摘要:

Tin Oxide Surfaces Part 3.-Infrared Study of the Adsorption of Some Small Organic Molecules on Tin(1v) Oxide BY EDWARD W. THORNTON AND PHILIP G. HARRISON* Department of Chemistry, University of Nottingham, University Park, Nottingham NG7 2RD Received 17th February, 1975 The tin@) oxide surface is strongly oxidising towards smaI1, reactive organic molecules. Methanol is chemisorbed to give methoxy groups but these are readily oxidised to a surface formate at temperatures >320 K. Acetone and acetaldehyde are adsorbed predominantly as acetates but some evidence is found for an en01 form which may be responsible for the rapid deuterium exchange between hexa- deuteroacetone and surface hydroxyl groups. The spectra of the formate and acetate structures are confirmed by adsorption of formic and acetic acids, where evidence is also found for undis- sociated acid molecules bonded to the surface.As expected the acetate is thermally more stable than the formate under vacuum. Alkyltin alkoxides react with a number of multiply bonded reagents to give adducts which are useful in organic synthesis, and often function as catalysts for the addition of alcohols to a number of unsaturated acceptor molecules.1 It was therefore thought profitable to exchange the surface hydroxyl groups on tin@) oxide for alkoxide groups, to generate an alkoxylated surface which might exhibit similar catalytic quali- ties. The adsorption characteristics of normal alcohols on several oxides have been investigated 2-5 and it has been found that several species including adsorbed alcohol, alkoxide groups and carboxylates can occur.The latter is usually formed from the alkoxide on heating. The adsorption of acids can lead to the formation of surface carboxylates ti or surface ester groupings.’ EXPERIMENTAL The apparatus, oxide sample, deuterium oxide, oxygen and carbon dioxide have already been described. MATERIALS Methanol (H.B.L.) was distilled from phosphorus pentoxide in an argon atmosphere Dimethyl carbonate (B.D.H.) was redistilled (362-363 K). Acetaldehyde was redistilled in an argon atmosphere (293.8 K). Acetone (B.D.H., AristaR grade) and hexadeuteroacetone (Fluorochem Ltd., isotopic purity 399.6 %) were used as supplied. Acetic acid (B.D.H. AnalaR grade, 399.7 %) and formic acid (East Anglia Chemicals, 98-100 %) were dried over anhydrous copper sulphate.All the materials were stored in sample bulbs attached to the grease free section of the vacuum line and were degassed in a series of freeze-thaw cycles prior to use. In addition the acetaldehyde, acetone and hexadeuteroacetone were cooled to 195 K before use. PROCEDURE Self-supporting discs of tin(1v) oxide ( N 1 1.5 mg cm-2) were evacuated (- 1 0-4 N m-2) for various times at room temperature (r.t.) before the appropriate vapour was allowed to 2468E . W . THORNTON A N D P . G . HARRISON 2469 expand into the cell compartment. During the adsorption process the disc was maintained at r.t. or at the ambient temperature of the spectrometer beam (a.b.t.) which was constant at about 320 K. The samples were then evacuated under various conditions. During the recording of spectra the reference beam was attenuated arbitrarily and the gain setting was varied to compromise between noise and pen response. The numbers on the ordinate axes of the diagrams are the transmittance at the highest wavenumber shown and except for displacing the ordinate axes, all spectra were traced from the originals.In fig. 2 and 4 the '' dot-dash " lines represent re-attenuation of the reference beam. RESULTS INTERPRETATION OF SPECTRAL DATA METHANOL A N D DIMETHYLCARBONATE ADSORPTION The spectra obtained from the chemisorption of both methanol and dimethyl carbonate on to tin(1v) oxide at G373 K are composed of the superimposed spectra of surface methoxy and formate groups, but at temperatures >373 only bands characteristic of the latter species are formed.The surface methoxy groups are characterised by bands at 1465-1470 cm-l [6,,(CH,)] and 1340-1360 cm-', which is due to 6,,,,(CH3) and S,ym(COO) of the surface formate group which overlap. The latter 58 7 2 t - I I I I I I 5 0 0 1403 1 2 0 0 wavenumber /cm- FIG. 1.-Infrared spectra of dimethyl carbonate chemisorbed on tin(xv) oxide. (1) Tin(1v) oxide disc evacuated (2 h, 293 K), reacted with dimethyl carbonate (2.0 kN m-2, 2 h, 293 K) and re- evacuated (10 min, 293 K). (2) During subsequent exposure to carbon dioxide (8.0 kN m-2, 320 K). (3) Disc reacted with dimethyl carbonate (2.0 kN m-2, 2 h, 320 K) and evacuated (15 min, 320 K). (4) During subsequent exposure to carbon dioxide (1.33 k N m-2, 320 K) and (5) after evacuation at 459 K for 1 h.2470 TIN OXIDE SURFACES band occurs at 1340 cm-l on samples containing no methoxy groups.No evidence is found for CH30H molecules bonded to the surface after evacuation at 320 K or for dimethyl carbonate after evacuation at 320 K. The bands observed at 1550, 1380 and 1340-1360 cm-' are readily assigned as v,,(COO), Si.,,(CH3) and v,,,(COO), respectively, of a surface formate group. More surface methoxy and formate groups were produced by treatment with dimethyl carbonate than with methanol, and more formate formed at 320 K [fig. 1(3)] than at 293 K [fig. 1(1)]. Above 320 K methoxy groups were progressively oxidised to formate, all methoxy groups being lost at 463 K ; the formate ion decomposed near 500 K, close to the decomposition temperature of ti@) formate. t - I I I I I I I 8 0 0 1600 1400 1200 wavenurnber Icm-I FIG.2.4nfrared spectra of organic carbonyl compounds chemisorbed on tin(rv) oxide. Adsorption conditions (pressure/N m-* ; temperature/K ; duration min) and evacuation conditions (temperature K ; durationlmin). (1) CHJCHO (532, 293, 30), (293, 30); (2) (CH,),CO (133, 320, 15), (320, 10); (3) (CD3),C0 (133, 320, 15), (320, 30) ; (4) (CD& CO on a fully deuterated surface (2.4 x lo3, 320, 15), (320, 180). Exposure of the dimethyl carbonate treated tin@) oxide disc to carbon dioxide (8.0 kN m-2) produced strong bands in the spectrum at 1550-1560, 1470 and 1360 cm-1 [fig. 1(2)], which may be assigned as v,,(COO), 6(CH3) and vsy,(COO), respec- tively, of a surface methyl carbonate species.The positions of the bands are similar to those observed for the same species on magnesium oxide.5 The bands persisted when the C 0 2 pressure was reduced to 1.33 N m-2 [fig. 1(4)] and the original spectrum was restored by pumping for a few minutes at 1.33 x N m-2. COz affected the spectrum in a similar fashion after evacuation of the sample at 373 K (1 h) and 423 K (2 h), but after removal of methoxy groups by evacuation of the sample at 459 K (1 h) the spectrum was unaffected by COz [fig. 1(5)].E . W . THORNTON A N D P . G. HARRISON 247 I The 1465cm-l band [S(CH,)] of the surface methoxy groups was removed by exchange with D20 (at 348 K), and simultaneously bands appeared at 2930 and 2820 cm-l replacing the broad hydroxyl stretching band.Admission of methanol vapour (4.0 kN m-2, a.b.t,) to a sample which had been fully deuterated and evacuated at 446 K resulted in a new, broad -OH stretching band centred at 3300 cm-l and a decrease in the intensity of the -OD band. In addition, new bands were observed at 2930, 2820, 1550, 1465, 1380 and 1340 cm-1 due to the formation of surface 8 8 81 wavenumber /cm-’ FIG. 3.-Infrared spectra of tin(rv) oxide discs during the adsorption of organic carbonyl compounds at 320 K. (1) CH3CH0 at 523 N m-’, (2) (CH3)&0 at 2.4 kN m-’ and (3) (CD3)2C0 at 2.4 kN m-’. All samples evacuated at 293 K prior to admission of adsorbate gas. methoxy and formate groups. Repeated treatment with methanol at a.b.t. re- established the broad -OH band in the spectrum and removed the -OD band.The 2930 and 2820 cm-1 bands were obscured, but the other bands below 2000 cm-l were not affected to any detectabIe extent. The occurrence of two bands in the C-H stretching region on deuteration of the partially methoxylated surface is unusual. Since the methoxy groups are removed under these conditions, the only species remaining on the surface are the formate ion and the species which gives rise to the band at 1445 cm-l. The most likely species responsible for this band is a surface carbonate.* The formate ion would be expected to give rise to only one C-H stretching vibration (2841 cm-’ for HCOUNa and 2803 cm-l for HCOO- in solu- tion).l’ Hence the 2820 cm-1 band can be assigned to the C-H stretching funda- mental of the formate ion, and the 2930 cm-l band is assigned to a combination band of the formate ion (1380 + 1550 = 2930 cm-l).Strong coupling of the v,,(COO) and 6,,(CH) vibrations might be expected since the formate ion is small and planar, both factors which facilitate the occurrence of combination and overtone bands. Small hydrocarbon molecules generally have been observed to exhibit unusually intense overtone and combination bands as a result of Fermi resonance of overtones with the C-H stretching fundarnentak62472 TIN OXlDE SURFACES FORMIC AND ACETIC ACID ADSORPTION Tin(1v) oxide discs which had been exposed to formic or acetic acid vapours gave very intense absorption spectra in the range 1650-1300 cm-l. Some of the results are summarised in table 1 . The assignment of bahds due to surface formate groups has been discussed above, and a comparison of these with model tin formates is given in table 2.Discs treated with formic acid give similar spectra to that observed on methanol or dimethyl carbonate chemisorption except that the antisymmetric (COO) stretching band is significantly broader at lower temperatures, extending to about 1640 cm-'. A band also occurs at 1250 cm-' (320 K evacuation), weakening and shifting to 1280-1290 cm-' after evacuation at 320-458 K. These bands are most probably due to the presence of formic acid molecules strongly held on the surface. The 1250 cm-l band is best interpreted as a predominantly C-0 stretching band, and in view of the large shift to lower energy of the carbonyl stretching band (A? = 110-120 cm-l from the pure liquid) the formic acid is probably coordinated through the carbonyl oxygen to Lewis acid sites on the surface, rather than merely hydrogen bonded to surface hydroxyl groups.Nevertheless, hydrogen bonding is indicated by the absence of the C-0-H deformation band which is observed in the spectrum of formic acid in the gas phase. 1 I I i I I 1 6 0 0 1400 120 3 1000 wavenumber/cm-l FIG. 4.4nfrared spectra of acetic acid (NaCl windows) adsorbed onto tin(1v) oxide. (1) " Starting surface ", disc evacuated at room temperature for 2 h. (2) Disc exposed to acetic acid vapour (266 N m-2, 320 K, 2 min) and evacuated at 320 K for 15 min and (3) at 503 K for 30 min. With acetic acid adsorption, bands due to both surface acetate groups and mole- cular acetic acid species bound to the surface are observed at lower temperatures ( < a 8 K) [fig. 4(2)].At temperatures 2 468 K, bands due to surface acetate groups only are observed [fig. 4(3)]. A full assignment of the bands due to surface acetate groups together with that for tin(@ acetate for comparison is presented in table 3. After evacuation for 30 min at 503 K, the bands at 1160, 1044, 1025 and 923 cm-lE . W. THORNTON A N D P. G. HARRISON 2473 TABLE 1 .