituent in the g-position (3JCO7Hb=5.0 ± 5.8 Hz).103, 104 In contrast to dienones 126a ± h, which exist exclusively in the open form, the g-substituted compounds 126i ± k and 127a,b containing a six-membered ring exist in some solvents in a dynamic equilibrium between the Z-isomer of the d-amino dien- one form and the corresponding 2H-pyran.Evidence for the existence of both forms is provided by the detection of two sets of signals in the 1Hand 13C NMRspectra as well as by absorption bands from both valence isomers in the UV spectra.103 ± 106 O O a NMe2 O R 126i ± k (10% ± 55%) O Y Me2N125 R3 Y=NMe2, OMe; R=Me (126i, 127a), OEt (126j ), Cl (126k, 127b). Bis(o-dimethylamino) polyenones 128 and 129 containing Me, Ph, Cl and OEt in the g,g 0-positions were synthesised by the reaction of the corresponding substituted acetals or acetal-amin- als 125 with acetone 107, 108 and cyclopentanone.103, 104, 109 Bis(3-dimethylaminopropenylidene)cyclohexanone 130 was obtained using fluorine-substituted acetals 125.110 O NMe2 Me2N 128 R R R=Me, Ph.O Me2N NMe2 R R 129 R=Me, OEt, Cl. O NMe2 Me2N F F 130 R3 R2 R1 Me Cl Br OEt (CH2)2 (CH2)2 (CH2)2 (CH2)2 R O NMe2 b d g NMe2NMe2 R 127a,b (60% ± 65%) R O NMe2 1031 Compound 131 which is formed first was isolated in the form of its valence isomer, viz., 2H-pyran 132, in the reaction of a-phenyl-b-dimethylaminoacrolein acetals or acetal-aminals with cyclohexanone.108 O NMe2 80 ± 100 8C, 2 h Y Me2N + Ph O NMe2 Me2N Ph Ph 131 (25%) Ph O NMe2 Ph 132 NMe2 Y=OMe, NMe2.d-Amino dienones 133 containing heterocyclic and heteroor- ganic substituents were synthesised 111 using b-dimethylamino- acrolein acetals or acetal-aminals 121. O O 121 R1 R1 Me2N R2CH2 133 (41% ± 87%) R2 , , R1=Ph, PhCH=CH, , C5H5FeC5H4; N N O R2=H, Me, Ph. Azaacetal 134 was used for the synthesis of unsaturated amino ketones of various structures. Thus the condensation of cyclo- alkanones 119c,d with the azaacetal 134 is accompanied by the cleavage of the C=N bond resulting in non-symmetrical bis(di- methylaminoalkenylidene)alkanones 135.112 O OMe 80 8C Me2N OMe N + Me2N 134 119c,d X O Me2N NMe2 X 135a,b (75% ± 80%) X=CH2 (119c, 135a), (CH2)2 (119d, 135b).The cleavage of theC=Nbond also takes place in the reaction of the azaacetal 134 with cycloalkanones and acetone at an equimolar ratio of the reagents.112 O OMe PriOH, 80 8C, 1 h OMe CH2R2 R1CH2 + Me2N N 134 O NMe2 R1CH2 (38% ± 43%) R2 R1=R2=H;R17R2=(CH2)2, (CH2)3, (CH2)4.1032The ketones 136 and 137 containing an azamethine group were synthesised at an cycloalkanone : azaacetal 134 ratio of 1 : 2 (PriOH, 80 8C, 5 h).112 O O 134 N N NMe2 Me2N136 (20%) O O 134 N NMe2 Me2N137 (41%) III. Conjugated o-amino polyenones 1. Condensation of a,b-unsaturated ketones with b-dimethylaminoacrolein acetals and acetal-aminals b-Dimethylaminoacrolein 121 acetals and acetal-aminals mani- fest exceptionally high reactivity.They react with cyclic and acyclic unsaturated ketones 138a,d and their vinilogues 138b,c,e containing no more than three double bonds at both the a-methyl (or a-methylene) group and the methyl group separated from the carbonyl function by double bonds to give bis(dimethylamino)- polyunsaturated ketones 139 ± 141 (Scheme 1).113 O NMe2 + Me Y R1CH2 Me2N n 121 R2 138a ± e O NMe2 Me2N n 139a ± e R2 R1 Y=NMe2 , OMe. n Yield of 139 (%) R2 Compounds 138, 139 R1H H 0 54 H H 1 40 H H 2 27 0 75 1 48 (CH2)3 (CH2)3 abcde In undeca-3,5,7,9-tetraen-2-one (142), the mobility of hydro- gen atoms of the methyl group separated from the carbonyl function by four double bonds is insufficient for the condensation with compounds 121; only the methyl group in the a-position O NMe2 20 8C Me Me Y + Me2N 121 142 O Me Me2N 143 (25%) Y=NMe2, OMe.relative to the carbonyl group reacts to yield ketone 143 contain- ing one 3-dimethylaminopropenylidene group.113 O Me Me Me2N NMe2 O Y Me2N 121 Me Me2N 141 (50%) Y=NMe2, OMe. Yu V Smirnova, Zh A Krasnaya In conjugated ketones 138a e, the mobility of hydrogen atoms of the methyl or methylene group vicinal to the carbonyl group is higher than that of the methyl group separated from the carbonyl group by one or several double bonds. This allows one to carry out condensation under milder conditions, e.g., at equimolar ratios of reagents and without heating, which results in selective formation of reaction products involving the a-methylene group.113 O Me Me2N n R2 R1 n=0±2; R1=R2=H; R1±R2=(CH2)3.Unfortunately, this method is confined mainly to the synthesis of non-symmetrical polyunsaturated bis(3-dimethylaminoprope- nylidene)alkanones. Moreover, this requires difficultly accessible unsaturated ketones as starting compounds. 2. Condensation of acetals and acetal-aminals of b-dimethylaminoacrolein vinilogues with ketones The synthesis of symmetrical polyconjugated a,a 0-bis(dimethyl- aminoalkenylidene)alkanones is based on the use of acetals and acetal-aminals of conjugated o-dimethylaminopolyenals 144 (n=2, 3) which react with different ketones at both methyl or methylene groups to give bis(dimethylamino)polyunsaturated ketones 145a ± e.114 O NMe2 n Y CH2R2 R1CH2 + Me2N 144 O n Me2N n NMe2 145a ± e R2 R1 Y=OMe, NMe2.n Yield of 145 (%) R2 R1 Compounds 145 (CH2)2 (CH2)3 H H 2 98 2 62 2 60 H H 3 3 3 12 abcde (CH2)2 In the case of compounds 144 (where n=3), the reaction proceeds under more drastic conditions where the polyunsatu- rated ketones formed are unstable. The corresponding derivatives 145d,e could be isolated only in reactions with acetone and cyclopentanone in low yields.114 The condensation of acetals or acetal-aminals of d-dimethyl- aminodienal 144 (n=2) with enamino, dienamino and trien- amino ketones was used for the synthesis of non-symmetrical polyunsaturated bis(dimethylamino) ketones 146 ± 148 (Scheme 2).114, 115 It should be noted that compounds 144 (where n=2) are characterised by sufficiently high reactivities to react with 2-ethyl- idenecyclohexanone (149) at both the a-methylene and the methyl Scheme 1 O NMe2 140 (34%) O NMe2Methods of synthesis of conjugated o-amino ketones O Me NMe2 Me2N O NMe2 NMe2 Me2N Y Me2N 144 O Me Me2N Me Me Me2N Y=NMe2, OMe.group separated from the carbonyl function by a double bond to yield the derivative 150.114 O NMe2 90 ± 100 8C Me Y + Me2N 149 144 O NMe2 Me2N 150 (25%) Y=NMe2, OMe. With a change in reaction conditions (viz., with equimolar ratio of the reactants, gentle, short-term heating), the reaction of acetals and acetal-aminals of o-dimethylaminopolyenals 144 with ketones occurs in a regioselective manner and yields the poly- unsaturated ketones 151 ± 153 which contain only one dimethyl- amino group.116 O O Me Me2N Me2N CH2R2 n R1 152 151 n=2, 3; R1=R2=H; n=2, 3; R17R2=(CH2)2; n=2, R17R2=(CH2)3.O NMe2 153 3. Condensation of ketones with polymethine salts Yet another well-studied approach to the synthesis of o-amino polyenones, is the condensation of ketones with trimethine or pentamethine salts in the presence of bases, such as NaNH2,117 MeOK, MeONa or ButOK,87, 111, 118 ± 120 NaH or LDA.121 ± 124 As a rule, this reaction is carried out in triethylamine or pyridine.O + NaH, Py, N2 CH2R2 Me2N NMe2 Cl7+ R1CH2 O Me2N CH2R2 R1 119 (66% ± 91%) R1=Me, Et; R2=H, Me; R17R2=(CH2)2, (CH2)3. d-Dimethylaminoalka-1,3-dienyl ketones 154 having substitu- ents in the aminopolyene chain were prepared by the condensation 1033 Scheme 2 O NMe2 146 (25%) O 147 (30%) NMe2 O Me NMe2 148 (14%) Me of aromatic and heteroaromatic methyl ketones with the substi- tuted trimethine salts 155.111, 118, 120 R1 O + NaMe, Py R3 Me Me2N NMe2 ClO¡4 + R2 155 O R1 R3 Me2N R2 154 (68% ± 93%) . ; R2=H, OMe, Ph; R3=Ph, R1=H, Ph, S S The reaction of the pentamethine salt 156 with methyl ketones makes it possible to synthesise o-aminopolyunsaturated ketones 157 with a more extended chromophore.111 O + NaOMe, Py + Me2N NMe2 ClO¡4R Me 156 O R Me2N 157 (68% ± 93%) .R=But, Ph, 4-MeC6H4, 4-MeOC6H4, 4-BrC6H4, S The condensation of isophorone with the pentamethine salt 156 involves the methyl group (rather than the sterically shielded a-methylene group), which affords the aminopolyunsaturated ketone 158 where NMe2 and CO groups are separated by four double bonds.111 O Me Me NMe2 158 (65%) The reaction of tri- and pentamethine salts 159 with ketones in pyridine in the presence of MeONa (2 equiv.) or ButOK was used 119 to obtain the symmetrical diamino polyenones 120a,c,d and 160a,b. The yields of x-dimethylamino trienones 160a,b do not exceed 35%. O + NaOMe or ButOK, Py + I7 PhMeN NMePh + n CH2R2 R1CH2 154 1591034 Me O n n N PhMeN NMePh R2 R1 120a,c,d; 160a,b 169 NMe2 n Yield (%) R2 Compounds R1 High reactivities of compounds 121 enable their reactions with the methyl group separated from the carbonyl group by a double bond or by a heterocyclic fragment to give the o-amino poly- enones 170 ± 172.113, 125, 126 NMe2 H H 1 55 1 83 1 65 2 26 2 35 120a 120c 120d 160a 160b (CH2)2 (CH2)3 (CH2)2 (CH2)3 R2 The non-symmetrical cyclic polyunsaturated ketones 162 were prepared from cyclic enamino ketones 161.87 O O O R1 + 2 M NaOMe, Py NR1R2 170 PhMeN NMePh I7+ 161 (CH2)n R1=Me, OH; R2=H, OMe.159 O Me2N NR1R2 PhMeN (CH2)n 162 (86% ± 90%) O X=O, S. R1=H, Me; R2=Ph, Me.O IV. Conjugated o-amino ketones containing a heterocyclic fragment O Me2N 172 R R=H, Me, Ph. The synthesis of conjugated o-amino ketones containing a hetero- cyclic fragment which strongly affects their physicochemical properties has been carried out in the last few years. * * * Thus the condensation of heterocyclic ketones containing an acetyl or an activated methylene group with acetals and acetal- aminals 121 or 144a,b was used to obtain d-amino dienones 163a,113 164, 165a, 166a,109 167a, 168a 125 and x-amino trienones 163b, 165b, 166b, 167b, 168b 125 and 169 116 containing a hetero- cyclic fragment. HN NMe2 Me2N Me X n 164 163a ± d O O In conclusion, it may be said that the o-amino polyenones 168, 169 and 170 containing a coumarin fragment are very valuable, since coumarin derivatives constitute an important group of organic luminophores and laser dyes which effectively generate radiation in the region of 400 ± 560 nm.127, 128 In addition, they are widely employed in medicine and biochemistry as fluorescent compounds.129 2,6-Bis(4-dimethylaminoalka-1,3-dienyl)-4H- pyran-4-ones (172) appeared to be very interesting reagents from the spectral and synthetic standpoints.126 X=O, S: n = 1 (a), 2 (b); X=O: n = 2 (c), 3 (d).Me O References HN n NMe2 N NMe2 O S S n 165a,b Ph N 166a,b n=1(a), 2 (b). n = 1 (a), 2 (b). O NMe2 n O O Et2N 167a,b n = 1 (a), 2 (b). O NMe2 n O N O168a,b n = 1 (a), 2 (b). 1. B B Aleksandrov (Ed.) Enaminy v Organicheskom Sinteze (Enamines in Organic Synthesis) (Sverdlovsk: Urals Branch Academy of Sciences of the USSR, 1989) 2.J V Greenhill Chem. Soc. Rev. 277 (1977) 3. Ya F Freimanis Khimiya Enaminoketonov, Enaminoiminov i Enaminotionov (The Chemistry of Enamino Ketones, Enaminoimines and Enaminothions) (Zinatne: Riga, 1974) 4. M Azzaro, S Geribaldi, B Videau Synthesis 880 (1981) 5. B F Kukharev, V K Stankevich, G R Klimenko, N A Lobanova Izv. Akad. Nauk, Ser. Khim. 2228 (1997) a 6. K I Pashkevich, V I Filyakova, Yu N Sheinker, O S Anisimova, I Ya Postovskii, E F Kuleshova Izv. Akad. Nauk SSSR, Ser. Khim. 2087 (1979) a 7. K I Pashkevich, V I Filyakova, I Ya Postovskii Izv. Akad. Nauk SSSR, Ser. Khim. 2346 (1981) a 8. V I Filyakova, V G Ratner, N S Karpenko, K I Pashkevich Izv.Akad. Nauk, Ser. Khim. 2278 (1996) a 9. E J Cone, R H Garner, A W Hayes J. Org. Chem. 37 4436 (1972) 10. P G Baraldi, D Simoni, S Manfredini Synthesis 902 (1983) 11. F Texier-Boullet, B Klein, J Hamelin Synthesis 409 (1986) Yu V Smirnova, Zh A Krasnaya X NMe2 171 NMe2 RMethods of synthesis of conjugated o-amino ketones 12. M E F Braibante, H S Braibante, L Missio, A Andricopulo Synthesis 898 (1994) 13. B Rechshteiner, F Texier-Boullet, J Hamelin Tetrahedron Lett. 34 5071 (1993) 14. P Ruault, J-F Pilard, B Touaux, F Texier-Boullet, J Hamelin Synlett 935 (1994) 15. F Effenberger, R Maier, K-H SchoÈ nwaÈ lder, T Ziegler Chem. Ber. 115 2766 (1982) 16. L-F Tietze, H Meier, E Voû Synthesis 274 (1988) 17.M Hojo, R Masuda, Y Kokuryo, H Shioda, S Matsuo Chem. Lett. 499 (1976) 18. M Hojo, R Masuda, E Okada Synthesis 1013 (1986) 19. L F Tietze, A Bergmann, G Brill, K BruÈ ggemann, U Hartfiel, E Voû Chem. Ber. 122 83 (1989) 20. L F Tietze, R Schimpf, J Wichmann Chem. Ber. 125 2571 (1992) 21. S V Pazenok, I I Gerus, E A Chaika, L M Yagupol'skii Zh. Org. Khim. 25 379 (1989) b 22. M Hojo, R Masuda, E Okada, S Sakawuchi, H Narumija, K Morimoto Tetrahedron Lett. 30 6173 (1989) 23. J Staunton, Comprehensive Organic Chemistry (Eds. D Barton, WD Ollis) (Oxford: Pergamon Press, 1979) Vol. 4, p. 659 24. I P Lokot', F S Pashkovskii, F A Lakhvich Zh. Org. Khim. 34 1407 (1998) b 25. V Ya Sosnovskikh Izv. Akad. Nauk, Ser. Khim.2263 (1997) a 26. S V Pazenok, I I Gerus, M G Gorbunova, E A Chaika Zh. Org. Khim. 25 1560 (1989) b 27. M Hojo, R Masuda, E Okada Synthesis 425 (1990) 28. A A Akhrem, F A Lakhvich, V A Khripach, I B Klebanovich Tetrahedron Lett. 3983 (1976) 29. D N McGregor, U Corbin, J E Swigor, L C Cheney Tetrahedron 25 389 (1969) 30. S Auricchio, S Morrocchi, A Ricca Tetrahedron Lett. 2793 (1974) 31. F A Lakhvich, V A Kozinets, Ya M Katok, E V Koroleva Zh. Org. Khim. 34 1254 (1998) b 32. A Alberola, A M Gonzalez, M A Laguna, F J Pulido Synth. Commun. 16 673 (1986) 33. M T De la Cal, B I Cristobal, P Cuadrado, A M Gonzalez, F J Pulido Synth. Commun. 19 1039 (1989) 34. K I Pashkevich, A Ya Aizikovich Dokl. Akad. Nauk SSSR 224 618 (1979) c 35. V Ya Sosnovskikh, T P Kostrikova Zh.Org. Khim. 22 883 (1986) b 36. V Ya Sosnovskikh, I S Ovsyannikov Zh. Org. Khim. 26 2086 (1990) b 37. V Ya Sosnovskikh, I S Ovsyannikov Zh. Org. Khim. 29 259 (1993) b 38. V Ya Sosnovskikh Zh. Org. Khim. 34 829 (1998) b 39. V Ya Sosnovskikh, I S Ovsyannikov Zh. Org. Khim. 29 89 (1993) b 40. V Ya Sosnovskikh Mendeleev Commun. 189 (1996) 41. V Ya Sosnovskikh,M Yu Mel'nikov Khim. Geterotsikl. Soedin. 1003 (1996) d 42. V Ya Sosnovskikh,M Yu Mel'nikov Zh. Org. Khim. 34 303 (1998) b 43. V Ya Sosnovskikh Zh. Org. Khim. 28 1307 (1992) b 44. V Ya Sosnovskikh,M Yu Mel'nikov, I A Kovaleva Izv. Akad. Nauk, Ser. Khim. 2305 (1998) a 45. V Ya Sosnovskikh,M Yu Mel'nikov, V A Kutsenko Zh. Org. Khim. 32 1440 (1996) b 46.V Ya Sosnovskikh,M Yu Mel'nikov, V A Kutsenko Izv. Akad. Nauk, Ser. Khim. 1866 (1996) a 47. Y L Chen, P S Mariano, G M Little, D O'Brien, P L Huesmann J. Org. Chem. 46 4643 (1981) 48. G Bartoli,M Bosco, C Cimarelli, R Dalpozzo,M Guerra, G Palmieri J. Chem. Soc., Perkin Trans. 2 649 (1992) 49. G Bartoli,M Bosco, C Cimarelli, R Dalpozzo, G Palmieri J. Chem. Soc., Perkin Trans. 1 2095 (1992) 50. G Bartoli,M Bosco, C Cimarelli, R Dalpozzo, G De Munno, G Guercio, G Palmieri J. Org. Chem. 57 6020 (1992) 51. G Bartoli,M Bosco, C Cimarelli, R Dalpozzo, G Palmieri Tetrahedron 49 2521 (1993) 52. G Bartoli,M Bosco, C Cimarelli, R Dalpozzo, G Palmieri Synlett 229 (1991) 53. C Cimarelli, G Palmieri, M Camalli Tetrahedron Asymmetry 7 2099 (1996) 54.G Bartoli,M Bosco, C Cimarelli, G Guercio, G Palmieri J. Chem. Soc., Perkin Trans. 1 2081 (1993) 55. G Bartoli,M Bosco, A Guerrieri, R Dalpozzo, A De Nino, E Iantorno, G Palmieri Gazz. Chem. Ital. 126 25 (1996) 1035 56. G Bartoli,M Bosco, R Dalpozzo, A De Nino, G Palmieri Tetrahedron 50 9831 (1994) 57. G Bartoli,M Bosco, R Dalpozzo, A De Nino, E Iantorno, A Tagarelli, G Palmieri Tetrahedron 52 9179 (1996) 58. R Dalpozzo,A DeNino, E Iantorno,G Bartoli,M Bosco, L Sambri Tetrahedron 53 2585 (1997) 59. R F Abdulla, R S Brinkmeyer Tetrahedron 35 1675 (1979) 60. V G Granik, A M Zhidkova, R G Glushkov Usp. Khim. 46 685 (1977) [Russ. Chem. Rev. 46 361 (1977)] 61. N Anand, L Singh Tetrahedron 44 5975 (1988) 62. V G Granik Khim. Geterotsikl. Soedin.762 (1992) d 63. H Meerwein, P Borner, O Fuchs, H J Sasse, H Schrodt, J Sprille Chem. Ber. 89 2060 (1956) 64. W Kantlehner, P Speh Chem. Ber. 105 1340 (1972) 65. H Bredereck, F Effenberger,H P Beyerlin Chem. Ber. 97 3081 (1964) 66. J Gloede, L Haase, H Gross Z. Chem. 9 201 (1969) 67. V G Granik,M K Polievktov, R G Glushkov Zh. Org. Khim. 7 1431 (1971) b 68. V G Granik, N B Marchenko, L I Budanova, V A Kuzovkin, T F Vlasova, O S Anisimova, R G Glushkov Zh. Org. Khim. 11 1829 (1975) b 69. V G Granik, A K Shanazarov, N P Solov'eva, V V Chistyakov, I V Persianova, Yu N Sheinker Khim. Geterotsikl. Soedin. 1470 (1987) d 70. A K Shanazarov, N P Solov'eva, V V Chistyakov, Yu N Sheinker, V G Granik Khim. Geterotsikl. Soedin. 1477 (1987) d 71.A K Shanazarov, N P Solov'eva, V V Chistyakov, V G Granik Khim. Geterotsikl. Soedin. 218 (1991) d 72. N Takeuchi, N Okada, S Tobinaga Chem. Pharm. Bull. 31 4355 (1983) 73. D L Jameson, L E Guise Tetrahedron Lett. 32 1999 (1991) 74. T W Bell, A Firestone J. Am. Chem. Soc. 108 8109 (1986) 75. W D Jones Jr , R A Schnettler, E W Huber J. Heterocycl. Chem. 27 511 (1990) 76. T V Golovko, N P Solov'eva, G A Bogdanova, Yu N Sheinker, V G Granik Khim. Geterotsikl. Soedin. 1190 (1991) d 77. A L Mikhal'chuk, O V Gulyakevich Zh. Org. Khim. 26 1804 (1990) b 78. A L Mikhal'chuk Zh. Obshch. Khim. 61 261 (1991) e 79. A L Mikhal'chuk, O V Gulyakevich Zh. Org. Khim. 31 151 (1995) b 80. Zh A Krasnaya, T S Stytsenko, V S Bogdanov, A S Dvornikov Khim. Geterotsikl.Soedin. 1325 (1988) d 81. H Bredereck, G Simchen, B Funke Chem. Ber. 104 2709 (1971) 82. A K Shanazarov, V V Chistyakov, V G Granik Khim. Geterotsikl. Soedin. 127 (1986) d 83. V G Granik, N B Marchenko, E O Sochneva, T F Vlasova, A B Grigor'ev, M K Polievktov, R G Glushkov Khim. Geterotsikl. Soedin. 1505 (1976) d 84. USSR P. 176 922; Byull. Izobret. (24) 26 (1965) 85. E C Taylor, J S Skotnicki Synth. Commun. 13 1137 (1983) 86. Zh A Krasnaya, T S Stytsenko, E P Prokof'ev, V F Kucherov Izv. Akad. Nauk SSSR, Ser. Khim. 116 (1978) a 87. Yu L Slominskii, I D Radchenko, S V Popov, A I Tolmachev Zh. Org. Khim. 19 2134 (1983) b 88. T Guvigny, H Normant Bull. Soc. Chim. Fr. 515 (1960) 89. Zh A Krasnaya, T S Stytsenko, E P Prokof'ev, V F Kucherov Izv.Akad. Nauk SSSR, Ser. Khim. 2511 (1975) a 90. W Schroth, G W Fisher Chem. Ber. 102 575 (1969) 91. Zh A Krasnaya, T S Stytsenko, E P Prokof'ev, V F Kucherov Izv. Akad. Nauk SSSR, Ser. Khim. 2008 (1973) a 92. USSR P. 479 758; Byull. Izobret. (29) 77 (1975) 93. Zh A Krasnaya, T S Stytsenko, E P Prokof'ev, V A Petukhov, V F Kucherov Izv. Akad. Nauk SSSR, Ser. Khim. 595 (1976) a 94. Zh A Krasnaya, E P Prokof'ev, V F Kucherov Izv. Akad. Nauk SSSR, Ser. Khim. 1373 (1980) a 95. T S Stytsenko, Candidate Thesis in Chemical Sciences, Institute of Organic Chemistry, Academy of Sciences of the USSR, Moscow, 1977 96. Zh A Krasnaya, T S Stytsenko Izv. Akad. Nauk SSSR, Ser. Khim. 850 (1983) a 97. A S Tatikolov, Doctoral Thesis in Chemical Sciences, Institute of Biochemical Physics, Russian Academy of Sciences, Moscow, 1996 98.L A Shvedova, A S Tatikolov, V A Kuz'min, Zh A Krasnaya, A R Bekker Dokl. Akad. Nauk SSSR 276 654 (1984) e 99. L A Shvedova, A S Tatikolov, A P Darmanyan, V A Kuz'min, Zh A Krasnaya Dokl. Akad. Nauk SSSR 276 164 (1984) eYu V Smirnova, Zh A Krasnaya 1036 100. Zh A Krasnaya, T S Stytsenko, E P Prokof'ev, V F Kucherov Izv. Akad. Nauk SSSR, Ser. Khim. 392 (1978) a 101. Yu A Fanov, Zh A Krasnaya, T S Stytsenko, V I Slovetskii, E I Isaev Izv. Akad. Nauk SSSR, Ser. Khim. 493 (1989) a 102. Zh A Krasnaya, T S Stytsenko, V S Bogdanov Izv. Akad. Nauk SSSR, Ser. Khim. 106 (1988) a 103. Zh A Krasnaya, T S Stytsenko, V S Bogdanov, E D Daeva, A S Dvornikov Izv. Akad. Nauk SSSR, Ser. Khim. 1075 (1985) a 104. Zh A Krasnaya, T S Stytsenko, V S Bogdanov, A S Dvornikov Izv. Akad. Nauk SSSR, Ser. Khim. 1323 (1989) a 105. Zh A Krasnaya, E P Prokof'ev, I P Yakovlev, E D Lubuzh Izv. Akad. Nauk SSSR, Ser. Khim. 2325 (1980) a 106. E P Prokof'ev, Zh A Krasnaya Izv. Akad. Nauk SSSR, Ser. Khim. 2284 (1980) a 107. Zh A Krasnaya, E P Prokof'ev, V F Kucherov Izv. Akad. Nauk SSSR, Ser. Khim. 123 (1978) a 108. Zh A Krasnaya, V F Kucherov Izv. Akad. Nauk SSSR, Ser. Khim. 1064 (1980) a 109. Zh A Krasnaya, T S Stytsenko Izv. Akad. Nauk SSSR, Ser. Khim. 821 (1987) a 110. Zh A Krasnaya, T S Stytsenko, V S Bogdanov, N V Monich, M M Kul'chitskii, S V Pazenok, L M Yagupol'skii Izv. Akad. Nauk SSSR, Ser. Khim. 636 (1989) a 111. Ch Jutz, R-M Warner, A Kraatz, H-G LoÈ bering Liebigs Ann. Chem. 874 (1975) 112. Zh A Krasnaya, V S Bogdanov Izv. Akad. Nauk SSSR, Ser. Khim. 2348 (1991) a 113. Zh A Krasnaya, T S Stytsenko, E P Prokof'ev, V F Kucherov Izv. Akad. Nauk SSSR, Ser. Khim. 1362 (1980) a 114. Zh A Krasnaya, T S Stytsenko, B M Uzhinov, S A Krashakov, V S Bogdanov Izv. Akad. Nauk SSSR, Ser. Khim. 2084 (1983) a 115. A S Tatikolov, V A Kuz'min, Zh A Krasnaya, Yu V Smirnova Izv. Akad. Nauk, Ser. Khim. 1293 (1999) a 116. Zh A Krasnaya, T S Stytsenko Izv. Akad. Nauk SSSR, Ser. Khim. 855 (1983) a 117. Z Arnold, J Zemlicka Collect. Czech. Chem. Commun. 25 1302 (1960) 118. C Jutz, R M Wagner Angew. Chem. 84 299 (1972) 119. Yu L Slominskii, I D Radchenko Ukr. Khim. Zh. 43 263 (1977) 120. A I Pavlyuchenko, E I Kovshev, V V Titov Khim. Geterotsikl. Soedin. 85 (1981) d 121. V Nair, C S Cooper Tetrahedron Lett. 21 3155 (1980) 122. V Nair, C S Cooper J. Org. Chem. 46 4759 (1981) 123. V Nair, T S Jahnke Tetrahedron Lett. 25 3547 (1984) 124. V Nair, T S Jahnke Synthesis 424 (1984) 125. Zh A Krasnaya, Yu V Smirnova, A S Tatikolov, V S Bogdanov, E V Nikishova, V L Savel'ev Izv. Akad. Nauk, Ser. Khim. 537 (1995) a 126. Zh A Krasnaya, Yu V Smirnova, A S Tatikolov, V A Kuz'min Izv. Akad. Nauk, Ser. Khim. 1340 (1999) a 127. L K Denisov, B M Uzhinov Khim. Geterotsikl. Soedin. 723 (1981) d 128. B M Krasovitskii, G M Boloshin Organicheskie Lyuminofory (Organic Luminophores) (Moscow: Khimiya, 1984) 129. S Yudenfel'd Fluorestsentnyi Analiz v Biologii (Fluorescent Analysis in Biology) (Translated into Russian; Moscow: Mir, 1965) a�Russ. Chem. Bull. (Engl. Transl.) b�Russ. J. Org. Chem. (Engl. Transl. ) c�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) d�Chem. Heterocycl. Compd. (Engl. Transl.) e�Russ. J. Gen. Chem. (Engl

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

Russian Chemical Reviews 69 (12) 1037 ± 1056 (2000) Optical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov Contents I. Introduction II. Molecular and electronic structure of dioxins III. Vibrational states IV. Singlet electronically excited states V. Luminescence VI. Spectroscopic analytical investigations VII. Conclusion Abstract. theoretical and spectra optical the on data Published Published data on the optical spectra and theoretical investigations of derivatives polychlorinated dioxins, of investigations of dioxins, polychlorinated derivatives of dibenzo- dibenzo- p gener- are toxins, environmental well-known are which -dioxin, -dioxin, which are well-known environmental toxins, are gener- alised.the of structures electronic and geometrical The alised. The geometrical and electronic structures of the molecules molecules and are states excited electronically and vibrational their and their vibrational and electronically excited states are consid- consid- ered. chlorine of positions and number the of influence The ered. The influence of the number and positions of chlorine atoms atoms in UV and (IR spectra optical the on molecules dioxin in dioxin molecules on the optical spectra (IR andUV absorption, absorption, fluorescence on particular, in spectra), phosphorescence and fluorescence and phosphorescence spectra), in particular, on the the low-temperature of spectra phosphorescence fine-structure low-temperature fine-structure phosphorescence spectra of solid solid solutions triplet of deactivation photophysical the on and solutions and on the photophysical deactivation of triplet elec- elec- tronic of use the of results The discussed.are states tronic states are discussed. The results of the use of optical optical spectroscopy dioxins of quantification and identification for spectroscopy for identification and quantification of dioxins are are presented. 108 includes bibliography The presented. The bibliography includes 108 references. references. I. Introduction Compounds of the dibenzo-p-dioxin series, also called dioxins, became widely known and fell into disrepute after high toxicity of some polychlorinated dibenzo-p-dioxins (PCDD) had been estab- lished.1±5 Due to their high stability, dioxins formed in industrial processes can be retained in soil and water for long periods, enter living organisms and be accumulated in them.Due to the extreme hazard of dioxin ecotoxicants, investigation of their physicochem- ical properties and development of highly sensitive methods for their detection and quantification have become urgent. Several reviews devoted to the methods and results of deter- mination of PCDD in environmental objects and biological tissues have been published;6±10 however, they mainly discuss the data of GC/MS methods of analysis, while the data of spectroscopic studies used for analytical purposes are scanty. Since the mid-1980s, the physicochemical properties of PCDD have been studied more vigorously, in particular, by spectroscopic methods. These methods possess a great analytical potential; they can be used to identify various PCDD in different media and to perform quantitative analysis. Futhermore, spectroscopic meth- E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov State Scientific Centre of the Russian Federation `L Ya Karpov Institute of Physical Chemistry', Vorontsovo pole 10, 103064 Moscow, Russian Federation.Fax (7-095) 975 24 50. Tel. (7-095) 917 39 03 (1 21). E-mail: gast@cc.nifhi.ac.ru (E A Gastilovich) Received 3 May 2000 Uspekhi Khimii 69 (12) 1128 ± 1148 (2000); translated by Z P Bobkova #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n12ABEH000602 1037 1037 1039 1043 1046 1052 1054 ods can also be used to solve a fundamental problem, namely, to elucidate the relationship between the structures and physico- chemical properties of PCDD molecules.They provide valuable information about the interaction of the molecule with light, its steric and electronic structure, various types of intra- and inter- molecular interactions, etc. With the progress in elucidation of the structure and physicochemical properties of PCDD, the mecha- nism of toxic action and the pathways of deactivation of these toxins would gradually become clear. The present review is, in our opinion, the first attempt to survey the published data obtained by both experimental and theoretical methods of optical spectroscopy on the structure, vibrational and electronic states of dioxin molecules and the physicochemical properties of dioxins.Data from IR, Raman, electronic absorption, fluorescence and phosphorescence spectra are discussed, in particular, fine-structure electronic-vibrational (vibronic) phosphorescence spectra at temperatures of 77 and 4.2K are presented. The manifestation of vibronic and vibronic- spin-orbit couplings in the fine structure of vibronic spectra, the change of the nuclear configuration of molecules in the triplet excited state, the rate constants of radiative deactivation of singlet and triplet electronic states, and the effect of the number and the positions of chlorine atoms in the PCDD molecule on the phosphorescence spectra and the phosphorescence rate constants are discussed.The last Section contains the results of the use of optical spectra (UV and IR absorption, fluorescence, phosphor- escence) for identification of dioxins and determination of their quantity. II. Molecular and electronic structure of dioxins 1. Molecular structure Dibenzo-p-dioxin (DD) is an aromatic cyclic ether consisting of two benzene rings connected through oxygen bridges. 1 9 z (1) O 8 23 7 y (2) O 4 6 x (3) DD X-Ray diffraction data for DD11 ± 13 and for five of its chloro- derivatives [mono-, di-, tetra-, hexa-, and octachlorodioxins (OCDD) 14 ± 19] have been described in the literature. In the crystals, the dioxin molecules are nearly planar, although some atoms including Cl atoms deflect from the plane.Thus in di- 16 and1038 hexachlorodioxin molecules,17 the angle between the C7Cl bond and the plane amounts to 3 ± 4 8, while in mono-,14, 16 octa- 16, 18 and tetrachlorodioxins,15, 16, 19 this angle does not exceed 2 8. The C7Cl bond length in mono-, di- and hexachlorodioxin molecules varies from 1.69 to 1.75A. The geometric parameters of the central fragment of the molecules differ insignificantly: the C7O bond length is equal to 1.36 ± 1.38A and the angle between the C7O bonds (dCOC) is*116 8. Thus, the size of the COC angle in dioxins is intermediate between those in anisole (110 8) and diphenyl oxide (120 8).20 According to X-ray diffraction data, the central fragment of dioxin molecules can be considered to be planar.In the 2,3,7,8- tetrachlorodibenzo-p-dioxin (b-TCDD),{ the oxygen atoms deflect slightly from the plane through the four carbon atoms in different directions, thus forming a `chair' configuration of the central fragment.16 However, the angle of deflection is only 0.8 8. In the DD molecule,11 all six atoms of the central fragment lie in one plane but this plane as a whole is somewhat (by 0.5 8) rotated around axis 1. The C7O bond length in dibenzo-p-dioxin is 1.383A, the internal and external CCO angles are 122 8 and 118 8, and the central COC angle is 116 8. Note that the geometric parameters determined from X-ray diffraction data refer to molecules forming the crystal lattice.No experimental data for these parameters in free dioxin molecules have been reported in the literature. In several studies, the geometric structures of dioxins have been determined theoretically, the geometry being optimised using semiempirical 21 and ab initio 22, 23 calculations. According to some data,21, 22 the DD and b-TCDD molecules are planar and the calculated values of bond lengths and angles are close to those found experimentally.15 According to another publication,23 the DD molecule is non-planar: the benzene rings are rotated with respect to the axis which passes through oxygen atoms (axis z) to form a dihedral angle equal to 167.93 8 (subsequently, this molecular configuration is called a `butterfly'). According to a calculation 23 carried out in the same approximation, the cation derived from the DD molecule is planar in the ground electronic state.In an earlier publication,21 DD has also been suggested to have a `butterfly' configuration. Thus the dipole moment of a solution of DD in benzene was found to be 0.55D (rather than zero). The molecule was modelled as the `butterfly' configuration and the dihedral angle was estimated from the dipole moment. It was found to be 164 8. Note that this conclusion cannot be regarded as unambiguous because the molecule could acquire a dipole moment due to interaction with the solvent without losing planarity; for example, it could occur as a configuration with C2u symmetry with retention of the C2(z) symmetry axis.The problem of the structure of dioxins attracts attention because the toxicity of PCDD is supposed to be correlated with either planar or non-planar structure of their molecules.24 Grainger et al.24 proposed determination of the molecular config- uration on the basis of the dCOC angle: if this angle is 109 ± 111 8, the molecular is non-planar. For some PCDD, the COC angles were found 24 using approximate calculation procedures based on a triatomic model and using spectroscopic data on the frequency of the stretching antisymmetrical vibrations of C7O7C groups. It was concluded that the PCDD molecule becomes more planar as the number of a-chlorine atoms decreases. However, this conclusion should be treated critically because the triatomic model is too rough to provide an adequate description of the vibrations of a polyatomic molecule.The authors of most spectroscopic studies of dioxins give preference to the planar configuration, relying on the experimen- tal data on molecular geometry; however, some researchers decide in favour of the `butterfly'. { In considering the spectral properties of polyaromatic derivatives, for example, substituted naphthalenes, anthracenes, 9,10-anthraquinones and so on, the positions of substituents are designated by Greek letters. E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov The highest-symmetry group to which planar dioxin mole- cules can belong is the D2h group (this group of symmetry corresponds to the planar molecules of DD, b-TCDD, 1,4,6,9-Cl4DD and OCDD).A molecule with the `butterfly' configuration has C2u symmetry in which the following symmetry elements are retained: the C2(x) axis and two symmetry planes s1 and s2 , perpendicular to the Cartesian coordinates z and y, respectively. The C2u(D2h) groups of symmetry are matched by the following correlation of irreversible representations: (1) A1(Ag+B3u), B2(B2u+B1g), B1(B1u+B2g), A2(B3g+Au). To answer the question about the configuration of the dioxin molecules mentioned above, it is expedient to use correlation (1) and employ selection rules when analysing spectroscopic data (see Sections III and IV). 2. Electronic structure of molecules The ground electronic state (S0) of the planar DD molecule is characterised by the following charges (re) on the atoms of the central fragment, oxygen and four carbons: re(O)&70.37 and re(C)&+0.22 (in the electron charge units) on each of equivalent atoms.The calculation was carried out by the INDO/S semi- empirical approximation. The charges on the C7H group atoms are small in magnitude (*0.05 ± 0.09 units) and have opposite signs (for carbon re<0). The presence of four Cl atoms in the planar b-TCDD molecule does not change the charge distribution in the central fragment with respect to that in DD. Charge redistribution in b-TCDD takes place on the atoms of the C7Cl groups � the negative charge on each chlorine atom [re(Cl)&70.12] reverses the sign of the charge on the corresponding b-carbon atoms [re(C)&+0.11]. The changes in the charge on the C7H group atoms are reduced to a slight increase in their magnitude re (to *0.08 ± 0.12), their signs being retained.The same pattern of charge distribution is observed for OCDD: re(O)&70.34, re(C)&+0.20 in the central fragment, while for C7Cl groups with the chlorine atoms in the a- and b-positions re(Cl)&7(0.09 ± 0.10) and re(C)&+(0.07 ± 0.09). Thus, an increase in the number of Cl atoms in dioxin molecules (from zero to eight) barely influences the charge distribution on the atoms of the central six-membered fragment of the molecules in the ground electronic state S0 . A theoretical study 25 of the electron-donating properties of DD, b-TCDD and 1,4,6,9-Cl4DD molecules has also demonstrated that the p-donor properties of these molecules (as well as their geometric structure) depend only slightly on the positions of Cl atoms. An important characteristic of the ground electronic state S0 of a molecule is the ionisation potential (IP).The first IP for several dioxin molecules (DD, 2-ClDD, 2,8-Cl2DD) cooled in a supersonic jet of argon were determined experimentally by the resonance-enhanced two-colour two-photon ionisation techni- que.23, 26 For nine dioxins containing zero to eight chlorine atoms in the molecule, a semiempirical calculation 23 of the IP values in the Koopmans approximation was performed. It was found that, as the number of chlorine atoms in the molecule increases, the IP value increases monotonically by approximately 0.1 eV for each additional Cl atom.Indeed, the IP for DD, b-TCDD and OCDD are 7.58, 7.99 and 8.42 eV, respectively. It is of interest that the pattern of dependence of the IP on the number of Cl atoms in dioxins differs substantially from that observed in the series of chloro-substituted benzene derivatives in which the number of Cl atoms ranges from zero to six.23, 27 In the case of chlorobenzenes, the IP value barely changes upon an increase in the number of Cl atoms (9.25 and 9.19 eV for benzene and hexachlorobenzene, respectively). This difference between the patterns of IP variation for these series of compounds was attributed 23 to the presence of oxygen in dioxin molecules. According to the Koopmans approximation, the IP values are determined by the energy of the highest occupied molecular orbital (with allowance for an adjusting parameter selectedOptical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives 70.22 0.00 0.40 70.20 B1g(B2) p*-MO50 0.00 0.00 70.35 0.36 Au(A2) p*-MO49 70.36 0.09 70.27 0.32 B3u(A1) p-MO48 0.41 70.21 0.21 0.14 B2g(B1) p-MO47 Figure 1.Classification of four MO of the DD molecule according to irreducible representations of the D2h(C2u) groups and their representa- tion as a LCAO. The numerical values for the coefficients 28 at the 2px-AO are given; the dark and light circles correspond to different signs of the coefficients at these AO.23 empirically for a particular series of molecules, for example, for dioxins 23).Figure 1 shows a p-MO (MO48) of the DD molecule for which the `butterfly' configuration has been accepted.23 From representation of the structure of this MO48 as a linear combina- tion of atomic orbitals (AO), one can see that the oxygen atoms can actually affect substantially the IP in the DD molecule. In connection with the results of studies of the dioxin IP, it is of interest to mention the data obtained in the calculations of the electronic states in the planar dioxin molecules 28, 29 concerning the influence of Cl atoms on the structure of the highest occupied p-MO(B3u), which determines the theoretical IP value. It was found that for this MO, the coefficients at the 2px-AO centered at theOandC atoms in the b-TCDD andOCDDmolecules withD2h symmetry differ only slightly (by hundredths of a unit) from the corresponding values for the DD molecule (see MO48 in Fig.1). For OCDD, the coefficients at the 2px-AO of the chlorine atoms proved to be relatively low: +0.13 for b-Cl and 70.03 for a-Cl. Since the coefficients at the 2px-AO of oxygen for MO in DD, b-TCDD and OCDD do not depend (with an accuracy of<3%) on the number of chlorine atoms in the molecule, in our opinion, the difference between the pattes of variation of IP in chlor- obenzenes and dioxins can hardly be definitely attributed to the influence of the O atoms, as has been proposed in a publication.23 Figure 1 also shows characteristics of a lower occupiedMO47, the removal of an electron from which 23 transforms the DD molecule into a cation, as well as characteristics of the vacant MO49 and MO50 .The electronic configurations corresponding to the transition of an electron from MO48 to MO50 and from MO47 to MO49 determine the properties mainly of the lower singlet electronically excited state S1 having B2(B2u) symmetry. This is a resonance state in two-photon two-colour resonance-enhanced ionisation experiments.23 Analysis of the signs of the coefficients at 2px-AO (see Fig. 1) centred on the atoms of the C7O7C groups for fourMOled the researchers to the conclusion23 that the DD molecule has different configurations in different electronic states. For instance, on passing from the S0 state with a `butterfly' configuration to the cationic form or to the S1 state, the DD molecule is believed to become flatter, due to the increase in the bond orders in these states.However, the arguments presented in the study in question 23 seem inconclusive. The spectroscopic data concerning the problem of the configuration of the DD molecule 1039 in different electronic states (S0 and S1) are presented in Section IV. III. Vibrational states 1. Experimental data The molecules of dioxins contain 22 atoms and have 60 normal modes. In planar molecules, 41 modes are in-plane vibrations and 19 modes are out-of-plane vibrations. In the infrared, Raman, fluorescence and phosphorescence spectra, which display the vibrations of molecules in the ground electronic state, only vibrations with particular types of symmetry can be observed, depending on the symmetry of the molecule, in accordance with the selection rules.In molecules with D2h symmetry, vibrations are classified into eight representations (types of symmetry) of this group: 11Ag+10B3g+10B1u+10B2u+4B1g+5B2g+5Au+5B3u . The last four types are out-of-plane vibrations (which nor- mally correspond to frequencies of < 1000 cm71). Vibrations with g type of symmetry are active in the Raman spectra, while u-type vibrations (except for Au) are IR active. The vibrations active in the phosphorescence spectra are totally symmetric Ag and non-totally symmetric vibrations of either g or u type, depending on the type of symmetry of the electronically excited state.Of the set of non-totally symmetric vibrations, out-of-plane vibrations are usually responsible for the most intense lines in the pp* phosphorescence spectra of heteroaromatic compounds 30 (it is these spectra that are typical of dioxins). Thus, in order to gain exhaustive information concerning the vibrations of dioxin molecules with a symmetry higher than Cs, various spectroscopic methods should be used. Meanwhile, IR spectra of dioxins have been studied most extensively due to the necessity of identifying these hazardous compounds. The molecules of chlorinated dioxins contain heavy Cl atoms; the frequencies of the fundamental vibrations of these molecules vary 22 over a broad range,*3300 ± 20 cm71.However, as a rule, the IR spectra of dioxins are measured 2, 24, 31 ± 37 in a narrower frequency range, namely, above 800 cm71. Data on the IR vibration frequencies of PCDD molecules in KBr pellets,2, 31 in the gas phase,24, 34 and under conditions of matrix isolation (argon matrix at 12 K) 33, 35 ± 37 have been reported. In particular, the spectra of 22 isomers of tetrachloro-substituted DD in an argon matrix and in the vapour phase were recorded.33 The attention was focused on the IR spectra of the most toxic b-TCDD.2, 32 ± 34 The IR spectrum of this compound in a KBr pellet was measured 2 in a wider frequency range (4000 ± 375 cm71) than the spectra of other tetrachloro-substituted DD. Figure 2 shows a fragment of this spectrum.37 0 20 40 1400 1200 1000 800 400 600 n /cm71 Figure 2.IR absorption spectrum of b-TCDD in a KBr pellet.2 The arrows mark the spectral regions in which analytical lines were identified in a study with matrix isolation.37 Transmission (%) 60 80 1001040The researchers were especially interested in the fact 22, 34 that the IR spectra of tetrachloro-substituted DD recorded under various experimental conditions exhibited doublet splitting of some lines. In the case of b-TCDD, this splitting is observed in the region of 1480, 1315 and 1115 cm71 (see Fig. 2). The results of a theoretical study of the possible causes of this effect in b-TCDD gave rise to the conclusion 22 that Fermi resonance is the main reason.If the molecular configuration of any dioxin has an inversion centre, only about a half of the fundamental modes are displayed in the IR spectra, while the other frequencies are active in the Raman spectra. The most comprehensive IR and Raman data have been reported for DD38, 39 and OCDD40 (Raman spectra of OCDD are shown in Fig. 3). For other PCDD, to the best of our knowledge, no data on the Raman spectra are available. Irel q k x rj Zd Za c* 60 120 180 l m t* rp* o u e2 ZdZac* 1200 800 400 1600 Dn /cm71 Figure 3. Raman spectrum of the OCDD powder, lexcit=488 nm.40 The spectral lines are marked by the letter designations of vibrations used in Tables 1 and 2. The asterisk marks the vibration frequencies for which a different interpretation is possible (see Table 4).The frequencies of the fundamental modes of the molecules of some dioxins [DD, b-TCDD, 1,3,6,8-tetrachlorodibenzo-p-dioxin (a,b-TCDD), OCDD] in the ground electronic state were also determined by analysis of the fine-structure phosphorescence spectra 39, 41 ± 43 (the results of analysis of these spectra are outlined in Section V). For DD, 2,7-Cl2DD, 2,8-Cl2DD and 2,3-Cl2DD molecules cooled in a supersonic jet, the vibrational structure of the first singlet electronically excited (S1) state was obtained using the resonance two-photon ionisation method.23, 26, 44 ± 46 On the basis of a study of the IR spectra, an empirical interpretation of a number of absorption bands of dioxins has been proposed.2, 24, 31, 37 As an example, we cite the empirical interpretation of seven frequencies corresponding to the analytical lines in the b-TCDD spectrum (see Fig.2):37 1575, 1489, 1471 cm71 are skeletal in-plane vibrations of the C=C groups in the aromatic fragments; 1312 cm71 is an antisymmetric stretching vibration of the C7O7C groups; 1176 and 1117 cm71 are in-plane deformation and stretching (breathing) vibrations of the aromatic fragments; and 879 cm71 is an out-of- plane vibration of the C7H groups. Studies of the vibrations of dioxins resorting to the data of theoretical calculations 22, 39 ± 41 have shown that this assignment has to be corrected. 