-INFRARED SPECTRA (1 100-1 800 cm-I) OF SOME ORGANIC MATERIALS ADSORBED ON TIN(IV) OXIDE AS A FUNCTION OF EVACUATION TEMPERATURE evacuation band positionsjcm-1 and assignments temp./ K a.b.t. r.t. 373 459 493 r.t. 373 448 468 516 573 625 a.b.t. 398 458 488 a.b.t. a.b.t. a.b.t. 373 483 628 time/ 0.25 1465 14 1470 1 1470 1 2 h 4 C H 3 y*25 1 1 li 1.5 } I 15 3 1 0.5 0.5 3 0.5 0.3 0.25 ~~ - HCOO- 1550, 1380, 1360-1340 1550, 1380, 1360-1340 1550, 1380, 1340 1550, 1382, 1340 1550, 1382, 1340 CH3COO- 1520-1 530.1430, 1347 1520-1530,1430, 1347 1530, 1425, 1347 1500-1650,1350 1500-1650, 1380, 1350 1530-1580, 1382,1340 1550, 1382, 1340 1520,1425, 1347 1530, 1425, 1347 1505,1415 J 1505,1415 1505, 1415 1500,1411 others b 1445 1445, 1425 1445, 1420 1625, 1380, 1320 1625, 1380, 1320 1625,1320 1250 1280 1289 1320 1320,1240 1247, 1347 1245, 1347 All gases adsorbed at room temperature on tin(xv) oxide discs evacuated at room temperature ; b see discussion for some possible assignments ; c Deuteroacetate. TABLE 2.-INFRARED SPECTRA OF MODEL METHOXY AND FORMATE STRUCTURES band positions/cm-l and assignments compound Sn(OOCH), Sn(OOCH), Cl(CH3)2Sn(00CH) HCOONa HCOONa(aq.soln). Sn(OMe), (CH3)2Sn(OOCH)2 MeOH on Sn02 methoxy 4 C H d VB(C00) 1339 1311 1368 1390 1373 1366 1351 1450(sh) 1430 1465(s) 1 340 1360(s) ~~ ~ formate ion %(COO) 1563 1608 1588 1595 1567 1585 1550 d i.p.(CH) 1385 1370 1404 1373 1325 1377 1383 1382 TABLE 3 .-FUNDAMENTAL VIBRATIONS OF ACETATES wavenumber/cm-I ref. 9 9 10 10 11 11 this work this work frequency v2 v3 v4 V8 V9 V l O v13 v14 Sn(OOCCH3)z 1335 1395 930 1530 1429 1015 1441 1046 this work 1347 1425* 923 1530 1425* 1025 1425* 1044 * These bands lie very close together and cannot be resolved by the spectrometer used study. in this2474 TIN OXIDE SURFACES are shifted to 1048, 1030, 940 and 925cm-l, and a weak broad band appeared at 1250 cm-l (cf.background). Additional bands observed at lower temperatures (table 1) at 1625, 1320 and 1160 cm-1 can be assigned to carbonyl stretching, &OH) and v(C-0), respectively, of the surface molecular acetic acid species, although in the case of monomeric carboxylic acids much coupling occurs between the latter two modes. The band observed at 1380 cm-' at low temperatures may also be associated with the adsorbed acetic acid, but assignment is difficult without further evidence. ACETALDEHYDE, ACETONE AND HEXADEUTEROACETONE ADSORPTION Fig. 2 shows the spectra obtained when tin(1v) oxide discs were exposed to acetalde- hyde, acetone or hexadeuteroacetone vapour and then evacuated at room temperature before recording the spectrum. The results are also summarised in table 1.Fig. 2( 1) shows clearly that acetaldehyde is oxidised to an acetate during chemisorption by tin@) oxide at 320 K, the spectrum being very similar to the spectra produced after acetic acid adsorption. An additional weak band also occurs at 1320 cm-l. With acetone, an acetate is also the major chemisorption product [fig. 2(2)], additional bands occurring at 1320 and 1240 cm-1 (both readily removed by evacuation above 373 K). When a fully hydroxylated disc was exposed to hexadeuteroacetone (2.4 kN m-2, 320 K, 0.5 h) and then evacuated (0.5 h, 320 K) the spectrum showed absorption bands at 2950(sh), 2912,2870(sh), 2700-2000 (very strong), 1510-1520, 1417, 1350 and 1250 cm-' (sh = shoulder) [compare fig. 2(3)] and the broad hydroxyl stretching band was almost entirely removed, indicating the occurrence of rapid deuterium exchange.Exposure of a fully deuterated tin(rv) oxide disc to (CD,),CO under the same condi- tions resulted in new bands at 1505, 1415, 1347 and 1247 cm-l [fig. 2(4)] the last two being eliminated by evacuation at 483 K for 20 min. The bands at 1505 and 1415 cm-l are readily assigned as the antisymmetric and symmetric (COO) vibrations, respectively, of a surface deuteroacetate. The shifts to lower frequency of these bands (Avas : 1530 -+ 1505 cm-' and AVS : 2425 -+ 1415 cm-I) from the corresponding acetate (in particular Avas > Avs) are consistent with the shifts observed for main group acetates and for a surface acetate on alumina.2 The spectra do not however help in assigning the positions of the methyl deformation frequencies. The sorption of hexadeuterioacetone on to a hydroxylated sample offers an intermediate position where v,,(COO) is at 1510-1520cm-' and v,,,(COO) is at 1417cm-', indicating a partially deuterated system.The symmetric methyl deformation mode is present at 1347 cm-' but very much reduced in intensity, and the presence of some protons is indicated by a very weak band at about 1347 cm-l in the case of a " fully-deuterated " system. A band not associated with the deuteroacetate is present at 1240-1250 cm-l. The spectra of tin(1v) oxide discs in contact with CH,CHO, (CH3)&0 or (CD3)2C0 vapours differed from those described above. In addition to bands due to the pro- ducts of chemisorption and the superimposed gas phase compounds, new adsorption maxima were observed at 1615 cm-1 (CH,CHO), 1690, 1600 cm-l [(CH,),CO] and 1680, 1590 cm-l [(CD,),CO]. These bands (fig.3) were all removed by evacuation at 320 K for 15 min, and during the adsorption process the weak 1320 cm-' band for acetaldehyde and acetone was increased in intensity, the 1240 cm-1 band (acetone and hexadeuteroacetone) being obscured by a strong gas-phase band. DISCUSSION The mechanism of formation of surface methoxy groups from methanol is un- certain, but the two most likely mechanisms are esterification :E. W. THORNTON AND P. G . HARRISON or dissociative chemisorption : 2475 (ir (ii) Since no molecular water is produced during the chemisorption process, the most likely mechanism is (ii). A similar type of mechanism has been proposed for the formation of methoxy groups on MgO and for the formation of surface esters (from acetic and propionic acids) on GeO,.’ In addition to the formation of surface methoxy and formate groups on treatment of the oxide disc with methanol, the observation of a hydroxyl stretching band replacing v(0D) in the spectrum on treating a fully deuterated surface with methanol indicates the occurrence of the ready OD-OH exchange equilibrium : \ \ / / -SnOD + CH30H + -SnOH + CH30D.Repeated treatment with methanol resulted in the total replacement Qf surface -OD groups. A similar dissociative chemisorption mechanism may be proposed for the forma- tion of surface methoxy groups from dimethyl carbonate : H 3CO., \ h c=o H3CO I \-/sn \,/Sn\ I I /s ”\ o/5n \ The intermediate surface methyl carbonate is known to be unstable, dissociating on evacuation to form additional surface methoxy groups.The methoxy groups are very unstable towards oxidation to surface formate groups. Indeed, it was not possible to produce a sample containing methoxy groups2476 TIN OXIDE SURFACES without the formation of some surface formate. This oxidation reaction takes place at temperatures 2 320 K and after evacuation at 459 K only formate ions exist on the surface of the sample. The spectrum due to the formate ion increases in intensity as the evacuation temperature is increased, probably owing to the collection on the disc of material (HCOOH, CH30H, HCHO, etc.) desorbing from the glass vacuum line and from the silica furnace.The temperature at which formate species are removed from the spectrum is between 439 and 548 K, which is reasonably consistent with the decomposi- tion temperature of tin(I1) formate (47 1-473 K). l4 The surface methoxy groups are easily hydrolysed, the exchange reaction with D20 proceeding rapidly at 348 K. In the presence of a pressure of carbon dioxide, the methoxy groups undergo reaction 0 II to form surface methyl carbonate species : I The reaction probably proceeds by initial coordination of the C 0 2 to Lewis acid tin sites adjacent to the surface methoxy groups, followed by nucleophilic attack of the methoxy oxygen at the electrophilic centre of the coordinated CO,. The reaction with CO, is reversible, the surface methyl carbonate dissociating on evacuation to regenerate the surface methoxy groups.This type of reaction is well known for organotin(1v) alkoxides with a variety of unsaturated acceptor molecules. In particular, organic isocyanates react to form N-stannylmethylcarbamates : R,SnOMe + R'N=C=O + R,Sn*NR'CO-OMe. Addition of ethyl isocyanate to a partially methoxylated tin(1v) oxide disc resulted only in the formation of a strong spectrum characteristic of 1,3-diethylurea, due to hydrolysis of the isocyanate by molecular water on the surface : EtNCO + H,O(ads) + EtNH, + CO, EtNCO I .1 (EtNH),CO rather than the formation of a surface carbamate species. The same results were also obtained when an untreated disc was dosed with ethyl i~0cyanate.l~ A structure of the type ,sn\o/sn\E. W.THORNTON AND P. G. HARRISON 2477 involving coordination of the carboxylic acid molecule to Lewis acid tin sites via the carbonyl oxygen together with additional hydrogen bonding may be inferred from the available data for the surface molecular carboxylic acid species, but there is no real evidence to show the orientation of the acid molecule relative to the surface plane. Molecular formic acid desorbs between 458 and 488 K, and acetic acid at between 448 and 468 K. Chemisorption of CH,CHO, (CH&CO or (CD&CO on to tin@) oxide proceeds with simultaneous oxidation generating surface (deutero)acetate groups. Rapid deuterium exchange occurs between a hydroxylated disc and hexadeuteroacetone, C-H stretching bands being observed at 2950(sh), 2912 and 2870(sh) cm-l.Since the system is at least partially deuterated, interpretation of these bands is difficult. As the reaction is faster than the reaction with D20, it would appear that adsorbed hexadeuteroacetone has a deuterium atom which is readily available for exchange, and the most likely species responsible for this is a surface-adsorbed enol : adsorption (CD3)2CO ~ + CD,-C=CD,(ads). I OD Some confirmation of this is seen in fig. 3 which shows spectra recorded during the adsorption processes. A superimposed gas phase spectrum is present, but two new bands are found for acetone and hexadeuteroacetone. One of these can readily be assigned to a hydrogen-bonded carbonyl of adsorbed ketone (1690 and 1680 cm-l, respectively). The band situated just above 1600 cm-' for all three adsorbates may be assigned to a C=C stretching band of the enol form, the bands observed between 1350 and 1200 cm-l being modes associated with the C-0-H grouping.Changes observed on deuteration for these bands lend support to this hypothesis, and are assigned as 6(COH) and v(C-0), respectively, although as noted previously for acetic acid, strong coupling between these modes is expected. The enol forms are all desorbed by evacuation at about 373 K, and the acetates formed show similar desorption characteristics to the acetate formed by acetic acid. We thank the Tin Research Institute and the S.R.C. for support in the form of a C.A.S.E. Award (to E. W. T.). A. J. Bloodworth, A. G. Davies and S. C. Vasishtha, J. Chem. SOC. C, 1967,1309 and references therein. R. G. Greenler, J. Chem. Phys., 1962,37,2094. R. 0. Kagel, J. Phys. Chem., 1967, 71, 844. A. A. Davydov, V. M. Shchekochikin, P. M. Zaitsev, Yu. M. Shehekochikin and N. P. Keir, Kinetics and Catalysis, 1971, 12, 61 1. R. 0. Kagel and R. G. Greenler, J. Chem. Phys., 1968, 49, 1638. L. H. Little, Infrared Spectra of Adsorbed Species (Academic Press, New York, 1967). J. C. McManus and M. J. D. Low, J. Phys. Chem., 1968,72,2378. E. W. Thornton and P. G. Harrison, J.C.S. Faraday I, 1975, 71, 461. J. D. Donaldson, J. F. Knifton and S. D. Ross, Spectrochim. Acta, 1964, 20, 847. l o R. Okawara, D. E. Webster and E. G. Rochow, J. Amer. Chem. Soc., 1960,82,3287. K. Nakamoto, Infrured Spectru of Inorganic and Coordination Compounds (Wiley, London, 1970). l2 A. G. Davies and P. G. Harrison, J. Chem. SOC. C, 1967, 1313. l 3 E. W. Thornton and P. G. Harrison, unpublished data. l4 J. D. Donaldson and J. F. Knifton, J. Chem. SOC., 1964,4801. D. Hadzi and M. Pintar, Spectrochim. Acta, 1958, 12, 162.

ISSN:0300-9599    

DOI:10.1039/F19757102468

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Simultaneous Measurement of Infrared Spectra and Adsorption Isotherins for the Adsorption of Phenol on Silica at the Solid/Liquid Interface BY KENNETH MARSHALL AND COLIN H. ROCHESTER* Department of Chemistry, University of Nottingham, Nottingham NG7 2RD Received 15th April, 1975 An infrared cell is described with which spectra of species adsorbed at the solid/liquid interface and the corresponding adsorption isotherms can be determined simultaneously. The system has been tested by a study of the adsorption of phenol onto silica immersed in carbon tetrachloride. Isolated hydroxyl groups on the oxide surface form hydrogen bonds with adsorbed phenol molecules. The combined adsorption isotherm and spectroscopic results have enabled an estimate to be made of the popblation of isolated hydroxyl groups on the sample of silica studied.A simple cell used for the measurement of infrared spectra of several hydrogen bond acceptor molecules ' 7 adsorbed on silica at the solid/ liquid interface has been described previous1y.l Although easy to use the cell had some disadvantages which made it unsuitable for the simultaneous determination of adsorption isotherms and infrared spectra. The liquid system was neither thermo- statted nor stirred and quantitative measurement of the uptake of the adsorbate by the oxide disc was impossible. The aim of the present work was to construct and test an infrared cell in which the liquid phase was adequately thermostatted and stirred, and differences between the added and equilibrium concentrations of adsorbate in solution could be determined reliably.Carbon tetrachloride was chosen as solvent because of its spectroscopic properties. Isotherms for the adsorption of phenol on silica immersed in carbon tetrachloride have been reported by Boehm and Gromes and by Davis et aL5 and carboxylic acids 2 - EXPERIMENTAL The infrared cell is shown in fig. 1. The basic design of that part of the cell which contained the oxide disc immersed in solutions of adsorbate was identical to that of the cell previously described.l Unless otherwise stated the materials used in the canstruction of the two cells were the same. The base section of the new cell was made from stainless steel because brass, which was used before, had occasionally reacted with the adsorbates being studied. The use of a copper gasket as a seal between the flanges holding the fluorite optical windows (diameter 4 cm) considerably improved both the dynamic and static vacuums attainable. There was no detectable contamination of the solution or oxide sample resulting from the use of copper, rather than Teflon,l as the gasket material.A subsidiary optical compartment was constructed from two stainless steel flanges into which fluorite windows of diameter 3 cm were fixed with Araldite and which were bolted together via a copper gasket. The main and subsidiary optical compartments were made demountable for ease af internal cleaning. The subsidiary compartment contained adsorbate in carbon tetrachloride at the same solution composition as was contained in the main compartment in equilibrium with the oxide sample.Equilibrium was maintained throughout the system by the anticlockwise (fig. 1) circulation of solution through the two compartments in series from a Pyrex reservoir containing a rotating paddle which provided the pumping mechanism, Solution contained 2478K . MARSHALL AND C. H. ROCHESTER 2479 in the reservoir was thermostatted by a glass jacket through which water was passed from a conventional thermostat bath. Temperature probes in the form of copper-constantan thermocouples in the reservoir and at the base of the cell showed that the temperature of the circulating solution varied at 25°C by < +0.2"C. Unlike the systems involving a reference cell adopted by Low and Hasegawa 6*7 the spacing of the main and subsidiary cell compart- ments was such that when the main compartment was positioned in the sample beam of the infrared spectrometer then the subsidiary cell was clear of the spectrometer reference beam, The reference beam passed through the space between the two compartments. The aim was to measure the separate spectra of the oxide disc immersed in a solution and of the copper.qa5ke.t f l u o r i t e window; stirrer motor t h z r rnoco u p ie , well --I-- ---per mar e r t mognzt --Te I f o r4 bea r i n q soft iror! piece CI amping w e II2480 solution itself. The optical path lengths of the main and subsidiary compartments were 2.8 mm and 3.8 mm respectively. The purification of carbon tetrachloride, the preparation and admission of solutions to the infrared cell, the preparation and pretreatment of silica discs (surface area 176 m2 g-I), and the measurement of infrared spectra of the discs immersed in solutions of adsorbate were as before.' Phenol was recrystallised twice from light petroleum, b.p.40-60°C. Spectra were measured of phenol solutions of known concentration in the subsidiary cell compartment. Linear calibration graphs of the absorbances of the bands at 3608 and 3045 cm-' due to OH and CH stretching vibrations, respectively, against phenol concentra- tion were plotted. For the 3608 cm-I band the graph was only linear up to - 18 mmol dm-3 as phenol was partially dimerized at higher concentrations. Equilibrium concentrations of phenol solutions in adsorption experiments were thence deduced from the corresponding spectra of the solutions in contact with oxide discs.The differences between the added and equilibrium concentrations gave a measure of the amount of phenol adsorbed by the oxide sample and enabled adsorption isotherms to be calculated. Spectra of the oxide immersed in phenol solutions were also recorded. The optimum weight of the disc for spectroscopic study was - 80 mg. This weight was insufficient to give a reliably measurable difference between the added and equilibrium concentrations of phenol in solution. A second heavier ( ~ 0 . 6 5 g) disc of silica was therefore supported directly above the first disc such that the former did not interfere with the optical beam. The two discs were subjected to identical thermal pretreatment conditions and both were completely immersed in the liquid phase during a series of adsorption experiments.The attainment of equilibrium was established by confirmation that spectra of the oxide disc and the circulating solution did not vary with time. ADSORPTION OF PHENOL ON Si02 RESULTS Spectra of a silica disc immersed in solutions of phenol in carbon tetrachloride are shown in fig. 2. The band at 3686 cm-1 due to isolated silanol groups perturbed by immersion of the oxide in carbon tetrachloride was broadened and shifted to -3400 cm-' when phenol was adsorbed onto the surface. The exact position of 4 600 3500 3000 2500 FIG. 2.-Spectra (c)-(f) immersed wavenumber /cm-' of silica (a) after evacuation at 480'C for 64 h, (b) immersed in carbon tetrachloride, in solutions of phenol at equilibrium concentrations of 3.3,10.3, 19.7 and 40.8 mmol d ~ n - ~ respectively.K .MARSHALL AND C. H. ROCHESTER 248 1 the broad maximum shifted from 3430 to 3390 cm-I as the equilibrium concentration of phenol in solution was increased. The loss of intensity of the band at 3686 cm-I was a linear function (fig. 3) of the increase in the intensity of the broader band at 0.6[ I I I f 0.2 0.4 0.6 0.8 optical density at 3400 cm-I FIG. 3.-Linear relationship (slope -1.1) between the intensities of the infrared bands due to free and perturbed silanol groups on silica when phenol adsorbs from carbon tetrachloride solution. -3400 cm-l. These spectroscopic changes are characteristic of a hydrogen bonding interaction between surface silanol groups and adsorbed phenol molecules.A band at 3613 cm-l is assigned to the OH stretching vibration of phenol monomer in solution. Phenol existed entirely as its monomer at concentrations up to - 18 mmoldm-3. At higher concentrations some dimer (band at 3485cm-') was also phenol adsorbed/mg g-' FIG. 4.-Linear relationship between the growth of the absorption band due to perturbed sikmol groups and the weight of phenol adsorbed.2482 present. The highest concentration studied was 40 mmol dm-3 for which 3 % of the phenol existed as dimer. Infrared bands at 3080, 3049 and 3024cm-l were due to the CH stretching vibrations of phenol molecules both in solution and in the adsorbed state. The positions and relative intensities of the three bands were identical to those for phenol in solution in carbon tetrachloride.The bands were apparently unaffected by the adsorption of phenol on the surface of silica. The variation in the intensity of the broad infrared band at -3400 cm-l as a function of the equilibrium concentration of phenol in solution was of similar form to the corresponding variation in the weight of phenol adsorbed per gramme of silica. The latter was evaluated from the differences between the added and equilibrium concentrations of phenol. The similar shapes of the two curves are emphasized by the approximately linear relationship (fig. 4) between the weight of phenol adsorbed and the increase in intensity of the absorption maximum at 3400cm-'. Adsorbed phenol was completely desorbed from the surface of silica by flushing a disc immersed in a phenol solution with pure carbon tetrachloride solvent.The disc remained in contact with liquid phase throughout the flushing procedure. There was no spectroscopic evidence for chemisorption of phenol on the oxide surface. ADSORPTION OF PHENOL ON Si02 DISCUSSION The perturbation of the infrared band at 3686 cm-l and the concomitant growth of a broad band at -3400 cm-1 which occurred when phenol adsorbed onto silica showed that adsorbed phenol molecules were involved in hydrogen bonding inter- actions with isolated surface silanol groups. In contrast to the corresponding results for cyclohexanone the linearity of the plot in fig. 3 suggests that phenol was not adsorbed onto adjacent interacting silanol groups at low surface coverages. Furthermore the linear relatioilship (fig.4) between the growth of the broad band at -3400 cm-l and the weight of phenol adsorbed suggests that for concentrations <40 mmol dm-3 there was no phenol adsorbed onto sites other than surface silanol groups. The shifts to -3400 crn-l of the infrared band due to isolated silanol groups is too great to be caused by hydrogen bonding interactions between surface hydroxyl groups and the benzene rings of adsorbed phenol molecules.' Isolated silanol groups were hydrogen bonded to the hydroxyl groups of phenol molecules. A single inter- action may be represented by structure A (R = phenyl) in which the figures represent the shifts AVOH in the positions of the bands due to the stretching vibrations of silanol and phenol hydroxyl groups which occurred on adsorption. The shifts may be compared with those for n-decanol and n-propanol (structure B ; R' = alkyl) for which two broad infrwed bands appeared on adsorption of the alcohol molecules.2 For phenol only one broad band appeared presumably because the absorption maxima for the perturbed silanol and phenolic hydroxyl groups coincided.The difference between the shifts for the alcohols and phenols may be rationalized as follows. Phenol is a stronger acid but a weaker base than either of the alcohol^.^ For two hydrogen bond donors the one of higher acidity gives the greater shift AVOHcm-' on forming a hydrogen bond to a particular acceptor Similarly the shift AVOH for alcohols or phenols acting as hydrogen bond acceptors is greater the more basic is the acceptor molecule.In structure A the hydrogen bond a would therefore be expected to be stronger and the hydrogen bond b would be weaker than in structure B. In general the shifts in the bands due to OH-stretching vibrations are greater if the hydroxyl groups are acting as hydrogen bond donors rather than acceptors.10 Thus the shift ABoN for the adsorbate molecules will be primarily influenced by the strength of hydrogen bond a but for the silanol groups the prime and linolenic acidK . MARSHALL A N D C. H. ROCHESTER 2483 influence will be hydrogen bond b. It follows that AToH for the adsorbate molecules should be greater and AToH for the silanol groups should be less when phenol rather than n-propanol or n-decanol are adsorbed on silica. These conclusions are consistent with the experimental observations.R 1 213 cm - l/o..**.. H H R’ I I 76 cm - l/o*=*.. H H a ’***..0/286 cm- 1 a “.