2. Calculation and interpretation of vibrations The calculations for vibrations in the harmonic approximation were performed with the assumption of a planar configuration of the molecule.For several tetrachloro-substituted DD isomers (b-TCDD, 1,4,6,9-Cl4DD, 1,3,7,8-Cl4DD, 1,4,7,8-Cl4DD) 22 and a number of hepta- and hexa-chloro-substitutedPCDDisomers,47 an ab initio quantum-chemical calculation of vibrations was performed using scaling factors. For a number of molecules E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov having the same D2h symmetry but different numbers of chlorine atoms, DD, b-TCDD and OCDD, the normal modes were calculated 39 ± 41 by the classical method.48 When applying this method to dioxins, the potential field of a molecule with a similar structure (xanthone) was chosen as the base.49, 50 As the hydrogen atoms were replaced by chlorine atoms, the increments 39 corre- sponding to the force constants of the C7C(Cl)7C groups in chlorinated benzenes were introduced without variations. Scaling factors for some natural vibrational coordinates of related mole- cules were used in the ab initio calculation;22, 51 for PCDD, they were not varied either.The results of calculations were used to determine the frequencies of all the normal modes of the dioxins studied;22, 39 ± 41 an interpretation for each vibration was proposed. In view of the fact that each vibration of the molecule usually has a complex form (combination of the amplitudes of variation of all bond lengths and bond angles in the molecule), different approaches to the interpretation have been proposed;50, 51 they were presented, for example, in Refs 22 and 40.One approach 51 makes use of the results of calculations of the distribution of the total energy of vibrations over atomic groups. In this case, each calculated vibration frequency in dioxins 22 is usually matched to vibrations of several atomic groups. For instance, the intense band with a maximum at about 1480 cm71 (see Fig. 2) in the IR spectrum of b-TCDD was assigned 22 to the skeletal vibrations of the aromatic fragments and the deformation vibrations of the CH groups in the plane of the molecule. In another approach,50 for each calculated normal mode of dioxins,40 data on the local type of symmetry were taken into account [within the framework of the D6h(D2h) groups of symmetry for the movement of atoms in the separate six-membered fragment of the molecule 50], in addi- tion to the relative vibration amplitudes of separate groups of atoms (Table 1).In Table 1, the vibration modes were assigned to the vibrations of particular atomic groups ± the benzene fragment Table 1. Letter designations 40 of the vibrations of dioxins and their assignment. Designation of vibrations a Atomic group In-plane vibrations C=C stretching (BF) k [e2g(ag)], l [e2g(b3g)], m [e1u(b1u)], n [e1u(b2u)], o [b2u(b2u)], r [a1g(ag)] p [b1u(b1u)], s [e2g(b3g)], t [e2g(ag)] CCC deformation (BF) q (ag), q (b2u), z (b3g), z (b1u) u (ag), c (b1u), c (b3g) CO stretching (CF) COC, OCC deformation (CF) a [b2u(b2u)], b [e1u(b1u)], d [e1u(b2u)], e [a2g(b3g)] CCH deformation b e CCl stretching Z CCCl deformation c Out-of-plane (deformation) vibrations u [b2g(b2g)], y [e2u(b3u)], o [e2u(au)] CCCC (BF) f [e2u(b3u)], g [a2u(b3u)], i [e1g(b2g)], j [b2g(b2g)] CHb r CCl c w (b2g), w (b3u) CCOC (CF) x (b1g), x (b3u), h (au) BF+CFd Note.BF and CF are the benzene and central fragments. a The local types of symmetry [G(G0)] of the D6h (D2h) point group are given in brackets. b For terminal atomic groups, the designation [G(G0)] has a meaning if all their four atoms in a fragment are identical. c For C7Cl groups, subscripts can be added to the designations Z and r corresponding to the letter designations for the C7Hgroups (to provide a more detailed interpretation of vibrations) if the fragment contains four chlorine atoms.d Vibrations of the benzene fragments mutual with respect to one another.Optical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives (local symmetry D6h) and the central fragment (local symmetry D2h) and to vibrations of the terminal (C7H and C7Cl) atomic groups. A more detailed interpretation of vibrations according to their form 40 was performed taking into account the definite alternation of the signs of the amplitudes characterising the changes in the bond lengths and angles in the given molecular fragment; this characteristic is reflected in a definite local (for each fragment) type of symmetry of theD6h(D2h) point group. This type of analysis of the forms of vibrations makes it possible to distinguish between normal modes assigned roughly to the same atomic groups (for example, to six-membered carbon fragments).This individual form of a normal mode of dioxin was designated by a special letter symbol (in particular, the known Whiffen designations of the vibrations of benzene and monosubstituted benzene derivatives were used). Tables 2 ± 5 contain the calculated values for all the vibration frequencies 22, 40 in the DD, b-TCDD and OCDD molecules distributed over types of symmetry of the D2h point group, together with the corresponding experimental values found from vibrational spectra (IR and Raman) or fine-structure phosphor- escence spectra.In the case of b-TCDD, the calculated vibration frequencies determined by different methods 22, 40 are in fairly good agreement; however, the use of different methods for the Table 2. Frequencies of the in-plane vibrations of the DD and PCDD molecules calculated theoretically and obtained experimentally from the Raman and phosphorescence spectra. Design- ations of vibrations Vibrations of Ag symmetry knoqae1 e2 drtuZd Za Vibrations of B3g symmetry lmzebe3 e4 pscZb Ze Note. From here on, vibration frequencies are given in cm71; experimental vibration frequencies determined from Raman (R) or IR and phosphor- escence (phos) spectra are given in parentheses. The asterisks mark the unambiguous assignments of frequencies to in-plane or out-of-plane vibrations.The Table does not contain the stretching vibrations of the C7H groups. The pairs of designations e1, e2 or e3, e4 imply different signs of the amplitudes of the changes in the C7Cl bond lengths in each BF. a From the data of Ref. 40. b From the data of Ref. 39. c From the data of Ref. 22. nDD (R) a nb-TCDD (phos) a, b nb-TCDD c nOCDD (R) a 1586 (1600) 1449 (1442) 1296 (1284) 1222 (1236) 1119 (1100) 1602 (1621) 1470 (1500) 1310 1208 (1226) 1152 (1154) 662 (654) 1023 (1030) 735 (726) 613 (564) 363 (398) 745 (739) 232 (206) 586 (548) 167 (149) 1596 1380 1209 1263 1610 (1586) 1452 1317 1254 (1192) 1112 (1094) 883 (898*) 547 (534) 477 (450*) 972 (985) 595 637 (628*) 236 383 (383*) Table 3.Frequencies of the in-plane vibrations of DD and PCDD molecules calculated theoretically and found experimentally from IR spectra. Design- ations of vibrations Vibrations of B2u symmetry knoqae1 e2 drtZd Za Vibrations of B1u symmetry lmzebe3 e4 pscZb Ze 864 (850) 996 (1003) 443 332 83 178 699 (666) a According to Ref. 40. b According to Ref. 39. c According to Ref. 22. 1566 (1590) 1346 1281 (1280) 1256 (1250) 1595 1483 1296 1226 1111 898 733 (754*) 661 analysis of the forms of vibrations 50, 51 precludes a proper comparison of various approaches to the interpretation. It is worth noting that the approach 50 taking into account the local symmetry of vibrations (unlike that described in Refs 22 and 51) allows correlation of the vibrations of molecules which belong 741 254 555 183 325 (345) 242 (274*) 601 (594) 146 (162) 169 (181) Table 4.Frequencies of the out-of-plane vibrations of DD and PCDD molecules calculated theoretically and found experimentally from Raman or phosphorescence spectra. 1578 (1550) 1392 (1425) 1218 Design- ations of vibrations 1574 1358 1154 1218 Vibrations of B2g symmetry 986 538 642 227 391 730 974 324 (358*) 525 198 (220*) 162 786 jiri rj wuo 533 725 (748) 258 (278) Vibrations of B1g symmetry f 943 895 (867) 874 g 744 (764) rg 322 (383*) 372 140 rf 323 (274*) y 430 (450*) 435 (441) 437 570 x 223 (242) 119 134 67 (80) a According to Ref.40. b According to Ref. 39. c According to Ref. 22. nDD a nb-TCDD a, b 1564 (1568) 1446 (1456) 1271 (1312) 1290 (1326) 1584 (1590) 1466 (1495) 1279 (1300) 1293 (1395) 1151 (1154) 1106 (1114) 864 (876) 432 (450) 680 (680) 1023 (1030) 827 (830) 598 (610) 177 1610 1345 (1393) 1186 (1172) 1267 (1242) 1622 (1628) 1457 (1462) 1196 (1200) 1327 (1305) 1147 (1120) 914 (876) 883 (854) 676 245 834 (789) 303 119 512 (500*) nDD (R) a nb-TCDD (phos) a, b nb-TCDD c nOCDD (R) a 890 977 861 (898*) 115 600 (628*) 730 309 1041 nOCDD a nb-TCDD c 1543 (1550) 1366 (1390) 1256 (1298) 1279 (1320) 1557 1470 1269 1303 813 (840) 979 (987) 1107 870 445 686 452 308 203 168 198 1594 1388 (1423) 1194 (1196) 1629 1387 1170 1234 930 786 326 111 497 846 110 358 85 (97) 601 754 (754*) 199 (220*) 610 648 2771042 Table 5.Frequencies of the out-of-plane vibrations of DD and PCDD molecules calculated theoretically and found experimentally from IR spectra.nOCDD a nDD a nb-TCDD c nb-TCDD a, b Design- ations of vibrations Vibrations of Au symmetry 845 891 977 861 216 228 334 91 767 602 38 629 610 48 744 599 55 741 521 144 jiri rj uohVibrations of B3u symmetry 875 899 (890) 945 (922) 750 (743) 168 205 444 383 19 515 (500*) 403 (390) 33 511 (450) 357 70 fgrg rf wyx 89 213 606 449 31 a According to Ref.40. b According to Ref. 39. c According to Ref. 22. to different point groups of symmetry. For example, this approach was used to perform a correlation of all the vibrations in mono-, di- and tetrachloro-substituted 9,10-anthraquinones.52 In addi- tion, in experimental and theoretical studies of the vibrations of the DD and b-TCDD molecules as well as 9,10-anthraquinone and tetra-b-chloro-substituted 9,10-anthraquinone molecules, it has been established that the frequencies and forms of vibrations for these molecules are very close.39 Therefore, interpretation of the vibrations of chlorinated 9,10-anthraquinones, which are more readily available, can be taken as the base for the inves- tigation of the corresponding dioxins.52 The variation of frequencies of all 60 vibrations following an increase in the number of Cl atoms in the series of DD, b-TCDD and OCDD, the molecules of which correspond to the same group of symmetry D2h, was discussed in detail.40 We shall briefly note some regularities 40 of the variation of the frequencies of vibra- tions with the same form in the DD, b-TCDD and OCDD molecules.In-plane vibrations. The high-frequency in-plane vibrations of the carbon skeleton (in the range of 1650 ± 1200 cm71), which can be assigned to the k, l, m, n, o stretching modes (of the benzene fragments) and q, z stretching modes (of the central fragment), change insignificantly, on the average by 5%, as the H atoms are replaced by Cl atoms (see Tables 2, 3). In the case of the two other stretching vibrations (Ag and B2u) of the benzene fragments, designated by r (which should be matched to the only totally symmetric vibration r of the core of benzene molecule in the region of 995 cm71), the vibration frequencies found forDDandOCDD differ appreciably, by a factor of about two.Note that the vibrations r are characterised by low frequencies also in other molecules with benzene side fragments, even when they contain no heavy atoms.50 For example,53 in anthracene nr=753 (Ag) and 808 cm71 (B2u). The frequencies of all [except for one, s (B3g)] low-frequency deformation vibrations of the carbon skeleton of the molecule (p, s, t, c, u) also change substantially, by approximately 200 ± 300 cm71 (by up to 65%), when all the H atoms have been replaced by Cl atoms.The stretching vibrations of the C7Cl (e) groups of the OCDD molecule were matched to frequencies in the region of 1000 ± 750 cm71. In many studies, considerable attention is devoted to the identification of the IR bands caused by the presence of the E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov C7O ether groups in the molecules.20 For example, in the case of diphenyl oxide, which contains a C7O7C group, the fre- quency nas C¡O=1236 cm71 was assigned to antisymmetric stretch- ing vibrations of this group. In molecules with D2h symmetry, the q(B2u) and z(B3g) modes can be matched to these vibrations (see Tables 2, 3).In various PCDD, frequencies in the region of 1330 ± 1280 cm71 were attributed to an antisymmetric vibration. It is noteworthy that the vibration frequencies of the C7C groups in the anthracene molecule with the rigid six-membered central fragment are close to the vibration frequencies of the C7O ether groups in DD. The frequencies (in cm71) of the stretching vibrations of the C7O or C7C groups of the central fragment of DD, 9,10-anthraquinone and anthracene are presented below for comparison.50, 53 DD Anthracene 9,10-Anthraquinone Ag B3g B1u 1260 1376 1274 1398 1148 1284 1335 1173 1226 1317 1200 1395 B2u In the DD molecule, as in other molecules having a similar structure and containing eight H atoms in the benzene fragments, each of the eight in-plane deformation vibrations of the C7C7H groups belongs to a particular frequency region (see Tables 2, 3), namely, 1360 ± 1240 cm71 (e are high-frequency vibrations) and 1150 ± 1000 cm71 (a, b, d is a group of relatively low-frequency vibrations). For similar deformation vibrations of the C7C7Cl (Z) groups of the OCDD molecule, two appreciably different frequency regions can also be distinguished. The range of 800 ± 700 cm71 is typical of the analogues of the vibrations e, while 200 ± 150 cm71 is the range where the analogues of the vibrations a, b and d show themselves.The empirical interpretation of the spectrum does not always coincide with the interpretation based on calculations. Below we present the assignment of vibrations according to their form based on the use of the local types of symmetry for several vibration frequencies of the b-TCDD molecule, which differs from the empirical assignment (presented at the end of Section III.1).37 Thus the frequency of 1312 cm71 (B2u) refers to the vibration o of the benzene fragments (with the local type of symmetry b2u); 1172 (B1u) and 1114 cm71 (B2u) correspond to the stretching vibrations z of the central fragment and e of the C7Cl groups, respectively.Afrequency of 450 cm71 was assigned to the `breathing vibration' r of the benzene molecules (the vibration B2u with the local symmetry a1g), while a frequency of 1326 cm71 was attributed to the antisymmetric vibration of the C7O7C groups.Out-of-plane vibrations. The main distinctive feature of most of the 11 out-of-plane vibrations of the carbon skeleton of the DD, b-TCDD and OCDD molecules is that their frequencies increase rather than decrease upon an increase in the number of Cl atoms in the molecule (see Tables 4, 5). Indeed, in the region above 300 cm71, the frequencies of only two, the highest-frequency out-of-plane vibrations of OCDD (of the u type, assigned to the benzene fragments) change relatively slightly (by 3%± 4%) with respect to the frequencies of the same vibrations in b-TCDD and DD. The other vibration frequencies increase appreciably, by 13%± 33%, which implies displacement of the bands in the spectrum by 70 ± 140 cm71.A different pattern of variation is observed for the other four lower-frequency vibrations of the molecular skeleton (below 300 cm71), namely, one vibration o, which was assigned to the benzene moieties, and three vibrations x and h, assigned to the central moiety. These frequencies, conversely, tend to decrease (by 25% to 75%). It should be mentioned that the vibrations x and h can be related to the rotations of two benzene fragments with respect to one another (Fig. 4). The molecule thus assumes a `butterfly' (x, B3u), `chair' (x, B1g) or `propeller' (h, Au) configuration. For comparison, Fig. 4 shows the form of the high-frequency out-of-Optical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives x (B3u) w (B3u) x (B1g) h (Au) Figure 4.Forms of some out-of-plane vibrations of the DD molecule of D2h symmetry (for vibration frequencies, see Tables 4, 5). The signs and the magnitudes of the displacements of atoms with respect to the plane of the molecule are marked by circles of different sizes and colours. plane vibration w (B3u). It can be seen that both x and w vibrations of the B3u type can be attributed to the molecule in the `butterfly' configuration but x is related to a larger extent to the rotations of the benzene fragments and w is related to the shifts of oxygen atoms. The four out-of-plane vibrations of the C7Cl groups (r) of the b-TCDD molecule show themselves at low frequencies, about *100, 215 and 320 cm71.An increase in the number of chlorine atoms does not result in any significant increase in the frequency range for the eight vibrations r of the OCDD molecule; these vibrations remain in approximately the same frequency regions: *90 ± 140,*215 and*360 cm71. Note that the eight out-of-plane vibrations of the C7C7H groups in DD, like those in other molecules having a similar structure and containing eight H atoms in the benzene fragments, correspond to a definite frequency region: 980 ± 940 cm71 (j, f ), *860 cm71 (i), 750 ± 680 cm71 (g). However, it was found that similar vibrations of theC7C7Cl groups (r with the correspond- ing subscripts) of the OCDD molecule cannot be matched successively to higher or lower frequencies (see Tables 4, 5).For example, both rj and rg vibrations in OCDD are characterised by low frequencies. Molecular configuration. To decide between the planar and non-planar (`butterfly') configurations of the DD and OCDD molecules, we shall consider the structure of their Raman spectra taking into account the selection rules for the D2h and C2u groups of symmetry and the correlation of their irreducible representa- tions (1). If the molecule has the `butterfly' configuration, i.e. belongs to the C2u symmetry group, all the vibrations should be active in the Raman spectra, including the u-type vibrations, which are forbidden in the case of the D2h group [see expression (1)]. In particular, the out-of-plane B3u vibrations of theD2h group become totally symmetric.The results of interpretation of the Raman spectra of two dioxins, DD and OCDD, demonstrate that all lines in the spectra of these compounds can be assigned to g-type vibrations, which implies a planar configuration of the molecule (see Fig. 3, Tables 2, 4). IV. Singlet electronically excited states 1. Experimental data Dioxins, like most benzene derivatives, are colourless, i.e. they do not absorb light in the visible region. The UV absorption spectra of dioxins [Sm(pp*)/S0] have been studied in solution 2, 54 ± 57 and in the vapour phase.58 Tables listing the absorption maxima 1043 and the molar extinction coefficients (emax) have been reported;2 the absorption curves for DD and its chloro-derivatives (b-TCDD, 1,2,3,7,8-Cl5DD and OCDD) can be found in several publications.55, 56, 58 The absorption spectra of DD and PCDD occur in the region of 200 ± 350 nm and consist of two bands � a weak broad long- wavelength band (A) in the region of 280 ± 320 nm with a maximum at *300 ± 304 nm and a short-wavelength narrow intense band (B) with a maximum at about *220 ± 240 nm.It can be seen from the absorption curves of some dioxins 56, 58 that the high-frequency absorption band has one more, even shorter- wavelength maximum (C); for example, in the spectra of solutions of DD, it is separated from band B by *5.5 nm.55 As examples, Figs 5 and 6 show the absorption spectra of a solution ofDDand of the b-TCDD vapour.1074e /litre mol71 cm71 4 1 3 B1u (0.213)1B2u (0.504) 2 1B1u (0.018) 1 1 1 1B3g 1B3g 1 Ag B2u (0.014) Ag 200 250 300 l /nm Figure 5. Electronic absorption spectrum of a solution of DD in iso- pentane at 77K56 and results of a quantum-chemical calculation 29 (vertical lines) of the energy and the type of symmetry of electronic pp* states and the oscillator strengths of transitions (given in parentheses). Dipole-inactive transitions are marked by arrows. The calculated spec- trum as a whole is shifted by*5000 cm71 to lower frequencies. The ratio of the molar extinction coefficients of bands B and A, eB / eA, for solutions of DD in isopentane 56 is equal to *10, while for b-TCDD and OCDD vapours,58 eB / eA=9 and 35. In particular,58 for OCDD, eA=0.056104 and eB= 1.736104 litre mol71 cm71.It should be noted that the data on the molar extinction coefficients are presented in few publications and, moreover, the reported e values are not always close in magnitude. For example, for DD in isopentane,57 chloroform 2 and heptane,21 eA=0.376104, 0.376104 and 0.746104 litre mol71 cm71; for b-TCDD in chloroform 2 and in the vapour phase,58 eA=0.56 and 0.416104 litre mol71 cm71. The eA values reported for vapour and for chloroform solutions of OCDD2, 58 differ by a factor of five. These differences may be due to both the interaction of dioxins with a solvent and the error in the determination of the true concentration of dioxins in the samples studied. The spectra of chloroform solutions of DD and several PCDD2 showed an unusual intensity ratio of the absorption bands, namely, the short-wavelength band B was found to be less intense than band A.For example, eB / eA=0.28, 0.53 and 5.5 for DD, b-TCDD and OCDD [for comparison, for DD in isopentane, eB / eA&10 (see Ref. 56)]. In our opinion, this effect is caused by the disturbing influence of the solvent, which also absorbs in the short-wavelength region of the spectrum (<240 nm).59 The absorption properties of DD are similar to those of other aromatic ethers. The spectra of ethers usually contain two absorption bands; the long-wavelength band with e&(0.16 ± 0.3)6104 litre mol71 cm71 is located at about10441074e /litre mol71 cm71 4 1B2u (0.569) 3 1B1u (0.470) 21 1B3g 220 260 Figure 6.Electronic absorption spectrumof b-TCDD vapours 58 and the results of a quantum-chemical calculation 29 of the energy and type of symmetry of electronic pp* states and the oscillator strengths of tran- sitions. For designations, see Fig. 5. 250 ± 290 nm and the short-wavelength band with e& (0.8 ± 1.8)6104 litre mol71 cm71 occurs in the region of 210 ± 240 nm.20 Both bands are associated with transitions from the electronic states corresponding to the benzene chromophore perturbed by the heteroatom of the ether group. As the number of chlorine atoms in dioxin molecules increases from zero to eight, the long-wavelength band in the absorption spectra of solutions shifts bathochromically.2 Thus in the spectra of chloroform solutions of DD, b-TCDD and OCDD, bands A are responsible for maxima at 293, 310 and 318 nm.However, in the case of vapours, the positions of the maxima of bands A and B virtually do not change�in the series b-TCDD, 1,2,3,7,8-Cl5DD and OCDD, band A has a maximum at about 303 ± 305 nm, while the maximum of band B is at*223 nm.58 In the UV absorption spectra ofDDin isopentane recorded at 77 K, the long-wavelength band has several maxima,56 which can be attributed to the vibrational structure. The vibrational fine structure of the electronically excited state S1 for the DD, 2,7- Cl2DD, 2,8-Cl2DD and b-TCDD molecules was obtained using the method of resonance-enhanced two-photon ionisation in a supersonic argon jet.44 ± 46 The vibrational structure of the S1 state of the DD molecule 23 is shown in Fig.7. In this type of spectrum of any dioxin, the first (lowest-frequency) line was assigned to the purely electronic transition S1/S0 . As a result, the following energies of the S1 state were determined for these dioxins:23 33774 (DD), 32805 (2,7-Cl2DD), 32729 (2,8-Cl2DD) and 32100 cm71 (b-TCDD). These frequencies are shifted by *750 cm71 to shorter wavelengths 23 with respect to those in the spectra of isooctane solutions. The fine structure of the spectra of the S1/S0 electronic transition has not been interpreted yet. It should be noted that the results of full interpretation of the spectra performed for many other molecules are widely used, for example, to determine the symmetry of the first electronically excited state, to study the mechanisms of intramolecular interactions, to decide between the retention and the change of the symmetry of the nuclear config- uration of the molecule upon transition from the ground state to the electronically excited state and to gain quantitative data concerning the change in the geometrical parameters of the molecule in the electronically excited state.Some results of the interpretation of the spectra of dioxins for the lower triplet state are presented in Section V. 1B3g 1B2u (0.023) 1B1u (0.008) 1Ag 300 l /nm E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov Irel 22 0-0 94 83 45 33800 Figure 7.Vibrational structure of the electronically excited state S1 (pp*) of DD molecules cooled in a supersonic jet of argon recorded using resonance-enhanced two-photon ionisation.23 The spectrum shown in Fig. 7 has prompted several conclu- sions 23 (which we consider erroneous) concerning the nuclear configurations of this molecule in the ground and in the S1 electronic states. For example, the most intense line of this spectrum (vibration frequency 22 cm71) was assigned 23 to a `butterfly' type vibration in the electronically excited state S1 [the form and the frequency of this vibration, x(B3u), in the ground electronic state are presented in Fig. 4 and in Tables 4, 5). Since this spectrum contained numerous low-frequency lines and the researchers found it to be simito the corresponding spectrum of dihydroanthracene,60 they 23 arrived at the conclusion that the configurations of the DD molecule in the S1 and S0 states are different. It was concluded that the molecule has a `butterfly' configuration in the S0 state and becomes flatter in the S1 state.However, this conclusion appears doubtful. Indeed, the ratio of the intensities of the strong and weak lines in the spectrum of DD (see Fig. 7) can be distorted due to the known saturation effect.61 In addition, the strong line, assigned to a `butterfly' type vibration, does not form an intense progression (unlike that in the spectrum of the dihydroanthracene molecule): the lines corresponding to the frequencies 2262, 2263, etc., are either weak or missing.Meanwhile, it is intense lines due to vibration overtones in the electronic-vibrational spectra that normally point to a change in the nuclear configuration of the molecule.62, 63 2. Calculation of the electronic states In order to interpret the observed electronic spectra of several dioxins (DD, b-TCDD, a,b-TCDD and OCDD), semiempirical calculations for electronically excited states have been carried out in the PPP, CNDO/S and INDO/S approximations taking into account singly excited electronic configurations.21, 29, 43, 56, 64, 65 The calculated spectra of DD and b-TCDD, in comparison with the experimental spectra, are presented in Figs 5 and 6. It can be seen that, although the absolute magnitudes of energies in the calculated spectra usually prove to be overestimated, the calcu- lations represent satisfactorily the general pattern of the absorp- tion spectrum, i.e.the relative arrangement and the relative intensities of the weak and strong bands. According to calculations,29, 43, 65 the spectral region corre- sponding to the long-wavelength band A contains the frequencies of the first four electronic transitions from S0 to pp* states, namely, B3g, B2u, Ag and B1u but only two of these states are dipole-active. Note that the mutual arrangement of the energy levels corresponding to the first three electronic states changes 181 204 126 172 34000 33900 n /cm71Optical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives depending on the parametrisation accepted in the semiempirical calculations; thus, the S1 state can belong to the B3g , B2u and Ag types.To the best of our knowledge, no experimental data on the type of symmetry of the S1 state are available. The S4 state in dioxins has B1u symmetry. The short-wavelength absorption bands B and C are due to the third and fourth dipole-allowed transitions from pp* states with theB2u andB1u types of symmetry, respectively (see Figs 5, 6). The calculations of electronic states showed that the structure of the pp*-absorption spectrum indicated above does not change substantially in the sequence DD, b-TCDD, a,b-TCDD and OCDD; the increase in the number of chlorine atoms in the molecule changes somewhat the energies of these states (this is in agreement with the experimental results pointing to the small difference between the positions of the band maxima in the UV spectra of various dioxins).The slight influence of the increase in the number of chlorine atoms on the pp*-absorption spectrum is also known for other molecules (for example, for chlorinated 9,10- anthraquinones 66). It should be mentioned that the calculation, which describes satisfactorily the relative oscillator strengths for transitions, reproduces poorly, as a rule, their absolute magnitudes. For example, according to experimental data (eA=0.376104 litre mol71 cm71), band A of the DD molecule corresponds to the oscillator strength fA = 0.067.29 Various semiempirical cal- culations which used different molecular models and different approximations or parametrisations gave the following fA values for DD corresponding to the sum of the oscillator strengths of the first two dipole-active transitions: fA=0.368,21 0.032 29 and 0.17.64, 65 The calculated 64, 65 relative fB / fA values for the DD, b-TCDD and OCDD molecules (fB /fA=9.8, 6.2 and 70, respec- tively) agree in order of magnitude with the ratios of extinction coefficients found experimentally (10, 9 and 35, respectively 56, 58).Thus, the calculation of the electronic states for planar models of the molecules confirms the experimental data 56, 58 concerning the substantial decrease in the absorption cross-section at 305 nm for OCDD molecule relative to the DD or b-TCDD molecules.The A±C bands in the singlet ± singlet pp*-absorption spec- trum of DD (or other dioxins) can be matched to the correspond- ing bands in the benzene spectrum. Scheme 1 shows three transitions from states with different types of symmetry and the corresponding absorption maxima for benzene.67 Scheme 1 1 1 E1u, S3B1u, S2 1B2u, S1 Ag, S0 The B2u / S0 and B1u / S0 transitions are dipole-inactive in the D6h symmetry group and are allowed only upon vibronic (Herzberg ± Teller) coupling. The B1u and B2u states of benzene should be matched to four states of theDDmolecule with D2h symmetry, which contains two benzene fragments, namely, B2u and Ag (analogues of B2u in benzene) and B1u and B3g (analogues of B1u in benzene).In the D2h symmetry group, the B2u and B1u states are dipole-active but (in line with their origin) they are characterised by low oscillator strengths. These four transitions correspond to one weak, the longest-wavelength band A in dioxins (see Figs 5, 6). The band with a maximum at 180 nm in the spectrum of benzene is due to the dipole-active E1u / S0 transition. In molecules with D2h symmetry, for example in DD, two (according to the number of benzene fragments) electronic states E1u of benzene are matched by the B2u, B1u, Ag and B3g states. The u-type states (B2u and B1u) in DD and in other dioxins should be identified with intense absorption bands B and C with maxima at about *250 nm. According to calculations, they are charac- terised by great oscillator strengths of transitions.180 nm 200 nm 260 nm 1045 The analogy between the electronic states of benzene and dioxins is confirmed by calculations of the corresponding p-MO of the molecule.28 Indeed, for at least six lower electronic states, the symmetry of the p-MO of benzene is approximately retained in the benzene fragments of DD. This is reflected, for example, in Fig. 1, which shows the distribution of the coefficients at the 2px-AO centred on the atoms of the six-membered benzene frag- ments of the molecule for four MO. It can be seen that this distribution is similar to the known distribution of the coefficients in the vacant (E1g) and occupied (E2u) p-MO of benzene.The oxygen atoms have little influence on the local symmetry of the MO under study, although the lone electron pairs (l-elec- trons) of oxygen, which occupy the 2px-AO, also contribute to p-conjugation (when this is allowed by the type of symmetry of the whole MO of the molecule). They bring about the greatest perturbation of occupied rather than of vacant MO of diox- ins.28, 56 For the highest occupied MO of the DD molecule, this contribution is 26%; in the S1 state, 0.15 of an electron charge is transferred from the oxygen atoms to the carbon atoms of the benzene fragments. The effect of l-electrons on the S1 state normally shows itself as a substantial decrease in the energy of the plp* state with respect to the pp* state and as an increase in the corresponding dipole moment of the transition.For dioxins, this influence of the l-electrons of oxygen and chlorine on the S1 state is slight; hence, the plp* state can be designated by pp*. In dioxins, the lone electron pairs (n electrons), which occupy 2pz- or 2py-AO, form very deeply located MO and the absorption bands corresponding to the np*/S0 transitions should occur in the vacuum UV region. According to calculations,65 the spectral region between absorption bands A and B of dioxins, as in the case of benzene, contains bands corresponding to the ps*/S0 tran- sitions. The analogy between the electronic transitions in DD and benzene (with allowance for the slight influence of chlorine atoms on the spectrum) accounts for the distribution of band intensity in the UV spectra of dioxins.The same analogy can be drawn between the pp* electronic states of benzene, DD and other molecules, for example, 9,10- anthraquinone and anthracene, which contain, like DD, two benzene side fragments. The DD, 9,10-anthraquinone and anthra- cene molecule differ in the central fragments. Comparison of the spectral and theoretical data concerning their pp* states demon- strates the influence of the electronic structure of the central fragment on the mutual arrangement of the first four energy levels. Recall that the first four electronic transitions in DD are responsible for one band A. According to theoretical and exper- imental data,68 in 9,10-anthraquinone, the energy level of the state analogous to the S4(B1u) state ofDDproves to be shifted to short- wavelength region with respect to the first states B2u, B3g and Ag to such an extent that a separate band with a maximum at*270 nm is attributed to the corresponding transition.Conversely, in the anthracene molecule, the energy level of the state analogous to S4(B1u) in DD is shifted to longer wavelengths with respect to the other three states; thus, it becomes the S1(B1u) state and the onset of the band corresponds to a wavelength of*280 nm.53, 69 Let us briefly summarise the data on the singlet ± singlet transitions in dioxins. The long-wavelength band in the absorp- tion spectrum has a combination structure in the sense that it `hides' transitions from the four electronic states having different types of symmetry: B1u, B3g, Ag and B2u, where B1u corresponds to the S4 state.Variation of the number of chlorine atoms from zero (DD) to eight (OCDD) does not change generally this structure of the band but results in a nearly parallel shift of the energy levels of the corresponding states towards lower frequencies. The mutual arrangement of the energy levels of the B1u and B2u states changes substantially only in a series of molecules having similar structures but containing different atomic groups in the central fragment, for example, in the series of DD, 9,10-anthraquinone and anthracene, in which the above-mentioned states are of the same nature. Thus,1046 the structure of the long-wavelength band in the spectra of dioxins is determined by the presence of the ether C7O7C groups in the central fragment.The energy levels calculated for the B2u, B3g and Ag states in dioxins are close; therefore, additional investigations are required to determine the type of symmetry of the S1 state of a particular molecule. For example, it would be useful to obtain and study the fine vibrational structure of the fluorescence spectrum. The low intensity of the long-wavelength band in the spectra of dioxins is due to its origin: the electronic transitions corre- sponding to this band are analogues of the first two, dipole- inactive transitions in benzene. The change in the number of chlorine atoms generally has little influence on the intensity distribution in the UV spectra of dioxins.V. Luminescence Dioxins belong to the class of luminescing compounds. Lumines- cence is due to the S1 ? S0 transition from the lower singlet electronically excited state (fluorescence) and to the T1 ? S0 transition from the lower triplet state (phosphorescence). Solu- tions of dioxins fluoresce in the region of 310 ± 400 nm; when frozen (77 K), they display a bright long-living blue to light-blue phosphorescence. As an example, Fig. 8 presents the emission spectrum of a solution of DD in isopentane recorded at 77 K.56 This spectrum demonstrates an important feature characteristic of both DD and PCDD, namely, very low intensity of the fluores- cence band 1 with respect to the phosphorescence band 2.Irel 2 1 400 350 450 l /nm Figure 8. Luminescence spectra of a solution of DD in isopentane at 77 K;56 (1) fluorescence, (2) phosphorescence. 1. Photophysical characteristics of luminescence In this Section, the quantitative data on the deactivation of the first singlet and triplet electronically excited states of dioxins are presented. We discuss experimental values for the quantum yields and the decay times of fluorescence (ffl and tfl) and phosphor- escence (fph and tph), the rate constants for deactivation of the lower singlet and triplet electronically excited states (kfl and kph), theoretical estimates of the kph values, and the results of calcu- lations of the influence of the number of chlorine atoms in the molecule on kph and tph .a. Fluorescence The fluorescence spectra of solutions of DD and PCDD exhibit a structureless band (see Fig. 8). The use of selective laser excitation of fluorescence, which often gives rise to a fine (vibrational) structure in the spectra of solutions,70 still has not allowed recording of structured spectra for the molecules under study (b-TCDD, 1,3,7,8-Cl4DD and OCDD).58 E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov The most complete experimental data, reported in the liter- ature, on the deactivation of the S1 state for solutions of DD in isopentane 29, 56 and ethanol 71 are presented below. Ref. kfl/ s71 tfl/ ns ffl 3.5 <1 56 71 29 0.003 0.000 7 7 0.96106 71.66107 For b-TCDD, an approximate estimate gave tfl&10 ns (vapour, 473 K) 58 and kfl&4.06107 s71 (vapour and a solution in CHCl3, 77 K).29 The difference between the given ffl and tfl values determined for DD in the publications cited is due to the fact that the measurements were performed near the limit of the capacity of the equipment employed and to the influence of the solvent.The kfl values for isopentane solutions of DD have been determined by different methods.29, 56 For DD (as well as for b-TCDD), the kfl values for the S1(B2u)?S0 transition were estimated 29 using the calculated data on the ratio of the oscillator strengths of the 1B2u / S0 and 1B1u / S0 transitions, which are responsible for the long-wavelength band A in the absorption spectra, and experimental data on the integrated intensity of this band (see Figs 5, 6).The kfl values found in this way are over- estimated because the integrated extinction coefficient of band A can contain contributions of the transitions caused by the Herz- berg ± Teller vibronic coupling. In another method,56 the kfl value was estimated from the expression kfl=ffl / tfl. It is difficult to explain this great (by an order of magnitude) discrepancy of the kfl values found for DD. The main reason may be as follows. One value 29 corresponds to the dipole-allowed 1B2u?S0 transition. The other, smaller kfl value 56 might refer to the emission from a g-type rather than u-type S1 singlet state, for example, to the 1Ag?S0 transition, i.e. to a transition allowed only with account of the Herzberg ± Teller intramolecular inter- actions.Some possible errors both in spectral measurements and in quantum-chemical calculations should also be taken into account. An interesting dependence of the fluorescence and phosphor- escence quantum yields on the structural factor for the series DD, dihydroanthracene and xanthene has been found.56 DD Xanthene Dihydroanthracene ffl 0.003 0.4 0.26 0.14 0.26 0.14 fph Despite the fact that the dihydroanthracene and xanthene molecules have the same luminescence quantum yields, the ffl value forDDis much smaller. Meanwhile, the fph value forDDis almost 3 times as great as those for dihydroanthracene or xan- thene. According to calculations,56 the S1 state in the xanthene and DD molecules can be assigned to the plp* type of orbital; the change in the p-electron density on the oxygen atoms in this state is nearly the same for these two molecules.Therefore, the sharp (by two orders of magnitude) decrease in the ffl value forDDwith respect to xanthene is believed to be due to the higher rate constant for the S1 Tj intersystem crossing (kisc) in the case of DD. It is known that the kisc values for the transfer of energy between states of different orbital natures, pp* and np* (or sp*), are several orders of magnitude higher than those for states with the same orbital nature (pp* and pp*).72 For this reason, the higher kisc value found forDDwas interpreted 56 by assuming that only DD posesses a Tj state of the np* type the energy of which is very close to the energy of S1.This arrangement is believed to be due to the lowering of the energy levels of the np* states in this molecule (containing two oxygen atoms) with respect to those in the xanthene molecule (with one oxygen atom). However, it follows from the calculations for the np*, ps* and sp* states that the energies of the corresponding triplet levels of the DDOptical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives molecule are more than 10 000 cm71 higher than the energy of the S1 state.29, 65 These states cannot ensure high kisc. The difference between the ffl values in xanthene and DD might be due to the difference between the symmetries of the xanthene (C2u) and DD (D2h) molecules, in particular, to the fact that the S1 state in the DD molecule might correspond to the g type: the first line in the spectrum of the S1/S0 transition shown in Fig.7 has a relatively low intensity. In our opinion, the question of what is the reason for the low ffl value in DD remains open. b. Phosphorescence Phosphorescence is manifested as a relatively intense band in the spectrum shown in Fig. 8. Using specific experimental conditions, spectra with clear-cut vibrational structure of this band have been obtained for DD and PCDD (see Section V.2). For a number of dioxins, the phosphorescence lifetimes tph at 77K have been measured.2, 43, 56, 64, 65, 73 For example, in the case of DD, tph=0.54 s (77 K, a solution in a 1 : 5 : 5 mixture of ethanol, isopentane and ether) 2 and 0.7 s (77 K, a solution in isopentane).56 For chloro-substituted derivatives, the phosphor- escence lifetimes are shorter than for unsubstituted DD.2 In the case of PCDD, the tph values found by different researchers can be substantially different.Thus for a number of dioxins (ranging from mono- to octa-substituted derivatives), tph values varying from 0.29 to 0.075 s have been reported.2 Accord- ing to another publication,73 the tph values for another series of tetra- and penta-substituted dioxins in n-hexane vary over the 1.55 ± 1.03 s range. The tph values presented in the two publica- tions2, 73 for the only compound common to both of them, 1,2,3,4- Cl4DD, differ by more than an order of magnitude � 0.057 and 1.03 s.Both groups of researchers noted little differences in the tph values within the series of dioxins they studied. Shorter tph values were found 2 for those dioxins which contain several (3 or 4) Cl atoms in only one benzene fragment. The great discrepancies between tph found for PCDD in different studies are apparently due to some systematic error of measurements (rather than to the use of different solvents). For OCDD, b-TCDD and a,b-TCDD as solutions in n-hexane, the times tph have been measured (see below) at 77 K.43, 64, 65 These values were found to be in satisfactory agreement with tph measured for solutions in a 1 : 5 : 5 mixture of ethanol, isopentane and ether.2 OCDD a,b-TCDD b-TCDD n-Hexane 0.30 0.30 0.20 tph/ s Solvent mixture 0.25 0.25 7 tph/ s The radiative T ? S0 transitions are allowed due to intra- molecular spin-orbit (SO) and vibronic-spin-orbit (VSO) cou- plings.74, 75 Correspondingly, the rate constant krad for the radiative deactivation of the triplet state can be represented as krad=kSO+kVSO; in the case of phosphorescence, T=T1 and krad=kph.It has been proposed to use the data on the distribution of the intensity in fine-structure electronic-vibrational phosphor- escence spectra for estimating the kVSO to kSO ratio.76 These estimates obtained for the T1 states of DD, b-TCDD and OCDD show 42 that the magnitudes of the phosphorescence rate constants kph for these molecules are stipulated by different mechanisms of intramolecular interactions.More precisely, this constant for DD is determined by only vibronic-spin-orbit cou- pling, while in b-TCDD and OCDD, they are determined by both spin-orbit and vibronic-spin-orbit couplings (see Section V.2). Let us dwell now on the data on kph. It follows fromthe tph and fph values measured for the DD molecule 56 that kph=0.57 s71, kSO=0; hence, kph=kVSO. For PCDD, no experimental data on kph are currently available, although the tph values are known; they can be found from the relation 1047 1 tph à kph á kTnr , where kph is the rate constant for the radiative and kTnr is the rate constant for the non-radiative deactivation of the T1 state. For several dioxin molecules, the kSO, kVSO and kph values have been estimated theoretically.43, 64, 65 Quantum-chemical calculations were employed to estimate the kSO values for DD, b-TCDD, a,b-TCDD and OCDD; it was found 43, 64, 65 that the number and the arrangement of Cl atoms in the molecule influences kSO values for the 3B1u states.First, it was concluded that each heteroatom (O, a-Cl or b-Cl) in these molecules makes an additive contribution to the dipole moment of the P0-0 transition (the square of which is related to kSO). Second, a cancellation effect of the a-Cl and b-Cl atoms on the P0-0 value was found. As a result, it was found that kSO(b-TCDD) / kSO(OCDD)=3.6, i.e. the decrease in the num- ber of Cl atoms in the molecules of b-TCDD with respect to that in OCDD (by a factor of two) induces a substantial increase rather than a decrease in the kSO component of the krad rate constant.It was concluded 43, 65 that the cancellation effect of several hetero- atoms (chemically identical but differing in the position in the molecule) on the spin-orbit coupling, found in the study of dioxins, could be useful in elucidating the reasons for the known abnormal influence 77, 78 of an internal heavy atom on the kph and tph values in some organic molecules. nr The use of calculated kSO values and experimental data on the kSO / kVSO ratio gave the following estimates for the rate constants kph for b-TCDD and OCDD: 7 and 5 s71. Modelling of the kT values made it possible to determine theoretically 64 the ratio of tph for these molecules [tph(OCDD) / tph(b-TCDD)=1.60.2], which is consistent with the experimental value 1.5 (the exper- imental data on tph were mentioned above).The theoretical absolute magnitudes of tph for these molecules also proved to be close to experimental values. For instance, for OCDD, the calculated value was tph = 0.10 ± 0.20 s. 2. Fine-structure phosphorescence spectra In this Section, we consider the results of investigations of fine- structure phosphorescence spectra of dioxins concerning the energies and types of symmetry of lower triplet electronic states, the frequencies and types of symmetry of vibrations (in particular, low-frequency vibrations), and the contributions of intramolecu- lar spin-orbit and vibronic-spin-orbit couplings to the phosphor- escence rate constant.The results of theoretical calculations for the change in the electron density and the estimates of the change in the molecular geometry in the lower excited triplet pp* state with respect to the ground electronic state are presented. a. Spectra at 77K Below we consider experimental data on the influence of solvents (non-branched alkanes) on the pattern of the low-temperature phosphorescence spectra of dioxins. The fine structure of broad-band phosphorescence spectra (see, for example, Fig. 8) can be recorded under specific condi- tions, which include conducting the experiment at low temper- atures, selection of the appropriate solvent and, in some cases, selection of a particular wavelength of the exciting light.70, 79 These experimental conditions should ensure low electron ± pho- non interactions and slight inhomogeneous or homogeneous spectral broadening.70, 80 A well-known method for recording the structured luminescence spectra of complex organic com- pounds is the Shpol'skii method,79 based on the use of n-alkanes (n-hexane, n-heptane, n-octane, etc.), which form a polycrystal- line material at low temperatures, as solvents. These solvents ensure slight inhomogeneous broadening of the spectrum.Upon deep cooling of solutions of luminescent molecules in these solvents down to the boiling point of liquid nitren (77 K) or liquid helium (4.2 K), the electron ± phonon interaction and homogeneous broadening of the spectrum decrease.80 As a con-1048 sequence, the electronic-vibrational (vibronic) emission spectra consist of a large number of quasi-lines (lines).These spectra are referred to as quasi-line or fine-structure spectra. The vibrational structure of the phosphorescence spectra of DD and PCDD was revealed at 77K for solutions in a 1 : 5 : 5 mixture of ethanol, isopentane and ether 2 and in normal hydro- carbons.81, 82 When the Shpol'skii method is used to make the structure of the spectrum as clear as possible at a specified temperature, it is important to select an appropriate solvent. The Bolotnikova rule 83 concerning the commensurability of the linear dimensions of the molecules of the solvent and of the compound under study serves as a criterion for choosing the solvent.Since dioxins are highly toxic, the problem of selecting the solvent has been studied in detail only for DD.81 It was established that the most clear spectrum is observed in n-hexane and n-heptane matrices (Fig. 9). The spectral patterns in these solvents are more structured than in a 1 : 5 : 5 mixture of ethanol, isopentane and ether. The structured phosphorescence spectra of solid solutions exhibit so-called components of the Shpol'skii multiplet, which are due to luminescence of different centres. Each of the lumines- cence centres is matched by lines for transition to vibrational sublevels of the ground electronic state (this is the vibronic structure of the spectrum) and by a line corresponding to a purely electronic transition (if it is dipole-allowed) with the frequency n0-0.As a consequence, the spectrum often has a very intricate structure. At 77 K, the lines are normally markedly broadened due to the phonon wing and they are often called quasi-lines.62, 70 Table 6 lists the quasi-line maxima (1 ± 9) in the spectra ofDD and PCDD;82 only for 1,2,3,4,6-Cl5DD was it impossible to obtain a quasi-line spectrum in solid solutions in hydrocarbons. The spectrum of this compound was a broad structureless band with a maximum at 480 nm. The same spectral pattern with a maximum at 485 nm was obtained for solutions of 1,2,3,4-Cl4DD in a 1 : 5 : 5 mixture of ethanol, isopentane and ether and in n-hexane at 77 K.2, 73 Now we consider briefly the possible reason for the absence of structure and the bathochromic shift of the phosphorescence band in the spectra of solutions of 1,2,3,4-Cl4DD2, 73 and 1,2,3,4,6- Cl5DD.82 A specific feature of these molecules is that one benzene ring in them is fully chlorinated, while the other contains either no chlorine atoms or only one atom.Molecules with so asymmetric an arrangement of chlorine atoms are polar. It has been shown that polar molecules of oxygen-containing compounds dimerise in solutions in saturated hydrocarbons.84 Thus the broad structureless bands in the phosphorescence spectra of 1,2,3,4-Cl4DDand 1,2,3,4,6-Cl5DDmay belong to their dimers. However, to elucidate more definitely the reason for the structure- Table 6.Positions of the quasi-line maxima in the phosphorescence spectra of solutions of dioxins in n-hexane at 77K.82 Compound l/ nm 1 2 3 4 5 6 7 8 9 400 402 407 408 412 410 417 415 415 396 399 406 407 410 407 412 411 413 394 396 402 403 406 406 410 409 410 480 408 414 419 420 424 430 415 417 423 DD 2-ClDD 2,3-Cl2DD 2,3,8-Cl3DD 2,3,7,9-Cl4DD 2,3,8,9-Cl4DD b-TCDD a,b-TCDD 1,2,3,7,8-Cl5DD 1,2,3,4,6-Cl5DD 1,2,3,4,7,8-Cl6DD 1,2,3,6,7,8-Cl6DD OCDD Note. The numerals 1 to 9 mark the numbers of lines in the spectra. The wavelengths corresponding to the most intense quasi-lines in the phosphorescence spectra are typed in bold. E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov 3 4 a 1 5 6 7 8 2 bc 400 Figure 9.Phosphorescence spectrum of solutions of DD in n-hexane (a), n-heptane (b) and n-octane (c) at 77K.81 The l values in the maxima of the numbered lines are given in Table 6. 402 408 412 410 417 417 421 417 421 420 414 429 418 428 434 438 426 442 412 410 416 415 420 428 433 419 436 426 440 436 446 449 443 443 9 450 l /nm 441 430 430 427 424 438 447 437 440 432 434 439 441 433 451 454 462Optical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives less spectral pattern and the bathochromic shift of the phosphor- escence band in the spectra of these compounds with respect to the spectra of other PCDD, additional investigations are needed.Data on the positions of the quasi-line maxima in the spectra and on the distribution of the relative intensities of quasi-lines led to the conclusion 82, 85 that these spectra are highly individual and could be used for identification of various dioxins, in particular, isomers (see Section VI). b. Spectra at 4.2K Study of the phosphorescence spectra of a number of dioxins at 4.2K (see compounds in Table 6) showed 82 that, as the temper- ature decreases from 77 to 4.2 K, the half-width of the quasi-lines decreases by an order of magnitude (to 0.2 ± 0.5 nm), while the peak intensity of lines increases. In addition, whereas the spectra recorded at the boiling point of liquid nitrogen consist, on the average, of 10 ± 15 quasi-lines, the spectra at 4.7K contain several dozens of quasi-lines. At low temperatures (for example, 4.2 K), the vibronic struc- ture of the emission spectra is due to the transitions from the zero vibrational sublevel of the electronically excited state to various vibrational sublevels of the ground electronic state S0.This spectrum contains vibration frequencies which are detected in the vibrational spectra (IR and Raman). As an example, Figure 10 shows a fine-structure phosphorescence spectrum of a solution of DD, while Table 7 presents the assignment of the structure of this spectrum,41 namely, determination of the com- ponents of the Shpol'skii multiplet, the frequencies of purely electronic transitions n0-0, the frequencies of the fundamental vibrations and the sum vibration frequencies.A similar analysis of the spectral structure has been carried out for solutions of b-TCDD, a,b-TCDD and OCDD.39, 42, 43 The T1?S0 transition, which is spin forbidden, becomes allowed as a result of spin-orbit and vibronic-spin-orbit cou- plings.74, 75 As a consequence, the fine-structure electronic spectra do not exhibit all vibrations but only totally symmetric vibrations and some types of non-totally symmetric ones. By taking account of the characteristic matrix elements 72 for the spin-orbit coupling and of the dipole moments of the 1sp*?S0 and 1pp*?S0 transitions, which influence the line intensity, it can be con- cluded 30 that the out-of-plane vibrations may be responsible for intense lines in the pp* phosphorescence spectra.Vibronic lines of Irel 10 12 8 3 5 13 2 18 14 15 6 16 4 1 390 400 Figure 10. Phosphorescence spectrum ofDDin n-hexane at 4.2 K.41 The interpretation of the spectrum is given in Table 7. The arrows mark the positions of the Shpol'skii multiplet components corresponding to a purely electronic transition. 1049 this type are actually observed in the phosphorescence spectra of dioxins (see Figs 10, 11 and Table 7). Determination of the symmetry of the T1 state in dioxins is difficult. According to quantum-chemical calculations, 29, 65 the two pp* triplet states lowest in energy (more precisely, plp*) correspond to electronic functions of g and u types, the T state of the u type being analogous in orbital structure to the fourth pp* singlet state S4(B1u), the band of which is `hidden' in the long- wavelength absorption band (see Figs 5, 6).The difference between the energies of the two T states is small (*1000 cm71); hence, variation of the positions of chlorine atoms in PCDD molecules can change the mutual arrangement of these triplet states. However, the accuracy of calculations does not suffice for deciding which of the states with the calculated types of symmetry B3g and B1u is the lowest (T1). Moreover, the pp* phosphorescence spectra of dioxins the molecules of which have an inversion centre can exhibit intense lines corresponding to out-of-plane vibrations for both symmetry forbidden (when T1 belongs to the g type of symmetry) and symmetry allowed (when T1 belongs to the u type of symmetry) transitions. The question of the type of symmetry of the T1 state can be answered by determining the type of symmetry of non-totally symmetric vibrations the frequencies of which n(phos) are observed in the phosphorescence spectra. For the T1?S0 tran- sitions from the states corresponding to g or u symmetry, non- totally symmetric (out-of-plane) vibrations of only a particular type of symmetry, either u or g, respectively, can manifest themselves, in conformity with the selection rules.75 This regu- larity was used 39, 41 ± 43 to determine the type of symmetry of the T1 state of dioxins.These states in DD, b-TCDD andOCDDwere assigned to different (B3g and B1u) types of symmetry (Fig. 12). The experimental values of the energies (n0-0) of the T1 state and the corresponding kSO/kVSO ratios for the components of the phosphorescence rate constants have also been determined for these compounds.42 OCDD DD b-TCDD n0-0 /cm71 23836 1 : 3.3 24410 1 : 0.7 25595 0 : 1 kSO : kVSO The n0-0 frequencies of purely electronic transitions, T1 ? S0, either allowed or forbidden from the orbital symmetry standpoint, 22 46 49 37 44 25 20 31 33 28 430 420 410 l /nm1050Irel 149 (Z) 441 (y) 383 (n, r) 206 (t) 410 Figure 11. Phosphorescence spectra of b-TCDD in n-hexane at 4.2 K.39 The arrows mark the positions of the Shpol'skii multiplet components corresponding to a purely electronic transition.In the line maxima of the most intense multiplet components, the frequencies (cm71) of only fundamental modes are given. The assignment of vibrations is given in parentheses (see Table 1). Table 7. Positions of the line maxima (nN/ cm71) and interpretation of the phosphorescence spectra of DD in n-hexane (see Fig. 10). Line number N nN/ cm71 Components of Dn a the Shpol'skii multiplet 00 189 188 332 329 III III III III III III III II II II II III II II I 25628 25595 25439 25407 25296 25266 25188 25141 25110 25052 25019 24887 24856 24814 24719 24639 24535 24484 24453 24418 24383 24325 12345678910 11 12 13 14 15 16 17 18 19 20 21 487 485 576 576 741 739 781 876 956 1060 1144 1142 1177 1212 1303 1300 II 24295 22 1331 1464 II I 24264 24164 23 24 n5+n10 (722) ???n4+n11 (1027) n3+n12 (1146) ?n4+n12 (1146) n5+n11 (1026) n4+n13 (1227) ?n5+n12 (1149) n4+n9+n10 (715) n5+n12 (1146) n5+n13 (1215) n5+n13 (1210) n2+n14 (1620) ?n5+n9+n10 (719) ?n3+n14 (1621) n3+n14 (1619) n4+n14 (1624) n4+n14 (1623) ?n5+n14 (1618) n5+n14 (1616) aDn=n0-07nN, where n0-0 is the frequency of the purely electronic transition.b The ordinal number L of the fundamental mode found in the phosphorescence spectrum is given.The frequencies nL, separated from the sum frequencies, with the numbers L=9 ± 14 correspond to totally symmetric fundamental vibrations (their frequencies are given in parentheses); the frequencies nL with the numbers L=1 ± 8 were assigned to out-of-plane vibrations of the B3u and Au types of symmetry (their letter designations are given in parenheses). E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov 1236 (q) 628 (s, w) 654 (e)739 (r) 1284 (o) 1100 (a) 548 (u) 1442 (n) 867 (f) 985 (e) 430 420 Interpretation b Line number N nN/ cm71 Components of the Shpol'skii multiplet 24134 24116 24081 24035 23992 23964 23931 23873 23830 23792 23766 23738 25 26 27 28 29 30 31 32 33 34 35 36 II II II II II II II II II II II III II III 23710 23672 23646 37 38 39 0-0 0-0 n1 (h) n1 (h) n2 (y) n2 (y) 0-0 ? n3 (w) n3 (w) n4 (o) n4 (o) n5 (g) n5 (g) n6 (u) n7 (i) n8 (j) n3+n9 (575) n4+n9 (568) n4+n9 (566) ?n3+n10 (727) n4+n10 (727) n5+n9 (562) n4+n10 (724) n5+n9 (561) ?n5+n10 (723) II II II III III II III 23610 23571 23544 23520 23491 23428 23396 23312 23269 23240 40 41 42 43 44 45 46 47 48 49 1600 (k) 440 l /nm Interpretation b Dn a 1461 1479 1514 1560 1603 1631 1664 1722 1765 1803 1829 1890 1857 1885 1956 1949 1985 2024 2051 2108 2104 2200 2199 2283 2359 2355Optical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives OCDD DD b-TCDD 3B1u 3B3g pp* 3B3g pp* pp* 3 3 B1u B3g 3B1u pp* pp* pp* Figure 12.Symmetry of the T1 and T2 electronic states.42 decrease somewhat in the sequence DD, b-TCDD and OCDD39, 41 ± 43 (the values presented above are the n0-0 frequen- cies of one or several components of the Shpol'skii multiplet, which accounts for the most intense vibronic lines in the spec- trum). The out-of-plane vibrations of dioxins are characterised by low frequencies which are unknown from IR or Raman spectra for the vast majority of dioxins. Therefore, in the absence of exper- imental data, the symmetry types of the vibrations the frequencies of which n(phos) can be seen in the phosphorescence spectra can be determined by comparing the calculated vibration frequencies with n(phos).Generally, this comparison provides a satisfactory result (see the phosphorescence spectrum of b-TCDD shown in Fig. 11 and the data on the frequencies of the fundamental modes in b-TCDD listed in Tables 2 ± 5). The most complicated problem is to interpret the phosphor- escence spectra corresponding to a forbidden purely electronic transition because in this spectrum, it is difficult to find lines due to the components of the Shpol'skii multiplet and n0-0 (see, for example, Fig. 10 and Table 7, data for DD). The positions of lines corresponding to the n0-0 frequencies are determined from the n(phos) values for non-totally symmetric vibrations. Therefore, to interpret these spectra, reliable experimental (from IR or Raman spectra) or theoretical data on the low-frequency vibrations of the molecule are needed.According to the results of quantum-chemical calcula- tions,43, 65 the T2(B1u) state of the DD molecule and the T1(B1u) state of the b-TCDD, a,b-TCDD and OCDD molecules are very close in orbital structure and are charge transfer states. For b-TCDD, the changes in the p-electron density on atoms and on valence bonds in the T1 state with respect to those in the S0 state (Fig. 13) have been calculated.86 It was concluded that in this state, charge transfer takes place from all heteroatoms of the molecule to the carbon atoms; the electron density on the C7O bonds in the central fragment of the molecule increases, which implies that the bonds are shortened. The data on the n0-0 frequencies corresponding to the singlet S1 state 23 and the triplet T1 state (see Fig.12) ofDDand b-TCDD molecules in solutions can be used to estimate the difference between the energies of these states � 7430 and 6940 cm71. These values are typical of aromatic molecules with an electron- H H Cl Cl O C C 70.034 70.032 C C C C 0.030 0.024 70.139 0.029 C C C C 70.158 C 70.025 Cl C Cl O 70.037 0.000 H H Figure 13. Changes in the electron density on the bonds and atoms of the b-TCDD molecule in the lower triplet excited state 3B1u (pp*).86 The values for only one of the equivalent atoms are presented.1051 donating group;69 this is consistent with the theoretical data on the pattern of electron density redistribution in the S1 and T1 states. The distribution of the intensity of vibronic lines in the fine- structure phosphorescence spectra depends on the intramolecular interactions.62, 63, 87 The relative intensities of the vibronic lines with maxima at n0-07nN corresponding to the fundamental modes with the frequencies nN were measured in the phosphor- escence spectra of DD, b-TCDD, a,b-TCDD and OCDD.39, 41 ± 43 The intensity distribution of the lines corresponding to totally symmetric vibrations is determined by the Franck ± Condon interaction, which results in a changed equilibrium geometry of the molecule in the electronically excited state.The symmetry of the nuclear configuration typical of the S0 state is retained. The change in the geometry induces a shift of the minimum of the electron potential (Del) along the normal vibration coordinates Q.62The Del values were determined experimentally for the T1(3B3g) electronic state of DD28 and for the T1(3B1u) electronic state of b-TCDD .86 In the case of b-TC, the highest Del value corresponds to the normal coordinate with the index Z, which describes bending of the C7Cl bond in the plane of the molecule (see the letter designations of vibrations in Table 1). In each of these molecules, the highest Del value for vibrations involving the benzene fragment corresponds to the normal coordinate Qk, which is typical of electronic states of the B3g or B1u types (but not Ag or B2u).28, 30 Using a simple theoretical model of the Franck ± Condon interactions,28 which reflects satisfactorily the experimental Del values, the changes in the geometric parameters in the benzene fragments of DD and b-TCDD in the T1 state have been calculated.28, 86 It was found that the changes in the lengths of all bonds parallel to the z axis of the molecule [C(2)7C(3) and the three other] are about an order of magnitude greater than the changes in the other bond lengths in the benzene fragments.This is consistent with the results obtained by INDO/S calculations of electronic states (see Fig. 13).The changes in the bond lengths are nearly the same in each molecule and are equal to 0.021 and 0.027A for DD and b-TCDD, respectively. In b-TCDD, the C7Cl bonds are rotated in the molecular plane through 0.9 8, according to an approximate estimate.39 The intensity of the region in the phosphorescence spectrum associated with the purely electronic transition and with totally symmetric vibrations is determined by the spin-orbit coupling, while the intensity of the other spectral region is governed by the vibronic-spin-orbit coupling.75 Determination of the relative intensities of some lines in the spectrum provides 76 an experimen- tal estimate of the contributions of spin-orbit and vibronic-spin- orbit couplings to the phosphorescence rate constant (kph=kSO+kVSO).The kSO / kVSO values were estimated for the T1 state of DD, b-TCDD, a,b-TCDD and OCDD mole- cules.39, 42, 43 The quantitative results obtained 39, 41 ± 43 indicate that an increase in the number of Cl atoms in the molecule (b-TCDD and OCDD) sharply increases the contribution of the vibronic-spin-orbit coupling (with respect to the spin-orbit cou- pling) to the constant kph. The lines corresponding to non-totally symmetric modes do not form progressions if they have originated due to the Herz- berg ± Teller interactions, i.e. if no change in the nuclear config- uration of the molecule related to the change of its symmetry typical of the S0 state is involved. However, apparently, it would be possible to establish correlations between the relative intensities of the vibronic lines caused by Herzberg ± Teller interactions and the probable destruction processes of dioxins in photochemial reactions in the electronic state T1.For instance, some correlations have been found between the intensity of vibronic lines related to non-totally symmetric vibrations of a particular form and the structural changes in the molecules of carbonyl-containing com- pounds upon photochemical reactions in the triplet (np*) and (pp*) electronically excited states.881052 VI. Spectroscopic analytical investigations It has been mentioned in the Introduction that dioxins are formed as trace side products in some industrial processes including chemical (production of chlorinated phenols, herbicides and so on) and pulp-and-paper (paper bleaching) processes and during the operation of garbage-incineration plants.This is accompanied by dioxin pollution of the environment. Polychlorinated dibenzo- p-dioxins are highly stable compounds. They do not decompose at high temperatures (up to 750 8C); the half-life of PCDD in soil is about 10 years.89 Due to their stability, the dioxin toxicants are accumulated in soil, water, plants and animals, which creates hazard for humans and other living organisms, because dioxins are many orders of magnitude more toxic than commonly known poisons (such as potassium cyanide or curare). Realisation of the dioxin menace stimulated the governments of many countries to adopt special programmes on the protection of the environment from these toxic compounds.One aspect of the environmental protection is the development of highly sensitive methods for determination and quantitative analysis of traces of PCDD. Since dioxins are very hazardous compounds, methods for their analysis should be highly sensitive, at the picogram level. Yet another, equally significant requirement to the methods of dioxin analysis is high selectivity. This requirement originates due to the fact that dioxins having different structures differ markedly in toxic properties.4, 90, 91 Figure 14 presents the results of an experiment with guinea- pigs. It demonstrates how sharply (by approximately four orders of magnitude) the toxicity ofPCDDcontaining from seven to four chlorine atoms increases as the a-Cl atoms are successively replaced by hydrogen atoms.The effect of PCDD on organisms is appreciably variable; however, generally, the regularity pre- sented in Fig. 14 holds. Thus b-TCDD is considered to be the most toxic. At present, toxicity equivalent factors (TEF) 91 common to humans and mammals have been established (for b-TCDD, TEF=1). Apart from b-TCDD, six more PCDD containing four chlorine atoms in the b-positions are considered to be toxic. For example, for 1,2,3,7,8-Cl5DD and the three isomers of hexachloro-substituted dibenzo-p-dioxins, TEF=1 and 0.1, respectively; for the octasubstituted dibenzo-p-dioxin, TEF=0.0001. Therefore, in monitoring the dioxin contamina- tion, the problem of identification of the most hazardousPCDDis a crucial point.Identification is also important in connection with the need to determine precisely the reason (i.e. who is to blame) for the contamination and to control the process of dioxin destruc- tion. From the scientific viewpoint, this is important for determin- ing the mechanisms and the products of destruction of dioxins during incineration, photolysis and chemical or biological deacti- vation.7log LD50 (mg kg71) 3210 �1 �2 �3 �4 �5 711 2 3 4 N Figure 14. Toxicity of dioxins (the lethal dose LD50 for guinea-pigs) vs. the number (N) of hydrogen atoms occupying a-positions in the 1,2,3,4,6,7,8-Cl7DD (1), 1,2,3,6,7,8-Cl6DD and 1,2,3,7,8,9-Cl6DD (2), 1,2,3,4,7,8-Cl6DD (3), 1,2,3,7,8-Cl5DD (4) and b-TCDD (5) molecules.24 E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov The best developed and most widely used methods for analysis of dioxins not related to optical spectroscopy include various modifications of the chromatographic method,92 ± 95 chromato- mass spectrometry,96 ± 98 a method combining chromatographic separation of isomers with theirNMRdetection,99 etc.Description of all the methods used currently is beyond the scope of this review.We shall restrict themselves to a brief description of the optical methods of analysis employed for this purpose. Among optical methods, Fourier IR spectroscopy (FTIR) is used most extensively to analyse dioxins. Usually, this technique is combined with chromatographic methods.98 Thus even in the first publications 32, 34 dealing with analysis of dioxins by gas chroma- tography and FTIR (GC-FTIR method), using reflectance and absorption spectra, the researchers were able to identify all the 22 isomers of tetrachloro-substituted dioxin.This method permits detection of dioxins present in concentrations of several hundreds of nanograms. A higher sensitivity of the method can be attained by using matrix isolation of the analysed compounds (GC-MI-FTIR).100 In this case, the isomers separated by chromatography were present in an argon matrix at 12 K; this decreases the line half- width and, correspondingly, increases the peak intensity of lines. The IR spectra of all the 22 isomers of tetrachloro-substituted dioxin were recorded by this method,33, 35, 36 including the spectra of b-TCDD and its completely isotope-substituted analogue ([13C] b-TCDD), in which all the 12C atoms have been replaced by 13C.This technique is suitable for the analysis of samples with contents of tetrachloro-substituted DD of the order of hundreds of picograms.36, 100, 101 For example, the use oTCDD as the internal standard permitted determination of b-TCDD in a concentration of 15 ± 45 pg per g in an extract from a fish tissue.37 To identify dioxins on the basis of their IR spectra, several characteristic bands in the range of 1600 ± 800 cm71 are used.24, 33, 35 Figure 15 presents the IR spectrum of the extract from a fish tissue containing b-TCDD and an internal standard admixed.Generally, the positions of bands in the spectrum recorded by the GC-MI-FTIR method coincide with the positions of the corresponding bands in the spectrum of b-TCDD in a KBr pellet (see Fig. 2). Figure 16 shows an example of identification, based on IR bands, of two tetrachloro-substituted DD isomers which are poorly resolved by gas chromatography 102, 103 but are separated by gas liquid chromatography. In the frequency range of 1600 ± 800 cm71, the positions of the maxima of bands in the spectra of these isomers differ only slightly; however, the isomers can be identified based on the different relative intensities of bands. It has also been proposed 24, 35 to identify isomers using the Absorbance 0.005 0.004 0.003 0.002 0.001 0.000 1500 1400 1300 1200 1100 1000 900 n /cm71 Figure 15.Absorption spectrum of the extract from a fish tissue contain- ing b-TCDD and an admixture of [13C]b-TCDD added as the internal standard. The spectrum was recorded by the GC-MI-FTIR technique; the analytical lines of the b-TCDD toxin 37 are marked by arrows.Optical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives Absorbance 1600 Figure 16. Absorption spectrum recorded by the GC-MI-FTIR method for the isomers 1,2,4,8-Cl4DD (a) and 1,2,4,7-Cl4DD (b).35 The numerals 1 ± 5 mark the analytical lines the relative intensities of which at 1420 cm71 (lines 1, 2) and at 980 ± 820 cm71 (lines 3 ± 5) can be used to indentify the isomers.nas C¡O frequencies of the antisymmetric stretching vibrations of the C7O7C groups in the range of 1330 ± 1280 cm71. However, in our opinion, this spectral region is less reliable for identification because the semiempirical triatomic model considered 35 deter- mines the nas C¡O values too crudely; meanwhile, the difference between the experimental values of these frequencies for some pairs of isomers is very little (1 ± 2 cm71). The results of identification of PCDD isomers by four independent methods: chromatography, GC-MS, GC-FTIR and 1H and 13C NMR spectroscopy have been compared.104 For several isomers, the results of identification did not coincide. It has been proposed 22, 47 to use the frequencies of normal modes of PCDD isomers found theoretically (by ab initio calcu- lations) to identify the isomers based on their IR spectra.Now we shall mention several other spectral methods for determination of dioxins. For example, theoretical estimates of the limits of detection of dioxins by Raman spectroscopy (*2 pg) and by giant Raman scattering (GRS, *261074 pg) allow the GRS method to be regarded as a promising analytical techni- que.105 A simple photometric method based on the ability of PCDD to be protonated and to form coloured associates with transition metal anions has been proposed.106 The lower limit of the b-TCDD concentration in a hexane extract determined by this method is 20 pg cm73. The detection powers of the setups used in the studies of dioxins by UV absorption and fluorescence methods have been evaluated.58, 73 Fluorescence of dioxins shows itself as a broad band with a maximum in the region of 330 ± 390 nm.For n-hexane solutions of PCDD containing 5 to 8 Cl atoms, the detection limit of the fluorescence analysis amounts to 0.5 ± 5 ng cm73 at room temperature for excitation with the wavelength lexcit=230 ± 245 nm.73 This spectrum region contains the maximum of band B, the most intense band in the UV absorption of dioxins (see Figs 5, 6). In laser-induced fluorescence analysis of dioxin vapours (b-TCDD, 1,2,3,7,8-Cl5DD and OCDD) with a 1 (1428) 4 (963) 3 (974) 5 (842) 2 (1412) b 3 (980) 1 (1427) 2 (1412) 4 (960) 5 (824) 1000 1400 1200 n /cm71 lexcit=303 nm and at 170 ± 270 8C, the detection limit was approximately four orders of magnitude lower; it was 0.09 ± 0.13 pg cm73 (see Ref.58). The detection power of the setup using the UV absorption method in the short-wavelength region (*224 nm) has been eval- uated. The limit of detection of b-TCDD, 1,2,3,7,8-Cl5DD and OCDD in vapours amounted to 8 ± 90 pgcm73 (see Ref. 58). It is noteworthy that the methods of detection mentioned above are not selective because the UV absorption and fluorescence bands of dioxins are very broad and, moreover, the maxima of these bands shift only slightly upon the change in the number of chlorine atoms in PCDD molecules (see Section IV). The use of phosphorescence, more precisely, low-temperature quasi-line phosphorescence spectra of solid solutions, for identi- fication of dioxins has also been proposed.73, 82, 85 For example, it is known that the luminescence spectra recorded by this method (Shpol'skii method) have been used successfully to determine the content of polycyclic aromatic hydrocarbons as hazardous con- taminants in foodstuffs and in non-edible oils.107, 108 The phos- phorescence spectra of solid solutions of DD and many of PCDD in a 1 : 5 : 5 mixture of ethanol, isopentane and ether 73 at 77Kand in n-alkanes at 77 and 4.2K81, 82, 85 were shown to possess a quasi- line structure, the structures of spectra of different PCDD being substantially dissimilar (see Table 6).The differences are observed both in the positions of lines and in the intensity distribution.Thus, these spectra possess advantages over the structureless UV absorption and fluorescence spectra in the identification of com- pounds. Figure 17 presents the phosphorescence spectrum (4.2 K) of a mixture of b-TCDD and a,b-TCDD isomers with concentrations a Irel 1 1 1 b 2 2 2 c 1 1 2 1 2 2 415 410 420 Figure 17. Phosphorescence spectra of a,b-TCDD (a), b-TCDD (b) and their mixtures (c) in n-hexane at 4.2 K.82 In spectrum c, the analytical lines 1 and 2 corresponding to the two isomers (a and b) are marked. 