*-.O/~X cm- 1 I Si I Si (4 / I \ (B) / I \ The linearity of the plots in fig. 3 and 4 and the evidence that phenol was only adsorbed onto isolated surface silanol groups suggests that the present spectroscopic and adsorption isotherm data provide a method for the estimation of these groups. Combination of the two linear graphs shows that a 50 % reduction in the intensity of the infrared band at 3686 cm-l due to isolated silanol groups resulted from the adsorption of 28.2 mg of phenol per gramme of silica (surface area 176 m2 g-I). The total number noH of isolated hydroxyl groups per unit area on the oxide surface may be calculated via eqn (1) where NA is Avogadro’s number, A is the surface area of the oxide, M is the molecular weight of phenol, and w is the weight of phenol adsorbed on m g of silica to give 50 % coverage of the silanol sites.Hence noH was 2.05 nm-* for a silica sample which had been evacuated at 480°C. This figure is in good agreement with a value of - 2.2 nm-2 interpolated from data for the total surface silanol concentration. The small difference between the two values would be consistent with a small residual concentration of adjacent interacting surface silanol groups. The latter were respon- sible for the shoulder at -3680 cm-1 in fig. 2a. The adsorption isotherm at 25°C measured here is compared in fig. 5 with the isotherm at 35°C obtained by Davis et aL5 In the figure no is the total number of nOH = (2NAw/MmA) (1) 0 0.I 0.2 0.3 0.4 0.5 1Oy.x:) FIG. 5.-Isotherms for the adsorption of phenol on silica from carbon tetrachloride solution, 0 aerosil silica, 25”C, present study ; 0 silica gel, 35”C, ref. ( 5 ) (see text),2484 ADSORPTION OF PHENOL ON SiOz moles of solute plus solvent in contact with rn g of silica with surface area A, x i is the equilibrium mole fraction of phenol, and Axv is the difference between the added and equilibrium mole fractions of phenol. The points plotted for the isotherm at 35°C were calculated from the adsorption parameters deduced by Davis et aL5 from their experimental data, numerical details of which were not published. The curve therefore represents a smoothed isotherm. The two isotherms (fig. 5) are similar in shape but differ in magnitude.The difference must arise in part because of the significantly different properties, particularly with respect to pore structure, of silica gel and Aerosil silica. However the prime effect must be that the sample studied by Davis et aLs had been evacuated at 120°C prior to immersion in solutions of phenol and therefore retained a much higher residual concentration of surface silanol groups and molecular water than the oxide evacuated in the present study at 480°C. The authors thank Tioxide International Ltd. for the award of a Fellowship (to K. M.). D. M. Griffiths, K. Marshall and C. H. Rochester, J.C.S. Faraday I, 1974,70,400. K. Marshall and C. H. Rochester, Faraday Disc. Chem. SOC., 1975,59, in press. K. Marshall and C.H. Rochester, J.C.S. Faraduy I, 1975,71, 1754. H. P. Boehm and W. Gromes, Angew. Chem., 1959,71,65. K. M . C. Davis, J. A. Deucher and D. A. Ibbitson, J.C.S. Far&y I, 1973,69,1117. M. J. D. Low and M. Hasegawa, J. Colloid Interface Sci., 1968, 26,95. ’ M. Hasegawa and M. J. D. Low, J. Colloid Interface Sci., 1969, 29, 593. P. G. Rouxhet and R. E. Sempels, J.C.S. Faraduy I, 1974,70, 2021. C. H. Rochester in The Chemistry of the Hydroxyl Group, ed. S . Patai (Interscience, London, 1971), p. 327. lo A. Hall and J. L. Wood, Spectrochim. Acta, 1967, 23A, 2657. li R. Bode, H. Ferch and H. Fratzscher, Properties and Applications of Aerosil (manufacturers handbook, Degussa, Frankfurt). Simultaneous Measurement of Infrared Spectra and Adsorption Isotherins for the Adsorption of Phenol on Silica at the Solid/Liquid Interface BY KENNETH MARSHALL AND COLIN H.ROCHESTER* Department of Chemistry, University of Nottingham, Nottingham NG7 2RD Received 15th April, 1975 An infrared cell is described with which spectra of species adsorbed at the solid/liquid interface and the corresponding adsorption isotherms can be determined simultaneously. The system has been tested by a study of the adsorption of phenol onto silica immersed in carbon tetrachloride. Isolated hydroxyl groups on the oxide surface form hydrogen bonds with adsorbed phenol molecules. The combined adsorption isotherm and spectroscopic results have enabled an estimate to be made of the popblation of isolated hydroxyl groups on the sample of silica studied. A simple cell used for the measurement of infrared spectra of several hydrogen bond acceptor molecules ' 7 adsorbed on silica at the solid/ liquid interface has been described previous1y.l Although easy to use the cell had some disadvantages which made it unsuitable for the simultaneous determination of adsorption isotherms and infrared spectra.The liquid system was neither thermo- statted nor stirred and quantitative measurement of the uptake of the adsorbate by the oxide disc was impossible. The aim of the present work was to construct and test an infrared cell in which the liquid phase was adequately thermostatted and stirred, and differences between the added and equilibrium concentrations of adsorbate in solution could be determined reliably. Carbon tetrachloride was chosen as solvent because of its spectroscopic properties.Isotherms for the adsorption of phenol on silica immersed in carbon tetrachloride have been reported by Boehm and Gromes and by Davis et aL5 and carboxylic acids 2 - EXPERIMENTAL The infrared cell is shown in fig. 1. The basic design of that part of the cell which contained the oxide disc immersed in solutions of adsorbate was identical to that of the cell previously described.l Unless otherwise stated the materials used in the canstruction of the two cells were the same. The base section of the new cell was made from stainless steel because brass, which was used before, had occasionally reacted with the adsorbates being studied. The use of a copper gasket as a seal between the flanges holding the fluorite optical windows (diameter 4 cm) considerably improved both the dynamic and static vacuums attainable.There was no detectable contamination of the solution or oxide sample resulting from the use of copper, rather than Teflon,l as the gasket material. A subsidiary optical compartment was constructed from two stainless steel flanges into which fluorite windows of diameter 3 cm were fixed with Araldite and which were bolted together via a copper gasket. The main and subsidiary optical compartments were made demountable for ease af internal cleaning. The subsidiary compartment contained adsorbate in carbon tetrachloride at the same solution composition as was contained in the main compartment in equilibrium with the oxide sample. Equilibrium was maintained throughout the system by the anticlockwise (fig.1) circulation of solution through the two compartments in series from a Pyrex reservoir containing a rotating paddle which provided the pumping mechanism, Solution contained 2478K . MARSHALL AND C. H. ROCHESTER 2479 in the reservoir was thermostatted by a glass jacket through which water was passed from a conventional thermostat bath. Temperature probes in the form of copper-constantan thermocouples in the reservoir and at the base of the cell showed that the temperature of the circulating solution varied at 25°C by < +0.2"C. Unlike the systems involving a reference cell adopted by Low and Hasegawa 6*7 the spacing of the main and subsidiary cell compart- ments was such that when the main compartment was positioned in the sample beam of the infrared spectrometer then the subsidiary cell was clear of the spectrometer reference beam, The reference beam passed through the space between the two compartments.The aim was to measure the separate spectra of the oxide disc immersed in a solution and of the copper. qa5ke.t f l u o r i t e window; stirrer motor t h z r rnoco u p ie , well --I-- ---per mar e r t mognzt --Te I f o r4 bea r i n q soft iror! piece CI amping w e II2480 solution itself. The optical path lengths of the main and subsidiary compartments were 2.8 mm and 3.8 mm respectively. The purification of carbon tetrachloride, the preparation and admission of solutions to the infrared cell, the preparation and pretreatment of silica discs (surface area 176 m2 g-I), and the measurement of infrared spectra of the discs immersed in solutions of adsorbate were as before.' Phenol was recrystallised twice from light petroleum, b.p.40-60°C. Spectra were measured of phenol solutions of known concentration in the subsidiary cell compartment. Linear calibration graphs of the absorbances of the bands at 3608 and 3045 cm-' due to OH and CH stretching vibrations, respectively, against phenol concentra- tion were plotted. For the 3608 cm-I band the graph was only linear up to - 18 mmol dm-3 as phenol was partially dimerized at higher concentrations. Equilibrium concentrations of phenol solutions in adsorption experiments were thence deduced from the corresponding spectra of the solutions in contact with oxide discs. The differences between the added and equilibrium concentrations gave a measure of the amount of phenol adsorbed by the oxide sample and enabled adsorption isotherms to be calculated.Spectra of the oxide immersed in phenol solutions were also recorded. The optimum weight of the disc for spectroscopic study was - 80 mg. This weight was insufficient to give a reliably measurable difference between the added and equilibrium concentrations of phenol in solution. A second heavier ( ~ 0 . 6 5 g) disc of silica was therefore supported directly above the first disc such that the former did not interfere with the optical beam. The two discs were subjected to identical thermal pretreatment conditions and both were completely immersed in the liquid phase during a series of adsorption experiments.The attainment of equilibrium was established by confirmation that spectra of the oxide disc and the circulating solution did not vary with time. ADSORPTION OF PHENOL ON Si02 RESULTS Spectra of a silica disc immersed in solutions of phenol in carbon tetrachloride are shown in fig. 2. The band at 3686 cm-1 due to isolated silanol groups perturbed by immersion of the oxide in carbon tetrachloride was broadened and shifted to -3400 cm-' when phenol was adsorbed onto the surface. The exact position of 4 600 3500 3000 2500 FIG. 2.-Spectra (c)-(f) immersed wavenumber /cm-' of silica (a) after evacuation at 480'C for 64 h, (b) immersed in carbon tetrachloride, in solutions of phenol at equilibrium concentrations of 3.3,10.3, 19.7 and 40.8 mmol d ~ n - ~ respectively.K .MARSHALL AND C. H. ROCHESTER 248 1 the broad maximum shifted from 3430 to 3390 cm-I as the equilibrium concentration of phenol in solution was increased. The loss of intensity of the band at 3686 cm-I was a linear function (fig. 3) of the increase in the intensity of the broader band at 0.6[ I I I f 0.2 0.4 0.6 0.8 optical density at 3400 cm-I FIG. 3.-Linear relationship (slope -1.1) between the intensities of the infrared bands due to free and perturbed silanol groups on silica when phenol adsorbs from carbon tetrachloride solution. -3400 cm-l. These spectroscopic changes are characteristic of a hydrogen bonding interaction between surface silanol groups and adsorbed phenol molecules. A band at 3613 cm-l is assigned to the OH stretching vibration of phenol monomer in solution.Phenol existed entirely as its monomer at concentrations up to - 18 mmoldm-3. At higher concentrations some dimer (band at 3485cm-') was also phenol adsorbed/mg g-' FIG. 4.-Linear relationship between the growth of the absorption band due to perturbed sikmol groups and the weight of phenol adsorbed.2482 present. The highest concentration studied was 40 mmol dm-3 for which 3 % of the phenol existed as dimer. Infrared bands at 3080, 3049 and 3024cm-l were due to the CH stretching vibrations of phenol molecules both in solution and in the adsorbed state. The positions and relative intensities of the three bands were identical to those for phenol in solution in carbon tetrachloride. The bands were apparently unaffected by the adsorption of phenol on the surface of silica.The variation in the intensity of the broad infrared band at -3400 cm-l as a function of the equilibrium concentration of phenol in solution was of similar form to the corresponding variation in the weight of phenol adsorbed per gramme of silica. The latter was evaluated from the differences between the added and equilibrium concentrations of phenol. The similar shapes of the two curves are emphasized by the approximately linear relationship (fig. 4) between the weight of phenol adsorbed and the increase in intensity of the absorption maximum at 3400cm-'. Adsorbed phenol was completely desorbed from the surface of silica by flushing a disc immersed in a phenol solution with pure carbon tetrachloride solvent. The disc remained in contact with liquid phase throughout the flushing procedure.There was no spectroscopic evidence for chemisorption of phenol on the oxide surface. ADSORPTION OF PHENOL ON Si02 DISCUSSION The perturbation of the infrared band at 3686 cm-l and the concomitant growth of a broad band at -3400 cm-1 which occurred when phenol adsorbed onto silica showed that adsorbed phenol molecules were involved in hydrogen bonding inter- actions with isolated surface silanol groups. In contrast to the corresponding results for cyclohexanone the linearity of the plot in fig. 3 suggests that phenol was not adsorbed onto adjacent interacting silanol groups at low surface coverages. Furthermore the linear relatioilship (fig. 4) between the growth of the broad band at -3400 cm-l and the weight of phenol adsorbed suggests that for concentrations <40 mmol dm-3 there was no phenol adsorbed onto sites other than surface silanol groups.The shifts to -3400 crn-l of the infrared band due to isolated silanol groups is too great to be caused by hydrogen bonding interactions between surface hydroxyl groups and the benzene rings of adsorbed phenol molecules.' Isolated silanol groups were hydrogen bonded to the hydroxyl groups of phenol molecules. A single inter- action may be represented by structure A (R = phenyl) in which the figures represent the shifts AVOH in the positions of the bands due to the stretching vibrations of silanol and phenol hydroxyl groups which occurred on adsorption. The shifts may be compared with those for n-decanol and n-propanol (structure B ; R' = alkyl) for which two broad infrwed bands appeared on adsorption of the alcohol molecules.2 For phenol only one broad band appeared presumably because the absorption maxima for the perturbed silanol and phenolic hydroxyl groups coincided.The difference between the shifts for the alcohols and phenols may be rationalized as follows. Phenol is a stronger acid but a weaker base than either of the alcohol^.