1053 11 440 425 l /nm1054 of 4.561075 and 361075 mol litre71, respectively. The quasi- lines observed in the emission spectrum of a two-component solution can be assigned unambiguously to one of the mixture components, which is a clear illustration of the possibility of selective determination of isomers using these spectra.At 77 K, the phosphorescence spectra of PCDD isomers are less suitable for selective recording due to the greater (*2 ± 4 nm) half-width of bands. However, the simplicity of the equipment and the process of recording provides the possibility of developing a screening method for dioxin detection based on them. The detection capacity of the spectral setups used for the detection of dioxins on the basis of quasi-line phosphorescence spectra has been evaluated.73, 81, 82 At 77 K, for n-hexane solutions of dioxins containing 4 to 7 Cl atoms, the detection limit is 1 ± 12 ng cm73 (see Ref.73). The spectra of b-TCDD were found to be recorded reliably down to a concentration of 16 ng cm73 (see Ref. 82). When analysing microvolumes of b-TCDD with recording of the most intense line in the spectrum (l=410 nm, see Table 6), determination of its amounts not lower than 100 pg is possible. It is expected that the limit of detection of tetrachloro-substituted DD could be markedly decreased by decreasing the temperature to 4.2 K, due to the increase in the peak intensity of vibronic lines. The differences between the quasi-line phosphorescence spec- tra even for isomers with close structures and for PCDD with different numbers of Cl atoms gives hope that development of such a method would allow reliable identification of PCDD in complex mixtures.VII. Conclusion Since it is the molecular structure that accounts for the different photophysical properties of compounds, it is important to estab- lish the geometrical structure of dioxins. X-Ray diffraction data point to a planar or nearly planar structure of DD and PCDD molecules in the crystals. No direct experimental data on the structures of these molecules in the gas phase or in solutions are available to date. Only indirect methods of determination do exist, relying on the data on the dipole moment ofDDin solutions or on the results of theoretical calculations. Depending on the theoret- ical method of modelling employed, the equilibrium nuclear configuration of the molecule in the ground electronic state S0 is found to be either planar or non-planar.In the case where the dioxin molecules are found to be non-planar, some researchers concluded that, in particular, DD has a `butterfly' configuration. Data on the structures of dioxin molecules are contradictory. Some investigators who give credit to X-ray diffraction data believe that the dioxin molecules are planar; other researchers who rely on calculations consider that the molecules are non- planar. Note that the intramolecular vibration mode of the `butterfly' type shows itself at a low frequency for all dioxins. Dioxin molecules might occur in different configurations depend- ing on the experimental conditions (temperature, intermolecular interactions). In some studies, attempts are made to relate different toxicities of PCDD to either planar or non-planar geometrical congfigura- tion.Spectroscopic methods of investigation are especially prom- ising for the solution of the problem of the structure of dioxins (at least, four of them, DD, b-TCDD, 1,4,6,9-Cl4DD and OCDD) which belong to the D2h group of symmetry in the case of a planar configuration. Analysis of the IR and Raman spectra from the standpoint of selection rules could promote solution of the problem of the molecular configuration in the ground electronic state S0. This type of analysis of the fine vibrational structure of the spectra recorded by electronic-vibrational spectroscopy meth- ods (low-temperature UV absorption, hole burning, fluorescence, phosphorescence, excitation of luminescence) and by the reso- nance enhanced two-photon ionisation technique could facilitate solution of the problem (which has not yet been solved either) of the symmetry of the dioxin molecules in the S1 and T1 states.E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov However, no comprehensive experimental and theoretical studies of dioxin vibrations that could answer the questions about the molecular configuration have been reported so far. In view of the foregoing, one of the most important problems in the studies of dioxins is to investigate the intramolecular vibrations (both theoretically and experimentally). At present, almost no data on the Raman spectra are available.Despite the extensive use of IR spectroscopy for analytical purposes, data on low-frequency vibrations can scarcely be found in the literature. The results on the theoretical calculations that provide informa- tion on not only the frequencies but also the forms and types of symmetry of vibrations are also few. The data on the vibrations of dioxin molecules are also important for the interpretation of the fine vibrational structure of the electronic-vibrational spectra, in particular, phosphorescence spectra of solid solutions because the structure of these spectra is complicated by the presence of the Shpol'skii multiplet components. Interpretation of the vibrational structure of these spectra provides valuable information on the type of symmetry of the triplet electronic state, on the influence of chlorine atoms on the processes of photophysical deactivation of the electronically excited state and on the changes in the nuclear configuration of the molecule in this state.Fourier Transform IR spectroscopy with matrix isolation of the molecule (combined with chromatography) is currently the main spectral method used to detect and identify dioxins. We believe that the Shpol'skii method would become yet another technique for the detection and identification of dioxins (at present, it is used for polycyclic aromatic hydrocarbons), provided that the light-gathering power of the equipment would be increased somewhat, for instance, by using FT spectrometers. At presence, the results of investigation of the fine-structure phos- phorescence spectra of solutions of dioxins in non-branched alkanes at 77 and 4.2K (using a standard spectrometer with diffraction gratings) confirmed the good prospects for using this technique for analytical purposes as a highly sensitive and selective method.The knowledge of experimental photophysical characteristics of dioxins and their changes induced by particular factors (medium, temperature, etc.) is required not only for theoretical investigations of the mechanisms of deactivation of electronic states but also for the solution of some practical problems. Thus it is known that photochemical decomposition, which is supposed to proceed through a T1 state, is the most effective process for the destruction of these toxicants.However, the available data on the photophysical studies of dioxins are obviously insufficient to provide optimisation of the methods for their photochemical deactivation. These data are also needed for searching for the ways of increasing the sensitivity of the spectral luminescence methods of analysis. New data on the structure and properties of dioxins would help to disclose the mechanism of action of these ecotoxicants; they would enable further development and perfection of both analytical methods for the detection of dioxins and methods of environmental protection. The review was supported by the Russian Foundation for Basic Research (Project No. 99-03-33215). References 1. L Friedman, D Firestone,W Horwitz, D Banes,M Anstead, G Shue J.Ass. Offic. Anal. Chem. 42 12 (1959) 2. A E Pohland, G C Yang J. Agricult. Food Chem. 20 1093 (1972) 3. B M Shepard, A L Young, in Human and Environment Risks of Chlorinated Dioxins and Related Compounds (Eds R E Tucker, A L Young, A P Gray) (New York: Plenum, 1983) 4. S Safe, O Hutzinger (Eds) Environmental Toxin. Ser. 3 (Berlin: Springer, 1990) Pt. 1 5. L A Fedorov Dioksiny kak Ekologicheskaya Opasnost', Retrospektiva i Perspektivy (Dioxins as Ecological Danger, Retrospective and Perspectives) (Moscow: Nauka, 1993)Optical spectra and photophysical properties of polychlorinated dibenzo-p-dioxin derivatives 46. R Zimmermann,U Boesl, C Weickhardt,D Lenoir, K-W Schramm, A Kettrup, E W Schlag Chemosphere 29 1877 (1994) 47.S Sommer, R Kamps, S Schumm, K F Kleinermanns Anal. Chem. 69 1113 (1997) 48. M V Vol'kenshtein, L A Gribov, M A El'yashevich, B I Stepanov Kolebaniya Molekul (Oscillation of Molecules) (Moscow: Nauka, 1972) 49. E A Gastilovich, V G Klimenko, K A Mishenina Zh. Fiz. Khim. 56 2789 (1982) d 50. YaMKolotyrkin (Ed.) Kolebatel'nye Spektry Mnogoatomnykh Molekul (Vibrational Spectra of Polyatomic Molecules) (Moscow: Nauka, 1986) 51. P Pulay, G Fogarasi, G Pongor, J E Boggs, A Vargha J. Am. Chem. Soc. 105 7037 (1983) 52. E A Gastilovich, N V Korol'kova, V G Klimenko Opt. Spektrosk. 85 764 (1998) c 53. V I Baranov, A N Solov'ev Zh. Fiz. Khim. 59 1720 (1985) d 54. R J Wratten,M A Ali Mol. Phys. 13 233 (1967) 55.R H Stehl, R R Papenfuss, R A Bredenweg, R W Roberts Adv. Chem. Ser. 120 119 (1973) 56. M B Ryzhikov, A N Rodionov, A N Stepanov Zh. Fiz. Khim. 63 2515 (1989) d 57. M B Ryzhikov, A N Rodionov, V G Klimenko, V F Traven' Metalloorg. Khim. 2 898 (1989) e 58. D J Funk, R C Oldennborg, D-P Dauton, J P Lacosse, J A Draves, T J Logan Appl. Spectrosc. 49 105 (1995) 59. E S Stern, C J Timmons Electronic Absorption Spectroscopy in Organic Chemistry (London: Edward Arnold, 1970) 60. Y-D Shin, H Saigusa, M Z Zgierski, F Zerbetto, E C Lim J. Chem. Phys. 94 3511 (1991) 61. T V Plakhotnik, R I Personov, E A Gastilovich Chem. Phys. 150 429 (1991) 62. K K Rebane Impurity Spectra of Solids (New York: Plenum, 1970) 63. I S Osad'ko Usp. Fiz. Nauk 128 31 (1979) f 64.E A Gastilovich, S A Serov, N V Korol'kova, V G Klimenko Opt. Spektrosk. 88 357 (2000) c 65. E A Gastilovich, S A Serov, N V Korol'kova, V G Klimenko J. Mol. Struct. 553 243 (2000) 66. N S Strokach, D N Shigorin, N A Shcheglova Elektronno-Koleba- tel'nye Spektry Mnogoatomnykh Molekul (Electronic Vibrational Spectra of Polyatomic Molecules) (Moscow: Nauka, 1982) 67. M F El'yashevich Atomnaya i Molekulyarnaya Spektroskopiya (Atomic and Molecular Spectroscopy) (Moscow: Fizmatgiz, 1962) 68. N V Korol'kova, E A Gastilovich Zh. Fiz. Khim. 71 1060 (1997) d 69. R N Nurmukhametov Pogloshchenie i Lyuminestsentsiya Aromati- cheskikh Soedinenii (Absorption and Luminescence of Aromatic Compounds) (Moscow: Khimiya, 1971) 70. M Orrit, J Bernard, R I Personov J.Phys. Chem. 97 10256 (1993) 71. J-M Bonnier, P J Jardon J. Chim. Phys. Phys.-Chim. Biol. 68 432 (1971) 72. V G Plotnikov Usp. Khim. 49 327 (1980) [Russ. Chem. Rev. 49 172 (1980) 73. L M Khasawnen, G D Winefordner Talanta 35 267 (1988) 74. A C Albrecht J. Chem. Phys. 38 354 (1963) 75. R M Hochstrasser Molecular Aspects of Symmetry (New York, Amsterdam: Benjamin INC, 1966) 76. E A Gastilovich, V G Klimenko, B V Ni Opt. Spektrosk. 78 770 (1995) c 77. K Hamanoue, T Nakayama, I Tsujimoto, S Miki, K Ushida J. Phys. Chem. 99 5802 (1995) 78. S Dobrin, P Kaszynski, S Ikeda, J Waluk Chem. Phys. 216 179 (1997) 79. E V Shpol'skii Usp. Fiz. Nauk 80 255 (1963) f 80. J L Skenner, W E Moerner J. Phys. Chem. 100 13251 (1996) 81.R N Nurmukhametov, V G Klimenko, I L Belaits, A S Lebedev Zh. Fiz. Khim. 71 148 (1997) d 82. V G Klimenko, R N Nurmukhametov J. Fluorescence 8 129 (1998) 83. T N Bolotnikova Opt. Spektrosk. 7 217 (1959) c 84. A M Rozen, V G Yurkin, Yu V Konovalov Zh. Fiz. Khim. 71 351 (1997) d 85. V G Klimenko, R N Nurmukhametov, A A Stekhin Dokl. Akad. Nauk 336 489 (1994) g 86. E A Gastilovich,N V Korol'kova, V G Klimenko Zh. Strukt. Khim. 6. A K Prokof'ev Usp. Khim. 59 1799 (1990) [Russ. Chem. Rev. 59 1051 (1990)] 7. L A Fedorov, B F Myasoedov Usp. Khim. 59 1818 (1990) [Russ. Chem. Rev. 59 1063 (1990)] 8. A D Kuntsevich, V F Golovkov, V R Rembovskii Usp. Khim. 65 29 (1996) [Russ. Chem. Rev. 65 24 (1996)] 9. N A Klyuev Zh. Anal. Khim. 51 163 (1996) a 10.V N Maisterenko, R Z Khalitov, G K Budnikov Ekologicheskii Monitoring Superekotoksikantov (Ecological Monitoring of Super- ecotoxicants) (Moscow: Khimiya, 1996) 11. M Senma, Z Taira, T Taga, K Osaki Cryst. Struct. Commun. 2 311 (1973) 12. AW Cordes, C K Fair Acta Crystallogr., Sect. B 30 1621 (1974) 13. P Singh, J D McKinney Acta Crystallogr., Sect. B 34 2956 (1978) 14. J S Cantrell, D W Tomlin, T A Beiter Chemosphere 19 183 (1989) 15. F P Boer, F P van Remoortere, P P North, M A Neuman Acta Crystallogr., Sect. B 28 1023 (1972) 16. F P Boer,MA Neuman, F P van Remoortere, P P North, HWRinn Adv. Chem. Ser. 120 14 (1973) 17. J S Cantrell, N C Webb, A J Mabis Acta Crystallogr., Sect. B 25 150 (1969) 18. C J Koester, J C Huffman, R A Hites Chemosphere 17 2419 (1988) 19.Qin Zhaohai, Chen Changying, Li Yanan, Li Weigang Wuli Huaxue Xuebao (Acta Phys.-Chim. Sin. ) 6 238 (1990); Chem. Abstr. 113 181 961 (1990) 20. N A Shimanko,M V Shishkina Infrakrasnye i Ul'trafioletovye Spektry Pogloshcheniya Aromaticheskikh Efirov (IR and UV Absorption Spectra of Aromatic Ethers) (Moscow: Nauka, 1987) 21. F P Colonna, G D Ano, V Galasso, K J Irgolic, C F King, G C Pappalardo J. Organomet. Chem. 146 235 (1978) 22. G Rauhut, P Pulay J. Am. Chem. Soc. 117 4167 (1995) 23. R Zimmermann, U Boesl, D Lenoir, A Kettrup, Th L Grebner, H J Neusser Int. J. Mass Spectrom. Ion Process. 145 97 (1995) 24. J Grainger, V V Reddy, D G Patterson Jr Appl. Spectrosc. 42 643 (1988) 25. A V Fokin, N P Vorob'eva, N I Raevskii, Yu A Borisov, A F Kolomiets Izv. Akad. Nauk SSSR, Ser. Khim. 2544 (1989) b 26. Th L Grebner, B Ernstberger, H J Neusser Int. J. Mass Spectrom. Ion Process. 136 101 (1994) 27. B Ruscic, L Klasinc, A Wolf, J V Knop J. Phys. Chem. 85 1490 (1981) 28. E A Gastilovich, V G Klimenko, B V Ni Opt. Spektrosk. 82 765 (1997) c 29. N V Korol'kova, V G Klimenko, E A Gastilovich Zh. Fiz. Khim. 72 1251 (1998) d 30. E A Gastilovich Kratkie Soobsh. Fiz. FIAN 76 (1996) 31. T Chen Jo-Yun J. Ass. Offic. Anal. Chem. 56 962 (1973) 32. D F Gurka, S Billets, J W Brasch, C J Riggle Anal. Chem. 57 1975 (1985) 33. D F Gurka, J W Brasch, R H Barnes, C J Riggle, S Bourne Appl. Spectrosc. 40 978 (1986) 34. J Grainger, L T Gelbaum Appl. Spectrosc. 41 809 (1987) 35. J Grainger, E Barnhart, D G Patterson Jr, D Presser Appl. Spectrosc. 42 321 (1988) 36. T T Holloway, B J Fairless, CE Freidine,HE Kimball,RDKloepfer, C J Wurrey, L A Jonooby, H G Palmer Appl. Spectrosc. 42 359 (1988) 37. M M Mossoba, R A Niemann, J-Y T Chen Anal. Chem. 61 1678 (1989) 38. E Koglin J. Mol. Struct. 173 369 (1988) 39. V G Klimenko, R N Nurmukhametov, E A Gastilovich Opt. Spektrosk. 86 239 (1999) c 40. V G Klimenko, R N Nurmukhametov, E A Gastilovich, S A Lebedev Opt. Spektrosk. 88 385 (2000) c 41. V G Klimenko, R N Nurmukhametov, E A Gastilovich Opt. Spektrosk. 83 92 (1997) c 42. V G Klimenko, R N Nurmukhametov, E A Gastilovich, S A Lebedev Opt. Spektrosk. 88 379 (2000) c 43. V G Klimenko, R N Nurmukhametov, S A Serov, E A Gastilovich Opt. Spektrosk. 89 49 (2000) c 44. C Weickhardt, R Zimmermann, U Boesl, E W Schlag Rapid Commun. Mass Spectrom. 7 183 (1993) 45. R Zimmermann, C Weickhardt,U Boesl,D Lenoir, K-W Schramm, A Kettrup, E W Schlag Organohalogen Compd. 11 91 (1993) 41 149 (2000) h 1055E A Gastilovich, V G Klimenko, N V Korol'kova, R N Nurmukhametov 1056 87. E A Gastilovich Usp. Fiz. Nauk 161 (7) 83 (1991) f 88. YaMKolotyrkin (Ed.) Elektronno-Vozbuzhdennye Sostoyaniya Mnogoatomnykh Molekul (Electronically Excited States of Polyatomic Molecules) (Moscow: Nauka, 1993) 89. G Ya Gerasimov Khim. Fiz. 18 (6) 87 (1999) i 90. G Choudhary, L H Keith, C Rappe Chlorinated Dioxins and Dibenzofurans in the Total Environment (Boston, MA: Butterworth, 1983) 91. M van den Berg, L Birnbaum, A T C Bosveld, B Brunstrom, Ph Cook,M Feeley, J P Giesy, A Hanberg, R Hasegawa, S W Kennedy, T Kubiak, J Ch Larsen, F X R van Leeuwen, A K D Liem, C Nolt, R E Peterson, L Poellinger, S Safe, D Schrenk, D Tillitt, M Tysklind, M Younes, F Waern, T Zacharewski Environ. Health Perspect. 106 775 (1998) 92. R K Mitchum, G F Moler,W A Korfmacher Anal. Chem. 52 2278 (1980) 93. J R Donnelly,W D Munslow J. Chromatogr. 392 51 (1987) 94. S Terabe, Y Miyashita, O Shibata, E R Barnhart, L Alexander, D G Patterson, B L Karger, N Tanaka, K Nosoya J. Chromatogr. 516 23 (1990) 95. K Kimata, K Hosoya, T Araki, N Tanaka, E R Barnhart, L R Alexander, S Sirimanne, P C McClure, J Grainger, D G Patterson Jr Anal. Chem. 65 2502 (1993) 96. D P Samsonov, V P Kiryukhin, N P Zhiryukhina Zh. Anal. Khim. 49 868 (1994) a 97. DP Samsonov,KYu Evdokimov,NP Zhiryukhina, V P Kiryukhin Zh. Anal. Khim. 53 754 (1998) a 98. N Ragunathan, K A Krock, C Klawun, T A Sasaki, C L Wilkins J. Chromatogr. A 856 349 (1999) 99. L T Gelbaum, D Ashley, D G Patterson Jr Chemosphere 17 551 (1988) 100. G T Reedy, D G Ettinger, J F Schneider, S Bourne Anal. Chem. 57 1602 (1985) 101. M M Mossoba, J T Chen,W C Brumley, S W Page Anal. Chem. 60 945 (1988) 102. T J Nestrick, L L Lamparski, R H Stehl Anal. Chem. 51 2273 (1979) 103. H R Buser, Ch Rappe Anal. Chem. 52 2257 (1980) 104. J Grainger, P C McClure, Z Lui, B Botero, S Sirimanne, D G Patterson Jr Chemosphere 32 13 (1996) 105. R M Lazorenko-Manevich Elektrokhimiya 33 827 (1997) j 106. A D Kuntsevich, E V Kuchanskii, I T Polyakov, V F Golovkov, A A Druzhinin, V N Syutkin Dokl. Akad. Nauk 335 326 (1994) g 107. T A Teplitskaya, T A Alekseeva,M M Val'dman Atlas Kvazili- neichatykh Spektrov Lyuminestsentsii Aromaticheskikh Molekul (Atlas of Quasi Line Spectra of Luminescence of Aromatic Molecules) (Moscow: Moscow State University, 1978) 108. C Gooijer, F Ariese, JWHofstraat (Eds) Spol'skii Spectroscopy and Other Site-Selection Methods (Applications in Environmental Analysis, Bioanalitical Chemistry and Chemical Physics) (New York: Wiley, 2000) a�J. Anal. Chem. (Engl. Transl.) b�Russ. Chem. Bull. (Engl. Transl.) c�Opt. Spectroscop. (Engl. Transl.) d�Russ. J. Phys. Chem. (Engl. Transl.) e�Russ. J. Organomet. Chem. (Engl. Transl.) f�Physics-Uspekhi (Engl. Transl.) g�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) h�RChem. (Engl. Transl.) i�Chem. Phys. Rep. (Engl. Transl.) j�Russ. J. Electrochem. (Engl. Tran

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

Russian Chemical Reviews 69 (12) 1057 ± 1082 (2000) Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method V N Vertoprakhov, S A Krupoder Contents I. Introduction II. Equipment used in CVD processes III. Barrier layers preventing the diffusion of copper atoms into the substrate IV. The starting copper-containing compounds: copper(II) derivatives V. The starting copper-containing compounds: copper(I) derivatives VI. The problem of selective deposition of copper films VII. Conclusion Abstract. thin of preparation the of aspects chemical main The The main chemical aspects of the preparation of thin copper films from the vapour of monovalent and divalent copper copper films from the vapour of monovalent and divalent copper derivatives are technique CVD the in precursors as derivatives as precursors in the CVD technique are considered. considered.Data of properties the and synthesis of methods the on Data on the methods of synthesis and the properties of various various types described and generalised are compounds these of types of these compounds are generalised and described system- system- atically. copper of mechanisms decomposition possible The atically. The possible decomposition mechanisms of copper com- com- pounds prospects The discussed. are conditions CVD under pounds under CVD conditions are discussed. The prospects of of using copper thin of preparation the for technology CVD the using theCVDtechnology for the preparation of thin copper films films in outlined.are devices microelectronic of circuits integrated in integrated circuits of microelectronic devices are outlined. The The bibliography references 432 includes bibliography includes 432 references. I. Introduction Micro- and nanoelectronics are the most promising fields for the application of thin copper films. Such films are used as conducting paths in large-scale and very-large-scale integrated circuits (LIC and VLIC, respectively); they make an alternative to aluminium films. It is known that non-doped aluminium films used as electric wiring in LIC and VLIC have a number of drawbacks. First, discontinuity of aluminium films is observed during operation of integrated circuits because of the large atom electromigration coefficient, which was shown by calculations to be due to relaxation of the bulk upon passage of a flow of vacancies.1 On the other hand, the electromigration coefficients in the layers of copper atoms are a few orders of magnitude smaller than in aluminium atom layers.Second, the specific electric resistance (SER) of copper is 30% smaller than that of aluminium. The addition of copper or silicon (2 at.% ± 4 at.%) to aluminium films does not eliminate all the drawbacks of aluminium as a material for interconnections in LIC and VLIC.2±9 In addition, thin copper films are used in devices which employ the effect of the gross magnetic resistance in polycrystal- line Co/Cu multilayers. The thickness of the cobalt and copper V N Vertoprakhov, S A Krupoder Institute of Inorganic Chemistry, Siberian Branch of the Russian Academy of Sciences, prosp.Acad. Lavrent'eva 3, 630090 Novosibirsk, Russian Federation. Fax (7-383) 234 44 89. Tel. (7-383) 234 16 46. E-mail: nikiphorovich@mail.ru (V N Vertoprakhov), magadanian@yahoo.com (S A Krupoder) Received 30 July 2000 Uspekhi Khimii 69 (12) 1149 ± 1177 (2000); translated by S S Veselyi #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n12ABEH000572 1057 1058 1058 1060 1068 1076 1077 layers ranges from a few fractions of a nanometre to several nanometres.10 ± 14 Copper films are deposited onto substrates. Silicon and silicon dioxide are the most common substrate materials. The diffusion coefficients of copper into silicon and silicon dioxide are high, they are a few orders of magnitude higher than those of other metals.The use of barrier layers in order to decrease this coefficient was suggested. At first, it was attempted to solve this problem by means of PVD (Physical Vapour Deposition) technology, using such methods as thermal evaporation, plasma spraying, magnet- ron or laser scattering, as well as by spray deposition of thin layers of other metals (Cr, Co, W, Ti, Ta) on an SiO2 surface. However, little progress has been made in this direction. Later, metal nitrides used as the barrier layer materials showed better performance than the pure metals. The common drawback of PVD technology is the necessity of heating a substance to high temperatures (above the vaporisation temperature).Therefore, intense development of CVD (Chemical Vapour Deposition) technology has started in recent years; this technique involves chemical methods of deposition of thin films from the gas phase using metal-containing compounds. Chemical methods have a number of advantages over physical methods. These include relatively mild conditions for film deposition, comparatively low vaporisation temperatures of the starting compounds, conditions favourable for good adhesion of the films to the substrate, wider technological prospects for the development of microelectronic devices with required parameters. This review considers the chemical methods used in CVD technology for the preparation of copper films from volatile copper(II) and copper(I) derivatives.The main attention is given to work published within the period from 1991 to 1999, although earlier fundamental studies have also been taken into account. The majority of compounds which we name `copper-contain- ing compounds' (CCC) correspond to copper complexes with copper ± oxygen bonds. As to the specific CCC types, the review mentions virtually all classes of compounds considered as promis- ing starting compounds for CVD technology. Many of these compounds have not yet found application in the synthesis of copper films. Nevertheless, we felt it worthwhile to mention them as well, at least because it is quite possible that other CCC of the same classes may be suitable for CVD processes.1058 II. Equipment used in CVD processes A vast literature (see, e.g., Refs 15 ± 28) is devoted to CVD technology.In essence, the CVD process involves evaporation of volatile metal-containing compounds followed by transfer of their vapour to a reaction compartment, where the starting compound is decomposed and the decomposition products are deposited onto a substrate in the form of a film. Most commonly, decom- position is accomplished by means of thermolysis, sometimes along with exposure to plasma or irradiation. Side products are removed from the reaction zone in the form of vapour. If the process is carried out with the use of an inert gas as the carrier for the vapour of the starting compound, it is usually purified from oxygen and other admixtures by passing it prelimi- narily through adsorbents or a membrane.Hydrogen is purified by passing it through palladium.29 Liquid compounds are injected using devices based on controlling the mass consumption rate.30, 31 Various designs of reactors have been developed for the CVD processes.{ As an example, Fig. 1 shows a simple reactor scheme for pyrolytic deposition of films by evaporation of metal-contain- ing compounds followed by transportation of the vapour to the deposition zone. 1 2 3 45 6 7 9 8 Figure 1. The simplest CVD reactor scheme;40 (1) flow controller, (2) reactor, (3) substrate, (4) heater, (5) thermocouple, (6) CCC source, (7) manometer, (8) pump, (9) liquid nitrogen trap.Many reactor designs and film deposition processes have been patented.35, 41 ± 44 A number of firms manufacture serial reactors, e.g., MR-200 (Cambridge Instruments, Great Britain),45 EMCORE (GS/3300 firm, USA) with a rotating disk and uni- formity of the substrate temperature to within 1 8C,46 EPISON (Thomas Swan Ltd, Great Britain).47, 48 { They can arbitrarily be divided in two groups: vertical 32 ± 35 and horizontal.35 ± 39 A review of the equipment used before 1972 for the deposition of films and coatings by CVD can be found in Ref. 15 (see p. 180 ff ). V N Vertoprakhov, S A Krupoder High-performance technological equipment for the prepara- tion of thin layers by CVD has been developed at the Samsung company's institute (Republic of Korea).This equipment makes it possible to manufacture more than 10 000 devices (diodes, photo diodes, etc.) simultaneously.36 The design of aCVDreactor in which hydrogen plasma is used was reported.49 High-quality copper films with a SER of 2.0 mO cm are obtained in such reactors. In recent years, combined instruments (of the so-called cluster type) for the deposition of any type of thin films by physical or chemical methods, including CVD technology, have appeared (see, e.g., Ref. 50). These instruments allow one to deposit multi- layer coatings from different materials.51 An example of such an instrument is shown in Fig. 2. 1 4 3 2 6 5 8 7 Figure 2. Block diagram of a `cluster' apparatus for the preparation of thin films;52 (1, 2) process modules, (3) observation module, (4) multifunctional module (exposure, polishing, etc.), (5) sample preparation module, (6, 8) measuring modules, (7) support inlet.CVD methods are successfully used if high-quality and uni- form films (both with respect to the composition and thickness) of large area are required. For this purpose, the substrates in the reactor are rotated; laminar gas streams in the reactor and uniform heating are provided. In this way, thickness uniformity within 2% and chemical composition uniformity within 1% can be achieved over the entire film surface.53 ± 56 Visualisation of convective flows in the growth chamber can be provided by laser illumination of dispersed particles or by creation of a 3Dimage by interference holography.The production of integrated circuits requires the preparation of planar deposited copper layers. For this purpose, the technol- ogy of chemical and mechanical polishing of copper films has been developed which includes treatment of the film surface with a suspension of fine SiO2 and Al2O3 particles; benzotriazole is added as an oxidant if necessary. In the latter case, the rate of film surface polishing depends on the concentration of benzotria- zole.57, 58 The waste gases are utilised using catalytic combustion, physical adsorption with activated carbon with or without oxida- tion and dry purification in reaction columns.59, 60 The CS-GmbH company (Germany) has developed three types of gas absorbents: absorbing cartridges (Cleanvent), washing and purifying systems (Cleansorb) and emergency absorption systems (Clean-Protect).The Cleanvent cartridges are intended for in situ removal of hazardous gases from gas distribution systems.61 III. Barrier layers preventing the diffusion of copper atoms into the substrate Diffusion of copper atoms into the substrate during the operation of microelectronic devices impairs their characteristics. This can be prevented by means of barrier layers between the copper film and the substrate. These layers consist of certain metals and/or binary compounds and can be obtained using PVD or CVDPreparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method technologies.CVD favours the formation of amorphous films, as shown, e.g., for tantalum nitride.62 1. Barrier layers prepared by PVD methods Originally, diffusion of copper atoms into silicon as the most common substrate was prevented by thin barrier layers of dia- mond-like carbon and metal films obtained by PVD methods.63, 64 Films of various binary compounds, primarily titanium, tungsten and tantalum silicides, borides and nitrides, are also suitable for this purpose.65 ± 85 Tungsten and tantalum are regarded as the best metals for the deposition of barrier layers by PVD, as they have the smallest self- diffusion coefficients in the series Ti>Cr>Nb>Mo> Ta*W.64 After deposition, the layers are exposed to annealing in hydrogen for 1 h at 600 8C.64 Metal film barrier layers were gradually replaced by films of titanium and tungsten silicides and nitrides.65 ± 74 In this case, the thermal stabilities of the layers increased to 500 8C (WSix multi- layer) 68 and 700 8C (WSiN/WSix/W multilayer).69 The layers of solid solutions of titanium in tungsten, TiW (30 at.% Ti), have somewhat different characteristics. If the barrier layers are 100 nm thick, the Cu/TiW/Si multilayers are stable up to 775 8C, and their stabilities increase to 850 8C upon keeping in air.In the latter case, 0.2 at.% of oxygen was detected in the film.72 The pure tantalum film 10 nm thick is an efficient barrier against the diffusion of copper into the silicon substrate up to 500 8C.69 The films of tantalum nitrides, viz., Ta2 N, TaN, and those of the Ta36Si44N50 alloy provide even more reliable protec- tion against penetration of copper atoms into the silicon sub- strate.75 ± 77, 79, 82 The characteristics of barrier layers are listed in Table 1.2. Barrier layers prepared by the CVD method The compositions and properties of barrier layers obtained by thermolysis using the CVD technology depend both on the Table 1. Parameters of barrier multilayers preventing the diffusion of copper atoms into the silicon substrate. Multilayers Cu/(DLC)/Si Cu/M/Si, M=Cr, Ti, Nb, Mo, Ta,W Cu/Cr/Si Cr/Cu/Co/Au Cu/CrNx /Si Cu/Cr/Si Cu/WSix /Si Cu/WSiN/WSix/W/Si Cu/WSiN/WSix /Si Cu/TiB2/Si Cu/TiW/Si Cu/TiW/Al/SiO2/Si Cu/TiN/Si Cu/TiN/Si Cu/TaN/Si Cu/TaN/Si Cu/TaN/SiO2/Si Cu/Ta2N/Si Cu/Ta/Si Cu/Ta+CeO2/Si Note.The following designations are used: l is barrier layer thickness; T1 is the upper limit of the stability temperature range; T2 is the annealing temperature where new phases are formed; DLC is diamond-like carbon; PACVD is plasma-activated CVD method; LP is low pressure. Deposition method PECVD CVD """"PACVD ""LPCVD PECVD PVD LPCVD CVD PVD """"" 1059 conditions of the process and on the properties of the starting volatile compounds. Generally, the CVD method allows one to work with a wide range of starting volatile compounds of metals (for example, 15 starting volatile compounds have been listed 78 for the preparation of just one compound, TiN); layers with required properties can be obtained by varying the process parameters (vapour phase composition, deposition temperature, pressure).Table 2 lists the starting compounds and some characteristics ofCVDprocesses for the preparation of thin barrier layers of TiN, TaN, Ta, Ti, Cr, and Co. It should be noted that PVD methods for the preparation of thin layers from the gas phase (spraying of compounds by bombardment with high-energy electrons and ions, magnetron and radio-frequency spray deposition, etc.) require that higher energy was applied to the substance for evaporation. The growth of films generally starts with the formation of separate islets- droplets, which subsequently merge to form a continuous film.Polycrystalline films are mostly formed; this is undesirable in terms of the diffusion of copper atoms, as the grain boundaries have higher diffusion coefficients than the grain bulk. On the other hand, CVD of compounds from the gas phase generally occurs at lower temperatures. In this case, amorphous films are formed. The diffusion coefficients of copper atoms through the bulk of such films are smaller than through the grain boundaries. This is one of the advantages of chemical methods for film deposition from the gas phase over physical methods. Yet another advantage of CVD methods is the possibility of selective processes, i.e., preferential deposition, in particular of copper, on one substrate material or surface area with certain mechanical characteristics with respect to another material or surface area.This is a very useful feature for the production of microelectronic devices, as it can replace lithographic template preparation in the creation of an integrated circuit. l /nm Ref. T2 /8C T1 /8C 63 64 400 600 100 60 77 400 400 70 500 500 650 500 650 775 450/500 450 550 500 10 ± 40 320 50 50 50 450 50 150 100 50/100 40 50 180 65 66 67 67 68 69 70 71 72 73 74 83 75 78 81 77 80 84, 85 500 500 800 7550 650 650 7850 777950 7 7 800 800 700 630 7 700/600 550 320 ± 630 800 25/10 50 20 71060 Table 2.Process parameters for CVD deposition of TiN, TaN, Ti, Ta, Cr and Co barrier layers. Starting compound T /8C Ti(NMe2)4 Ti(NEt2)4 (But2N)Ta(NEt2)3 745 60 77777740 ± 50 7777285 80 ± 82 25 ± 35 77777 CpTi(C7H7) (Cp)2Ti(N3)2 Ti(OPri)4 Ta(NEt2)5 Ar2CrI (C6H5COOEt)2Cr Co2(CO)8 CpCoEt2 Co(acac)2 Co(acac)3 Co(CO)3NO Note. The following designations are used: T is the vaporisation temperature of the compound; Tg is carrier gas temperature; P is pressure in the reactor. a In plasma. b Under UV irradiation. IV. The starting copper-containing compounds: copper(II) derivatives Volatile copper(II)-containing compounds had been synthesised long before the CVD technology appeared. Of these, chelate complexes of copper(II) with b-diketones 1 have been studied most thoroughly.Copper(II) acetylacetonate, Cu(acac)2 (1, R=R0=Me) syn- thesised in the past century,106 was the first compound of this series which may be regarded as an appropriate volatile CCC for CVD. By the early 1960s, the number of copper(II) b-diketonates with various starting b-diketones exceeded sixty.107 ± 109 R R0 CH HCC O O C Cu C O O CR0 R 1 Initially, the application of these compounds was limited to their use as convenient labels for characterisation of the corre- sponding b-diketones and in gas chromatography for the separa- tion of metals.109 As a matter of fact, the use of copper b-diketonates in CVD technologies began in 1965 when the first study devoted to the CVD synthesis of thin copper films based on b-diketonates appeared.110 To date, there are several reviews dealing with copper b-diketonates as volatile CCC (see, for example, Refs 26 and 111), but they did not cover the work published in recent years.Certainly, not all copper(II) b-diketonates and their deriva- tives are suitable for CVD processes. The synthetic accessibility of the starting b-diketone is of certain importance, but the practical value of a particular copper(II) b-diketonate is determined by the combination of its physicochemical properties. The relationship between the volatilities of metal b-diketo- nates (including that of copper) and their structures was analysed by Mazurenko et al.112, 113 (see also the review by Tsyganova and Dyagileva 114).The main criteria of stability of these compounds in the vapour phase have been stated in these reports. It is Tg /8C 100 ± 500 350 a 200 ± 450 300 ± 580 377 300 ± 350 250 ± 500 a 425 300 ± 580 425 450 ± 650 300 ± 600 a 350 b 500 ± 550 350 ± 400 320 260 200 ± 300 240 ± 400 320 ± 350 320 ± 350 350 ± 480 V N Vertoprakhov, S A Krupoder Ref. Substrate P /Torr Carrier gas Si SiO2/Si SiO2/Si 1 ± 20 10.8 ± 5 NH3 He, N2 , Ar Ar NH3 NH3 NH3 N2, H2, NH3 Ar, NH3 NH3 NH3 N2 H2, N2, NH3 N2 N2, H2 TiSi2 , Al, Cu InP Si Si Si Si Si Si Si steel 72 ± 7.5 0.5 ± 5 10 710 0.2 777 SiO2 , Si SiO2 , Cu, Al H2 H2 86, 87 88 89 7 7 84 90 75 87, 91 92 93 92 94 95 96 97 7 7 7 98 7 7 7 99 7 7 7 100 101, 102 103 104 104 105 0.01 ± 0.1 7 7 7 glass 7 7 " 1.5 SiO2/Si Ar, H2 believed 113 that the rather high volatilities of these compounds are due to shielding of the coordination unit by the highly developed `periphery' of the molecules, which decreases the contributions of the electrostatic and intermolecular interactions to the intermolecular potential.However, volatility is a necessary but insufficient condition for the applicability of compounds in CVD technology. The basic requirements to the starting CCC are as follows:113 the ability to undergo sublimation in low vacuum at as low temperatures as possible, providing as high a vapour pressure as possible; the ability to undergo sublimation and resublimation without decom- position; the ability to undergo thermolytic decomposition on a substrate surface at as low a substrate temperature as possible with a minimum possible content of admixtures in the resulting copper film.Analysis of these requirements makes it possible to conclude that the thermolytic stability of the starting compound should have an optimum. In the best case, an insufficiently stable compound will reach the substrate surface as decomposition products, which is unfavourable for reproducibility of the process. If the compound is very stable, high substrate temperature (or additional energy in the form of radiation) is inevitably required for thermolysis, which generally affects adversely the film quality.The starting volatile copper(II) derivatives with optimum technological properties for the CVD method have been sought among several basic groups of compounds considered below. 1. Copper(II) b-diketonates with alkyl groups The main advantage of copper(II) b-diketonates with alkyl groups is their synthetic accessibility. However, they were not the first compounds used for the development of the CVD preparation of copper films; their polyfluorinated analogues were the first sub- stances.110, 115, 116 This is most probably due to the rather low volatilities of this type of copper(II) b-diketonates.117 The b-dike- tones required for the preparation of metal b-diketonates can be synthesised rather easily by the Claisen condensation of the corresponding ketones with esters in the presence of nucleophilic reagents.118Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method EtONa RCOCH2COR0 RCOOEt +CH3COR0 7EtOH The suitability of copper(II) acetylacetonate Cu(acac)2 and copper(II) dipivaloylmethanate Cu(DPM)2 in CVD processes was considered. The latter compound is more volatile than Cu(acac)2 because the b-diketone has a branched hydrocarbon chain (see 1, R=R0=But).116 Substituted copper(II) acetoacetates 117 (Section IV.4a) pro- vide an alternative to these compounds in the hydrocarbon series.a.Synthesis and physicochemical properties The main method for the synthesis of Cu(acac)2 and other copper(II) b-diketonates with alkyl groups involves the reaction of the corresponding b-diketone or its hydrate with a hot aqueous solution of copper(II) acetate 119 or a mixture of a water-soluble copper(II) salt with sodium acetate.120 Copper b-diketonate hydrates are usually obtained in this way; anhydrous compounds are obtained by heating or by keeping the hydrates overH2SO4.120 It was noted, however, that the formation of a hydrate is untypical of Cu(acac)2.121 Direct reaction of copper(II) acetate with an anhydrous b-diketone, which occurs in several minutes in boiling b-diketone, is also possible.119 Cu(RCOCH2COR0)2 Cu(Ac)2+2RCOCH2OR0 72H(Ac) The structure and spectral properties of the simplest b-diket- onate, viz., Cu(acac)2, have been studied rather thoroughly: all bond lengths and bond angles in the Cu(acac)2 molecule have been determined;122 quantum-chemical calculations of the electronic structure of Cu(acac)2 have been carried out;123, 124 the ESR spectra of this compound have been recorded (see, e.g., Refs 125 and 126); a detailed assignment of its IR spectral bands has been performed.127 When deciding whether a compound can be used in CVD, its molecular structure and thermogravimetric data are of special importance.It was shown that in Cu(acac)2 and other copper(II) b-diketonates, the separate molecules form infinite chains through intermolecular copper ± chelate ring carbon bonds 128, 129 with a bond length of 0.301 ± 0.312 nm.This interaction complements the coordination of copper atoms to give a strongly stretched tetragonal bipyramid. The relative instabilities of these intermo- lecular bonds can be a reason for the volatility of this class of compound.113 The thermodynamic parameters of the Cu(acac)2 molecule are listed in Ref. 130. The vapour pressures of this compound were measured at various temperatures (for example, it is*0.3 Torr at 150 8C) and the temperature dependence of the vapour pressure was obtained.131 It was shown by thermogravimetry that Cu(acac)2 is sublimed completely and does not undergo decom- position in an inert atmosphere.132 On the other hand, the results of thermogravimetric analysis of Cu(acac)2 in air suggest that a single-step thermal decomposition occurs at 3058C.133 The usual temperature range of Cu(acac)2 sublimation in a CVD process is 100 ± 170 8C.134 Dipivaloylmethane H(DPM) and Cu(DPM)2 were first obtained by condensation of ketones with esters in the presence of NaNH2.121 Simpler methods for the synthesis of Cu(DPM)2, e.g., the reaction of H(DPM) with an aqueous solution of CuSO4 in the presence of sodium acetate, were developed later.135 The properties of Cu(DPM)2 have also been studied rather thor- oughly: the crystal structure has been determined;136 thermogra- vimetric data have been obtained showing that in a nitrogen atmosphere Cu(DPM)2 sublimes completely at 240 ± 245 8C;131, 132 its sublimation enthalpy (74.8 kJ mol71) and entropy (132.2 J mol71 K71) have been determined.135 Analysis of thermogravimetric curves for a wide range of cop- per(II) b-diketonates with alkyl groups 131, 132 showed that 1061 Cu(DPM)2 is actually the most volatile compound in the series where both ligands are identical (the compounds with different ligands are less accessible synthetically).b. The mechanisms of decomposition of copper-containing compounds on substrate surfaces The mechanisms of formation of copper films from copper(II) b-diketonates with alkyl groups in CVD processes have been studied in various respects. A number of studies (see, e.g., Refs 137, 138 and references cited therein) deal mainly with the physical aspects of the growth of copper films and its mathematical simulation. Turgambayeva et al.139, 140 used mass-spectrometric data to analyse the thermolysis of Cu(acac)2 molecules in the vapour phase in a vacuum as well as in a stream of oxygen and hydrogen.It was shown that, depending on the temperature, thermolysis can result in various volatile organic products, such as H(acac), acetone, ethyl methyl ketone, etc. The kinetic and activation parameters of the thermal decomposition of Cu(acac)2 in the vapour phase were reported;140, 141 this process was described satisfactorily by a first-order kinetic equation. It was concluded on the basis of IR spectra of Cu(acac)2 that the decomposition starts with destabilisation of the coordination unit at the copper7oxy- gen bonds of the complex.142 A similar study of the kinetics and thermal decomposition products in the vapour phase was also carried out for Cu(DPM)2.143 It is very important to have data on the mechanism of thermolytic decomposition of compounds directly on the sub- strate in order to bring the CVD processes to perfection.One of the first schemes suggested for the description of Cu(acac)2 decomposition on heated substrate surfaces involved a radical mechanism:16 (acac)Cu.+(acac). , Cu(acac)2 Cu+(acac). . (acac)Cu. It is considered that subsequent decomposition of the acetyl- acetone radical will produce contaminants (admixtures of carbon and copper oxides) in the film. A somewhat different scheme of Cu(DPM)2 decomposition is described in Refs 144 ± 146.Astudy of the thermolysis mechanism by in situ Fourier IR spectroscopy at a substrate temperature of 500 8C and a pressure of 20 Torr showed that the process starts with the cleavage of the C7C bonds of the chelate ring accom- panied by simultaneous partial elimination of the tert-butyl groups.As a result, isobutene is formed. The thermolysis is abruptly accelerated at 600 ± 700 8C; the films formed at temper- atures above 600 8C are considerably contaminated with carbon. The overall decomposition equation is: T Cu[(CH3)3COCH2OC(CH3)3] Cu+2C+4CO: + + 4H2C=C(CH3)2:+3H2: . The interaction of Cu(DPM)2 with a substrate surface was studied by IR spectroscopy during the decomposition of the vapour in the presence of H2O on the (100) planes of SiO2 crystals (see Refs 147 and 148) and SrTiO3 crystals (see Ref.149) with formation of a copper film. It was shown that the film formation on the SiO2 surface starts with stoichiometric chemisorption of the copper(II) b-diketonate involving reactive surfaceOHgroups. This reaction occurs even at 50 8C and results in the disappearance of the peak assigned to OH groups in the IR spectrum.148 It was found that one Cu(DPM)2 molecule at the surface is coordinated with two or three OH groups. In the presence of water vapour providing adsorbed water on the substrate or upon direct participation of OH group protons, decomposition of Cu(DPM)2 with elimination of one H(DPM) molecule and formation of copper(I) dipivaloylmethanate Cu(DPM) occur, as confirmed by XPS data.1062 Cu(DPM)2 (ads)+H2O(ads) H(DPM):+OH¡ÖadsÜ +Cu(DPM)(ads) As shown in more detail for polyfluorinated analogues (see Section IV.2.b), the intermediate copper(I) compound can undergo both reduction and disproportionation to give metallic copper and the original copper(II) compound.This reaction on the SiO2 surface also gives copper(I) oxide.147 Deposition of copper onto the SrTiO3 substrate was preceded by bombardment with Ar+ ions giving Ti3+ ions in the near- surface layer, which favoured the adsorption of the H2O mole- cules. The mechanism of Cu(DPM)2 transformations on the SrTiO3 substrate is similar to that described above for the SiO2 substrate.149 Improvement of copper film characteristics in the CVD process based on copper(II) b-diketonates with alkyl groups, in particular a decrease in the SER, is reached under additional UV irradiation.Zama et al.150 believe that in this case, UV irradiation favours the migration of the copper atoms over the growing copper film surface, thus decreasing the grain size and, hence, producing a smoother surface. c. Preparation and characteristics of copper films The use of copper(II) acetylacetonate as the starting CCC for the preparation of thin copper films by CVD began in the early 1990s,134, 151 though some earlier works on this process (mostly patents) are known (see Ref. 16). Astudy ofCVDof copper films from Cu(acac)2 vapour onto a SiO2/Si substrate at 220 ± 250 8C (with an argon ± hydrogen mixture as the carrier gas) showed that the structure and mor- phology of copper films are primarily determined by two factors, viz., the deposition rate and the partial pressure of hydrogen in the carrier gas.151 In a later study,152 polycrystalline copper films were obtained from Cu(acac)2 vapour in pure hydrogen.The prepara- tion of copper films with satisfactory characteristics from the same CCC but at higher substrate temperatures (Ts) of 300 ± 450 8C (Table 3) has been patented.153 A high temperature is required for CVD with Cu(acac)2 as the starting compound. It was suggested that it may be decreased (by *75 8) and the film deposition rate increased by using UV irradiation.154 Yet another way for the activation of the CVD process involves the plasma-activated decomposition of Cu(acac)2 vapour in hydrogen (PACVD), which allows generation of reactive hydrogen species at rather low temperatures.155 For example, smooth thin copper films with a low SER (see Table 3) were obtained from Cu(acac)2 vapour.134 The decomposition was carried out in the presence of hydrogen (carrier gas) up to a substrate temperature of 300 8C; it was found that at higher temperatures decomposition does not depend on the nature of the carrier gas.This is apparently due to the fact that the amount of hydrogen formed upon intense pyrolysis in this temperature range is sufficient for the reduction of copper.146 However, the presence of even small concentrations of O2 results in an abrupt increase in the SER of the films.Table 3. CVD conditions and properties of copper films obtained from copper(II) b-diketonates. Starting MCC Carrier gas T /8C Ts /8C Cu(acac)2 Cu(DPM)2 220 ± 250 220 ± 250 350 ± 400 <300 a 350 220 ± 400 250 ± 350 250 ± 350 b H2/Ar H2 H2 H2 H2 H2/N2 H2 H2 777110 ± 170 110 120 77 a In plasma. b Under UV irradiation. V N Vertoprakhov, S A Krupoder If Cu(DPM)2 vapour is used, the copper film growth on a quartz substrate starts at 350 8C.133, 156 The polycrystalline films obtained by this method 133 were preferentially oriented along the [111] direction and had a thickness >300 nm and an SER of 10 mO cm (see Table 3).According to XPS data, carbon and oxygen were the main admixtures in this case (their total content was up to 10%); their simultaneous presence in the film could result from the reaction of copper with CO formed upon decom- position.146 The specific electric resistance of copper films obtained 116 from Cu(DPM)2 vapour increased continuously from 8.1 to 41.2 mO cm (see Table 3) as their thickness decreased from 60 to 40.9 nm. It was noted that the films were smooth and mirror-like up to 300 8C but became dim as the temperature was increased. If the samples are cooled in a hydrogen atmosphere, carbon and oxygen admixtures in the film are virtually absent. On the other hand, cooling of the films in a nitrogen atmosphere results in a rather high content of the carbon admixture (12% ± 19%). The factors determining the growth rate, the adhesion to the substrate and the morphology of the copper films obtained from Cu(DPM)2 vapour in the CVD process were considered in detail by Dutta et al.157 The effect of additional UV irradiation on such films was discussed:150 at a constant substrate temperature, addi- tional photostimulation of thermolysis decreased significantly the SER of the films (see Table 3).The following variant of the CVD process has been sug- gested:158 the vapour of the starting Cu(DPM)2 was carried with a stream of nitrogen containing hydrogen and was then decom- posed on a borosilicate glass substrate in the temperature range from 220 to 400 8C.In this case, the deposition of copper films was rather fast for Cu(DPM)2 (26 nm min71) at a relatively low SER (1.7 ± 2.7 mO cm at thicknesses550 nm). 2. Polyfluorinated copper(II) b-diketonates It is known that incorporation of fluorine atoms into molecules of organic compounds generally increases their volatilities, despite the essential increase in the molecular masses. The nature of this effect is still uncertain. It is assumed that it can be due to weakening of the van der Waals interactions 111, 159 and (in some cases) due to the disappearance of hydrogen bonds upon replace- ment of hydrogen atoms in the molecule by fluorine atoms.109 However, this can hardly be regarded as a comprehensive explan- ation. Copper (II) b-diketonates also obey this rule.132, 159, 160 For example, replacement of methyl groups in the Cu(acac)2 molecule by trifluoromethyl groups results in a 104-fold increase in satu- rated vapour pressure of the chelate at constant temperature.160 In addition, the thermal decomposition of polyfluorinated copper(II) b-diketonates results in films with a lower content of impurities (in particular, carbon).In most cases, the formation of copper fluorides is not observed even under rather drastic deposition conditions,161 though the formation of CuF as a labile intermedi- ate product is possible. This can be due both to the relative strength of the C7F bond in comparison with the C7O and Ref. Substrate P /Torr SER /mO cm 151 152 153 134 133 116 150 150 SiO2/Si SiO2/Si SiO2/Si SiO2/Si SiO2 SiO2/Si sapphire " 777720 5 ± 10 77 71.7 71.8 10 8.1 ± 41.2 4.6 1.7Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method C7C bonds and to the high stabilities and volatilities of the polyfluorinated organic products of thermolysis. Owing to these properties, polyfluorinated copper(II) b-diketonates became the first CCC utilised in the CVD technology.110 Copper(II) hexafluoroacetylacetonate Cu(HFA)2 (1, R=R0=CF3) is the polyfluorinated copper(II) b-diketonate studied most thoroughly and used most widely in the CVD process.111 Copper(II) trifluoroacetylacetonate Cu(TFA)2 (1, R=CH3, R0=CF3) was also considered as a starting CCC.The use of copper(II) pivaloyltrifluoroacetylmethanate Cu(PTA)2 and some other substances of this series was much less common. a. Synthesis and physicochemical properties Polyfluorinated copper(II) b-diketonates are synthesised from the corresponding polyfluoro-b-diketones. These starting com- pounds were difficult to prepare 162 until they became available by the Claisen condensation.163, 164 CF3COCH2COR CF3COOEt+CH3COR EtONa 7EtOH R=CH3, CF3 , etc. This reaction became a reliable source of trifluoro- and hexafluoroacetylacetone and a series of their analogues. The classical method for the synthesis of copper(II) hexa- fluoroacetylacetonate Cu(HFA)2 involves the reaction of cop- per(II) nitrate with hexafluoroacetylacetone in aqueous solution in the presence of sodium acetate (molar ratio 1 : 2 : 2).120 This results in the precipitation of green crystals, the composition of which was determined differently by different research- ers.119, 165, 166 It is believed120 that these are the crystal hydrates Cu(HFA)2 .2H2O which do not decompose on sublimation. However, this conclusion was disputed later 167 and it was noted that the amount of water of crystallisation in the Cu(HFA)2 hydrate is not constant. According to the results of Funck and Ortolano,167 the green crystals are monohydrates (see also Ref. 168), since the dihydrate isolated by them was a thermally unstable yellowish-green compound. Difficulties in assigning Cu(HFA)2 .nH2O obtained by this method to mono- and dihydrates were also encountered later.161 Yet another method for the preparation of Cu(HFA)2 is the reaction of the ammonia complex of H(HFA) with CuCl2 in aqueous solution.169 The main product in the form of bluish-green crystals was characterised as a dihydrate, its melting point coincided with that reported previously.120 Nevertheless, these data disagree with other reported results.167 Yet another method for the synthesis of high-purity Cu(HFA)2 with the use of H(HFA) and copper(I) oxide was patented.170 A drawback of the monohydrate Cu(HFA)2 .H2O as the starting compound in the CVD processes is its ability to undergo decomposition with elimination of water in the reactor at low pressures in a carrier gas stream.171 Therefore, anhydrous Cu(HFA)2 is extensively used in these processes; under standard conditions, it has the form of purple crystals which are rather hygroscopic in air.120 Copper(II) trifluoroacetylacetonate Cu(TFA)2 was first syn- thesised by Schultz and Larsen 172 from the corresponding b-dike- tone (see also later studies 164, 173).A procedure for its preparation with the use of ammonium b-diketonate was reported.169 Many copper(II) b-diketonates fluorine-containing [including Cu(TFA)2] and the corresponding b-diketones have been obtained by Park et al.174 Generally, Cu(TFA)2 and Cu(HFA)2 are much more volatile than their analogues with alkyl groups described in the previous section.175 ± 177 Futhermore, they were found to be the most volatile in a large series of fluorine-containing copper(II) b-diket- onates, as indicated by thermogravimetric data.178 The vapour pressure of Cu(HFA)2 measured by various methods 179, 180 is 3 Torr at room temperature and atmospheric pressure.181 1063 The initial study of the decomposition of Cu(TFA)2 and Cu(HFA)2 .H2O vapour in the CVD process showed that the sublimation of these CCC occurs at 85 and 65 8C, respectively.110 The standard enthalpy of evaporation of Cu(HFA)2 is 13.20.8 kcal mol71 (see Ref.180). The b-diketonates of divalent copper containing the n-hepta- fluoropropyl group are rather exotic but quite suitable for the preparation of thin copper films by CVD. These compounds include such chelates as copper 6,6,7,7,8,8,8-heptafluoro-2,2- dimethyloctane-3,5-dionate Cu(FOD)2 (1, R=But, R0= n-C3F7) 182 and copper 1,1,1,2,2,3,3,7,7,8,8,9,9,9-tetradecafluoro- nonane-4,6-dionate Cu(FND)2 (1, R=R0=n-C3F7).These b-diketonates can be synthesised from b-diketones by the usual method.183 Stabnikov et al.184 studied the possibility of synthesis of mixed-ligand copper(II) b-diketonates where the copper atom is simultaneously bound to diketones with alkyl (acac, DPM, etc.) and fluorinated alkyl groups (TFA, HFA). The crystal and molecular structures of some of these compounds were studied.160, 185 ± 187 It was shown that they have intermediate properties between those of `homogeneous' bisdiketonates incor- porating the corresponding ligands; in the vapour phase, mixed- ligand compounds can undergo disproportionation to give homo- geneous bisdiketonates. It was therefore concluded that these compounds have no advantages which would make them promis- ing for practical use in CVD technologies.184 b. Mechanisms of decomposition on the substrate surface The thermal decomposition of Cu(HFA)2 vapour in CVD proc- esses and its products have been studied in detail.139, 188, 189 It was shown that H(HFA) is the main decomposition product of Cu(HFA)2 in the vapour phase, while thermal decomposition on a SiO2 surface at 480 8Ccan give either a CuO film (in air) or a film of metallic copper with small admixtures of Cu2O and HFA decomposition products (in an argon atmosphere).The main decomposition product in air at 800 8C is Cu2O in the form of finely dispersed particles. As shown by Lecohier et al.,190 the growth of a copper film from Cu(HFA)2 vapour involves several stages. During an induc- tion period, the duration of which depends on the properties of the substrate surface and on the deposition method, film growth is not observed. Subsequently, the formation of isolated copper islets starts. Their merging results in a sharp decrease in the SER. Cu(b-diket)2 + H2 Cu+2 H(b-diket): . Van Hemert et al.110 suggested a mechanism of copper reduction from its chelate in a hydrogen atmosphere; this mech- anism can be expressed by the overall equation In the case of copper(II) b-diketonate hydrates, the formation of precursors in hydrated form was postulated.A similar decomposition equation was suggested for the thermolysis of Cu(HFA)2 at relatively low temperatures (below 340 8C).180 Thus, it is believed that thermolysis of the compounds in question starts with the cleavage of the copper7oxygen bonds. It was noted that at higher temperatures (500 ± 600 8C) cleavage of the C7C bonds in the ligand also occurs and results in carbon admixtures in the film. These concepts are basically correct but they do not provide a sufficiently full description of the processes occurring when the molecules of the starting CCC reach a heated substrate surface. A detailed scheme of these processes was developed somewhat later and generalised by Kim et al.191 The following assumptions were used in the description of the mechanism of transformations in the CVD processes: the CCC and hydrogen molecules are adsorbed on different sites of the substrate surface; the adsorption of molecules of CCC competes with that of the original b-diketone formed upon reduction; the reduction occurs on the substrate rather than in the gas phase.1064A scheme consisting of five basic steps was suggested based on these assumptions: (1) adsorption of molecules of CCC with their partial dissoci- ation Cu(HFA)A*+(HFA) A* ; Cu(HFA)2+2 A* (2) disproportionation of partially adsorbed copper(I)-con- taining intermediates 2Cu(HFA) A* Cu+Cu(HFA)2 A* , or Cu+(HFA) A* ; Cu(HFA) A* 2HB* ; H2+2B* H(HFA) A*+B* ; (HFA) A*+HB* H(HFA) A* H(HFA):+A* .(2 0) progression of dissociation of the intermediate copper(I) derivative (3) dissociative adsorption of hydrogen (4) formation of a volatile product, viz., a b-diketone, on the surface (5) desorption of the b-diketone Here A* and B* indicate the active sites of, initially, the substrate and, later, the growing copper film capable of chem- isorption.192 Step 4 is the limiting step of the entire process.192, 193 A somewhat simplified version of this mechanism was sug- gested earlier.194 It was noted that the irreversibility of the entire process begins in the step where metallic copper adsorbed on the substrate appears. The copper atoms form microclusters on the substrate surface, which consist of 8 ± 12 copper atoms and are clearly separated from the organic ligands.195 Certainly, the temperature of the process and the material of the substrate affect the course of the steps considered.A study of the thermodynamic parameters of the thermal decomposition of Cu(HFA)2 and the phase compositions of the copper films formed in a rather broad temperature range showed that copper(I) fluoride can be formed in one of the intermediate steps at rather low temperatures. This is not taken into account in the scheme suggested above. As the temperature increases, the amount of CuF decreases until it disappears completely.196 The effect of the substrate material on the mechanism of thermolysis of Cu(HFA)2 vapour was analysed in detail.197 It was noted that decomposition of Cu(HFA)2 in a stream of H2 on a TiN surface can result in replacement of fluorine atoms in the trifluoromethyl group by hydrogen atoms.In the case of the SiO2 substrate, a noticeable amount of CF3COOH is formed, which is the product of cleavage of the C7C bonds in the HFA fragment. Let us outline the applicability limits of the mechanism considered. Two types of CVD processes are distinguished, depending on the conditions.198 At rather low temperatures and partial vapour pressures of the starting CCC, the process is primarily characterised by reactions on the substrate surface occurring in accordance with the mechanism described above (the so-called reaction-limited conditions). At high substrate temperatures and pressures in the reactor and at high vapour pressures of the substance, this mechanism does not operate and the film deposition process is characterised by the transport of molecules of CCC to the substrate (transport-limited conditions).These conditions generally result in film heterogeneity.198 In order to reveal the details of the thermal decomposition mechanism of Cu(HFA)2 on various substrates, some authors used ultrahigh vacuum conditions (41077 Torr). Such condi- tions are unlikely to be suitable for industrial CVD processes, but we feel that the results obtained should be discussed. If thermal decomposition of Cu(HFA)2 vapour duringCVDis carried out in ultrahigh vacuum, the process mechanism changes V N Vertoprakhov, S A Krupoder essentially.In this case, disproportionation of the intermediate copper(I) derivative does not occur.199 For example, the Cu(HFA)2 molecules are quickly desorbed from the TiN substrate at as low temperature as 75 8C, and further decomposition of the intermediate [reaction (2 0), see above] becomes the dominating process. At 200 8C, decomposition of the adsorbed ligand begins to give CO, CO2 and CF4; as a result, *30% of Cu from the starting material remains on the substrate. Disproportionation of copper(I) intermediates is not observed, either, in the deposition of copper films onto polycrys- talline tantalum substrates from Cu(HFA)2 vapour.200 It was noted that adsorption of CCC on a tantalum surface starts at 158 8C; formation of a copper film occurs at temperatures 4117 8C, while diffusion of copper into tantalum is observed at 477 8C.The adsorption of Cu(HFA)2 molecules and their subsequent decomposition on the (100) and (111) surfaces of copper single crystals in ultrahigh vacuum (10710 Torr) have been studied.201 It was shown that under these conditions, surface migration of ligands of the adsorbed molecules of CCC with formation of isolated HFA groups adsorbed on the substrate parallel to its surface occurs even at low temperatures (*773 8C). At 102 8C, their decomposition gives CO, CO2 (quickly desorbed from the substrate), CF3 and the ketenylidene fragments C=C=O adsorbed through the terminal carbon atom. At temperatures above 227 8C, the CF3 species are desorbed, while the C=C=O fragments undergo decarbonylation; the carbon remaining on the surface is the basic source of film contamination.As shown for the adsorption of Cu(HFA)2 on a platinum surface under similar conditions,202 the orientation of HFA groups depends on the thickness of the layer of molecules of CCC adsorbed on the substrate. Ligand orientation parallel to the substrate surface is characteristic of very thin films; as their thickness increases, the ligands are `reoriented' to become perpen- dicular to the substrate. It was also noted that difluorocarbene CF2 is the most strongly adsorbed species, which is desorbed in ultrahigh vacuum only at 577 8C. The nature of interaction of Cu(HFA)2 and the b-diketone H(HFA) itself with the surface of the SiO2 substrate was studied in detail.203 The experiments were carried out in ultrahigh vacuum (561078 Torr). In this case, everything is determined by the presence of OH groups, both isolated and connected by hydrogen bonds, and siloxane bonds, Si7O7Si, on the surface of the substrate.It was shown that preferential adsorption of Cu(HFA)2 molecules at 25 8C occurs with involvement of isolated surface OH groups. On heating the substrate (studied in the range up to 400 8C), decomposition of adsorbed Cu(HFA)2 was observed; in this case, the OH groups involved in the adsorption were not recovered. Farkas et al.203 reported experiments aimed at decreasing the density of OH groups on a SiO2 surface as a result of its thermal treatment.It was found that in this case strong adsorption of the Cu(HFA)2 molecules involves surface siloxane bonds, which are cleaved irreversibly upon subsequent heating and thermolysis of Cu(HFA)2. However, desorption of the decomposition products occurs at rather high temperatures, viz., about 400 8C irrespective of pretreatment of the SiO2 surface. The mechanism suggested by the authors involves direct axial coordination of the copper(II) atom with the non-shared electron pairs of the oxygen atoms of the OH groups or the Si7O7Si bonds, giving a five-coordinate complex. A similar explanation was suggested previously for the interaction of Cu(acac)2 with a SiO2 surface.204 The deposition mechanism of copper films from Cu(HFA)2 vapour in the presence of water vapour was considered in detail.205 ± 208 It was assumed that intermediate formation of a Cu(HFA)2 mono- or dihydrate occurs in the gas phase H2O Cu(HFA)2 .H2O(vapour). Cu(HFA)2 (vapour)Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method In the resulting complex (under gas-phase conditions) an intramolecular transfer of H2O protons to the ligand occurs, which weakens the copper7oxygen bonds. Adsorption of this complex on the surface results in elimination of the hydrated enol form of the b-diketone Cu(HFA)2 .H2(ads) Cu(HFA)2 .H2(vapour) 2+ Cu(ads)+2 (HFA) .H2(ads) . The adsorbed Cu2+ ions are reduced under the action of hydrogen. Since reduction is the limiting step of the process, the side formation of copper(I) oxide is observed at low deposition temperatures.209, 210 It was found with the use of the 18O isotope that the formation of an oxide phase involves the oxygen atoms of H2O rather than those of the H(HFA) ligand.209 However, in general, the addition of water vapour to the carrier gas in an optimum temperature mode does not result in contamination of films with copper oxides.This is believed to be due to the fact that the reduction of copper 2Cu+H2O Cu2O+H2 is an exothermic process (the equilibrium constant is 0.786108 at 325 8C).205 The presence of water vapour in the carrier gas improves the morphological characteristics of the films. This can be due to the fact that the nucleation rate of copper is determined by the rather low activation energy for decomposition of the adsorbed Cu(HFA)2 hydrate and hence almost does not depend on the properties of the substrate surface.Owing to the high density of the nucleation centres determined by the amount of the adsorbed hydrate, smooth films with good morphology are formed. On the other hand, the formation of film nuclei in usual systems is limited by a surface reaction between the adsorbed molecules of anhy- drous CCC and hydrogen, the rate of which depends on the density of the adsorption centres on the substrate surface. In this case, the film is very sensitive to the quality of the substrate.205 The role of molecules of an alcohol in the deposition of copper films from vapour of solutions of the starting CCC in the alcohol is somewhat similar to the role of water molecules.Like water, methanol and ethanol in the vapour phase can be coordinated with the copper(II) ion in copper b-diketonate; as a result, molecules of the complex of copper b-diketonate with the corresponding alcohol are deposited on the substrate.208 The proton of the alcohol molecule in the adsorbed complex can then be transferred to the adsorbed HFA ligand obtained due to the CCC decomposition; this favours the desorption of an H(HFA) molecule from the substrate surface. An adsorbed alkoxy group that remains on the surface can be removed by two pathways: either by elimination of the hydrogen atom from the a-carbon atom to give a volatile carbonyl compound, or by the reaction with the adsorbed hydrogen atom from the carrier gas to give the alcohol.211, 212 In the former case, the alcohol acts as a reducing agent, while in the latter case, as a reduction catalyst.Propan-2-ol is the most efficient compound,213 in spite of the fact that propan- 2-olic complexes of copper b-diketonates are the least stable complexes with alcohols.208 The mechanism of photochemical deposition of copper films from the vapour of CCC under laser irradiation of substrates was suggested and considered in detail by Ehrlich and Tsao.214 Special attention was given to film growth rate simulation. A study by Zheng et al.215 dealt with the mechanism of Cu(HFA)2 vapour decomposition in a stream of hydrogen under exposure to plasma.It was shown that under these conditions preferential consecutive elimination of CF3 groups from the ligand is observed, rather than elimination of the entire ligand with cleavage of the Cu7O bonds. Nevertheless, this decomposi- tion pathway also gives only volatile organic products which do not contaminate the film obtained. 