^ For two hydrogen bond donors the one of higher acidity gives the greater shift AVOHcm-' on forming a hydrogen bond to a particular acceptor Similarly the shift AVOH for alcohols or phenols acting as hydrogen bond acceptors is greater the more basic is the acceptor molecule. In structure A the hydrogen bond a would therefore be expected to be stronger and the hydrogen bond b would be weaker than in structure B. In general the shifts in the bands due to OH-stretching vibrations are greater if the hydroxyl groups are acting as hydrogen bond donors rather than acceptors.10 Thus the shift ABoN for the adsorbate molecules will be primarily influenced by the strength of hydrogen bond a but for the silanol groups the prime and linolenic acidK .MARSHALL A N D C. H. ROCHESTER 2483 influence will be hydrogen bond b. It follows that AToH for the adsorbate molecules should be greater and AToH for the silanol groups should be less when phenol rather than n-propanol or n-decanol are adsorbed on silica. These conclusions are consistent with the experimental observations. R 1 213 cm - l/o..**.. H H R’ I I 76 cm - l/o*=*..H H a ’***..0/286 cm- 1 a “.*-.O/~X cm- 1 I Si I Si (4 / I \ (B) / I \ The linearity of the plots in fig. 3 and 4 and the evidence that phenol was only adsorbed onto isolated surface silanol groups suggests that the present spectroscopic and adsorption isotherm data provide a method for the estimation of these groups. Combination of the two linear graphs shows that a 50 % reduction in the intensity of the infrared band at 3686 cm-l due to isolated silanol groups resulted from the adsorption of 28.2 mg of phenol per gramme of silica (surface area 176 m2 g-I). The total number noH of isolated hydroxyl groups per unit area on the oxide surface may be calculated via eqn (1) where NA is Avogadro’s number, A is the surface area of the oxide, M is the molecular weight of phenol, and w is the weight of phenol adsorbed on m g of silica to give 50 % coverage of the silanol sites.Hence noH was 2.05 nm-* for a silica sample which had been evacuated at 480°C. This figure is in good agreement with a value of - 2.2 nm-2 interpolated from data for the total surface silanol concentration. The small difference between the two values would be consistent with a small residual concentration of adjacent interacting surface silanol groups. The latter were respon- sible for the shoulder at -3680 cm-1 in fig. 2a. The adsorption isotherm at 25°C measured here is compared in fig. 5 with the isotherm at 35°C obtained by Davis et aL5 In the figure no is the total number of nOH = (2NAw/MmA) (1) 0 0. I 0.2 0.3 0.4 0.5 1Oy.x:) FIG.5.-Isotherms for the adsorption of phenol on silica from carbon tetrachloride solution, 0 aerosil silica, 25”C, present study ; 0 silica gel, 35”C, ref. ( 5 ) (see text),2484 ADSORPTION OF PHENOL ON SiOz moles of solute plus solvent in contact with rn g of silica with surface area A, x i is the equilibrium mole fraction of phenol, and Axv is the difference between the added and equilibrium mole fractions of phenol. The points plotted for the isotherm at 35°C were calculated from the adsorption parameters deduced by Davis et aL5 from their experimental data, numerical details of which were not published. The curve therefore represents a smoothed isotherm. The two isotherms (fig. 5) are similar in shape but differ in magnitude. The difference must arise in part because of the significantly different properties, particularly with respect to pore structure, of silica gel and Aerosil silica. However the prime effect must be that the sample studied by Davis et aLs had been evacuated at 120°C prior to immersion in solutions of phenol and therefore retained a much higher residual concentration of surface silanol groups and molecular water than the oxide evacuated in the present study at 480°C. The authors thank Tioxide International Ltd. for the award of a Fellowship (to K. M.). D. M. Griffiths, K. Marshall and C. H. Rochester, J.C.S. Faraday I, 1974,70,400. K. Marshall and C. H. Rochester, Faraday Disc. Chem. SOC., 1975,59, in press. K. Marshall and C. H. Rochester, J.C.S. Faraduy I, 1975,71, 1754. H. P. Boehm and W. Gromes, Angew. Chem., 1959,71,65. K. M . C. Davis, J. A. Deucher and D. A. Ibbitson, J.C.S. Far&y I, 1973,69,1117. M. J. D. Low and M. Hasegawa, J. Colloid Interface Sci., 1968, 26,95. ’ M. Hasegawa and M. J. D. Low, J. Colloid Interface Sci., 1969, 29, 593. P. G. Rouxhet and R. E. Sempels, J.C.S. Faraduy I, 1974,70, 2021. C. H. Rochester in The Chemistry of the Hydroxyl Group, ed. S . Patai (Interscience, London, 1971), p. 327. lo A. Hall and J. L. Wood, Spectrochim. Acta, 1967, 23A, 2657. li R. Bode, H. Ferch and H. Fratzscher, Properties and Applications of Aerosil (manufacturers handbook, Degussa, Frankfurt).

ISSN:0300-9599    

DOI:10.1039/F19757102478

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Thermal Decomposition of 3-Ethyl-3-methyloxetan and 3,3-Diethyloxetan BY ALLAN D. CLEMEN'TS? HENRY M. FREY" AND JEREMY G. FREY Chemistry Department, University of Reading, Whiteknights, Reading RG6 2AD Received 22nd April, 1975 The thermal decompositions of 3-ethyl-3-methyloxetan and 3,3-diethyloxetan have been followed in the gas phase. Both decompositions are homogeneous, kinetically first order and probably unimolecular. In the temperature range 407 to 448°C the 3-ethyl-3-methyl compound yields 2- methylbut-1-ene and formaldehyde and the rate constants fit the Arrhenius equation log kl/s-' = 15.357kO.151-(251 230f2020) J mol-l/RTInlO. Similarly it was found that 3,3-diethyloxetan gave 2-ethylbut-1-ene and formaldehyde and in the tem- perature range 402 to 463°C the rate constants fitted the Arrhenius equation log kz/s-l = 15.297 2 0.063 - (249 8502 840) J mol-l/RTlnlO.The decompositions probably proceed by a biradical mechanism. Until recently there was little published work on the pyrolysis of oxetans which contrasts sharply with the data available on cyclobutanes. Early work by Bittker and Walters on oxetan itself was followed twelve years later by a study of the 3,3-dimethyl compound.2 The fact that the rates of decomposition of these oxetans and the corresponding cyclobutanes are essentially identical has been noted. Also, the Arrhenius parameters for the decomposition of both dimethyl derivatives are identical within experimental error. However, there is a difference in the reported parameters for oxetan and cyclobutane which was felt unlikely and possibly indicated that both the reported A and Ea values for oxetan were too More recently Holbrook and Scott have studied the pyrolysis of cis- and trans- 2,3-dimethyloxetan.This is a somewhat more complex system owing to the occur- rence of a geometric isomerization and also because there are two distinct fragmenta- tion reactions. It was possible to obtain good data for the decomposition pathways. While the rates of decomposition are again close to those for the corresponding cyclobutanes, the agreement was not as strikingly close as for the other compounds mentioned. In particular, the trans-2,3-dimethyloxetan decomposes a little more slowly than oxetan itself, whereas trans-l,2-dimethylcyclobutane decomposes nearly twice as rapidly as cyclobutane. The results reported in this paper were obtained from studies on two oxetans which were carried out in an attempt to determine whether the 3,3-dimethyloxetan results were representative of this type of substitution.It was also hoped that more information about the nature of the activated complex might be obtained. EXPERIMENTAL 3-Ethyl-3-methyloxetan was prepared from diethyl carbonate and 2-ethyl-2-methylpro- pane-1,3-diol by the method of Casteignau et aL5 The oxetan was purified by fractional distillation and then by preparative gas chromatography using a column containing di- isodecyl phthalate as the liquid phase. Before kinetic runs it was dried over a molecuIar sieve type 4A. 24852486 THERMAL DECOMPOSITION OF OXETANS 3,3-Diethyloxetan was obtained from Professor G.Casteignau and, after simple distilla- tion, was found to be -99 % pure by gas chromatography. Purification by preparative chromatography yielded a product with total impurities of -0.2 %. All other materials were commercial samples. APPARATUS A conventional high vacuum static pyrolysis system was used. Teflon-glass greaseless stopcocks were employed throughout to minimise absorption problems and the entire gas handling system was maintained between 85 and 95OC. Pyrex reaction vessels both packed and unpacked were maintained at the required temperature (within 0.1OC) by immersion in a high temperature fused salt thermostat. The progress of the reaction was monitored in the majority of runs by pressure change and in a few cases by analysis using gas chromatography. For the former, a pressure transducer (Bell and Howell type 4-327-0003) connected to the reaction vessel was used and for the latter a Perkin Elmer F11 instrument equipped with a flame ionization detector and a gas sampling valve.Attempts to obtain reaction mixture compositions by direct analysis of gas samples were unsuccessful. The analytical reproducibility was very poor and we attribute this to adsorption of the relatively high boiling oxetans on the metal surfaces of the gas sampling system. Accordingly the entire reaction mixture, after a predetermined time in the heated reaction vessel, was condensed into 250 mm3 of p-xylene and aliquot por- tions of this mixture were analysed using liquid sampling. A silicone oil column gave com- pIete separation between reactants, products and the p-xylene.Chromatographic peaks were measured using either a ball and disc or an electronic integrator. RESULTS 3 -ETHYL- 3 -ME TH Y LOX E T A N The thermal decomposition of 3-ethyl-3-methyloxetan was investigated in the temperature range 407 to 448°C. Pyrolyses were carried out in a Pyrex reaction vessel that has been " aged " by treatment with 10 Torr (1 Torr = 133 N of hexamethyldisilazane at 430°C for 12 h. The only decomposition product observed by gas chromatography was 2-methylbut-1-ene. By analogy with the results of Walters on 3,3-dimethyloxetan it was assumed that one other product, formaldehyde was formed, which polymerises under our analytical conditions on the chromato- graphic column.This was confirmed by a U.V. analysis of the reaction mixture after pyrolysis, which showed the presence of formaldehyde quite unambiguously. The pressure change in the system was consistent with the stoichiometry (1). Plots of log (2Po-Pt) against time were linear up to at least 50 % decomposition (Po and P, are the initial pressure and the pressure at time t respectively). Some runs were followed for many half lives when it was found that P, = 2P0. Rate constants were obtained from a least squares analysis of the pressure plots. A series of runs was carried out at 434.1"C with initial reactant pressures in the range 7.2 to 12.5 Torr. Within experimental error these 7 runs yielded rate constants independent of the initial pressure. In another series of runs at 407.5"C, the progress of the reaction was monitored by both pressure change and gas chromatographic analysis.The pressure data yielded a rate constant of 1.187 x s-l (average from 5 runs) and the value obtained from the analytical results was 1.238 x s-l. The 4 % difference between these values is not considered significant in view of the errorsA. D . CLEMENTS, H. M. FREY AND J . G . FREY 2487 associated with the liquid sampling method employed. Finally, in another series a " packed " reaction vessel was used. This had a surface to volume ratio about 14 times that of the unpacked vessel and the reaction was monitored by gas chromatu- graphy. The measured rate constant was 1.250 x s-l. This is in good agree- ment with the value obtained by this method of analysis in the unpacked reaction vessel and indicates that there can be no appreciable surface component of the reaction.s-'. Another run with an initial oxetan pressure of 8.36 Torr but to which 20.5 Torr of propene had been added gave a rate constant of 1.570~ s-l. These results make it unlikely that there are any measurable radical decomposition pathways under the experimental conditions used. The rate constants obtained from 51 runs at 10 temperatures are given in table 1, in all cases throughout this paper the quoted errors are standard deviations. Two runs at 41 1.6"C yielded a value for the rate constant of 1.582fO.01 x TABLE 1 .