1065 The thermodynamics of decomposition of Cu(HFA)2 and its analogue, Cu(FOD)2, in the PACVD process has been consid- ered.216 The authors of the first studies on the PACVD method for the preparation of copper films assumed that the role of plasma involves the formation of a high concentration of reactive hydro- gen species.155, 216, 217 However, the details of the mechanism remained unclear.A model of decomposition of Cu(HFA)2 with the use of hydrogen plasma was described.193 It was noted that in this process the hydrogen plasma is mainly electropositive and consists of cations, i.e., H+, Há2 and Há3 , electrons and hydrogen atoms. The overall process occurs according to the five-step mechanism discussed above; however, both the hydrogen atoms adsorbed on the substrate and the hydrogen atoms from the gas phase can participate in the limiting fourth step. The film quality and the deposition rate are determined by the pressure in the reactor, the electric field power and the substrate temperature. At small power (up to 25 W), the following kinetic equation is valid: Rd=k[Cu(HFA)2]1/2[H] , where Rd is the rate of the direct deposition reaction, k is the rate constant of the limiting (fourth) step.At high electric field power, decomposition of HFA fragments starts, the process becomes more complicated and strong contamination of films with carbon and fluorine is observed.193 The decomposition mechanisms of other fluorinated cop- per(II) b-diketonates in CVD processes have almost not been studied. It was shown 189 that H(FOD) is the main product of Cu(FOD)2 decomposition in the vapour phase. One can also mention a study of the kinetics of thermal decomposition of Cu(PTA)2 vapour upon heating of the substrate with a CO2 laser;218 the observed phenomena are in full agreement with the concepts of the reaction-limited and transport-limited processes as a function of deposition conditions.Data are available on the products of Cu(FND)2 thermolysis and its behaviour under CVD conditions in the presence of vapour of water and alcohols.208 c. Preparation and characteristics of copper films In the first experiments on the preparation of thin copper films by CVD based on decomposition of Cu(TFA)2 and Cu(HFA)2 .H2O vapour in a stream of hydrogen under atmospheric pressure, the optimum temperature of glass substrates on which films are formed was found to be 250 8C. However, the characteristics of films were not reported.110 The decomposition was purely ther- mal, without additional activation with laser radiation or plasma. A little later, a similar procedure was patented by the same authors.219 The process of non-activated thermal deposition of copper films from Cu(HFA)2 vapour was also studied in a number of works (see, e.g., Refs 40, 115, 180, 191, 194, 219 ± 244).Typical conditions were used by Kim et al.:191 the pressure in the reactor was 10 Torr, the temperature of the TiSi2/Si substrate was 300 ± 400 8C, an H2/Ar mixture (3 : 1) was used as the carrier gas (Table 4). It was found that at low pressure and with this substrate material, no growth of the copper film on the substrate occurred at temperatures up to 300 8C. Above this temperature, the growth rate was 10 ± 100 nm min71. The film thickness depended linearly on the deposition rate, i.e., the deposition rate was practically the same in each experiment.As the hydrogen partial pressure in the carrier gas was decreased, a considerable decrease in the deposition rate was observed. The surface mor- phology of the films obtained under these conditions was studied in detail.220, 223 The deposition of copper films by decomposition of Cu(HFA)2 vapour on SiO2/Si substrates in the temperature range from 250 to 650 8C in the absence of hydrogen (with pure argon as the carrier gas) was studied by Temple et al.,180 who did not observe the formation of films at 250 8C, as opposed to the results of other studies.110 Pure copper films appeared in the1066 Table 4. CVD conditions and properties of copper films obtained from polyfluorinated copper(II) b-diketonates. Ts /8C T /8C Starting CCC�Cu(HFA)2 77 300 ± 400 300 ± 400 7 7340 ± 390 250 ± 300 250 160 ± 170 a 310 ± 390 310 ± 390 300 ± 400 800 b 400c 350 ± 410 310 ± 390 130 ± 200c 300 230 a 160 ± 170 a 225 a 97 ± 120 7780 ± 100 65 ± 85 65 ± 90 60 760 75 ± 94 65 ± 85 100 95 ± 98 7780 Starting CCC�Cu(FOD)2 <250 300 ± 400 77 H2 H2 a In plasma.b Under laser irradiation followed by cooling to 250 8C. c Under laser irradiation. temperature range of 340 ± 390 8C (see Table 4). Upon further heating of the substrate to 430 8C, the films containing 65% of carbon (25% of copper) were formed, while the films obtained at 500 8C consisted of pure carbon. In other studies, the temperature range for the formation of pure copper films by decomposition of Cu(HFA)2 vapour in argon was somewhat wider, viz., from 310 to 400 8C.221, 222 Films of rather high quality were obtained with a growth rate of 65 nm min71.It was noted that the substrate (TiN, TiW) surfaces had to be cleaned from oxide contaminants.222 Awaya and Arita 224 synthesised copper films from Cu(HFA)2 vapour on a Si substrate coated with a few buffer layers at once (the top layer consisted of chromium, see Table 4). Such films had low SER values and contained an admixture with chromium. In addition to hydrogen, water can be used as a reducing agent for Cu(HFA)2.171, 190, 205, 225, 226 In this case, the growth rate of copper films depends directly on the amount of water vapour in the gas phase.225 The presence of water vapour during the decomposition of Cu(HFA)2 vapour increases considerably the film growth rate (approximately by two orders of magnitude 171) but the selectivity of the process decreases (see Section VI).The specifics of the film growth kinetics in this method and the film morphology were considered in detail.190 It was reported 171 that the SER of the resulting films can be as low as*2.0 ± 2.1 mO cm while the deposition rate remains high. Another group of researchers who studied a similar process suggested the use of Cu(HFA)2 dihydrate rather than Cu(HFA)2 as the starting CCC.205 They reported a high deposition rate and good morphology of the copper film surfaces. However, the results of this study were poorly reproducible, probably due to the instability of the dihydrate; it was reported that this substance readily eliminates water molecules in the vapour phase.227 The use of anhydrous Cu(HFA)2 in the presence of water vapour gave films with low SER (deposition rates up to 90 nm min71).An interesting modification of the method considered above involves the deposition of copper films from the vapour phase of a CCC solution in an alcohol.211 It is known that anhydrous Cu(HFA)2 is freely soluble in alcohols (ethanol, propan-2-ol). In P /Torr Carrier gas 10 1078 ± 10 0.75 2 ± 10 1 ± 1.7 15 2 ± 20 0.15 Ar/H2 Ar7Ar H2/Ar H2/He H2 H2 H2/Ar H27 7 7 He/H2O H2/H2O H2/Ar H2/Ar H2 H2 H2 H2/Ar 0.075 15 2 ± 10 760 1 ± 10 2711 ± 10 1 ± 10 V N Vertoprakhov, S A Krupoder Ref.Substrate SER /mO cm 2.0 2.0 1.7 3 ± 7 1.9 2 ± 3 1.7 ± 1.9 2.50.3 2.00.1 TiSi2/Si TiN/Si TiW/Si Cr (on Si) SiO2/Si Si, SiO2 glass Si, Ti, W Au/Si SiO2 , TiSi2 SiO2/Si 73.60.1 2 191 221, 222 224 180 115 194 244 218 220 199 236 190 205 223 241 211 243 244 215 SiO2/Si Ta/SiO2 SiO2/Si, TiSi2 SiO2/Si TiN/Si SiO2 , Si, Al SiO2/Si W/Si Si 2.0 2.0 ± 2.3 72.5 ± 5 2.0 1.7 ± 1.9 1.7 182 182 3.0 1.8 SiO2 , Si, Al SiO2 , Si, Al this method, a stream of the carrier gas (H2) is passed through an *0.4 g ml71 alcoholic solution of Cu(HFA)2 (higher concen- trations result in an increased solution viscosity and, as a consequence, instability of the process).The alcohol vapour present in the gas phase serves as an additional reducing agent. The advantage of this technique is the significant (approx- imately fivefold 211) increase in the deposition rate of copper films in comparison with that in the usual process in a stream of hydrogen. However, the films obtained using this technique have higher porosity which inevitably impairs their properties. For example, the SER of copper films synthesised by Borgharkar et al.211 was 2.5 ± 5 mO cm (see Table 4) provided that their thickness was no less than 1000 nm. The characteristics of these films can be improved by addition of water vapour to the carrier gas; in this case, the SER of the films was 2.6 ± 2.9 mO cm.211 A technique for CVDof copper films fromCCCwith participation of a solvent has been patented.228 Attempts at performing the CVD process with Cu(HFA)2 with an excess of the starting b-diketone H(HFA), which in turn is a decomposition product of the precursor, were reported.It was noted 194 that the excess of H(HFA) decreased the copper film deposition rate. Nevertheless, if SiO2 templates deposited on a copper substrate were used for the preparation of thin copper strips from Cu(HFA)2, pretreatment of the template channels with H(HFA) vapour improved considerably the characteristics of the strips, but the presence of H(HFA) vapour during the deposition process is unacceptable as it caused an abrupt impair- ment of the strip morphology and a loss of selectivity.229 It is known that polyfluorinated copper(II) b-diketonates are photo-labile compounds.Their photochemistry in solution 230 and in the gas phase 231 ± 233 has been studied rather thoroughly. Laser irradiation of the substrate is a popular method for decom- position of the vapour of these compounds for deposition on substrates.181, 232, 234 ± 238 In this case, Cu(HFA)2 itself or its alcoholic solution (an `alcoholic' CVD technique) can be utilised for the preparation of films. This creates certain difficulties. First, it is difficult to estimate to what degree the observed process is pyrolytic and to whatPreparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method degree it is photochemical, as the substrate temperature is *250 8C,236 and it is difficult to estimate accurately the temper- ature of the process in laser experiments.239 It is known that the quantum yield of decomposition of copper(II) b-diketonate vapour to the metal is rather small; this determines the low film growth rates by the purely photochemical mechanism and the irregularity of their morphology.Second, photochemical decom- position can increase the content of the carbon admixture in the copper film; fluorine and oxygen are removed with volatile photolysis products.234 It was shown in experiments on copper film deposition under laser irradiation with a wavelength corre- sponding to the visible spectrum range that carbon is a photolytic admixture in this case; under these conditions, the carbon con- tamination of the films was absent.236 A thorough refinement of the photochemical process modes for the deposition of films was required to overcome the difficul- ties noted above.237, 240 In a study devoted to the effect of laser power on the growth rate of copper films from Cu(HFA)2 vapour,241 the deposition rate was increased from 20 to 120 nm min71 in a stream of an argon ± hydrogen mixture and from 4 to 27 nm min71 in a stream of argon, but the SER of the films were not reported.It was noted that in the presence of ethanol under laser activation conditions, as in the usual thermol- ysis, the growth rate of copper films increased.234, 238 The properties of films obtained with the use of a laser were analysed by Hoyle et al.236 The films were formed on a silicon substrate at *800 8C (the power density of laser radiation was 46105 W cm72). As the film grew, the substrate temperature decreased gradually to 250 8C, i.e., the optimum value for Cu(HFA)2 vapour decomposition.The films had good adhesion properties but rather high SER values (see Table 4). The preparation of film microstructures (copper microstrips) from Cu(HFA)2 vapour is a promising field of application of the CVD process with the use of laser radiation. The optimum conditions of the process for the preparation of strips 2 mm long were chosen and the factors determining their basic characteristics were studied.240 A milder way of photochemical decomposition of Cu(HFA)2 vapour is the use of UV irradiation.242 In this case, the level of contamination of copper films is much lower.Yet another method for activation of decomposition of Cu(HFA)2 vapour is the use of plasma.161, 243, 244 It was shown in the very first study on this technique that if the process is carried out in a stream of hydrogen, high-quality thin copper films could be obtained virtually without carbon admixtures by varying the source power and the substrate temperature.161 The SER of the best film samples, 60 ± 80 nm thick, were within 1.7 ± 2.4 mO cm; these values could not be reached with laser radiation.This can be due to the fact that efficient growth of films in the plasma method requires a considerably lower source power; an increase in the power immediately resulted in carbon contamination. Plasma activation was used to improve considerably the characteristics of the copper films obtained by the `alcoholic' method. For example, decomposition of the vapour of an etha- nolic solution of Cu(HFA)2 in a stream of hydrogen at a plasma source power of 45 W gave a copper film 63 nm thick with a low SER (see Table 4).243 An optimum content of hydrogen in the carrier gas is a prerequisite for the preparation of high-quality copper films by this method: at a hydrogen concentration lower than 88%, the SER of copper films increased to 9 ± 10 mO cm.An opposite relationship was observed for the grain size in the copper film: a decrease in the fraction of hydrogen in the carrier gas decreased the grain size from *100 (88% H2) to *10 nm (0% H2). The reasons for this phenomenon were not rationalised. Copper films with good characteristics were obtained 244 using a similar technique by decomposition of the vapour of Cu(HFA)2 solutions in ethanol and propan-2-ol at relatively low substrate temperatures and a power density of the plasma source of 0.13 ± 0.25 W cm72. A high film growth rate, viz., 25 nm min71, was reported; the SER of films 440 nm thick was 1.7 ± 1.9 mO cm. 1067 Polycrystalline copper films 500 ± 1000 nm thick were obtained from the vapour of a rather rare starting CCC, viz., Cu(FOD)2.182 The resulting copper film (argon as the carrier gas) contained up to 24% admixtures (C, O, F).Replacement of argon with hydro- gen resulted in a pure copper film with good characteristics; in this case, a decrease in the substrate temperature (to 250 8C and lower) increased the SER of the films considerably (see Table 4). 3. Complexes of copper(II) b-diketonates with additional ligands The copper atom in copper(II) b-diketonates is coordination- unsaturated, hence it is possible to synthesise adducts of copper(II) b-diketonates with various organic Lewis bases (primarily, nitro- gen- and oxygen-containing ones). Some of these adducts are volatile. However, the prospects of their use in the CVD technol- ogy as precursors are doubtful. Unlike copper(I) b-diketonates, copper(II) b-diketonates are convenient starting CCC for the CVD technology.As a rule, it is easier to change their properties by modifying the starting b-diketone than to try to obtain a stable adduct with an additional ligand which can be sublimed. Nevertheless, the use was suggested of some compounds of this type as volatile CCC, for example, the adduct of Cu(DPM)2 with o-phenanthroline (1 : 1), which is sublimed in vacuo at 130 8C without decomposition.245 Under the same conditions, Cu(DPM)2 itself is sublimed at 160 8C. It was reported that an attempt at sublimation of a similar Cu(acac)2 adduct resulted in its decomposition. The 1 : 1 adducts of Cu(HFA)2 and Cu(PTA)2 with o-phenan- throline and 2,2 0-bipyridyl are also volatile.246 However, the sublimation temperatures for all four adducts exceed 160 8C, which is higher than those for the original copper b-diketonates.More promising in this respect is the synthesis of a series of volatile adducts of Cu(HFA)2 with ethanolamines.247 As noted above, the addition of ethanol vapour to the carrier gas increased the copper film deposition rate. It was assumed that the ethanol- amine adducts obtained, which contained a `preformed' ethanol fragment, will maintain this advantage; on the other hand, the presence of the amino group should impart stability to the adduct under sublimation conditions. Preliminary CVD experiments on this type of copper(II) derivatives showed that the adduct of Cu(HFA)2 with 1-dimethyl- aminopropan-2-ol, which forms a copper film at a substrate temperature of 300 8C, has quite satisfactory characteristics suitable for CVD processes.247 However, the parameters of the process and the film characteristics were not reported.No films were obtained with the use of a similar adduct with ethanolamine. 4. Other classes of volatile compounds of copper(II) The range of volatile organic derivatives of copper(II) suitable for CVD processes is not limited to b-diketonates, but the choice is rather narrow. a. Chelate complexes, analogues of copper(II) b-diketonates Nitrogen-containing analogues of b-diketones (iminoketones) form volatile derivatives of copper(II) with physicochemical properties satisfactory for CVD processes.248 Nevertheless, we could find only one study devoted to the practical preparation of copper films from this class of compound.249 A new compound synthesised by the authors, viz., bis[4-(2,2,2-trifluoroethyl)imino- 1,1,1,5,5,5-hexafluoropentan-2-onato]copper(II) designated as Cu(nona-F)2, had a decomposition temperature of 270 8C.Cop- per films with satisfactory characteristics were obtained from this compound but no further developments were reported. Of compounds the structure of which is somewhat similar to that of copper b-diketonates, chelate complexes of copper(II) with analogues of ethyl acetoacetate (CH3COCH2COOR) can be considered as a real alternative to the former. A series of these compounds (R=Me, Et, But, CH2Ph and CH2CH2OCH3), their hydrates and adducts with alcohols were synthesised 250 using a1068 technique similar to that used for copper b-diketonates.120 The sublimation temperatures of these compounds were from 78 to 113 8C at 0.16 Torr.The crystal structure of Cu(CH3COCHCOOBut)2 monohydrate was determined.136 CVD experiments with compounds of this series at low pressures (0.1 ± 1 Torr) in the presence of hydrogen resulted in high-quality copper films with SER values ranging from 1.8 to 2.5 mO cm; in addition, the factors determining their adhesion to the substrate were studied.250, 251ASi substrate with a TiN or TiW barrier layer was used; the deposition rate varied from 0.5 to 20 nm min71 at substrate temperatures from 160 to 330 8C.Judging by the characteristics of the films obtained, Cu(CH3COCHCOOCH2CH2OCH3)2 was the best starting CCC in this series.250 If copper(II) ester chelates are used, copper films are formed at a relatively low substrate temperature (160 8C). For comparison: the minimum substrate temperature is 310 8C for Cu(HFA)2.220, 223 This is an important advantage of copper(II) ester chelates in comparison with copper(II) b-diketonates. No noticeable thermolysis of the adsorbed free ligands occurs at 160 8C; this ensures that the films are not contaminated. The surface morphology of copper films synthesised by the CVD method from Cu(CH3COCHCOOBut)2 vapour was studied in detail.252 Comparison of the deposition processes and micro- structures of copper films obtained from Cu(DPM)2 and Cu(CH3COCHCOOBut)2 showed that dense, crack-free films are formed from both CCC at substrate temperatures up to 350 8C, but the substrate temperature required for the deposition is about 100 8C lower in the case of Cu(CH3COCHCOOBut)2.253, 254 The growth rate of the copper film from Cu(CH3COCHCOOBut)2 vapour is several-fold smaller, but the film structure is much better than that of the film obtained from Cu(DPM)2 vapour.In another study by the same scientific group,117 the SER of copper films synthesised from Cu(CH3COCHCOOBut)2 at 225 8C was reported to be 2.5 ± 4.0 mO cm at a film thickness of 50 ± 120 nm. Cu(CH3COCHCOOEt)2 is also believed by many authors to be a promising starting CCC for the deposition of copper films.For example, dense fine-grain copper films 595 nm thick were obtained from this compound at 225 8C (on a SiO2 substrate) in a hydrogen atmosphere;255 the SER of these films was close to that of metallic copper. This compound was also used 256 in a CVD process with plasma activation giving a copper film >250 nm thick with a SER of 2.0 mO cm. The same work also reported the preparation of a 256 MbRAMchip byCVDusing decomposition of Cu(CH3COCHCOOEt)2 vapour. b. Copper(II) carboxylates Copper(II) carboxylates that are more readily accessible syntheti- cally than copper b-diketonates could become a real alternative to the latter in CVD processes. However, the use of this class of substances, first of all copper(II) formate Cu(HCOO)2, for the preparation of copper films was originally limited to deposition from liquids and aerosols.63, 257, 258 Nevertheless, it was shown subsequently that Cu(HCOO)2 hydrate can be directly used in the CVD process.259, 260 Prepara- tion of copper films from the vapour of this compound was carried out on Si(100) and TiN substrates at 200 ± 350 8C with and without a carrier gas (H2, Ar).The highest-quality crystalline copper films without oxygen admixtures were obtained at a substrate temperature of*300 8C. Of copper carboxylates intended as potential starting CCC, volatile complexes of copper(II) polyfluorocarboxylates with 1,4- dioxane were also obtained; these compounds can sublime with- out decomposition at 120 ± 130 8C (0.05 Torr).261 c.Copper(II) alkoxides and their derivatives Volatile alkyl derivatives of copper 262, 263 cannot be regarded as technologically suitable compounds, both with respect to their V N Vertoprakhov, S A Krupoder synthetic accessibility and properties. Alkoxide derivatives of divalent copper are much more stable and accessible. The simplest copper(II) alkoxides of the type Cu(OR)2, where R=Me, Et, etc., are unsuitable for CVD processes because they are polymeric non-volatile compounds.264 Compounds obtained by partial replacement of hydrogen atoms in Cu(II) alkoxides by fluorine atoms sublime at 80 ± 90 8C. Their thermal decomposi- tion occurs at 140 ± 200 8C,265 but these compounds have not been studied from the viewpoint of their use in CVD technology.Attempts were made to synthesise monomeric alkoxides as the starting compounds for CVD processes.266, 267 In fact, a series of convenient compounds [e.g., Cu(OCH2CH2NR2)2, where R=Me, Et] which sublime in vacuo at 60 ± 110 8C were obtained, but no experiments on the preparation of copper films using CVD from them as the starting CCC were carried out. An interesting development of this line of research was a rather simple synthesis of volatile complexes of copper(II) fluo- roalkoxides with various nitrogen-containing Lewis bases with general formula Cu(OR)2 . L, where R is hexafluoroisopropyl CH(CF3)2 (HFIP) and C(CH3)(CF3)2, while L=N,N,N0,N0- tetramethylenediamine (TMED) and 2,2 0-bipyridyl.268 These complexes are monomeric and undergo sublimation with partial decomposition. However, it was shown that decomposition of vaporised (HFIP)2Cu(TMED) complex on silicon substrates at 300 8C gives a copper film with an admixture of <1% carbon.However, in the authors' opinion,268 the search for starting compounds which would be more competitive with copper b-diketonates among this class of compounds is necessary. V. The starting copper-containing compounds: copper(I) derivatives Disproportionation to copper(II) derivatives and metallic copper is one of the main reactions of copper(I) compounds. It is no wonder that this process attracted the attention of researchers in the field of preparation of thin copper films by the CVD process.It is known that it is more difficult to synthesise volatile copper(I) derivatives than similar compounds containing cop- per(II). However, studies in the field of CVD processes with the use of monovalent copper compounds, which have started rather recently, are very intense.111, 269 This is due to the fact that thermally induced disproportionation has certain advantages over thermal decomposition. These advantages include the following: the process occurs with higher deposition rate and at lower temperatures; the ligands bound to the Cu(I) atoms are removed in a more favourable temperature range without decomposition to volatile products. This provides good technological characteristics: the process is safer as generally there is no need to use a carrier gas and hydrogen as the reducing agent; a high-purity film grows because it is not contaminated with carbon, fluorine and oxygen.These advan- tages expand the prospects and reduce the restrictions related to the design and control of the technological process of selective deposition of copper films. It should be noted that in the case of CVD processes involving copper(I) derivatives, a certain competition exists between the CCC with organic ligands and the starting inorganic CCC (usually, copper halides). Attempts were made to use copper(I) halides for the prepara- tion of copper films by the CVD method.270, 271 High evaporator temperatures are required in this case; in addition, the problem of film purity arises owing to the low volatility of the starting compounds.111 Copper(I) halides are more volatile than copper(II) halides.272 ± 274 For example, dense fine-grain copper films with SER of 1.9 mO cm devoid of iodine admixture were obtained from CuI (the temperature of the SiO2 substrate was 697 8C).274 However, too high substrate temperatures are required for CVD process based on monovalent copper halides; this limits essentially the use of these compounds.Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method 1.Copper(I) b-diketonates and their complexes Among volatile copper(I) compounds, as among copper(II) deriv- atives, b-diketonates are promising for CVD technology. These compounds are attractive due to their ability to undergo dispro- portionation at comparatively low temperatures to give the corresponding volatile copper(II) b-diketonates and metallic cop- per directly in the form of thin films Cu0 + CuII(b-diket)2 .2CuI(b-diket) However, the operation with copper(I) b-diketonates has specific difficulties. First, the contact of these compounds with atmospheric oxygen during synthesis results in their smooth oxidation to the corresponding copper(II) b-diketonates, which necessitates that their synthesis be carried out in an inert atmos- phere. Moreover, pure copper(I) b-diketonates are unstable and undergo disproportionation even at relatively low tempera- tures.275 This can be prevented by additional complexing of the coordination-unsaturated copper atom with various ligands (Lewis bases).Various organic and inorganic compounds (carbon monox- ide, alkynes, dienes, phosphines, amines) forming volatile com- plexes (b-diket)Cu . Ln can be used as additional ligands. The additional ligand is eliminated upon disproportionation of the complex to give a volatile neutral compound 2(b-diket)CuI . Ln Cu0 + CuII(b-diket)2 + 2n L:. Of numerous compounds of this type synthesised to date, only compounds with H(HFA) satisfy the CVD process requirements. Derivatives of other b-diketones are generally insufficiently stable or do not provide reproducible results in the film synthesis. For example, rather stable complexes of Cu(acac) with trimethylsilyl- alkynes and trimethylsilylalkadiynes were obtained.These com- plexes were suggested as volatile CCC,276 but CVD processes based on them have not been carried out. We believe that the problem of stability of complexes with copper(I) b-diketonates, in particular with Cu(HFA), with addi- tional ligands is a key problem for the evaluation of their possible use as the starting CCC in CVD technology. Thermodynamically stable complexes with high activation energies of thermolysis upon film deposition require high substrate temperatures or additional energy sources (plasma, laser radiation); this dimin- ishes the advantages of these CCC over similar copper(II) b-di- ketonates which are much more readily accessible synthetically. Cu(I) complexes of low stability cannot provide the deposition efficiency and the process reproducibility.Hence, there should be a certain optimum degree of stabilisation of the Cu(HFA) molecule by virtue of the p-donor ability of a Lewis base used as the additional ligand. In fact, of all Cu(HFA) . L complexes synthesised, derivatives considered as potential starting CCC for CVD processes include those which contain moderately strong p-donors (e.g., phosphines rather than amines) or compounds in which the p-donor ability of the multiple bond is somewhat weakened due to the incorporation of p-accepting trimethylsilyl groups [trimethylvinylsilane, bis(trimethylsilyl)acetylene].277 To date, several methods for the preparation of complexes of copper(I) b-diketonates with Lewis bases have been described.These include direct reaction with copper(I) oxide, the exchange reaction with sodium b-diketonates and metallation of b-dike- tones with monovalent copper compounds.278 ± 282 (1) Cu2O + 2H(b-diket) + 2nL 2(b-diket)Cu . Ln + H2O, (2) (b-diket)Cu . Ln + NaCl, ClCu . Ln + Na(b-diket) (Z5-C5H5)Cu . P(CH3)3 + H(b-diket) (3) (b-diket)Cu . P(CH3)3 + C5H6 . Each of these variants involves certain synthetic difficulties. For example, reaction (1) is, technically, the simplest method.282 However, water formed in the reaction favours undesirable disproportionation during the synthesis. In addition, it is difficult 1069 to monitor the stoichiometry of the products in this method.283 The order of addition of the reagents is very important, since the formation of free Cu(I) b-diketonate, which undergoes dispropor- tionation below 78 8C, should not be allowed.This is avoided by using an excess of the Lewis base. The additional problem in method (2) involves the presence of sodium cations, which can appear as admixtures in the films obtained from the CCC synthesised using this approach.111 The presence of these admixtures is absolutely unacceptable. The reaction according to the pathway (3) occurs without complications; however, the cyclopentadienyl complex of cop- per(I) required is the least accessible potential starting compound of all three pathways. a. Complexes of Cu(I) b-diketonates with alkenes and alkenylsilanes The stabilising effect of the double bond in alkenes with respect to copper(I) b-diketonates is generally sufficient for the formation of stable complexes that can serve as the starting CCC.This is exemplified in the complex of (HFA)Cu(I) with vinylcyclohexane (VCH) synthesised 284 according to the procedure described previously.282 Unfortunately, data on the physicochemical prop- erties of this compound were not reported. The substrate temper- ature during the film deposition was 100 ± 175 8C, which is lower than the temperatures for other CCC. The copper films obtained had low SER and virtually did not contain admixtures. Of copper(I) derivatives stabilised by double bonds of the additional ligand, complexes of Cu(HFA) with alkenyltrimethyl- silanes are among the most promising compounds for the syn- thesis of copper films. It is known that the trimethylsilyl group manifests a signifi- cant p-acceptor effect 277 which is unfavourable for the coordina- tion of the adjacentC=Cbond of the alkene with the monovalent copper ion.In fact, the compound Cu(HFA) with the simplest alkenyltrimethylsilane, viz., vinyltrimethylsilane (VTMS) 2 first described in 1991,285 is readily oxidised in air to give copper(II) derivatives and is quickly decomposed at*150 8C.286 CF3 H2C CH HC CF3 O C CuO C CH3 Si H3C CH3 2 The decomposition of (HFA)Cu(VTMS) was noted even on storage of this compound at room temperature.287 However, these drawbacks are compensated by the fact that this compound is a volatile liquid even at 20 8C.The vapour pressure of (HFA). .Cu(VTMS) is 0.14 Torr at 30 8C (see Ref. 288) and 0.3 Torr at 40 8C (see Ref. 289) and is considerably higher than the similar parameter for Cu(HFA)2 up to 95 8C.285 The thermolysis activation energies of (HFA)Cu(VTMS) for various deposition conditions and substrate materials have been reported.290, 291 Various aspects of substrate surface pretreatment for maximising the deposition efficiency of copper films from this compound were studied in detail for TiN.292 ± 296 The best charac- teristics of copper films obtained from vapour on this widespread material were achieved on substrate surfaces pretreated with WF6 vapour.295 On the whole, the CVD process involving (HFA)Cu(VTMS) has been studied rather thoroughly.An inert gas (He, Ar, N2) or its mixture with H2 was used as a carrier gas in most cases. It was shown 288 that the deposition rate of copper films increases with increase in the partial pressure of CCC vapour, reaching a maximum (36 nm min71) at a pressure of 5.7 mTorr. The SER of the copper films was *2.2 mO cm at thicknesses >500 nm. Similar values were obtained under the same or similar conditions in other studies.297, 2981070As mentioned above, the presence of a reducing gas (H2) in the thermolytic disproportionation of copper(I) derivatives is gener- ally unnecessary. Nevertheless, it was noted 296 that the character- istics of copper films can be improved essentially by additional annealing at 450 8C in a stream of Ar/H2 (9 : 1).This increases the grain size in the films, and their SER is decreased from 2.35 to 2.12 mO cm. Various aspects of the favourable effect of hydrogen as the carrier gas in the CVD process based on (HFA)Cu(VTMS) were considered by Lin and Chen.299 It was shown that the use of H2 provides denser copper films with lower SER values and increases the deposition rate. The film deposition rate can be increased additionally by substrate surface pretreatment with nitrogen plasma. Braeckelmann et al.300 studied the effect of the TiN substrate temperature in the range from 165 to 205 8C on the characteristics of copper films obtained from (HFA)Cu(VTMS) vapour. It was shown in particular that an increase in temperature decreased the SER of copper films from 2.15 to 1.8 mO cm and increased the deposition rate by a factor of 2.5.This is attributed to an increase in the CCC disproportionation rate. However, it was shown in a study of a similar process on W/Si substrates 301 that the SER of the copper films obtained in the substrate temperature range from 160 to 190 8C remained virtually unchanged. The temperature of 190 8C was found to be the optimum providing the maximum film growth rate (up to 500 nm min71). It was found in a study of the deposition of copper films on a TiN substrate from (HFA)Cu(VTMS) vapour in a wider temper- ature range (from 160 to 330 8C) 290 that below 200 8C the SER of the films is *2.0 mO cm. An increase in the temperature sharply impairs the morphology of the films and increases their SER.In addition to the substrate temperature, the partial pressure of the starting CCC vapour is an important factor determining the deposition rate and the resulting copper film parameters. A study of the deposition of copper films from (HFA)Cu(VTMS) vapour (TiN/Ti/Si substrate at a temperature of 170 8C) at an overall pressure of 0.5 Torr showed 302, 303 that under these conditions, an increase in the partial pressure of CCC vapour to 0.048 Torr increased the deposition rate monotonically up to 25 nm min71, and the SER of the film decreased monotonically to 2.00 mO cm. With an increase in the partial pressure, the copper films became denser and had fewer surface defects.At partial pressures of (HFA)Cu(VTMS) vapour above 0.048 Torr, the deposition rate and the SER of copper films became virtually constant. A similar dependence was obtained by Awaya and Arita.304 It was shown that at temperatures below 200 8C, the rate of deposition of copper films from (HFA)Cu(VTMS) vapour is proportional to the square root of the CCC vapour partial pressure. It was noted above that if copper(II) derivatives [for example, Cu(HFA)2] are used inCVDprocess, the addition of water vapour to the carrier gas increased the growth rate of copper films while their SER remained unchanged, and in some cases (not always) also increased the process selectivity.171, 190, 205, 225, 226 Similar effects were reported for (HFA)Cu(VTMS).Deposition of copper films from (HFA)Cu(VTMS) vapour on SiO2 and Pt/SiO2 substrates at 150 ± 250 8C gave copper films 800 nm thick with low SER in the absence of water vapour.305 The addition of water vapour did not accelerate the process in the case of the SiO2 substrate. In the case of Pt/SiO2 substrates, a small increase in the deposition rate was observed, but much smaller than that with the use of copper(II) derivatives under the same conditions. The copper films obtained in the presence of water vapour had rough surface structure and high SER (>20 mO cm). More successful experiments on the deposition of copper films from (HFA)Cu(VTMS) vapour in the presence of water were carried out by Gelatos et al.289 It was shown that at rather low partial pressure of water vapour (< 0.4 Torr), as in the case of Cu(HFA)2 vapour, a sharp increase in the copper film deposition rate was in fact observed.This effect was particularly well V N Vertoprakhov, S A Krupoder pronounced at PH2O<0.2 Torr. The specific electric resistance of the films obtained was 2.30.1 mO cm. Subsequent annealing decreased the SER to 1.90.1 mO cm, and the copper films had a considerably higher density than similar films obtained by CVD from (HFA)Cu(VTMS) vapour in the absence of water vapour. It was noted that such a low SER value was reached in the presence of water vapour at a film thickness of 50 nm, while under `anhydrous' deposition conditions, this required a film thickness of at least 300 nm.An increase in the water vapour partial pressure to 0.75 Torr impairs strongly the characteristics of the films.289, 306 The possibility of optimisation of a CVD process involving (HFA)Cu(VTMS) and water vapour was studied.307, 308 An important fact was noted: the preparation of copper films with low SER, good adhesion and high deposition rate requires that the injection of water vapour be carried out quickly in the beginning of the process. Copper films with SER of 1.9 mO cm (after annealing), and copper strips 0.4 mm wide and 1 mm thick with good characteristics were obtained under these conditions. The addition of small amounts of H(HFA) vapour in theCVD of copper films from (HFA)Cu(VTMS) vapour can improve substantially the copper film morphology.229 The presence of the b-diketone favours an accelerated and uniform growth of nucle- ation centres on the entire film surface, thus protecting it from cracks and roughening.However, the process loses its selectivity completely in this case. One more variant for the deposition of copper films by the CVD method from (HFA)Cu(VTMS) vapour involves the addi- tion of an excess of liquid VTMS to the starting CCC (40% ± 60% relative to the volume of the starting CCC) before the evapora- tion.309, 310 This enabled deposition rates of 40 ± 60 nm min71 to be achieved. It was noted that replacement of the carrier gas by hydrogen did not change essentially the film growth rate.310 The mean specific electric resistance of the copper films obtained, with thicknesses from 200 to 1800 nm, was 1.86 mO cm; additional annealing of copper films at 600 8C in a high vacuum (1077 Torr) decreased it to 1.76 ± 1.77 mO cm.309 Simultaneous use of H(HFA) and VTMS vapour is possi- ble.311 The H(HFA) vapour increases the copper film deposition rate, and the VTMS vapour accelerates the CCC evaporation.The rate of copper film deposition increases considerably if a CVD process with the use of (HFA)Cu(VTMS) vapour is carried out with simultaneous application of a shifting voltage of730 V to the substrate.312, 313 This effect is explained by reorientation of the dipole molecules of the starting CCC in the electric field stimulating their adhesion to the substrate. The decomposition of (HFA)Cu(VTMS) vapour can be addi- tionally promoted [as in the case of copper(II) derivatives as the starting CCC] by radiation.This variant can primarily be used in the preparation of microstructures from this CCC by lithographic methods. In the study of Ochiani et al.,314 deposition of copper films from (HFA)Cu(VTMS) vapour was carried out simultaneously with electron lithography. The copper film samples obtained had rather low SER. The formation of copper films from (HFA)Cu(VTMS) exposed to an electron beam has been studied.315 It was found that the energy threshold where decomposition starts is 40.5 eV and that decomposition occurs by dissociative electron addition. However, the films obtained by this method are considerably contaminated with carbon, apparently due to decomposition of HFA fragments under these rather drastic conditions.A study of the preparation of thin copper strips 3 mm wide by thermal decomposition of (HFA)Cu(VTMS) vapour under argon laser radiation 286 showed that in this case, as with Cu(HFA)2 vapour,225 the addition of water vapour to the vapour of the starting CCC affects favourably the deposition rate and the characteristics of the resulting copper microstructures. The SER values of the best copper strips obtained this way were close to the SER of metallic copper. Similar results were obtained underPreparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method analogous conditions;316 it was also noted that the morphology of the surface copper strips depends essentially on the laser radiation power.Moilanen et al.317 achieved rather high film deposition rates of up to 400 nm s71, but the SER of the strips increased somewhat (up to 3.7 mO cm). Thin copper strips (channels) up to 16 mm deep were obtained by CVD using decomposition of (HFA)Cu(VTMS) vapour.318 The specific electric resistance of the copper channels was 1.95 mO cm; annealing gave a value of 1.86 mO cm. Successful optimisation of preparation of copper strips with SER of42.0 mO cm using CVD from (HFA)Cu(VTMS) vapour was reported.319 Popovici et al.320 observed the terrace growth of copper films during the decomposition of (HFA)Cu(VTMS) vapour on Teflon substrates under excimer laser radiation (l=248 nm).It was assumed that growth of this type occurs owing to the interference between the Teflon support and the precipitated copper layers. In turn, the successful use of (HFA)Cu(VTMS) vapour as a convenient starting CCC for the CVD process has served as an incentive for the search for volatile Cu(HFA) complexes with similar compositions and physicochemical characteristics. As a result, new startingCCCwere reported, where allyltrimethylsilane (ATMS) (for example, compound 3) 321 ± 323 and vinylalkoxysi- lanes, such as vinyltrimethoxysilane (VTMOS),287 vinyl(dime- thoxy)methylsilane (VDMOMS)324 and vinyltriethoxysilane (VTEOS),324, 325 served as Lewis bases. These CCC are more stable than (HFA)Cu(VTMS).They are liquid at room temper- ature, but data on their physicochemical properties were not reported in these studies. CF3 H2C CH HC O C CuO CCF3 CH2 Si H3C CH3 CH3 3 If (HFA)Cu(VTMOS) vapour is used, the optimum temper- ature range for the TiN substrate, where copper films with the best characteristics are formed, is 175 ± 200 8C.287 The lowest SER of the films obtained under these conditions is 2.5 mO cm. Somewhat better results (deposition rate up to 100 nm min71 and lower SER of the films) were obtained in CVD processes based on (HFA)Cu(VDMOMS) and (HFA)Cu(VTEOS).324, 325 It was noted that the presence of water vapour during the deposition of films is necessary for the former compound.324 Very interesting results were obtained in CVD processes using (HFA)Cu(ATMS).It was found that, other conditions being equal, the incorporation of ATMS in the complex decreases considerably the temperatures for copper film deposition in comparison with those for similar compounds without ATMS. (HFA)Cu(ATMS) evaporates well at 40 ± 45 8C and forms copper films with low SER at substrate temperatures of 60 ± 170 8C.321 It was noted that the film parameters remain almost unchanged in this rather wide temperature range. A certain role in these features of CVD processes is apparently associated with the fact that (HFA)Cu(ATMS) is even less stable thermally than (HFA)Cu(VTMS) and (HFA)Cu(VTMOS), and hence it is more reactive.323, 326 At a temperature as low as 60 8C, the IR spectra of (HFA)Cu(ATMS) vapour contain bands corresponding to Cu(HFA)2.It is interesting that the morphology of the copper film surfaces obtained also depends on the reactivity of the starting compounds: for example, the grain size in the copper films obtained from (HFA)Cu(ATMS) vapour is much larger than that in the copper films obtained from (HFA)Cu(VTMS) vapour.323 1071 b. Complexes of copper(I) b-diketonates with isocyanides Of copper(I) b-diketonate complexes stabilised by interactions of the copper(I) ion with the double bond, those with tert-butyl isocyanide (ButNC) are considered as promising starting CCC for the preparation of copper films by CVD technology. It should be noted that the corresponding complexes of copper(I) derivatives with nitriles RCN decompose even at room temperature.327 It is known that the nitrile group is an efficient p-acceptor [the resonance constant is s0R=+0.08 (see Ref. 277)] and hence it cannot form stable complexes with the copper(I) ion.On the other hand, the corresponding effect of the isocyanide group is close to zero,277 and the additional p-donor tert-butyl group makes the complex sufficiently stable. Kruck and Terfloth 328 reported data on the synthesis [by method (2)] and the physicochemical properties of volatile (HFA)Cu . ButNC and (HFA)Cu . (ButNC)2 complexes (in partic- ular, the sublimation temperatures at 0.1 Torr are 55 and 95 8C, respectively). CVD processes with the use of vapour of these compounds gave copper films with good characteristics.329 c.Complexes of copper(I) b-diketonates with dienes Dienes form rather stable 1 : 1 complexes with the copper(I) ion. Copper(I) b-diketonates are usually stabilised by 1,5-cycloocta- diene (1,5-COD). Like complexes with VTMS, complexes of 1,5-COD with Cu(acac) and Cu(TFA) have low stability;282 on the contrary, (HFA)Cu(1,5-COD) is a very stable compound which sublimes without decomposition at 65 ± 70 8C and remains unchanged in air for a period of one year. The vapour pressure of this compound is*56 mTorr at 62 8C.330 The spatial structure of the (HFA)Cu(1,5-COD) molecule (Fig. 3) was determined 282 and analysed in detail.280 It was noted in the latter study that the actual coordination of the Cu(I) centre is more complex than Fig.3 shows and fits a 3+1 scheme. In other words, the 1,5-COD fragment is coordinated asymmetri- cally. Decomposition of vapour of this CCC at 200 8C resulted in copper film formation and evolution of volatile decomposition products, viz., Cu(HFA)2 and 1,5-COD.282 The preparation of copper films by CVD from (HFA)Cu(1,5- COD) vapour has been described in a number of papers.330 ± 335 For example, it was noted 330 that a temperature of 65 8C in the evaporator and a temperature of 70 8C of the reactor cell walls are sufficient for the process. Admixtures of carbon, oxygen and fluorine are virtually absent in the films (<1%). The specific electric resistance of films obtained from (HFA)Cu(1,5-COD) vapour depends primarily on the substrate temperature.At the optimum temperature (190 ± 210 8C), the SER of the films was 1.970.10 mO cm. Higher deposition rates (>15 nm min71) and larger film thicknesses (>250 nm) worsen this parameter. F(6) C(7) F(5) O(2) C(5) C(8) C(6) F(4) C(3) C(2) Cu C(9) C(13) F(1) O(1) C(10) C(1)C(4) C(12) F(2) C(11) F(3) Figure 3. (HFA)Cu(1,5-COD) molecule.1072(HFA)Cu(1,5-COD) is a promising starting compound in an aerosol modification of CVD technology.335 The films are depos- ited from an aerosol of a toluene solution of the starting compound. The aerosol is fed to the reactor cell in a stream of an inert gas (nitrogen); intermediate preheating of the vapour to 40 ± 100 8C is provided.Even at a substrate temperature of 120 8C, crystalline copper films were obtained; their SER were within 1.7 ± 3.5 mO cm under optimum process conditions (140 8C, deposition rate 80 nm min71). Copper films with good characteristics are obtained using (HFA)Cu(1,5-COD) as the startingCCCin the presence of CO.336 Carbon monoxide displaces the 1,5-COD ligand from the complex at 770 8C to give an intermediate unstable carbonyl compound, (HFA)Cu(CO), which in this case serves as the startingCCCin the CVD process. Isolation of this carbonyl complex in a pure state failed. 1,5-COD is not the only diene that can give volatile complexes with copper(I) b-diketonates applicable in CVD technology. In particular, a hydrocarbon which is rare but quite accessible synthetically, viz., 7-tert-butoxynorbornadiene 7-ButO-NBD, was suggested for this purpose 337 (for its synthesis, see Ref. 338).Using a procedure which is a modification of the method (2), Chi et al.337 obtained a series of complexes of the type (b-diket)Cu(7-ButO-NBD), where b-diket=acac, TFA, HFA, FOD, with sublimation temperatures from 75 to 90 8C at 0.1 Torr. The geometry of the molecules was determined for all of these complexes. However, only one of these compounds, namely, Cu(HFA)(7-ButO-NBD), was found to be suitable for the prep- aration of copper films. The specific electric resistance of the films was satisfactory at sufficiently high substrate temperatures; the SER of films increased abruptly (>1000 mO cm) and their sur- face morphology is deteriorated upon decrease in the temperature of the substrate to 170 8C.337 d.Complexes of copper(I) b-diketonates with alkynes It is known that the triple bond, particularly that in alkyl- substituted acetylenes, has a stronger p-donor effect than the double bond.277 Hence, alkynes should form sufficiently stable complexes with the copper(I) atom in b-diketonate derivatives. The first synthesis of volatile Z2-complexes of Cu(HFA) with alkynes, viz., but-2-yne, bis(trimethylsilyl)acetylene and diphenyl- acetylene, was carried out by the reaction of (HFA)Na with CuCl in the presence of excess alkyne.278 Volatile Cu(HFA) complexes with pent-2-yne 339 and hex-3-yne 340 were also obtained.A little later, a simple one-stage procedure was proposed for the synthesis of Cu(HFA) complexes with but-2-yne 341 and bis(trimethyl- silyl)acetylene.342 A series of compounds of this class containing silicon atoms at the triple bond were patented as the starting CCC for CVD technology.343 According to X-ray diffraction data, the copper(I) atom in the compounds in question has a flat trigonal configuration with coplanar diketonate and alkyne ligands (Fig. 4).278 The nature of bonding and the electron interaction of the fragments in this type of complex was studied in more detail using 1H and 13C NMR spectroscopy.281, 342 The possibility of the existence of equilibrium between mono- nuclear and binuclear forms of these complexes was also consid- ered,342 as some of them undergo chemical transformations in the bubbler of the CVD reactor at 45 ± 65 8C to give low-volatility products.340, 341 In this case, the corresponding binuclear com- plexes are obviously formed by m-Z2 bonding of the triple bond with the copper atom [the ratio alkyne : Cu(HFA)=1 : 2].Such complexes are also known in the chemistry of other transition metals.344, 345 However, in this case the complexes are extremely unstable 342 and can only be formed as labile intermediates. On the whole, the nature of this reaction has been studied insufficiently. The Cu(HFA) complex with but-2-yne is decomposed on the substrate at 150 ± 210 8C under the conditions of the CVD process to give dense mirror-like copper films with low SER.339 The rate of deposition of such films can be increased by addition of water V N Vertoprakhov, S A Krupoder F(2) F(1) F(3) O(1) C(5) C(9) C(1) C(2) C(8) C(3) Cu C(7) C(4) F(4) O(2) C(6) F(5) F(6) Figure 4.(HFA)Cu(CH3C:CCH3) molecule. vapour.306 The possibility of selective deposition of copper films starting from this compound was considered.346 An unusual modification of this type of the starting CCC was suggested by Schmidt and Behrens.347 They synthesised a volatile 2 : 1 Cu(HFA) complex with bis(tert-butylethynyl) sulfide. It was shown that the sulfur atom did not participate in coordination with the copper(I) ion. Similar Cu(I) complexes with other b-diketonates as the ligands were found to be unstable even at room temperature. e.Complexes of copper(I) b-diketonates with phosphines Yet another possibility to stabilise Cu(HFA) is the preparation of its complexes with derivatives of VA Group elements. Of these derivatives, trialkylphosphines are used most widely as additional ligands. However, the volatility of biscomplexes of the type (b-diket)Cu(PAlk3)2 synthesised first 348 was insufficient for their use in CVD technology. The syntheses of volatile monomeric 1 : 1 complexes of cop- per(I) b-diketonates with organic phosphines for CVD were first reported by Shin et al.349 ± 351 Initially, the reaction according to pathway (3) became the main method for their synthesis;351 the applicability of pathway (2) was shown later.279, 352 The com- pounds obtained were low-melting crystals capable of sublimation or viscous liquids.As in the case of similar complexes with 1,5-COD,280 the HFA derivative (Fig. 5) was the most stable of the series of compounds with the general formula (b-diket)Cu(PMe3), where b-diket= acac, TFA, HFA. The same is true for complexes with other trialkylphos- phines.352 To interpret this phenomenon and to study the nature of the interaction of the copper(I) ion with the phosphine molecule, photoelectron spectroscopic data and quantum-chem- ical calculations were used.353 C(4) C(8) C(1) O(1) C(6) C(2) P Cu C(5) O(2) C(3) C(7) Figure 5. (HFA)Cu(PMe3) molecule.Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method The vapour pressure of Cu(HFA)(PMe3) was measured at various temperatures and the logP71/T relationship was plot- ted.279 The corresponding vapour pressures of Cu(HFA)(PMe3) and Cu(HFA)2 were found to be similar.179 The first experiments on CVD of thin films with the use of phosphine complexes of copper(II) b-diketonates as CCC gave promising results.For example, a rather high deposition rate (up to 100 nm min71) was achieved at 150 8C on a Pt/SiO2 substrate resulting in dense mirror-like copper films with good adhesion.354 It was reported that the process can occur selectively.350 A more detailed study showed that phosphine complexes of Cu(HFA) have higher selectivity than similar complexes with 1,5-COD or alkynes under comparable conditions.Thin copper films were also deposited from Cu(HFA)(PMe3) vapour in the presence of water and organic solvents (alcohols, toluene, tetrahydrofuran).206 It was shown that this compound can react with water at room temperature with partial oxidation of copper(I) to copper(II). The use of alcohols results in reversible coordination of the alcohol molecules with copper(I) ions with formation of a more volatile alkoxide and, hence, leads to an increase in the deposition rate of copper films. 2. Volatile compounds of copper(I) devoid of b-diketonate ligands A series of volatile derivatives of monovalent copper includes CCC which are not inferior to b-diketonates in terms of phys- icochemical properties required in CVD technology.Such CCC can be considered as alternatives to the classical b-diketonate copper(I) complexes. An obvious direction of the molecular design of such compounds is some structural similarity to copper b-diketonates. In view of this, we included copper b-iminoketo- nates and b-oxocarboxylates in our consideration. Yet another direction in the creation of new CCC is the search for fragments capable of replacing completely the b-diketonate group while preserving the properties needed for CVD technol- ogy. The promising compounds of this type include alkoxides, amides, cyclopentadiene derivatives and heterocyclic compounds (e.g., pyrazole derivatives). a. Copper(I) b-iminoketonates and b-oxoacid derivatives Fluorinated b-iminoketones required for the synthesis of stable nitrogen-containing analogues of copper(I) b-diketonates are obtained by a two-step procedure from the corresponding b-dike- tones.355 This fact itself implies the low competitiveness of copper(I) b-iminoketonates synthesised from sodium b-iminoke- tones by reaction (2) in comparison with similar copper(I) b-diketonates.At room temperature, complexes of this series (with trialkyl- phosphines as additional ligands) are volatile liquids or solid compounds. A study of whether copper films can be obtained by CVD from vapours of copper(I) b-iminoketonates did not reveal any considerable advantages of these compounds over the more accessible oxygen analogues,355, 356 i.e., copper(I) b-diketonates.Only one study was published 357 reporting good results in the preparation of copper films by CVD with the use of a complex of copper(I) 3-oxobutanoate with trimethylphosphine which is rather stable and resistant against oxidation in air. b. Copper(I) alkoxides The main drawback of copper(I) alkoxides as potential CCC for CVD technology is their rather easy transition into non-volatile polymers. Nevertheless, successful attempts at the preparation of copper films from tetrameric copper(I) tert-butoxide [Cu(OBut)]4 were reported.358, 359 The pyrolysis temperature of this compound is rather high (*400 8C). As indicated by Lemmen et al.,360 this compound is rather sensitive to oxygen and atmospheric moisture (the films obtained contained from 2% to 5% of oxygen), there- fore it was not widely used in CVD technology.Hampden-Smith et al.361 synthesised a volatile monomeric complex of copper(I) tert-butoxide with trimethylphosphine 1073 (ButO)Cu(PMe3). The CVD method made it possible to obtain pure copper films containing no oxygen and phosphorus admix- tures from this compound.361 c. Amide derivatives of copper(I) In the series of volatile amide derivatives of copper(I), there is one unusual compound used as a CCC, viz., the tetrameric cluster {Cu[N(SiMe3)2]}4.362 This compound sublimes at rather high temperature (*180 8C). Nevertheless, it produces copper films with satisfactory characteristics. It is possible that further research on copper(I) amide derivatives will allow one to obtain com- pounds of this class that would be more volatile and, accordingly, more promising for CVD technology.d. Cyclopentadienyl derivatives of copper(I) Cyclopentadienyl derivatives of copper(I) compete satisfactorily with copper b-diketonates as starting compounds in CVD proc- esses. Cyclopentadienyl derivatives of copper(I) are usually stabi- lised by trialkylphosphines and isocyanides. Trialkylphosphine complexes of cyclopentadienylcopper(I) are usually synthesised using a procedure similar to the pathway (2) from cyclopentadienylsodium and halophosphine complexes of copper(I).363, 364 Their typical sublimation temperatures are 60 ± 70 8C, and the thermal decomposition temperatures are 120 ± 125 8C.365 The calculated vapour pressure of (Cp)Cu(PMe3) at 80 and 100 8C are 14 and 30 Torr, respectively.7 For similar platinum compounds, it was found that incorpo- ration of a methyl group in the cyclopentadiene ring increases considerably the volatility of the compound.364 In view of this, copper(I) methylcyclopentadienyl complexes with trialkylphos- phines were synthesised,363 but their sublimation temperatures and vapour pressures were almost the same as those of the non- substituted analogues.A study of the physicochemical properties showed that the methyl-substituted derivatives are less stable and undergo decomposition at temperatures 15 ± 20 8C lower than the non-substituted compounds. This is an advantage of methyl- substituted cyclopentadienyl derivatives of copper(I) for the CVD process.Recently, a number of methyl- and ethyl-substi- tuted cyclopentadienyl complexes of copper(I) with triisopropyl- phosphine were patented as promising volatile CCC for CVD technology.366 Originally, such complexes were used inCVD for the synthesis of mixed copper-containing semiconductor systems of the CuGaS2 type.365 The first studies on the CVD of copper films with the use of these complexes gave good results. For example, it was shown 367 that decomposition of (Cp)Cu(PAlk3) vapour (Alk=Me, Et, Bu) on substrates made of different materials in vacuo gave high-purity copper films with a SER that differed from the SER of bulk metallic copper by *10%. Based on (Cp)Cu(PMe3), copper films with SER of 1.8 ± 2.2 mO cm were obtained,361 but this value was achieved only after annealing.The morphology of the surface of copper films formed from (Cp)Cu(PEt3) vapour on SiO2 was considered in detail.368 The (Cp)Cu(PEt3) complex was used as the starting CCC in laser-activated preparation of copper strips and in thermal laser lithography.369 Irradiation of a thin layer of this compound with the light of an argon laser at l=514 nm gave a copper strip containing less than 10% carbon and with SER of *6.8 mO cm. The use of radiation excimer lasers with shorter wavelengths resulted in strips of copper ± carbon composites. The (Cp)Cu(PEt3) complex was used for the CVD synthesis of thin copper ± aluminium films in situ.7 In this method, the rather low concentration of copper (*1.4%) decreases greatly the SER of the basic aluminium film.The valuable property of (Cp)Cu(PEt3) is that, unlike the majority of the starting volatile CCC, it is chemically inert with respect to the second aluminium- containing starting compound used simultaneously, viz., dime- thylalane (CH3)2AlH. As in the case of Cu(HFA), tert-butyl isocyanide can be used for the stabilisation of copper(I) cyclopentadienyl complexes. The1074 (Cp)Cu(ButNC) compound was obtained for the first time using method (1) and was used as a catalyst of some organic reac- tions.370 Later, this substance was synthesised using method (2) and was suggested as a starting CCC for the preparation of copper films by CVD.328 Studies of the behaviour of this compound in CVD showed329, 371 that high-quality films free from carbon admixtures can in fact be obtained on its basis.The optimum substrate temperature range is 200 ± 350 8C,371 although notice- able decomposition of (Cp)Cu(ButNC) starts at 135 8C.328 e. Complexes of copper(I) with pyrazole derivatives Many volatile compounds are known in which the copper(I) ion is bound to heterocyclic fragments. However, only one study is devoted to the synthesis of copper(I) pyrazolyl borate complexes with triethylphosphine.372 At a substrate temperature of 150 8C, these compounds give polycrystalline copper films with good characteristics and high selectivity. Apparently, studies in this direction will continue.3. Mechanisms of decomposition of copper(I) derivatives on the substrate It is generally believed that copper films are formed from mono- valent copper ±HFA complexes due to their disproportionation followed by removal of the volatile product. The formation of Cu(HFA)2 from (HFA)Cu(VTMS) in the gas phase was proved by spectral methods.189, 373 The initial step of (HFA)Cu(VTMS) decomposition should apparently involve cleavage of the weakest bond in the molecule, i.e., the bond between the copper(I) ion and the VTMS molecule. The question is, where does this cleavage occur? There are two answers to this question: first, on the substrate surface; second, in the vapour phase. Evidence in favour of both variants can be found in the literature.The majority of the authors favour the first variant.201, 290, 374 For example, a comparative study of the Cu(HFA)2 and (HFA)- Cu(VTMS) decomposition by Fourier IR spectroscopy revealed bands unambiguously assigned to the VTMS ligand in the spectra of (HFA)Cu(VTMS) adsorbed on a Si(100) substrate at 7148 8C.201 However, these bands disappeared at 27 8C, which made it possible to conclude 201 that at this temperature VTMS was eliminated extremely quickly from the adsorbed compound. The disproportionation of the resulting Cu(HFA) molecules was proved and considered in detail.310, 374 ± 376 Thus, taking this into account, one can make the conclusion that the most probable mechanism of decomposition of (HFA)Cu(VTMS) vapour involves four steps: (1) adsorption of the entire molecule on the substrate (HFA)Cu(VTMS)(ads); (HFA)Cu(VTMS)(vapour) (2) elimination of VTMS followed by fast desorption (HFA)Cu(VTMS)(ads) (HFA)Cu(ads)+VTMS: ; Cu(ads)+Cu(HFA)2(ads); 2(HFA)Cu(ads) Cu(HFA)2(ads) Cu(HFA)2(vapour):.(3) bimolecular disproportionation of the intermediate formed (4) desorption of the volatile product of Cu(HFA)2 dispro- portionation However, studies of the kinetics of (HFA)Cu(VTMS) decom- position 377 gave evidence in favour of the second variant, i.e., molecular decomposition in the vapour phase. Hence, it is the intermediate Cu(HFA) that should be considered the actual starting compound in the copper film formation. The second mechanism, alternative to the one mentioned above, is as follows: (1) vapour-phase thermolysis (HFA)Cu(VTMS)(vapour) (HFA)Cu(vapour)+VTMS: ; V N Vertoprakhov, S A Krupoder (2) adsorption of the intermediate (HFA)Cu(ads); (HFA)Cu(vapour) (3) bimolecular disproportionation of the intermediate formed Cu(ads)+Cu(HFA)2(ads); 2(HFA)Cu(ads) (4) desorption of the volatile product of Cu(HFA)2 dispro- portionation Cu(HFA)2(ads) Cu(HFA)2(vapour):.The third and fourth steps are similar in both cases. It is possible that bimolecular disproportionation (third step) is not the only pathway for the transformation of the Cu(HFA) intermediate. A direct relationship between the activation energy of the copper nucleation from (HFA)Cu(VTMS) on the surface and the electric conductivity of the substrate was established; this is believed 378 to suggest a certain role of one-electron transfer from the surface to the adsorbed Cu(HFA), which was not taken into account in the mechanism considered above.One can imagine a situation (for example, at a sufficiently high temperature of the vapour flow) where a portion of molecules of the starting CCC are decomposed in the vapour phase, while the other portion do so on the substrate surface. The relative fractions of each process depend on the particular conditions of the process. The effect of the substrate temperature on the mechanism of low-vacuum deposition of the copper film from the (HFA)Cu(VTMS) vapour was studied in detail.298 It was shown that in the temperature range of 140 ± 220 8C, deposition obeys the Arrhenius law; at temperatures higher than 220 8C, the process switches to the saturation mode (the `transport-limited' situation).In this mode, the SER of the film and the content of contaminating carbon increase.379 In addition, the film formed becomes less uniform due to a decrease in mobility of copper atoms under the effect of the atoms of the quickly growing surface layers.380 The mechanism of the effect of partial pressure of (HFA)Cu(VTMS) vapour on the characteristics of films consid- ered by Lee et al.302 fits completely the deposition scheme described above. At low partial pressure of CCC vapour, the sites for the adsorption of the Cu(HFA) intermediates are preserved on the substrate surface.As the pressure is increased to a certain threshold, the surface becomes `adsorption-saturated' and the copper film deposition rate no longer depends on the partial pressure. As in a case of Cu(HFA)2, a special situation arises if CVD is carried out with the use of (HFA)Cu(VTMS) under ultrahigh vacuum (1079± 10710 Torr).199 Under these conditions, the first stage involves chemisorption of the entire CCC molecule on the substrate (TiN). At 27 8C, the CCC is completely transformed to the adsorbed Cu(HFA) intermediate, which does not dispropor- tionate upon heating but is decomposed to CO, CO2, CF4 and metallic copper. About 90% of the copper contained in the precursor remains on the substrate [*30% in the case of Cu(HFA)2 under the same conditions].At 280 8C, the diffusion of copper into the substrate starts. It is believed 306 that the favourable effect of water vapour on the deposition rate of copper films upon decomposition of (HFA)Cu(VTMS) vapour is associated with facilitation of the VTMS ligand elimination. The observed impairment of the characteristics of copper films at high partial pressure of water vapour 289, 306 occurs because of the side reaction of water with the adsorbed Cu(HFA) intermediate resulting in a Cu2Oadmixture in the films. Experiments with water labelled with the 18O isotope showed that it is the oxygen atoms from water molecules that are incorporated in the film.306 A mechanism of (HFA)Cu(VTMS) decomposition under ion beam impact to give copper films has been suggested.381 The decomposition mechanism of other copper(I) derivatives, such as Cu(HFA)(1,5-COD), and Cu(HFA) complexes with but-Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method 2-yne on a silicon substrate during the CVD process (*210 8C) is totally similar to the mechanism considered above 375, 382 (includ- ing that in the presence of water vapour 306).However, there are no data on the elimination of additional ligands (Lewis bases) from these compounds directly in the vapour phase. The mechanisms of transformations in thin copper films during their subsequent annealing 289, 309 have not been elucidated so far. The available data on the effect of annealing on the characteristics of films are rather contradictory.For example, it was reported that annealing of films decreased their SER consid- erably,289, 383 decreased it slightly 309 or did not affect it.384 This could result from processes affecting the SER in opposite ways: annealing increases the film crystallinity and, on the other hand, Table 5. CVD conditions and properties of copper films obtained from copper(I) compounds. Substrate P /Torr Barrier layer Starting CCC T /8C (HFA)Cu(VCH) (HFA)Cu(VTMS) TiN TiN TiN W, SiO2 , TiN 45 ± 55 7777 0.5 0.12 777 0.12 Si3N4 W, TiN, TiN, TiW SiO2 W, TiN, TiN 7 7777 2 TiN Si3N4 , SiO2 TiN TiN, Ti TiN, TiW 7750 740 0.5 33 (HFA)Cu(ATMS) 0.06 70 ± 80 7 740 ± 45 7 TiN, TiW TiN TiN TiN 60 ± 90 7 7 7 7 7 (HFA)Cu(VDMOMS) (HFA)Cu(VTEOS) (HFA)Cu(VTMOS) 0.07 (HFA)Cu(1,5-COD) 70 ± 80 60 ± 65 7 762 77 7TiN TiN 7Ag SiO2 760 SiO2 Pt W, SiO2 (HFA)Cu(CH3C:CCH3) 7 (ButO)Cu(PMe3) (Cp)Cu(PMe3) 7 (Cp)Cu(PEt3) 400 W, SiO2 SiO2 SiO2 W, SiO2 7 766 77180 (Cp)Cu(PBu3) [CuN(SiMe3)2]4 0.2 284 289 300 291 260, 308 311 309, 310 386 387 305 306, 388 290 389 314 316 206 317 379 296 302 378 390 288 377 284 321 322 324 324 325 287 351 330 334, 375 335 337 339 10 ± 12 (1.8 ± 2.2) e 361 367 361 363, 367 368 369 367 362 a Under a 30 ± 50 keV electron beam.b Under argon laser irradiation. c Under krypton laser irradiation. d Other supports were also used. e After annealing of the films at 600 8C. Si Si Si CoSi2 Si(100) 7 7 Si Si Al SiO2 Pt, SiO2 SiO2, Si SiO2, Si Si Si 7 7 Si Si 7Si Si Si 7 7 SiO2 , Pt 7 7 Si Si Si SiO2 Si(100) Si Si 7 7 Si Si(100) Si Si Si Si(111) Si, W (HFA)Cu(7-ButO-NBD)7 7 Pt, SiO2 Si SiO2 Si d 7 7 Pt, SiO2 Si Si(100) Si Si d Si 1075 favours their oxidation and enhances the diffusion of copper atoms into the substrate material.385 The mechanism of thermal decomposition of the vapour of copper(I) tert-butoxide tetramer on the substrate was studied in detail.359 It was shown that a homolytic process results in intermediate tert-butyl radicals and copper(I) oxide; the effect of the latter can explain the presence of oxygen in the films obtained.On the other hand, thermal decomposition of a vapourised phosphine complex of copper(I) tert-butoxide did not result in either the formation of Cu2O or an admixture of oxygen in the films. Data on the deposition conditions of copper films from copper(I) derivatives and the corresponding SER values are listed in Table 5. Ref. Carrier gas SER /mO cm Tp /8C Ar H2O 100 ± 175 200 ± 250 165 ± 205 160 ± 300 300 H2O, H2 H(HFA), VTMS VTMS, N2 220 ± 250 140 200 Ar/N2 H2O H2O N2 H2 Ar He H2 H2 H2O N2 Ar 1.9 ± 2.3 1.9 ± 2.3 1.8 ± 2.06 771.8 1.82 ± 1.86 73 ± 10 >20 2.0 ± 12.0 2.0 1.8 3.6 773.7 72.35 2.0 772.2 ± 2.3 72.0 1.7 ± 1.9 2.0 2.0 2.0 2.0 2.5 71.970.10 2.0 ± 2.4 N2 1.7 ± 3.5 <6 2.0 H2 <200 150 ± 200 a 150 ± 200 a 150 ± 200 b 150 ± 200 c 150 ± 200 b 100 ± 300 180 170 225 150 ± 250 130 ± 180 120 ± 180 75 ± 250 60 ± 170 275 195 195 180 ± 220 175 ± 200 160 ± 220 190 ± 210 180 ± 300 140 >200 210 260 ± 450 150 ± 220 260 150 ± 220 280 ± 550 280 ± 550 b 150 ± 220 145 2.0 1.8 ± 2.2 e 2.0 76.8 2.0 7 H21076 VI.The problem of selective deposition of copper films The selectivity of deposition of copper films by the CVD method from the vapour of copper(II) and copper(I) derivatives deserves a separate discussion.Let us note the short reviews 391, 392 on the subject. The selectivity of deposition implies preferential growth of the copper film on the surface of a certain material relative to the surface of some other material. This problem is important because at the present time there are no versatile methods for etching copper films covering the entire surface of a sample, which would allow one to obtain predetermined shapes on the surface required in the production of microelectronic devices. For example, a method for the dry etching of copper by consecutive treatment of its surface with H2O2 (which oxidises copper to Cu2O) and H(HFA) (which transforms Cu2O into a volatile copper b-diketonate) was suggested.388 A similar idea was implemented by Steger and Masel 393 who used H(HFA) vapour in a stream of O2.However, though high etching rates were achieved [*1000 nm min71 at 190 8C (see Ref. 388) and *1500 nm min71 at 350 8C (see Ref. 393)], these techniques did not become popular. This is also true for the other known etching methods requiring rather drastic conditions, such as a method with the use of SiCl4 in a stream of N2 (T>250 8C) 394 and treatment with a mixture of chlorine and PEt3 vapour in the presence of H(HFA) at 420 ± 540 8C.395 1. Deposition of copper films in `metal ± dielectric' systems As a rule, SiO2 is used as a dielectric for the deposition of copper films in `metal ± dielectric' systems. The growth of a copper film on a dielectric substrate surface primarily depends on the ease of adsorption of molecules of volatile CCC and compounds accompanying the process on the surface.The growth step following adsorption is nucleation, i.e., formation of primary grains of the growing film at the sites where adsorption is the most efficient and where decomposition of molecules of the CCC mostly starts. It was assumed that it is this step that determines the overall process selectivity.387 Hence, the simplest solution for the selectivity problem can involve suppressing the nucleation 396 or providing conditions where adsorption of molecules of the CCC on the dielectric surface is suppressed, while that on the metal surface is not suppressed. However, this solution also involves certain difficul- ties.The main of these results from the fact that, depending on the deposition conditions (temperature, pressure, carrier gas), the adsorption of molecules of CCC can be either physical or chemical.203 Therefore, a versatile method for the treatment of SiO2 substrate surfaces in order to minimise adsorption and achieve the maximum selectivity obviously cannot exist: every- thing depends on the particular CVD technology and the CCC used. The nature of selectivity in the deposition of copper films from Cu(HFA)2 vapour on Ag/SiO2 substrates was considered in detail by Cohen et al.376 It was assumed that selectivity is provided due to the existence of several paths of theCVDprocess: on the surface of silver, fast reduction of copper(II) to copper(I) occurs with desorption of the HFA ligand due to electron transfer from the metal surface, whereas on the surface of the SiO2 dielectric, the molecules of the starting CCC are adsorbed in the original state.The consecutive growth of film layers relative to the dielectric substrate can be regarded as a particular case of the above scheme. Deposition of copper films by decomposition of Cu(HFA)2 vapour in the presence of water vapour decreases greatly the selectivity of film growth in comparison with the `water-free' technique.171, 205 On the other hand, the presence of water vapour increases the growth rate. In order to keep this advantage and still preserve the process selectivity, the films are prepared at rather high substrate temperatures in the absence ofH2 (He, Ar as carrier gases).171, 190, 226 If high selectivity of the process is required, V N Vertoprakhov, S A Krupoder another variant is used in which such a substrate or barrier layer material is selected where water molecules are adsorbed poorly.205 Analysis of the dependence of the properties of the growing copper film, which is formed upon decomposition of (HFA)Cu(VTMS) vapour, on the thickness of the primary copper layers on SiO2 showed 397 that the highest deposition rate of copper is observed at a copper layer thickness of 4 nm, and the lowest SER of the film is observed at a thickness of 15 nm.The selectivity of the CVD process with the use of this CCC in the Cu/SiO2 system at temperatures below 200 8C depends on the carrier gas and on the CCC vapour partial pressure.304 Awaya et al.229 conducted a comparative study of the effect of the b-diketone excess on the rate and selectivity of CVD of copper strips upon decomposition of Cu(HFA)2 and (HFA)Cu(VTMS) vapour. The copper strips were deposited in the presence of water vapour on copper substrates with SiO2 templates previously sputter-deposited on the surface, at a substrate temperature of 390 8C.It was found that the optimum (fast and selective) process for the preparation of copper strips involves two stages. In the first stage, the substrate is treated with H(HFA) vapour in the presence of H2; after that, Cu(HFA)2 is fed into the reactor and reduced on the substrate in the presence ofH2.However, the presence of a free b-diketone in the stage of Cu(HFA)2 reduction results in the loss of selectivity. In the case where (HFA)Cu(VTMS) vapour is used in CVD processes, copper with a sputter-deposited SiO2 template was also used as the substrate; the substrate temperature was 160 ± 180 8C. It was shown that, unlike in the experiment with the Cu(HFA)2 vapour described above, the presence of H(HFA) in the stage of strip formation increases the rate of filling the template channel; however, this completely suppresses the process selectivity, as distinct strip edges could not be obtained. It is believed 229 that such a different behaviour of copper(I) and copper(II) compounds in this case is due to the difference between their decomposition mechanisms.Optimisation of the CVDprocess for film deposition with the use of vapour of copper(I) derivatives and similar templates was considered by Norman et al.398 Some increase in selectivity can be achieved by the preparation of copper strips on the same substrates, viz., Cu/SiO2 (template), from (HFA)Cu(VTMS) vapour. This is due to the fact that adsorption of the precursor molecules on SiO2, which decreases selectivity, depends directly on the presence of reactive hydroxy groups on the substrate surface.386, 399 Therefore, to exclude a decrease in selectivity, it was suggested 399 to pre-treat the SiO2 surface with dichlorodimethylsilane vapour.This compound eliminates the reactive OH groups from the SiO2 surface, thus creating unfavourable conditions for adsorption and film growth at these sites. Nevertheless, as the overall process selectivity is controlled by several factors (e.g., the substrate temperature and the dichlorodimethylsilane vapour pressure), it is difficult to obtain reproducible results. The selective deposition of copper films can also be provided by using thermally grown SiO2 as the substrate.400, 401 In this case, the presence ofH2 during the process [which is usually the case for processes based on Cu(II) derivatives] is an unfavourable factor for the selectivity, as it results in reduction of the SiO2 surface layer and generation of OH groups.402 A study of nucleation upon decomposition of (HFA)Cu(VTMS) on a SiO2 surface `freed' from the OH groups showed that under these conditions, an increase in the flow rate of the starting CCC virtually does not affect selectivity;403 the pressure in the reactor becomes the main parameter for control- ling selectivity.Cu(HFA)(1,5-COD) manifests certain selectivity with respect to the substrate material: the formation and growth of copper films on the surfaces of such dielectrics as SiO2 and Si3N4 occurs with much greater difficulty than on metal substrates.334 How- ever, this effect is not observed at sufficiently high temperatures (>200 8C).330Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method A very high selectivity in the deposition of copper films in `metal ± SiO2' systems is achieved if Cu(HFA)(PMe3) and Cu(HFA)(PEt3) complexes are used.331 An unexpectedly high selectivity with respect to SiO2 was shown by copper(I) pyrazolyl borate complexes as the starting CCC upon deposition on Pt/SiO2, Au/SiO2, Al/SiO2 and W/SiO2 substrates.372 However, this effect was not observed on the Pd/SiO2 substrate relative to SiO2.2. Use of substrates made of polymeric materials in CVD processes The use of polymers as substrate materials in CVD processes has certain advantages. For example, it was found that the chem- isorption of precursor molecules and the adhesion of copper films to polyimide substrates are much weaker than in the case of SiO2; this fact enables highly selective deposition.404 CVD processes with the use of Teflon for the preparation of copper films have become especially popular.Teflon is used as a dielectric layer with unique properties, such as high resistivity and small dielectric constant. When combined with the low SER of copper films, this affords a very promising basis for the creation of fast microelectronic devices with short signal delay times.405 In this case, selectivity is based on two processes. Laser irradiation of the Teflon surface results in cleavage of the polymer chains with simultaneous formation of cross-links between the chains and perfluorocarbons with short chains in the near-surface layer. Both of these processes result in a strong decrease in the copper adhesion on the irradiated substrate areas.406 Alkaline etching of Teflon results in partial defluorination of the surface layer to a depth of 5300 nm and, hence, formation of a rough surface enriched with carbon atoms, which favours the efficient adhesion of copper films.405, 407 Hence, a combination of irradi- ated and etched zones on a Teflon surface can serve as a template owing to purely mechanical properties. The principles of `Teflon technologies' were formulated in a series of papers.408 ± 410 The process usually includes three stages: alkaline etching, creation of a pattern by KrF laser radiation, and high-selectivity CVD of the film (the copper film remains only on the etched areas).In the best experiments, the deposition rate of up to 1 mm min71 was reached and the SER of the films was *1.8 mO cm.410 VII. Conclusion Until 1995 accumulation of basic experimental data and the accompanying development of concepts on the mechanisms of CVD of thin copper films with the use of volatile CCC had occurred. Subsequently, the research in this field reached a qualitatively different level. Over the last two or three years, a certain decrease in the number of original papers was accompanied by an increase in the number of patents obtained both in the fields of the new volatile CCC, the ways for the synthesis of the already known compounds, and the specific details of CVD processes. This is also true for copper(I) and copper(II) derivatives as the starting CCC: the former are attractive due to the low temperatures of evaporation (sublimation) and decomposition on a substrate and the high film deposition rates; the latter, due to the stability during storage and the rather simple synthesis.For example, copper(II) 6-ethyl-2,2-dimethyloctane-3,5-dio- nate was patented as a new starting volatile CCC.411 A series of patents were obtained for many compounds as potential starting CCC for CVD processes in the series of complexes of Cu(HFA) with silylalkenes 412 ± 415 and silylalkynes.416 Analysis of the com- position and properties of the compounds patented shows that by varying the electronic effects of the substituents at the silicon atom and in the diketonate fragment, it is possible to reach an optimum ratio between the stability of the compound and its ability to decompose at low temperatures.1077 Let us note some technological patents with the use of (HFA)Cu(VTMS). A two-stage CVD process based on this CCC with addition of H(HFA) vapour 417 andH2O vapour 418 has been patented. It should probably be expected that the number of patents in this field will increase. Yet another trend worthy of note is the fast progress in the technologies for the selective deposition of films. The `Teflon technologies' are a significant but not the only achievement in this field. The first specialised review devoted to these technologies 419 generalised the results obtained by that time. So far, a number of important studies have been completed (see, e.g., Refs 420 and 421) that allow one to state that new microelectronic devices have been created using these technologies.In our opinion, one more promising field for the application of CVD technologies based on volatile copper(II) and copper(I) compounds is the preparation of copper films doped with other metals within one technological process. For example, is was found that the doping of copper films with zirconium atoms provides a good protection against electromigration with an insignificant increase in SER.422 The addition of a small amount of Pd(HFA)2 to (HFA)Cu(VTMS) vapour results in a copper film containing *0.5% Pd, which is completely stable against oxida- tion in air up to 300 8C (see Refs 423 and 424).Attempts at synthesis of similar films doped with Co (see Refs 425 and 426) and Sn (see Ref. 427) were reported. Further development of studies in this field may well result in new materials with unique properties. A conceptually similar approach is the development of CVD methods for the integrated deposition (using the same equipment) of a buffer layer onto a silicon substrate and a copper film atop this layer. The first study using this approach made it possible to obtain a copper film with very good characteristics (SER=1.8 mO cm) from (HFA)Cu(VTMS) vapour on silicon substrates with a TiN layer sputter-deposited prior to copper deposition.389 Perhaps, the above study reflects the most fundamental trend in the field, i.e., the transition from separate experimental studies on the CVD preparation of copper films with the use of concrete starting volatile copper compounds to the creation of integrated CVD technologies with full optimisation of all the conditions of the process up to the preparation of the final product.Spectacular examples of this approach include: the complete optimisation of aCVDprocess based on the industrially produced CCC of the CupraSelect grade (`Schumacher', Germany),428 the study where the technological approach to super-resolution resists was suggested,392 and the study suggesting the full-scale CVD production of copper films with SER<2.0 mO cm, deposition rate >200 nm min71, good adhesion and selectivity, and with ensured reproducibility of the properties of the batch films.390 It is obvious that the number of studies in this field will increase. In addition, the CVD-based preparation of copper films can be a technological component of the production of more complex systems.Recently, a technique for optimalCVDof copper films as an intermediate stage in the production of absorbing layers for solar batteries based on CuInSe2 was developed.429 Considerable help in the development of CVD technology for the deposition of copper films is provided by the database for quantum-chemical calculations that has appeared recently. For example, the deposition of copper films from (HFA)Cu(VTMS) vapour in a stream of argon was simulated numerically on a computer using the PHOENICS-CVD program and this data- base.430 The calculations made it possible to optimise both the conditions of the process and the equipment.In conclusion of the topic of the development and creation of volatile copper(I) and copper(II) derivatives, it is necessary to mention the attempts to go from searches for new specific technologically promising substances to the development of a general methodology for the synthesis of starting volatile CCC, based both on the analysis of the crystal and molecular structures of the already existing CCC and on quantum-chemical calcula-1078 tions of the thermodynamic parameters of crystal ± vapour phase transitions. Developments of suitable algorithms are under way,431, 432 and it may well be that this approach has great prospects.References 1. T M Makhviladze, M E Saryshev, Y V Zhitnikov Trudy FTIAN 13 98 (1997) 2. G Hass,MH Francombe, RWHofman (Eds) Physics of Thin Films. Advances in Research and Development Vol. 7 (New York: Academic Press, 1973) 3. S Vaidya, D B Fraser,W S Lindenberger J. Appl. Phys. 51 4475 (1980) 4. A A Milgram J. Vac. Sci. Technol. B 1 490 (1983) 5. R A Levy, L C Parrillo, L J Lecheler, R V Knoell J. Electrochem. Soc. 132 159 (1985) 6. J D McBrayer,R M Swanson, T W Sigmon J. Electrochem. Soc. 133 1242 (1986) 7. T Katagiri, E Kondoh, N Takeyasu, T Nakano, H Yamamoto, T Ohta Jpn. J. Appl. Phys., Pt. 2 32 L1078 (1993) 8. E Kondoh,Y Kawano,N Takeyasu,T Ohta J. Electrochem. Soc. 141 3494 (1994) 9.N Awaya, T Ohno, Y Arita Oyo Butsuri 64 554 (1995); Chem. Abstr. 