-RATE CONSTANTS FOR THE DECOMPOSITION OF 3-ETHYL-3-METHYLOXETAN temperature/'C 407.5 41 1.6 415.8 420.2 104kl /s-~ 1.187f0.012 1.582+0.010 2.091+0.061 2.613&0.107 1 04k1 /s-l 3.8454 0.184 5.012k 0.139 6.273 0.171 8.667f 0.283 temperature/"C 443.3 448.1 10% 1 /s-1 lo.%+ 0.16 14.874 0.1 8 temperature/"<: 425.6 430.6 434.1 439.3 An Arrhenius plot of the results quoted in table 1 yielded a good straight line from which the Arrhenius parameters were obtained by least squares, viz., log kl/s-l = 15.357fO.151-(60 046+483) cal mol-'/RTln 10 log kl/s-' = 15.357+0.151-(251 23022020) J mol-'/RTln 10.3,3-DIETHY LOXETAN Most of the details of the decomposition of this compound were closely similar to those of the ethyl methyl compound. Initial runs showed that plots of log (2P0 -P,) against time were accurately linear up to 50 % pressure increase. Rate constants determined from such plots were independent of initial reactant pressure in the range 5 to 15 Torr. (The range of pressures that could be employed was limited at the lower end by the sensitivity of the pressure transducer and at the upper end by the relatively low volatility of the reactant.) By analogy with other oxetan decomposi- tions the expected stoichiometry for the decomposition is (2) : For this study the 2-ethylbut-1-ene was identified by gas chromatography by compari- son of retention times (with an authentic sample) on several different columns.Further quantitative analysis (using the liquid sampling technique) showed that 1 mole of the oxetan yielded 1 mole of the olefin. Formaldehyde was detected by U.V. analysis and by gas chromatographic analysis using a column packed with Poropak N, but it proved impossible to determine quantitatively (owing to poor reproducibility).How- ever, the absence of any other peaks on the chromatograms together with the observed pressure changes confirm the quantitative nature of the decomposition.A. COX AND T . J . KEMP 2493 LACTIC ACID An intense absorption was obtained at 77 K which we regard as due to a radical mixture. Its basic pattern is a 1 : 4 : 6 : 4 : 1 quintet, but the coupling constant varied somewhat from peak to peak, averaging to 1.67 mT which seems too low a figure to be attributed to CH,eHOH. Curiously, Poznyak et aZ.12 found aqueous ethylene glycol solutions of FeIrl complexes of lactate and mandelate ions to photodecompose at 77 K (indicated by loss of the e.s.r. line of FelI1) but saw no production of RcHOH radicals until the matrix was warmed to 140 K ; this " delayed action '' in radical production was rationalised by the authors in terms of an [Fell-RcHOH] adduct which they believe to show no e.s.r.absorption; this thermally dissociates at 140 K into Fe" + RCHOH. DISCUSSION Thermodynamically Fe"' is intermediate in activity between Ce" and Uv' on which we have previously concentrated. l-' Whilst CeIV photo-oxidations of organic substrates involve light absorption in the charge-transfer band of the CeIv-substrate complex (A,,, - 300 nm or 361 kJ mol-l) together with net reduction of CeIv to Ce"', for which E" = 1.70 V in HC104 solution (1 mol dm-,), equivalent to a free energy change of 164 kJ mol-', the corresponding figures for UO$+ are Amax = 400 nm (or 271 kJmol-l) and E" = 0.05 V (or AGO = 4.8 kJ mol-') and for Fe3+ in HC104 solutions with added substrate, Amax x 350 nm (or 309 kJ mol-l) and E" = 0.772 V (or AGO = 74.5 kJ mol-l).Both as regards the photochemical and electrochemical terms, then, Fe"' might be expected to show behaviour intermediate between Ce'" and Uvl in its photochemical interaction with organic substrates. In its interaction with the two simplest alcohols, excited iron(1rr) ion, denoted FeI1I4:, behaves both as CeIV* in abstracting a hydrogen atom from the hydroxylic carbon atom to give RCHOH (R = H or Me). This takes place with both FeCI, and Fe(ClO,),, indicating the process to be due to a genuine attack by FelI1* rather than by intermediate formation of a chlorine atom,6 followed by a secondary attack of the latter upon the matrix.While both of these mechanisms are conceivable for FeCl,, the CI. atom mechanism cannot operate in the case of Fe(C104),, nor can it explain the formation of readily detectable alkyl radicals as a principal path- way in the Fe"'* oxidation of tertiary alcohols which is found with both Fe(C104), and FeCl,. These originate from a C-C cleavage photoprocess previously found with CeIV*, but not with Uvl*, viz., and Uv'* R1R2R3COH + FelI1* + R1R2R3&H + Fel1 (1) (2) + - R1R2R3COH 4 H+ + R1* + R2R3C=0 Towards carboxylic acids, FelI1* shows ambivalent behaviour. [steps (1) and (2) may be concerted). The radical derived from RC02H, where R = Me, Et, (Me),CHCH, and CH2=CHCH2, is predomi- nantly or exclusively R e , which comes from a process of oxidative decarboxylation found with CeIV* and PbtV* ; l4 with isobutyric acid, however, the seven-line spectrum must be due to (Me),CC02H, i.e., the route is one of hydrogen-atom abstraction [the possibility that the primary isopropyl radical is highly reactive towards the labile tertiary hydrogen atom at 77K is ruled out by the observation of Me,CH under these conditions during CeIV* oxidation], Clearly FelI1* behaves very much like CeIV* [and unlike Uvr';] in these particular oxidations, with the exception of isobutyric acid towards which it behaves like Uvl*.A .D. CLEMENTS, H. M . FREY AND J . G . FREY 2489 The Arrhenius parameters and the rate constants (calculated from these para- meters) at 425°C for a number of oxetans and for cyclobutane and 1,l-dimethyl- cyclobutane are shown in table 3.The results in table 3 make the similarity between cyclobutane and oxetan decom- position very clear indeed. The most recent work on oxetan itself shows that the sur- mise that the previously reported Arrhenius parameters were too low was correct. The new values now fit the general pattern well. Within experimental error the parameters for cyclobutane and oxetan are the same. On the basis of a biradical mechanism this almost certainly means that in the oxetan decomposition both bi- radicals (I) and (11) contribute to the reaction. Also, the enthalpy changes associated with their formation from oxetan must be close to one another and also close to the corresponding enthalpy change for the formation of (111), the tetramethylene biradical, from cyclobutane.This is consistent with thermochemical calculations which yield values for the enthalpies of these reactions (molecule to biradical) AHzgs of 55.6, 56.2 and 55.3 kcal mol-l respectively. (1) (11) (IW (IV) (V) The substitution of geminal dimethyl groups in both cyclobutane and oxetan produces essentially the same rate enhancement. Further, in both cases, this appears to be almost entirely the result of a reduction in the energy of activation for the reaction. This is exactly what would be expected on the basis of a biradical mechanism and would be consistent with the intermediate formation of biradicals of the types (IV) and (V). Thermochemical calculations yield virtually the same enthalpy change in going from dimethyloxetan to (IV) as from dimethylcyclobutane to (V), which amounts to - 51.2 kcal mol-l.This is about 4 kcal mol-1 less than for the unsubstituted case which is greater than the observed differences in energies of activation between sub- stituted and unsubstituted molecules. If the barrier to recyclisation of the biradicals is either greater or approximately equal to that for decomposition (the most probable situation) then these calculations and comparisons with experiment indicate that the barriers are not identical for (11) and (IV) or (111) and (V). However, it is doubtful whether the thermochemical data are sufficiently precise to be certain. What is clear is that the present work has fully confirmed the earlier results on the 3,3-dimethyl- oxetan and the relationship between oxetan and cyclobutane decomposition.We thank Professor G. Casteignau for the sample of diethyloxetan. We are grateful to the S.R.C. for the award of a studentship (to A. D. C.). D. A. Bittker and W. D. Walters, J. Amer. Chem. Soc., 1955, 77, 2326. G. F. Cohoe and W. D. Walters, J. Phys. Chem., 1967,71,2326. H. M. Frey and R. Walsh, Chem. Rev., 1969, 69, 103. K. A, Holbrook and R. A. Scott, J.C.S. Faraday 1,1974,70,43. J. L. Halary, T. Yvernault and G. Casteignau, Bull. Soc. chirn. France, 1972, 12,4655. R. B. Woodward and R. Hoffmann, The Conservation of Orbital Symmetry (Academic Press, New York, 1970). K. A. Holbrook, personal communication ; K. A. Holbrook and R. A. Scott, J.C.S. Faraahy I, 1975,71,1849. C. T.Genaux and W. D. Walters, J. Amer. Chem. Soc., 1951,73,4497 ; R. W. Carr and W. D. Walters, J. Phys. Chem., 1963, 67, 1370. ’ J. A. Berson, J. Amer. Chem. Soc., 1972, 94, 8917 lo P. C. Rotoli, M.S. Thesis (University of Rochester, 1963). Thermal Decomposition of 3-Ethyl-3-methyloxetan and 3,3-Diethyloxetan BY ALLAN D. CLEMEN'TS? HENRY M. FREY" AND JEREMY G. FREY Chemistry Department, University of Reading, Whiteknights, Reading RG6 2AD Received 22nd April, 1975 The thermal decompositions of 3-ethyl-3-methyloxetan and 3,3-diethyloxetan have been followed in the gas phase. Both decompositions are homogeneous, kinetically first order and probably unimolecular. In the temperature range 407 to 448°C the 3-ethyl-3-methyl compound yields 2- methylbut-1-ene and formaldehyde and the rate constants fit the Arrhenius equation log kl/s-' = 15.357kO.151-(251 230f2020) J mol-l/RTInlO.Similarly it was found that 3,3-diethyloxetan gave 2-ethylbut-1-ene and formaldehyde and in the tem- perature range 402 to 463°C the rate constants fitted the Arrhenius equation log kz/s-l = 15.297 2 0.063 - (249 8502 840) J mol-l/RTlnlO. The decompositions probably proceed by a biradical mechanism. Until recently there was little published work on the pyrolysis of oxetans which contrasts sharply with the data available on cyclobutanes. Early work by Bittker and Walters on oxetan itself was followed twelve years later by a study of the 3,3-dimethyl compound.2 The fact that the rates of decomposition of these oxetans and the corresponding cyclobutanes are essentially identical has been noted.Also, the Arrhenius parameters for the decomposition of both dimethyl derivatives are identical within experimental error. However, there is a difference in the reported parameters for oxetan and cyclobutane which was felt unlikely and possibly indicated that both the reported A and Ea values for oxetan were too More recently Holbrook and Scott have studied the pyrolysis of cis- and trans- 2,3-dimethyloxetan. This is a somewhat more complex system owing to the occur- rence of a geometric isomerization and also because there are two distinct fragmenta- tion reactions. It was possible to obtain good data for the decomposition pathways. While the rates of decomposition are again close to those for the corresponding cyclobutanes, the agreement was not as strikingly close as for the other compounds mentioned.In particular, the trans-2,3-dimethyloxetan decomposes a little more slowly than oxetan itself, whereas trans-l,2-dimethylcyclobutane decomposes nearly twice as rapidly as cyclobutane. The results reported in this paper were obtained from studies on two oxetans which were carried out in an attempt to determine whether the 3,3-dimethyloxetan results were representative of this type of substitution. It was also hoped that more information about the nature of the activated complex might be obtained. EXPERIMENTAL 3-Ethyl-3-methyloxetan was prepared from diethyl carbonate and 2-ethyl-2-methylpro- pane-1,3-diol by the method of Casteignau et aL5 The oxetan was purified by fractional distillation and then by preparative gas chromatography using a column containing di- isodecyl phthalate as the liquid phase.Before kinetic runs it was dried over a molecuIar sieve type 4A. 24852486 THERMAL DECOMPOSITION OF OXETANS 3,3-Diethyloxetan was obtained from Professor G. Casteignau and, after simple distilla- tion, was found to be -99 % pure by gas chromatography. Purification by preparative chromatography yielded a product with total impurities of -0.2 %. All other materials were commercial samples. APPARATUS A conventional high vacuum static pyrolysis system was used. Teflon-glass greaseless stopcocks were employed throughout to minimise absorption problems and the entire gas handling system was maintained between 85 and 95OC. Pyrex reaction vessels both packed and unpacked were maintained at the required temperature (within 0.1OC) by immersion in a high temperature fused salt thermostat.The progress of the reaction was monitored in the majority of runs by pressure change and in a few cases by analysis using gas chromatography. For the former, a pressure transducer (Bell and Howell type 4-327-0003) connected to the reaction vessel was used and for the latter a Perkin Elmer F11 instrument equipped with a flame ionization detector and a gas sampling valve. Attempts to obtain reaction mixture compositions by direct analysis of gas samples were unsuccessful. The analytical reproducibility was very poor and we attribute this to adsorption of the relatively high boiling oxetans on the metal surfaces of the gas sampling system. Accordingly the entire reaction mixture, after a predetermined time in the heated reaction vessel, was condensed into 250 mm3 of p-xylene and aliquot por- tions of this mixture were analysed using liquid sampling.A silicone oil column gave com- pIete separation between reactants, products and the p-xylene. Chromatographic peaks were measured using either a ball and disc or an electronic integrator. RESULTS 3 -ETHYL- 3 -ME TH Y LOX E T A N The thermal decomposition of 3-ethyl-3-methyloxetan was investigated in the temperature range 407 to 448°C. Pyrolyses were carried out in a Pyrex reaction vessel that has been " aged " by treatment with 10 Torr (1 Torr = 133 N of hexamethyldisilazane at 430°C for 12 h. The only decomposition product observed by gas chromatography was 2-methylbut-1-ene.By analogy with the results of Walters on 3,3-dimethyloxetan it was assumed that one other product, formaldehyde was formed, which polymerises under our analytical conditions on the chromato- graphic column. This was confirmed by a U.V. analysis of the reaction mixture after pyrolysis, which showed the presence of formaldehyde quite unambiguously. The pressure change in the system was consistent with the stoichiometry (1). Plots of log (2Po-Pt) against time were linear up to at least 50 % decomposition (Po and P, are the initial pressure and the pressure at time t respectively). Some runs were followed for many half lives when it was found that P, = 2P0. Rate constants were obtained from a least squares analysis of the pressure plots.A series of runs was carried out at 434.1"C with initial reactant pressures in the range 7.2 to 12.5 Torr. Within experimental error these 7 runs yielded rate constants independent of the initial pressure. In another series of runs at 407.5"C, the progress of the reaction was monitored by both pressure change and gas chromatographic analysis. The pressure data yielded a rate constant of 1.187 x s-l (average from 5 runs) and the value obtained from the analytical results was 1.238 x s-l. The 4 % difference between these values is not considered significant in view of the errorsA. D . CLEMENTS, H. M. FREY AND J . G . FREY 2487 associated with the liquid sampling method employed. Finally, in another series a " packed " reaction vessel was used.This had a surface to volume ratio about 14 times that of the unpacked vessel and the reaction was monitored by gas chromatu- graphy. The measured rate constant was 1.250 x s-l. This is in good agree- ment with the value obtained by this method of analysis in the unpacked reaction vessel and indicates that there can be no appreciable surface component of the reaction. s-'. Another run with an initial oxetan pressure of 8.36 Torr but to which 20.5 Torr of propene had been added gave a rate constant of 1.570~ s-l. These results make it unlikely that there are any measurable radical decomposition pathways under the experimental conditions used. The rate constants obtained from 51 runs at 10 temperatures are given in table 1, in all cases throughout this paper the quoted errors are standard deviations.Two runs at 41 1.6"C yielded a value for the rate constant of 1.582fO.01 x TABLE 1 .-RATE CONSTANTS FOR THE DECOMPOSITION OF 3-ETHYL-3-METHYLOXETAN temperature/'C 407.5 41 1.6 415.8 420.2 104kl /s-~ 1.187f0.012 1.582+0.010 2.091+0.061 2.613&0.107 1 04k1 /s-l 3.8454 0.184 5.012k 0.139 6.273 0.171 8.667f 0.283 temperature/"C 443.3 448.1 10% 1 /s-1 lo.%+ 0.16 14.874 0.1 8 temperature/"<: 425.6 430.6 434.1 439.3 An Arrhenius plot of the results quoted in table 1 yielded a good straight line from which the Arrhenius parameters were obtained by least squares, viz., log kl/s-l = 15.357fO.151-(60 046+483) cal mol-'/RTln 10 log kl/s-' = 15.357+0.151-(251 23022020) J mol-'/RTln 10. 3,3-DIETHY LOXETAN Most of the details of the decomposition of this compound were closely similar to those of the ethyl methyl compound.Initial runs showed that plots of log (2P0 -P,) against time were accurately linear up to 50 % pressure increase. Rate constants determined from such plots were independent of initial reactant pressure in the range 5 to 15 Torr. (The range of pressures that could be employed was limited at the lower end by the sensitivity of the pressure transducer and at the upper end by the relatively low volatility of the reactant.) By analogy with other oxetan decomposi- tions the expected stoichiometry for the decomposition is (2) : For this study the 2-ethylbut-1-ene was identified by gas chromatography by compari- son of retention times (with an authentic sample) on several different columns.Further quantitative analysis (using the liquid sampling technique) showed that 1 mole of the oxetan yielded 1 mole of the olefin. Formaldehyde was detected by U.V. analysis and by gas chromatographic analysis using a column packed with Poropak N, but it proved impossible to determine quantitatively (owing to poor reproducibility). How- ever, the absence of any other peaks on the chromatograms together with the observed pressure changes confirm the quantitative nature of the decomposition.A. COX AND T . J . KEMP 2493 LACTIC ACID An intense absorption was obtained at 77 K which we regard as due to a radical mixture. Its basic pattern is a 1 : 4 : 6 : 4 : 1 quintet, but the coupling constant varied somewhat from peak to peak, averaging to 1.67 mT which seems too low a figure to be attributed to CH,eHOH.Curiously, Poznyak et aZ.12 found aqueous ethylene glycol solutions of FeIrl complexes of lactate and mandelate ions to photodecompose at 77 K (indicated by loss of the e.s.r. line of FelI1) but saw no production of RcHOH radicals until the matrix was warmed to 140 K ; this " delayed action '' in radical production was rationalised by the authors in terms of an [Fell-RcHOH] adduct which they believe to show no e.s.r. absorption; this thermally dissociates at 140 K into Fe" + RCHOH. DISCUSSION Thermodynamically Fe"' is intermediate in activity between Ce" and Uv' on which we have previously concentrated. l-' Whilst CeIV photo-oxidations of organic substrates involve light absorption in the charge-transfer band of the CeIv-substrate complex (A,,, - 300 nm or 361 kJ mol-l) together with net reduction of CeIv to Ce"', for which E" = 1.70 V in HC104 solution (1 mol dm-,), equivalent to a free energy change of 164 kJ mol-', the corresponding figures for UO$+ are Amax = 400 nm (or 271 kJmol-l) and E" = 0.05 V (or AGO = 4.8 kJ mol-') and for Fe3+ in HC104 solutions with added substrate, Amax x 350 nm (or 309 kJ mol-l) and E" = 0.772 V (or AGO = 74.5 kJ mol-l).Both as regards the photochemical and electrochemical terms, then, Fe"' might be expected to show behaviour intermediate between Ce'" and Uvl in its photochemical interaction with organic substrates. In its interaction with the two simplest alcohols, excited iron(1rr) ion, denoted FeI1I4:, behaves both as CeIV* in abstracting a hydrogen atom from the hydroxylic carbon atom to give RCHOH (R = H or Me).This takes place with both FeCI, and Fe(ClO,),, indicating the process to be due to a genuine attack by FelI1* rather than by intermediate formation of a chlorine atom,6 followed by a secondary attack of the latter upon the matrix. While both of these mechanisms are conceivable for FeCl,, the CI. atom mechanism cannot operate in the case of Fe(C104),, nor can it explain the formation of readily detectable alkyl radicals as a principal path- way in the Fe"'* oxidation of tertiary alcohols which is found with both Fe(C104), and FeCl,. These originate from a C-C cleavage photoprocess previously found with CeIV*, but not with Uvl*, viz., and Uv'* R1R2R3COH + FelI1* + R1R2R3&H + Fel1 (1) (2) + - R1R2R3COH 4 H+ + R1* + R2R3C=0 Towards carboxylic acids, FelI1* shows ambivalent behaviour.[steps (1) and (2) may be concerted). The radical derived from RC02H, where R = Me, Et, (Me),CHCH, and CH2=CHCH2, is predomi- nantly or exclusively R e , which comes from a process of oxidative decarboxylation found with CeIV* and PbtV* ; l4 with isobutyric acid, however, the seven-line spectrum must be due to (Me),CC02H, i.e., the route is one of hydrogen-atom abstraction [the possibility that the primary isopropyl radical is highly reactive towards the labile tertiary hydrogen atom at 77K is ruled out by the observation of Me,CH under these conditions during CeIV* oxidation], Clearly FelI1* behaves very much like CeIV* [and unlike Uvr';] in these particular oxidations, with the exception of isobutyric acid towards which it behaves like Uvl*.A .D. CLEMENTS, H. M . FREY AND J . G . FREY 2489 The Arrhenius parameters and the rate constants (calculated from these para- meters) at 425°C for a number of oxetans and for cyclobutane and 1,l-dimethyl- cyclobutane are shown in table 3. The results in table 3 make the similarity between cyclobutane and oxetan decom- position very clear indeed. The most recent work on oxetan itself shows that the sur- mise that the previously reported Arrhenius parameters were too low was correct. The new values now fit the general pattern well. Within experimental error the parameters for cyclobutane and oxetan are the same. On the basis of a biradical mechanism this almost certainly means that in the oxetan decomposition both bi- radicals (I) and (11) contribute to the reaction.Also, the enthalpy changes associated with their formation from oxetan must be close to one another and also close to the corresponding enthalpy change for the formation of (111), the tetramethylene biradical, from cyclobutane. This is consistent with thermochemical calculations which yield values for the enthalpies of these reactions (molecule to biradical) AHzgs of 55.6, 56.2 and 55.3 kcal mol-l respectively. (1) (11) (IW (IV) (V) The substitution of geminal dimethyl groups in both cyclobutane and oxetan produces essentially the same rate enhancement. Further, in both cases, this appears to be almost entirely the result of a reduction in the energy of activation for the reaction.This is exactly what would be expected on the basis of a biradical mechanism and would be consistent with the intermediate formation of biradicals of the types (IV) and (V). Thermochemical calculations yield virtually the same enthalpy change in going from dimethyloxetan to (IV) as from dimethylcyclobutane to (V), which amounts to - 51.2 kcal mol-l. This is about 4 kcal mol-1 less than for the unsubstituted case which is greater than the observed differences in energies of activation between sub- stituted and unsubstituted molecules. If the barrier to recyclisation of the biradicals is either greater or approximately equal to that for decomposition (the most probable situation) then these calculations and comparisons with experiment indicate that the barriers are not identical for (11) and (IV) or (111) and (V). However, it is doubtful whether the thermochemical data are sufficiently precise to be certain. What is clear is that the present work has fully confirmed the earlier results on the 3,3-dimethyl- oxetan and the relationship between oxetan and cyclobutane decomposition. We thank Professor G. Casteignau for the sample of diethyloxetan. We are grateful to the S.R.C. for the award of a studentship (to A. D. C.). D. A. Bittker and W. D. Walters, J. Amer. Chem. Soc., 1955, 77, 2326. G. F. Cohoe and W. D. Walters, J. Phys. Chem., 1967,71,2326. H. M. Frey and R. Walsh, Chem. Rev., 1969, 69, 103. K. A, Holbrook and R. A. Scott, J.C.S. Faraday 1,1974,70,43. J. L. Halary, T. Yvernault and G. Casteignau, Bull. Soc. chirn. France, 1972, 12,4655. R. B. Woodward and R. Hoffmann, The Conservation of Orbital Symmetry (Academic Press, New York, 1970). K. A. Holbrook, personal communication ; K. A. Holbrook and R. A. Scott, J.C.S. Faraahy I, 1975,71,1849. C. T. Genaux and W. D. Walters, J. Amer. Chem. Soc., 1951,73,4497 ; R. W. Carr and W. D. Walters, J. Phys. Chem., 1963, 67, 1370. ’ J. A. Berson, J. Amer. Chem. Soc., 1972, 94, 8917 lo P. C. Rotoli, M.S. Thesis (University of Rochester, 1963).