123 271 556 (1995) 10. J R Childress, C L Chien Phys. Rev. B, Condens. Matter 43 8089 (1991) 11. C Dorner,M Haidl, H Hoffmann J. Appl. Phys. 74 5886 (1993) 12. A Blondel, J P Meier, B Doudin, J P Ansermet Appl. Phys. Lett. 65 3019 (1994) 13. S T Roshchenko, I G Shipkova, A A Koz'ma, V I Pinegin Fiz. Met. Metalloved. 85 162 (1998) a 14. C Christides, S Stavroyiannis, N Boukos, A Travlos, D Niarchos J. Appl. Phys. 83 3724 (1998) 15. G A Razuvaev, B G Gribov, G A Domrachev, B A Salamatin Metalloorganicheskie Soedineniya v Elektronike (Organometallic Compounds in Electronics) (Moscow: Nauka, 1972) 16. B G Gribov, G A Domrachev, B V Zhuk, B S Kaverin, B I Kozyr- kin, V V Mel'nikov, O N Suvorova Osazhdenie Plenok i Pokrytii Razlozheniem Metalloorganicheskikh Soedinenii (Deposition of Films and Coatings by Decomposition of Organometallic Compounds) (Ed.G A Razuvaev) (Moscow: Nauka, 1981) 17. G A Domrachev, O N Suvorova, V A Varyukhin, in Teoreti- cheskaya i Prikladnaya Khimiya Beta-Diketonatov Metallov (Theo- retical and Practical Chemistry of Metal Beta-Diketonates) (Moscow: Nauka, 1985) p. 228 18. G A Razuvaev (Ed.) Primenenie Metalloorganicheskikh Soedinenii dlya Polucheniya Neorganicheskikh Pokrytii (Application of Organo- metallic Compounds for Preparation of Inorganic Coatings) (Moscow: Nauka, 1986) 19. A Sherman Chemical Vapor Deposition for Microelectronics (Park Ridge: Noyes Publ., 1987) 20.I P Herman Chem. Rev. 89 1323 (1989) 21. G B Stringfellow Organometallic Vapor-Phase Epitaxy: Theory and Practice (Boston: Academic Press, 1989) 22. G L Criffin, A W Maverick, in The Chemistry of Metal CVD (Eds T T Kodas, MJ Hampden-Smith) (Weinheim: VCH, 1994) p. 175 23. S P Murarka, SW Hymes Crit. Rev. Solid State Mater. Sci. 20 87 (1995) 24. E R Aicha, F D Barlow (Eds) Thin Films Technology. Handbook (London: McGraw-Hill, 1996) 25. A V Gelatos, A Jain, R Marsh, C J Mogab MRS Bull. 19 49 (1994) 26. P Doppelt, T H Baum MRS Bull. 19 41 (1994) 27. G A Domrachev, O N Suvorova Usp. Khim. 49 167 (1980) [Russ. Chem. Rev. 49 810 (1980) 28. N Hayasaka Jpn. Technol. Rev., Sect. A 32 196 (1998) 29. J S Roberts Platinum Met, Rev. 36 12 (1992) 30.H J Boer Solid State Technol. 39 149 (1996) 31. S Serghini-Monim, L L Coatsworth, P R Norton, R J Puddephatt Rev. Sci. Instrum. 67 3672 (1996) 32. D I Fotiadis, A M Kremer, D R McKenna, K F Jensen J. Cryst. Growth 85 154 (1987) V N Vertoprakhov, S A Krupoder 33. C Houtman, D B Graves, K F Jensen J. Electrochem. Soc. 133 961 (1986) 34. K F Jensen, D I Fotiadis, T J Mountziaris J. Cryst. Growth 107 1 (1991) 35. US P. 4 839 145; Ref. Zh. Elektron. 12 G 32P (1990) 36. N-H Kim Electronics 67 9 (1994) 37. P M Frijlink J. Cryst. Growth 93 207 (1988) 38. J van Suchtelen, J EMHogenkamp,WG HM van Sark, L J Giling J. Cryst. Growth. 93 201 (1988) 39. E Woelk, H Beneking J. Cryst. Growth 93 216 (1988) 40. B Lecohier, J-M Philippoz, H van den Bergh J.Vac. Sci. Technol., B 10 262 (1992) 41. US P. 4 948 623; Ref. Zh. Elektron. 7 G 522P (1991) 42. US P. 5 019 423; Ref. Zh. Elektron. 10 G 29P (1992) 43. US P. 5 186 756; Ref. Zh. Elektron. 7 G 38P (1994) 44. US P. 5 316 796; Ref. Zh. Elektron. 8 G 568P (1995) 45. N Hayafuji, K Mizuguchi, S Jchi, T Murotani J. Cryst. Growth 77 281 (1986) 46. G S Tompa,MA McKee, C Beckham, P A Zawadzki, JMColabella, R D Reinert, K Capuder, R A Stall, P E Norris J. Cryst. Growth 93 220 (1988) 47. A I Gurary,A G Thompson,R A Stall,W J Kroll,N E Schumaker J. Electron. Mater. 24 1637 (1995) 48. J P Stagg, J Christer J. Cryst. Growth 120 98 (1992) 49. M Shiratani, Y Watanabe Oyo Butsuri 68 299 (1999); Chem. Abstr. 130 24 575 (1999) 50.R Kroger,M Eizenberg, D Cong, N Yoshida, L J Chen, S Ramaswami, D Carl J. Electrochem. Soc. 146 3248 (1999) 51. S Riedel Microelectron. Eng. 50 533 (2000) 52. Y Shacham-Diamand, S Lopatin Electrochim. Acta 44 3639 (1999) 53. A A Arendarenko, A N Pavlov, G A Charikov Elektron. Promst 2 58 (1992) 54. MKondo,A Kuramata, T Fuji, C Anayama, J Okazaki,HSekiguchi, T Tanahashi, S Yamazaki, in The 6th International Conference on Metalloorganic Vapour Phase Epitaxy (Abstracts of Reports), Cambridge, 1992 p. 123 55. K Matsumoto, K Itoh, T Tabuchi, R Tsunoda J. Cryst. Growth 77 151 (1986) 56. Y Monteil,R Favre,A Bekkaoui, P Raffin, J Bouix J. Cryst. Growth 93 270 (1988) 57. R Carpio, J Farkas, R Jairath Thin Solid Films 266 238 (1995) 58. M T Wang,M S Tsai, C Lin,W T Tseng, T C Chang, L J Chen, M C Cheng Thin Solid Films 308 ± 309 518 (1997) 59.J C Lee, P D Moskowitz Sol. Cells 28 209 (1990) 60. N Proust, C Fresnais Rev. Tech. Thomson-CSF 23 619 (1991) 61. Semicond. Int. 15 L102 (1992); Ref. Zh. Elektron. 6 G 54 (1993) 62. A E Kaloyeros, X Chen, T Stark, K Kumar, S-C Seo, G G Peterson, H L Frisch, B Arkles, J Sullivan J. Electrochem. Soc. 146 170 (1999) 63. R Padiyath, J Seth, S V Babu L J Matienzo J. Appl. Phys. 73 2326 (1993) 64. H Ono, T Nakano, T Ohta Appl. Phys. Lett. 64 1511 (1994) 65. Y Ezer, J Harkonen, V Sokolov, J Saarilahti, J Kaitila, P Kuivalainen Mater. Res. Bull. 33 1331 (1998) 66. P Madakson J. Appl. Phys. 70 1374 (1991) 67. J-C Chuang, S-L Tu, M-C Chen J.Electrochem. Soc. 145 4290 (1998) 68. MT Wang, Y C Lin, J Y Lee, C C Wang, M-C Chen J. Electrochem. Soc. 145 4206 (1998) 69. M T Wang, L J Chen,M-C Chen J. Electrochem. Soc. 146 728 (1999) 70. M T Wang, Y C Lin, C C Wang, M-C Chen J. Electrochem. Soc. 146 1583 (1999) 71. C S Choi,G A Ruggles, A S Shah, G C Xing, CMOsburn, JD Hunn J. Electrochem. Soc. 138 3062 (1991) 72. S-Q Wang, S Suthar, C Hoeflich, B J Burrow J. Appl. Phys. 73 2301 (1993) 73. H Krautz, C H Wenzel, K Bornkessel, G Blasek Phys. Status Solidi A 110 K77 (1988) 74. J O Olowolafe, C J Mogab, R B Gregory,M Kottke J. Appl. Phys. 72 4099 (1992) 75. E Kalowa, J S Chen, J S Reid, P J Pokela, M A Nicolet J. Appl. Phys. 70 1369 (1991) 76. T Takewaki, H Yamada, T Shibata, T Ohmi, T Nitta Mater.Chem. Phys. 41 182 (1995) 77. K Holloway, P M Fryer, C Cabral, J M E Harper, P J Bailey, K H Kelleher J. Appl. Phys. 71 5433 (1992)Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method 78. X Sun, E Kolawa, J S Chen, J S Reid,M A Nicolet Thin Solid Films 236 347 (1993) 79. P L Pai, C H Ting IEEE Electron Device Lett. 10 423 (1989) 80. L A Clevenger, N A Bojarczuk, K Holloway, J M E Harper, C Cabral, R G Schad, F Cardone, L Stolt J. Appl. Phys. 73 300 (1993) 81. MT Wang,YC Lin,MCChen J. Electrochem. Soc. 145 2538 (1998) 82. M Takeyama, A Noya, T Sase, A Ohta, K Sasaki J. Vac. Sci. Technol. B 14 674 (1996) 83. N Awaya, H Inokawa, E Yamamoto, Y Okazaki,M Miyake, Y Arita, T Kobayashi IEEE Trans.Electron Devices ED-43 1206 (1996) 84. D-S Yoon, H-K Baik, S-M Lee J. Appl. Phys. 83 1333 (1998) 85. D-S Yoon, H-K Baik, S-M Lee J. Appl. Phys. 83 8074 (1998) 86. C I M A Spee, J P A M Driessen, A D Kuypers J. Phys. IV 5 C5-719 (1995) 87. US P. 5 192 589; Ref. Zh. Elektron. 8 G 234P (1994) 88. J-Y Yun, M-Y Park, F-W Rhee J. Electrochem. Soc. 146 1804 (1999) 89. S C Sun,M H Tsai Thin Solid Films 253 440 (1994) 90. L H Dubois, B R Zegarski, G S Girolami J. Electrochem. Soc. 139 3600 (1992) 91. J-Y Yun, M-Y Park, S-W Rhee J. Electrochem. Soc. 145 2453 (1998) 92. I J Raaijmakers Thin Solid Films 247 85 (1994) 93. K Ishihara, K Yamazaki, H Hamada, K Kamisako, Y Tarui Jpn. J. Appl. Phys., Pt. 1 29 2103 (1990) 94.M H Tsai, S C Sun, H T Chiu, C E Tsai, S H Chuang Appl. Phys. Lett. 67 1128 (1995) 95. R M Charatan,M E Gross, D J Eaglesham J. Appl. Phys. 76 4377 (1994) 96. KIkeda,MMaeda, Y Arita Jpn. J. Appl. Phys., Pt. 1 32 3085 (1993) 97. T Wierzchon, J R Sobiecki J. Phys. IV 5 C5-699 (1995) 98. G C Jun, S L Cho, K B Kim, H K Shin, D H Kim Jpn. J. Appl. Phys., Pt. 2 37 L30 (1998) 99. B G Gribov, N N Travkin, G M Tabrina, V P Rumyantseva, B A Salamatin, B I Kozyrkin, A S Pashinkin Dokl. Akad. Nauk SSSR 187 330 (1969) b 100. G A Razuvaev, G A Domrachev, V D Zinov'ev Dokl. Akad. Nauk SSSR 223 617 (1975) b 101. Yu A Kaplin, L S Chernysheva, G N Bortnikov, E V Mitrofanov Khimiya Elementoorganicheskikh Soedinenii (Mezhvuzovskii Sb.) [The Chemistry of Organoelement Compounds (Intercollegiate Collection)] (Gor'kii: Gor'kii State University, 1987) p.57 102. G J M Dormans,G J B M Mukes, E G J Staring J. Cryst. Growth 114 364 (1991) 103. Yu A Kaplin, L S Chernysheva, T V Guseva, S F Zhil'tsov Izv. Vyssh. Uchebn. Zaved., Khim. Khim. Tekhnol. 23 1121 (1980) 104. T Maruyama Jpn. J. Appl. Phys., Pt. 2 36 L705 (1997) 105. A R Ivanova, G Nuesca, X Chen, C Goldberg, A E Kaloyeros, B Arkles, J J Sullivan J. Electrochem. Soc. 146 2139 (1999) 106. A Combes C.R. Hebd. Seances Acad. Sci. 105 868 (1887) 107. J P Fackler Prog. Inorg. Chem. 7 384 (1966) 108. J P Fackler, F A Cotton Inorg. Chem. 2 102 (1963) 109. R W Moshier, R E Sievers Gas Chromatography of Metal Chelates (New York: Pergamon Press, 1965) 110.R L van Hemert, L B Spendlove, R E Sievers J. Electrochem. Soc. 112 1123 (1965) 111. T T Kodas,MJ Hampden-Smith (Eds) The Chemistry of Metal CVD (Weinheim: VCH, 1994) 112. E A Mazurenko, A I Gerasimchuk, Zh N Bublik Ukr. Khim. Zh. 57 1011 (1991) 113. E A Mazurenko, A I Gerasimchuk Ukr. Khim. Zh. 59 526 (1993) 114. E I Tsyganova, L M Dyagileva Usp. Khim. 65 334 (1996) [Russ. Chem. Rev. 65 315 (1996)] 115. A E Kaloyeros, A Feng, J Garhart, K C Brooks, S K Ghosh, A N Saxena, F Luehers J. Electron. Mater. 19 271 (1990) 116. P Martensson, J-O Carlsson J. Electrochem. Soc. 145 2926 (1998) 117. J Goswami, L Raghunathan, A Devi, S A Shivashankar, S Chandrasekaran, J. Mater. Sci. Lett. 15 573 (1996) 118. R.Adams (Ed.).Organic Reactions Vol. 1 (New York: Wiley, 1942) 119. RL Belford,AEMartell,MCalvin J. Inorg. Nucl. Chem. 2 11 (1956) 120. J A Bertrand, R I Kaplan Inorg. Chem. 5 489 (1966) 121. J T Adams, C R Hauser J. Am. Chem. Soc. 66 1220 (1944) 122. E A Shugam Dokl. Akad. Nauk SSSR 81 853 (1951) b 1079 123. G N La Mar Acta Chem. Scand. 20 1359 (1966) 124. F A Cotton, C B Harris, J J Wise Inorg. Chem. 6 909 (1967) 125. F A Cotton, J J Wise Inorg. Chem. 6 915 (1967) 126. S V Volkov, G M Larin, V Ya Zub, E A Mazurenko Koord. Khim. 9 26 (1983) c 127. C Duval, R Freymann, J Lecomte Bull. Soc. Chim. Fr. 106 (1952) 128. H-J Goetze, K Bloss, H Molketin Z. Phys. Chem. 73 314 (1970) 129. A Kito, Y Miyake Bull. Gov. Ind. Res. Inst. Osaka 25 123 (1974); Chem.Abstr. 82 25 392 (1975) 130. D G Batyr, G N Marchenko, Kh Sh Khariton, N S Mitsul, G P Pogonin, S F Borisov, in Issledovaniya po Khimii Koordinat- sionnykh Soedinenii i Fiziko-Khimicheskim Metodam Analiza (Investigation on the Chemistry of Coordination Compounds and Physicochemical Methods of Analysis) (Ed. A V Ablov) (Kishinev: Academy of Sciences, Mold. SSR, 1969) p. 24 131. A Kito,MNakane, Y Miyake Bunseku Kagaku 27 1 (1978); Chem. Abstr. 89 35 891 (1978) 132. T Ozawa Thermochim. Acta 174 185 (1991) 133. M Becht, K-H Dahmen, F Atamny, A Baiker Fresenius J. Anal. Chem., Pt. 1 353 718 (1995) 134. N Awaya, Y Arita Jpn. J. Appl. Phys., Pt. 1 30 1813 (1991) 135. Z Yuan, G Meng Huaxue Shiji 18 76 (1996); Chem. Abstr. 125 47 592 (1996) 136.S Patnaik, T N Guru Row, L Raghunathan, A Devi, J Goswami, S A Shrivashankar, S Chandrasekaran, W T Robinson Acta Crystallogr., Sect. C 52 891 (1996) 137. J-Y Kim, P J Reucroft, D-K Park Thin Solid Films 289 184 (1996) 138. D A Grigor'ev, S A Kukushkin Zh. Tekhn. Fiz. 68 111 (1998) d 139. A E Turgambayeva, A F Bykov, I K Igumenov J. Phys. IV 5 C5-221 (1995) 140. A E Turgambayeva, A F Bykov, I K Igumenov Thermochim. Acta, 256 443 (1995) 141. E I Tsyganova, G A Mazurenko, V N Drobotenko, L M Dyagileva, Yu A Aleksandrov Zh. Obshch. Khim. 62 499 (1992) e 142. S V Volkov, L I Zheleznova, E A Mazurenko Koord. Khim. 9 1338 (1983) c 143. A F Bykov, P P Semyannikov, I K Igumenov J. Therm. Anal. 38 1463 (1992) 144. S Haukka Appl.Surf. Sci. 112 23 (1997) 145. K Hanaoka, H Ohnishi, H Harima, K Tachibana Physica C 190 145 (1991) 146. K Hanaoka, H Ohnishi, K Tachibana Jpn. J. Appl. Phys., Pt. 1 32 4774 (1993) 147. R Sekine, M Kawai, T Hikita, T Hanada Surf. Sci. 242 508 (1991) 148. R Sekine, M Kawai Appl. Phys. Lett. 56 1466 (1990) 149. T Hikita,R Sekine, T Hanada,M Kawai Surf. Sci. 262 L139 (1992) 150. H Zama, T Miyake, T Hattori, S Oda Jpn. J. Appl. Phys., Pt. 2 31 L588 (1992) 151. Y Pauleau, A F Fasasi Chem. Mater. 3 45 (1991) 152. T Maruyama, T Shirai J. Mater. Sci. 30 5551 (1995) 153. US P. 2 881 514; Chem. Abstr. 53 12 152d (1959) 154. D Touneau, R Pierrinard, H Dallaporta, W Marine J. Phys. IV 5 C5-629 (1995) 155. H Li, E T Eisenbrown, A E Kaloyeros J. Vac.Sci. Technol., B 10 1337 (1992) 156. T Gerfin, M Becht, K-H Dahmen J. Phys. IV 3 345 (1993) 157. A Dutta, J Goswami, S A Shrivashankar Bull. Mater. Sci. 18 901 (1995) 158. T Maruyama, Y Ikuta J. Mater. Sci. 28 5540 (1993) 159. R E Sievers, B W Ponder,M L Morris, R W Moshier Inorg. Chem. 2 693 (1963) 160. I A Baidina, P A Stabnikov, I K Igumenov, S A Borisov Koord. Khim. 10 1699 (1984) c 161. C Oehr, H Suhr Appl. Phys., A, Solids Surf. 45 151 (1988) 162. F A Swarts Bull. Cl. Sci. Acad. R. Belg. 12 692 (1926) 163. A L Henne,M S Newman, L L Quill, R A Stanifort J. Am. Chem. Soc. 69 1819 (1947) 164. J C Reid, M Calvin J. Am. Chem. Soc. 72 2948 (1950) 165. J P Fackler, F A Cotton, D W Barnum Inorg. Chem. 2 97 (1963) 166. M L Morris, R W Moshier, R E Sievers Inorg.Chem. 2 411 (1963) 167. L L Funck, T R Ortolano Inorg. Chem. 7 567 (1968) 168. W R Walker, N C Li J. Inorg. Nucl. Chem. 27 2255 (1965) 169. M R Kidd, R S Sager,W H Watson Inorg. Chem. 6 946 (1967)1080 170. Jpn. P. 08 245 639; Chem. Abstr. 125 343 578 (1996) 171. B Lecohier, B Calpini, J-M Philippoz, H van den Bergh J. Appl. Phys. 72 2022 (1992) 172. B G Schultz, E M Larsen J. Am. Chem. Soc. 71 3250 (1949) 173. R N Haszeldine,W K R Musgrave, F Smith, L M Turton J. Chem. Soc. 609 (1951) 174. J D Park, H A Brown, J R Lacher J. Am. Chem. Soc. 75 4753 (1953) 175. S V Volkov, V Ya Zub, E A Mazurenko, G M Larin Koord. Khim. 8 1333 (1982) c 176. E W Berg, F R Hartlage Anal. Chim. Acta 34 46 (1966) 177. E W Berg, N M Herrera Anal.Chim. Acta 60 117 (1972) 178. A Kito,MNakane, Y Miyake Bunseku Kagaku 27 7 (1978); Chem. Abstr. 89 52 744 (1978) 179. W R Wolf, R E Sievers, G H Brown Inorg. Chem. 11 1995 (1972) 180. D Temple, A Reisman J. Electrochem. Soc. 136 3525 (1989) 181. C R Moylan, T H Baum, C R Jones Appl. Phys., A 40 1 (1986) 182. A E Kaloyeros, A N Saxena, K Brooks, S Ghosh, E Einsenbrawn MRS Symp. Proc. 181 79 (1990) 183. B S Richards, S L Cook, D L Pinch, G W Andrews J. Phys. IV 5 C5-407 (1995) 184. P A Stabnikov, I A Baidina, K V Kradenov, S A Gromilov, I K Igumenov Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. (5) 21 (1989) 185. I A Baidina, P A Stabnikov Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. (3) 29 (1990) 186. I A Baidina, S A Gromilov Zh.Strukt. Khim. 32 96 (1991) f 187. S A Gromilov, I A Baidina Zh. Strukt. Khim. 38 814 (1997) f 188. D N Armitage, N I Dunhill, R H West, J O Williams J. Cryst. Growth 108 683 (1991) 189. M B Naik, W N Gill, R H Wentorf, R R Reeves Thin Solid Films 262 60 (1995) 190. B Lecohier, B Calpini, J-M Philippoz, H van den Bergh, D Laub, P A Buffat J. Electrochem. Soc. 140 789 (1993) 191. D-H Kim, R H Wentorf,W N Gill J. Electrochem. Soc. 140 3267 (1993) 192. N S Borgharkar, G L Griffin J. Electrochem. Soc. 145 347 (1998) 193. S K Lakshmanan,W N Gill J. Electrochem. Soc. 145 694 (1998) 194. W G Lai, Y Xie,W C Griffin J. Electrochem. Soc. 138 3499 (1991) 195. J H Horton, J G Shapter, T Cheng,W N Lennard, P R Norton Surf. Sci. 375 171 (1997) 196.C-H Jun, Y-T Kim, J-T Baek, H-J Hoo, D-W Park, B-J Choi, D-R Kim Hanguk Chaelyo Hakhoechi 5 657 (1995); Chem. Abstr. 124 216 395 (1995) 197. K Hanaoka, K Tachibana, H Ohnishi Thin Solid Films 262 209 (1995) 198. J Wang, R B Little, W G Lai, G L Griffin Thin Solid Films 262 31 (1995) 199. V M Donnelly, M E Gross J. Vac. Sci. Technol. A 11 66 (1993) 200. G M Nuesca, J A Kelber Thin Solid Films 262 224 (1995) 201. G S Girolami, P M Jeffries, L H Dubois J. Am. Chem. Soc. 115 1015 (1993) 202. J E Parmeter J. Phys. Chem. 97 11530 (1993) 203. J Farkas,M J Hampden-Smith, T T Kodas J. Phys. Chem. 98. 6753 (1994) 204. J C Kenvin,M G White, M B Mitchel Langmuir 7 1198 (1991) 205. N Awaya, Y Arita Jpn. J. Appl. Phys., Pt. 1 32 3915 (1993) 206.C M Chiang, T M Miller, L H Dubois J. Phys. Chem. 97 11781 (1993) 207. J-Y Kim, H A Marzouk, P J Reucroft, C C Eloi, J D Robertson J. Appl. Phys. 78 245 (1995) 208. B Zheng, C Goldberg, E T Eisenbrown, J Lin, A E Kaloyeros, P J Toscano, S P Murarka, J F Loan, J Sullivan Mater. Chem. Phys. 41 173 (1995) 209. J Pinkas, J C Huffman, D V Baxter, MH Chisholm, K G Caulton Chem. Mater. 7 1589 (1995) 210. J-Y Kim, Y-K Lee, H-S Park, J-W Park, D-K Park, J H Joo, W H Lee, Y-K Ko, P J Reucroft, B R Cho Thin Solid Films 330 190 (1998) 211. N S Borgharkar, G L Griffin, H Fan, A W Maverick J. Electrochem. Soc. 146 1041 (1999) 212. N S Borgharkar, G L Griffin, A James, A W Maverick Thin Solid Films 320 86 (1998) 213. A E Kaloyeros, B Zheng, I Lou, J Lou, J W Hellgeth Thin Solid Films 262 20 (1995) V N Vertoprakhov, S A Krupoder 214.D J Ehrlich, J Y Tsao J. Vac. Sci. Technol. B 1 969 (1983) 215. B Zheng, G Braechelmann, K Kujawski, I Lou, S Lane, A E Kaloyeros J. Electrochem. Soc. 142 3896 (1995) 216. W W Lee, P S Locke Thin Solid Films 262 39 (1995) 217. E T Eisenbrown, B Zheng, C P Dundon, P J Ding, A E Kaloyeros Appl. Phys. Lett. 60 3126 (1992) 218. D M Mao, K Z Jin, Q Z Qin J. Appl. Phys. 71 6111 (1992) 219. US P. 3 356 527; Chem. Abstr. 68 P32 703 (1968) 220. D-H Kim, R H Wentorf,W N Gill J. Electrochem. Soc. 140 3273 (1993) 221. D-H Kim, R H Wentorf,W N Gill MRS Symp. Proc. 260 107 (1992) 222. D-H Kim, R H Wentorf,W N Gill J. Appl. Phys. 74 5164 (1993) 223.D-H Kim, R H Wentorf,W N Gill J. Vac. Sci. Technol. A 12 153 (1994) 224. N Awaya, Y Arita J. Electron. Mater. 21 959 (1992) 225. B Lecohier, J-M Philippoz, B Calpini, T Stumm, H van den Bergh Appl. Phys. Lett. 60 3114 (1992) 226. B Lecohier, J-M Philippoz, B Calpini, T H Stumm, H van den Bergh J. Phys. IV 1 C2-279 (1991) 227. L I Zheleznova, S V Volkov, E A Mazurenko, N V Gerasimenko Ukr. Khim. Zh. 55 122 (1989) 228. US P. 5 376 409; Chem. Abstr. 123 45 716 (1995) 229. N Awaya, K Ohno, Y Arita J. Electrochem. Soc. 142 3173 (1995) 230. H D Hafney, R L Lintvedt J. Am. Chem. Soc. 93 1623 (1971) 231. X Yang, Ch Kutal J. Am. Chem. Soc. 105 6038 (1983) 232. E E Marinero, C R Jones J. Chem. Phys. 82 1608 (1985) 233. D S Talaga, J I Zink Inorg.Chem. 35 5050 (1996) 234. F A Hoyle, R J Wilson, T H Baum J. Vac. Sci. Technol. A 4 2452 (1986) 235. R J Wilson, F A Hoyle Phys. Rev. Lett. 55 2184 (1985) 236. F A Hoyle, C R Jones, T Baum, C Pico C A Kovac Appl. Phys. Lett. 46 204 (1985) 237. G Auvert Appl. Surf. Sci. 43 47 (1989) 238. F A Hoyle Appl. Phys. A 41 315 (1986) 239. N S Gluck, Z Ying, C E Bartosch,W Ho J. Chem. Phys. 86 4957 (1987) 240. B Markwalder,M Vidmer, D Braichotte, H van den Bergh J. Appl. Phys. 65 2470 (1989) 241. Y D Chen, A Reisman, I Turlik, D Temple J. Electrochem. Soc. 142 3911 (1995) 242. C R Jones, F A Hoyle, C A Kovac, T H Baum Appl. Phys. Lett. 46 97 (1985) 243. M Shiratani, H Kawasaki, T Fukuzawa, T Kinoshita, Y Watanabe J. Phys. D, Appl. Phys.29 2754 (1996) 244. B Zheng, E T Eisenbrown, J Liu, A E Kaloyeros Appl. Phys. Lett. 61 2175 (1992) 245. S I Troyanov, O Yu Gorbenko, A A Bosak Polyhedron 16 1595 (1997) 246. O Yu Gorbenko, S I Troyanov, A Meetsma, A A Bosak Polyhedron 16 1999 (1997) 247. J Pinkas, J C Huffman, J C Bollinger, W E Streib, D V Baxter, M H Chisholm, K G Caulton Inorg. Chem. 36 2930 (1997) 248. E A Mazurenko, O A Posil'skii, V Ya Zub, A I Gerasimchuk Ukr. Khim. Zh. 58 16 (1992) 249. S M Fine, J A T Norman, B A Muratore, R L Iampietro MRS Symp. Proc. 204 415 (1991) 250. S Hwang, H Choi, I Shim Chem. Mater. 8 981 (1996) 251. G J Choi, Y S Cho Mater. Lett. 32 185 (1997) 252. G Ramaswami, A K Raychaudhuri, J Goswami, S A Shivashankar J. Appl. Phys. 82 3797 (1997) 253.J Goswami, S A Shivashankar, G Anathakrishna Thin Solid Films 305 52 (1997) 254. A Devi, J Goswami, R Lakshmi, S A Shivashankar, S Chandrasekaran J. Mater. Res. 13 687 (1998) 255. A Devi, S A Shrivashankar J. Mater. Sci. Lett. 17 367 (1998) 256. S T Hwang, I Shim, K O Lee, K S Kim, J H Kim, G J Choi, Y S Cho, H S Choi J. Mater. Res. 11 1051 (1996) 257. R Hoffman, B Lecohier, S Goldoni, H van den Bergh Appl. Surf. Sci. 43 54 (1989) 258. B Lecohier, H van den Bergh Appl. Surf. Sci. 43 61 (1989) 259. M-J Mouche, J-L Mermet, C Mathon, R Cimard Adv. Sci. Technol. 5 231 (1995)Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method 260. M-J Mouche, J-L Mermet,M Romand,M Charbonnier Thin Solid Films 262 1 (1995) 261.S A Krupoder, V S Danilovich,A O Miller,G G Furin J. Fluorine Chem. 73 13 (1995) 262. J J Jarvis, R Pearce,M F Lappert J. Chem. Soc., Dalton Trans. 999 (1977) 263. A Miyashita, A Yamamoto Bull. Chem. Soc. Jpn. 50 1102 (1977) 264. J V Singh, B P Baranval, R C Mehrotra Z. Anorg. Allg. Chem. 477 235 (1981) 265. M E Gross J. Electrochem. Soc. 138 2422 (1991) 266. H S Horovitz, S J McLain, A W Sleight, J D Druliner, P L Gai, M J van Kavelaar, J L Wagner Science 243 66 (1989) 267. S C Goel, K S Kramer, M Y Chiang, W E Buhro Polyhedron 9 611 (1990) 268. PMJeffries, SRWilson,G S Girolami Inorg. Chem. 31 4503 (1992) 269. M J Hampden-Smith, T T Kodas Polyhedron 14 699 (1995) 270. G D Kuznetsov, A A Badad-Zakhryapin, F Giod Prot.Met. 8 565 (1972) 271. J Gillardeau, R Hasson, J Oudar J. Cryst. Growth 2 149 (1968) 272. C Lampe-Onnerud, J. Cryst. Growth 121 223 (1992) 273. R Madar J. Microelectron. Eng. 19 571 (1992) 274. A Moller, R Kall, V Till, G Wortberg, G Adomeit J. Cryst. Growth 174 837 (1994) 275. R Nast, R Mohr, C Schultze Chem. Ber. 96 2127 (1963) 276. H Lang, K Kohler, M Buchner Chem. Ber. 128 525 (1995) 277. M Charton Prog. Phys. Org. Chem. 13 119 (1981) 278. K M Chi, H K Shin,M J Hampden-Smith, T T Kodas, E N Duesler Inorg. Chem. 30 4293 (1991) 279. H K Shin, K M Chi, J Farkas, M J Hampden-Smith, T T Kodas, E N Duesler Inorg. Chem. 31 424 (1992) 280. KMChi,HKShin,MJ Hampden-Smith, T T Kodas, E N Duesler Polyhedron 10 2293 (1991) 281.THBaum, CE Larson,G May J. Organomet. Chem. 425 189 (1992) 282. G Doyle, K A Eriksen, D van Engen Organometallics 4 830 (1985) 283. US P. 5 085 731; Chem. Abstr. 116 266 288 (1992) 284. S-W Kang, M-Y Park, S-W Rhee Electrochem. Solid-State Lett. 2 22 (1999) 285. J A T Norman, B A Muratore, P N Dyer, D A Roberts, A K Hochberg J. Phys. IV 1 C2-273 (1991) 286. M Widmer, H van den Bergh J. Appl. Phys. 77 5464 (1995) 287. E S Choi, S K Park, H K Shin, H H Lee Appl. Phys. Lett. 68 1017 (1996) 288. M B Naik, S K Lakshmanan, R H Wentorf, R R Reeves, W N Gill J. Cryst. Growth 193 133 (1998) 289. A V Gelatos, R Marsh,M Kottke, C J Mogab Appl. Phys. Lett. 63 2842 (1993) 290. W-J Lee, J S Min, S-K Rha, S-S Chun, C O Park J.Mater. Sci., Mater. Electron. 7 111 (1996) 291. J-C Chiou, Y-J Chen, M-C Chen J. Electron. Mater. 23 383 (1994) 292. K V Guinn, V M Donnelly, M E Gross, F A Baiocchi, I Petrov, J E Greene Surf. Sci. 295 219 (1993) 293. K V Guinn, V M Donnelly, M E Gross, F A Baiocchi, I Petrov, J E Greene MRS Symp. Proc. 282 379 (1993) 294. C-H Jun, Y-T Kim, J-T Back, H-J Yoo, J Hyung, D-R Kim J. Vac. Sci. Technol. A 14 3214 (1996) 295. D H Kim, Y J Lee, C O Park, J W Park, J J Kim Chem. Eng. Commun. 152 ± 153 307 (1996) 296. S-K Rha, W-J Lee, S-Y Lee, D-W Kim, C-O Park J. Mater. Sci., Mater. Electron. 8 217 (1997) 297. J Rober, C Kaufmann, T Gessner Appl. Surf. Sci. 91 134 (1995) 298. D E Boss, S Sambasinan, D Torok, S Meren Proc. Electrochem. Soc.96 564 (1996) 299. P-J Lin, M-C Chen J. Electron. Mater. 28 567 (1999) 300. G Braeckelmann, D Manger, A Burke, G G Peterson, A E Kaloyeros, C Reidsema, T R Omstead, J F Loan, J J Sullivan J. Vac. Sci. Technol. B 14 1828 (1996) 301. A Jain, K-M Chi, T T Kodas,M J Hampden-Smith J. Electrochem. Soc. 140 1434 (1993) 302. S-Y Lee, S-K Rha, W-J Lee, D-W Kim, J-C Huang, C-O Park Jpn. J. Appl. Phys., Pt. 1 36 5249 (1997) 303. W-J Lee, S-K Rha, S-Y Lee, D-W Kim, C-O Park Thin Solid Films 305 254 (1997) 304. N Awaya, Y Arita Thin Solid Films 262 12 (1995) 305. T H Stumm, H van den Bergh Mater. Sci. Eng. B 23 48 (1994) 1081 306. A Jain, T T Kodas, T S Corbitt, M J Hampden-Smith Chem. Mater. 8 1119 (1996) 307. J-L Mermet, M-J Mouche, F Pires, E Richard, J Torres, J Palleau, F Brand J.Phys. IV 5 C5-517 (1995) 308. M-J Mouche, J-L Mermet, F Pires, E Richard, J Torres, J Palleau, F Brand Appl. Surf. Sci. 91 129 (1995) 309. J E Parmeter, G A Petersen, P M Smith, C A Apblett, J S Reid, J A T Norman, A K Hochberg, D A Roberts, T R Omstead J. Vac. Sci. Technol. B 13 130 (1995) 310. GAPetersen, J E Parmeter, PMSmith, CA Arblett,MF Gonsales, P M Smith, T R Omstead J. Electrochem. Soc. 142 939 (1995) 311. JAT Norman,DARoberts,AKHochberg, P Smith,GAPetersen, J E Parmeter, C A Arblett Thin Solid Films 262 46 (1995) 312. W-J Lee, S-K Rha, S-Y Lee, D-W Kim, S-S Chun, C-O Park Mater. Res. Soc. Symp. Proc. 427 207 (1996) 313. W-J Lee, S-K Rha, S-Y Lee, C-O Park J. Electrochem. Soc. 14 683 (1997) 314.Y Ochiai, J Fujita, S Matsui J. Vac. Sci. Technol. B 14 3887 (1996) 315. S Mezhenny, I Lyubinetsky,W J Choyke, J T Yates J. Appl. Phys. 85 3368 (1999) 316. J Han, K F Jensen J. Appl. Phys. 75 2240 (1994) 317. H Moilanen, J Remes, S Leppavuori Phys. Scr. T69 237 (1997) 318. D Bollmann, R Merkel, A Klumpp Microelectron. Eng. 37/38 105 (1997) 319. A Burke, G Braechelmann, D Manger, E Eisenbrann, A E Kaloyeros, J P McVittie, J Han, D Bang, J F Loan, J J Sullivan J. Appl. Phys. 82 4651 (1997) 320. D Popovici, K Piyakis, E Sacher,M Meunier MRS Symp. Proc. 397 643 (1996) 321. M-Y Park, J-H Son, S-W Rhee Electrochem. Solid-State Lett. 1 32 (1998) 322. J-H Son, M-Y Park, S-W Rhee Thin Solid Films 335 229 (1998) 323. M-Y Park, J-H Son, S-W Kang, S-W Rhee J.Mater. Res. 14 975 (1999) 324. Y Senzaki,M Kobayashi, L J Charneski, T Nguyen J. Electrochem. Soc. 144 L154 (1997) 325. Y Senzaki J. Electrochem. Soc. 145 362 (1998) 326. M-Y Park, S-W Rhee Hanguk Chaelyo Hakhoechi 8 345 (1998); Chem. Abstr. 129 61 106 (1998) 327. F A Cotton, T J Marks J. Am. Chem. Soc. 92 5114 (1970) 328. T Kruck, C Terfloth Chem. Ber. 126 1109 (1993) 329. D Blessman, A Graefe, R Heinen, F Jansen, T Kruck, C Terfloth Mater. Sci. Eng. B 17 104 (1993) 330. S K Reynolds, C J Smart, E F Baran, T H Baum, C E Larson, P J Brock Appl. Phys. Lett. 59 2332 (1991) 331. H K Shin, K M Chi,M J Hampden-Smith, T T Kodas, M F Paffett, J D Farr Chem. Mater. 4 788 (1992) 332. A Jain, K M Chi, J Farkas, M J Hampden-Smith, T T Kodas, M F Paffett, J D Farr J.Mater. Res. 7 261 (1992) 333. R Kumar, F R Fronezek, A W Maverick,W G Lai, G L Griffin Chem. Mater. 4 577 (1992) 334. S L Cohen,M Liehr, S Kasi J. Vac. Sci. Technol. A 10 863 (1992) 335. C Roger, T S Corbitt, M J Hampden-Smith, T T Kodas Appl. Phys. Lett. 65 1021 (1994) 336. R Kumar, A W Maverick Chem. Mater. 5 251 (1993) 337. K-M Chi, H-C Hou, P-T Hung, S-M Peng, G-H Lee Organometallics 14 2641 (1995) 338. P R Story, S R Fahrenholtz, in Organometallic Syntheses. Collect. Vol. 5 (Eds J J Eisch, R B King) (New York: Wiley, 1973) p. 151 339. A Jain, K M Chi, T T Kodas,M J Hampden-Smith, J D Farr, M F Paffett Chem. Mater. 3 995 (1991) 340. T H Baum, C E Larson J. Electrochem. Soc.140 154 (1993) 341. T H Baum, C E Larson Chem. Mater. 4 365 (1992) 342. P Doppelt, T H Baum J. Organomet. Chem. 517 53 (1996) 343. US P. 5 187 300; Chem. Abstr. 119 83 398 (1993) 344. D M Hoffmann, R Hoffmann, C R Fizel J. Am. Chem. Soc. 104 3858 (1992) 345. D L Reger, M Huff Organometallics 11 69 (1992) 346. A Jain, J Farkas, T T Kodas, K M Chi,M J Hampden-Smith Appl. Phys. Lett. 61 2662 (1992) 347. G Schmidt, U Behrens J. Organomet. Chem. 503 101 (1995) 348. R Nast, W-H Lepel Chem. Ber. 102 3224 (1969) 349. H K Shin,M J Hampden-Smith, T T Kodas, E N Duesler, J D Farr,M F Paffett MRS Symp. Proc. 187 193 (1990)1082 350. H K Shin, K M Chi,M J Hampden-Smith, T T Kodas, J D Farr, M F Paffett MRS Symp. Proc. 204 421 (1991) 351. H K Shin,M J Hampden-Smith, E N Duesler, T T Kodas Polyhedron 10 645 (1991) 352. H K Shin,M J Hampden-Smith, E N Duesler, T T Kodas Can. J. Chem. 70 2954 (1992) 353. X Li, G M Bancroft, R J Puddephatt, Z Yuan, K H Tan Inorg. Chem. 35 5040 (1996) 354. H K Shin, K M Chi,M J Hampden-Smith, T T Kodas, J D Farr, M F Paffett Adv. Mater. 3 246 (1991) 355. H K Shin,M J Hampden-Smith, T T Kodas, A L Rheingold J. Chem. Soc., Chem. Commun. 217 (1992) 356. A Jain, H K Shin, K M Chi, M J Hampden-Smith, T T Kodas, J Farkas, M F Paffett Proc. SPIE - Int. Soc. Opt. Eng. 1596 23 (1991) 357. H S Choi, S T Hwang Chem. Mater. 10 2326 (1998) 358. P M Jeffries, G S Girolami Chem. Mater. 1 8 (1989) 359. L H Dubois, G S Girolami Chem. Mater. 4 1169 (1992) 360. T H Lemmen, G V Goeden, J C Huffman, R L Geerts, K G Caulton Inorg. Chem. 29 3680 (1990) 361. M J Hampden-Smith, T T Kodas,M Paffett, J D Farr, H K Shin Chem. Mater. 2 636 (1990) 362. A M James, R K Laxman, F R Fronezek, A W Maverick Inorg. Chem. 37 3785 (1991) 363. D B Beach, W F Kane, F K Legoues, C J Knors Adv. Metal. Microelectron. 81 73 (1990) 364. X Zilling, M J Strouse, D K Shuh, C B Knobler, H D Kaesh, R F Hicks,W R Williams J. Am. Chem. Soc. 111 8779 (1989) 365. K Hara, T Kojima,H Kukimoto Jpn. J. Appl. Phys., Pt. 2 26 L1107 (1987) 366. US P. 10 287 688; Chem. Abstr. 129 337 959 (1998) 367. D B Beach, F K Legoues, C K Hu Chem. Mater. 2 216 (1990) 368. S Chichibu, N Yoshida, H Higuchi, S Matsumoto Jpn. J. Appl. Phys., Pt. 2 31 L1778 (1992) 369. C G Dupuy, D B Beach, J E Hurst, J M Jasinski Chem. Mater. 1 16 (1989) 370. T Saegusa, Y Ito, S Tomita J. Am. Chem. Soc. 93 5656 (1971) 371. A Graefe, R Heinen, F Klein, T Kruck, M Sherer, M Schober Appl. Surf. Sci. 91 187 (1995) 372. E-C Plappert, T Stumm, H van den Bergh, R Hauert, K-H Dahmen Chem. Vap. Deposition 3 37 (1997) 373. C M Chiang, L H Dubois MRS Symp. Proc. 282 341 (1993) 374. L H Dubois, B R Zegarski J. Electrochem. Soc. 139 3295 (1992) 375. S L Cohen,M S Liehr, S Casi Appl. Phys. Lett. 60 50 (1992) 376. S L Cohen,M S Liehr, S Casi Appl. Phys. Lett. 60 1585 (1992) 377. Y K Chae, Y Shimogaki, H Komiyama J. Electrochem. Soc. 145 4226 (1998) 378. L-S Hong, M-Z Lin Jpn. J. Appl. Phys., Pt. 2 36 L711 (1997) 379. H-Y Yoen, Y-B Park, S-W Rhee J. Mater. Electron. 8 189 (1997) 380. S S Yoon, S W Kang, S-S Chun MRS Symp. Proc. 391 315 (1995) 381. T P Chiang,H H Sawin, C V Thompson J. Vac. Sci. Technol. A 15 3104 (1997) 382. J Farkas,M J Hampden-Smith, T T Kodas J. Phys. Chem. 98 6763 (1994) 383. A F Burnett, J M Cech J. Vac. Sci. Technol. A 11 2970 (1993) 384. B C Johnson J. Appl. Phys. 67 3018 (1990) 385. C-H Jun, Y-T Kim, J-T Back, D-R Kim Hanguk Chaelyo Hakhoechi 6 1233 (1996); Chem. Abstr. 126 257 153 (1996) 386. J B Webb, D Northcott, I Emesh Thin Solid Films 270 483 (1995) 387. J-C Chiou, J-C Juang, M-C Chen J. Electrochem. Soc. 142 177 (1995) 388. A Jain, T T Kodas, M J Hampden-Smith Proc. SPIE - Int. Soc. Opt. Eng. 2335 52 (1994) 389. K Numajiri, T Goya, R Tobe, O Okada, N Hosokawa, C Mu, N Cox, C Scott, J Yu Appl. Surf. Sci. 100/101 541 (1996) 390. T Nguyen, L J Charneski, S T Hsu J. Electrochem. Soc. 144 2829 (1997) 391. A E Kaloyeros, M A Fury MRS Bull. 18 22 (1993) 392. P Doppelt Coord. Chem. Rev., Pt. 2 178 ± 180 1785 (1998) 393. R Steger, R Masel Thin Solid Films 342 221 (1999) 394. K Ohno,M Sato, Y Arita Jpn. J. Appl. Phys., Pt. 2 28 L1070 (1989) 395. J Farkas, K MChi,MJ Hampden-Smith, T T Kodas, L H Dubois Mater. Sci. Eng. B 17 93 (1993) V N Vertoprakhov, S A Krupoder 396. Y-S Kim, D-J Kim, S-K Kwak, E-K Kim, S-K Min, D-G Jung J. Appl. Phys., Pt. 2 37 L462 (1998) 397. J-M Park, S Kim, D-J Choi, D-H Ko Yoop Hakhoechi 35 827 (1998); Chem. Abstr. 130 99 269 (1998) 398. J A T Norman, B A Muratore, P N Dyer, D A Roberts, A K Hochberg, L H Dubois Mater. Sci. Eng. B 17 87 (1993) 399. A Jain, T T Kodas, R Jairath,M J Hampden-Smith J. Vac. Sci. Technol. B 11 2107 (1993) 400. J A T Norman, D A Roberts, A K Hochberg MRS Symp. Proc. 282 347 (1993) 401. J A T Norman, D A Roberts, A K Hochberg Proc. Electrochem. Soc. 93-2 221 (1993) 402. K Hanaoka, H Ohnishi, K Tachibana J. Appl. Phys., Pt. 1 34 2430 (1995) 403. G Friese, A Abdul-Hak, B Schwierzi, U Hohne Microelectron. Eng. 37/38 157 (1997) 404. Y D Chen, A Reisman, I Turlik, D Tample J. Electrochem. Soc. 142 3903 (1995) 405. D Popovici,G Cheremuzkin,M Meunier, E Sacher Appl. Surf. Sci. 126 198 (1998) 406. R R Rye, R J Martinez J. Appl. Polym. Sci. 37 2529 (1989) 407. R R Rye, G W Arnold Langmuir 5 1331 (1989) 408. R R Rye, K M Chi, M J Hampden-Smith, T T Kodas J. Electrochem. Soc. 139 L60 (1992) 409. W L Perry,K M Chi, T T Kodas,M J Hampden-Smith,R R Rye Appl. Surf. Sci. 69 94 (1993) 410. W L Perry, A Jain, T T Kodas, M J Hampden-Smith Thin Solid Films 262 7 (1995) 411. Jpn. P. 09 53 177; Chem. Abstr. 126 271 165 (1997) 412. Jpn. P. 07 252 266; Chem. Abstr. 124 161 140 (1996) 413. Jpn. P. 09 151 189; Chem. Abstr. 127 95 403 (1997) 414. US P. 5 767 301; Chem. Abstr. 129 75 031 (1998) 415. Eur. P. 855 399; Chem. Abstr. 129 155 688 (1998) 416. Jpn. P. 07 215 982; Chem. Abstr. 124 133 010 (1996) 417. Jpn. P. 08 139 030; Chem. Abstr. 125 102 693 (1996) 418. US P. 5 744 192; Chem. Abstr. 128 283 647 (1998) 419. M J Hampden-Smith, T T Kodas, R R Rye Adv. Mater. 4 524 (1992) 420. R R Rye, A J Ricco,WL Perry, MJ Hampden-Smith, T T Kodas Plast. Eng. 43 15 (1998) 421. T Y Chen, C Combellas, P Doppelt, F Kanoufi, A Thiebault Chem. Vap. Deposition 5 185 (1999) 422. N Awaya, T Kobayashi Jpn. J. Appl. Phys., Pt. 1 37 1156 (1998) 423. P Atanasova, V Bhaskaran, T T Kodas,M J Hampden-Smith MRS Symp. Proc. 428 25 (1996) 424. V Bhaskaran, P Atanasova,M J Hampden-Smith, T T Kodas Chem. Mater. 9 2822 (1997) 425. S Gu, X Yao,M J Hampden-Smith, T T Kodas Chem. Mater. 10 2145 (1998) 426. S Gu, P Atanasova, M J Hampden-Smith, T T Kodas Thin Solid Films 340 45 (1999) 427. P Doppelt, T H Baum Thin Solid Films 270 480 (1995) 428. C Marsadal, E Richard, J-L Mermet, J Torres, J Palleau, B Alaux, M Bakli Microelectron. Eng. 33 3 (1997) 429. S K Ramaiah, R D Pilkington, A E Hill, R D Tomlinson J. Mater. Sci. Mater. Electron. 10 51 (1999) 430. H Wolf, J Rober, S Riedel, R Streiter, T Gessner Microelectron. Eng. 45 15 (1999) 431. W S Rees Proc. Electrochem. Soc. 96-5 225 (1996) 432. I K Igumenov, F A Kuznetsov, V R Belosludov Proc. Electrochem. Soc. 96-5 273 (1996) a�Phys. Met. Metallogr. (Engl. Transl.) b�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) c�Russ. J. Coord. Chem. (Engl. Transl.) d�Techn. Phys. (Engl. Transl.) e�Russ. J. Gen. Chem. (Engl. Transl.) f�Russ. J. Struct. Chem. (En

ISSN:0036-021X    

出版商:RSC    

年代:2000

数据来源: RSC