ISSN:0300-9599    

DOI:10.1039/F19757102485

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Electron Spin Resonance Studies of Photo-oxidation by Metal Ions in Rigid Media at Low Temperatures Part 5.-Photo-oxidation by the Iron(m) Ion BY ALAN COX AND TERENCE J. KEMP* Department of Molecular Sciences, University of Warwick, Coventry CV4 7AL Received 21.d May, 1975 Iron(ur), both as its perchlorate and chloride salts, is an effective photo-oxidant for a wide range of organic compounds to give radical species readily trapped and detected by e.s.r. spectroscopy at 77 K. Its general behaviour parallels that of cerium(rv) photo-oxidation in that C-C fission processes are prominent with tertiary alcohols and carboxylic acids, although with simple alcohols, amides and certain other types of molecule, abstraction of a hydrogen atom from an activated C-H bond is found. The closely similar behaviour of the two salts towards the compounds we have examined suggests that Cl atoms are not essential intermediates in FeJII photo-oxidations, as has been previously suggested, although there are photo-oxidations by FeCl, in which chlorine is incorporated into the isolabte product.In Parts 1-3 1-3 we showed that photolysis of solutions of cerium(1v) salts in various organic media at 77 K using light absorbed only by CeIV species (usually a charge-transfer complex with the solvent) leads to the efficient production of solvent- derived radicals readily detectable by e.s.r. spectroscopy. Experiments with organic substrates dissolved in dilute perchloric acid and solutions of cerium(1v) salts gave essentially similar results. In Part 4 the behaviour of uranyl salts was examined under similar experimental conditions ; again, organic radicals were produced, but often of a type different from those encountered in cerium(1v) photosensitisation.In view of the considerable literature existing on the photo-induced reactions of iron@) salts with organic substrates, we report here the results of a brief survey of the e.s.r. spectra generated by photo-oxidation of a representative variety of organic substrates, mostly by iron(rn) perchlorate, making comparison with the behaviour of CeIV and Uvl, and with previous data on FelIr where this is available. EXPERIMENTAL The procedures described in Part 1 were used. The radiation from a 100 W point- source Xe arc was filtered through Pyrex. Solutions were made up in neat solvent where the organic molecule of interest was a liquid, otherwise saturated aqueous solutions were used.Where iron(m) perchlorate was comparatively insoluble, iron(m) chloride was used instead. [Fe1Ir] was typically 0.05 mol dm-3. RESULTS ALCOHOLS METHANOL A 1 : 2 : 1 triplet was produced with QH = 1.98 mT and g = 2.002 4 which, by comparison with our previous data and those in the literature. is assigned to CH20H radical. This result agrees with that of Kryukov et aLY6 who used both perchlorate and chloride salts. 2490A . COX AND T . J . KEMP 249 1 ETHANOL A basic five-line spectrum of binomial intensity distribution was produced in agreement with the result of Kryukov et u Z . , ~ which is identical to that found in both CeIV and Uvl photo-oxidations,'* and which is assigned to CH,tHOH.As with Cexy, however, additional small peaks appeared near the maxima of the second and third absorption lines (numbering from low field) indicative of the presence of a second, minor radical. PROPAN-~-OL Photoreaction was much slower in this instance and the spectrum was far less intense than that found with either of the lower alcohols or using CeIV as oxidant1 It consisted of a central singlet flanked by a number of weaker, but sharp and repro- ducible peaks. This substrate gave complex behaviour with CeIV and (in the liquid state) yielded only a very weak spectrum with Uv* as photo-oxidant. t-BUTYL ALCOHOL A relatively sharp 1 : 3 : 3 : 1 quartet was obtained with aH - 2.2 mT. runs traces were apparent of an additional 1 : 2 I 1 triplet with aH 1.89 mT.species are assigned to CH3 and -CH,C(CH,),OH respectively. In some These 2 - MET HY L B U T A N - 2 - 0 L( t - A M Y L A 1, C OH 0 L) Photoreaction was again rather slower than with CeIV. At 77 K a very clear spectrum of ethyl radical gradually emerged ; on warming the samples to - 150 K the spectrum changed to a 1 : 2 : 1 triplet suggestive of attack of C2H5 upon substrate to yield a CH2X species. BUTAN-2-OL A radical mixture was obtained comprising a central singlet and other peaks typical of C2H5 radical (major coupling 2.6 mT). 3 -METHY LHEXAN- 3 -0L Quite rapid photolysis occurred to yield a pure spectrum of ethyl radical. BENZYL ALCOHOL Quite rapid photolysis gave a broad singlet showing hyperfine coupling suggestive of a benzyl-type radical (possibly C6H&HOH).CARBOXYLIC ACIDS ACETIC ACID At 77 K a radical mixture (of approximately equal concentrations) was obtained assigned to .CH3 and CH2C02H from the couplings and g-factors. On warming, the methyl radical disappeared leaving -CH,C02H which displayed the anisotropy apparent in the spectrum of Ayscough et a1.' PROPIONIC ACID The principal feature was a five-line spectrum of binomial intensity distribution together with two quite sharp additional peaks with al, 2.6 mT, suggestive of ethyl2492 E . S . R . OF PHOTO-OXIDATION PRODUCTS radical. The quintet coupling of 2.35 mT suggests assignment to MeCHC0,H (the remaining peaks of the ethyl radical are submerged under those of the quintet). ISOBUTYRIC ACID An intense, very well-defined seven-line spectrum was produced with a binomial intensity distribution and a, 2.06 mT, due clearly to MezeCO2H (and not the isoprop yI radical).ISOVALERIC ACID An intense five-line spectrum resulted (a, - 1.89 mT) with the appearance of that for y-irradiated isobutyl halides suggestive described by Ayscough and Thomson of the formation of isobutyl radicals. VINYLACETIC ACID A very intense five-line spectrum of binomial intensity distribution and with aH = 1.51 mT was given, suggesting the formation of ally1 radical. MISCELLANEOUS SUBSTRATES FORMAMIDE An intense five-line spectrum was obtained, identical in appearance to that obtained by Bosco et aL9 for matrix-isolated CONH2 radical. (One proton coupling is too small to be resolved and one line of the six expected is submerged in the broad central peak.) Our couplings are slightly larger, aN = 2.3 mT, aH = 3.5 mT. DIMETHYLFORMAMIDE A very intense 1 : 2 : 1 triplet was produced with a, = 1.89 mT due to the radical HCON(Me)CH2*, identical to that found on y-radiolysis of this material at 77 K," PENTAN-3-ONE A quintet of binomial intensity distribution was given with aH = 2.20 mT assigned to the radical MecHCOEt.ACETALDEHYDE The spectrum consisted of a basic singlet centred on g,, = 2.001 7 showing considerable hyperfine structure to both high and low field with coupling of - 1.0 mT. The nature of this radical is obscure ; MeCO has gav = 2.0005 while CH2CH0 has = 2.07 mT. TETRAHYDROFURAN AND 2-METHYLTETRAHYDROFURAN Weak spectra exhibiting fine structure were given.They did not correspond to those expected of the a-radical, produced during y-radiolysis,1° although we did find cerium(1v) ammonium nitrate photo-oxidation of tetrahydrofuran to produce the a-radical, with an intensity pattern 1 : 2 : 2 : 2 : 1 (aav 1.89 mT). The a-radical, MeCHOEt has been observed during FeC1, photo-oxidation of diethyl ether,l and we also found it during photo-oxidation by cerium(iv) ammonium nitrate at 77 K with a, = 1.98 mT.A. COX AND T . J . KEMP 2493 LACTIC ACID An intense absorption was obtained at 77 K which we regard as due to a radical mixture. Its basic pattern is a 1 : 4 : 6 : 4 : 1 quintet, but the coupling constant varied somewhat from peak to peak, averaging to 1.67 mT which seems too low a figure to be attributed to CH,eHOH.Curiously, Poznyak et aZ.12 found aqueous ethylene glycol solutions of FeIrl complexes of lactate and mandelate ions to photodecompose at 77 K (indicated by loss of the e.s.r. line of FelI1) but saw no production of RcHOH radicals until the matrix was warmed to 140 K ; this " delayed action '' in radical production was rationalised by the authors in terms of an [Fell-RcHOH] adduct which they believe to show no e.s.r. absorption; this thermally dissociates at 140 K into Fe" + RCHOH. DISCUSSION Thermodynamically Fe"' is intermediate in activity between Ce" and Uv' on which we have previously concentrated. l-' Whilst CeIV photo-oxidations of organic substrates involve light absorption in the charge-transfer band of the CeIv-substrate complex (A,,, - 300 nm or 361 kJ mol-l) together with net reduction of CeIv to Ce"', for which E" = 1.70 V in HC104 solution (1 mol dm-,), equivalent to a free energy change of 164 kJ mol-', the corresponding figures for UO$+ are Amax = 400 nm (or 271 kJmol-l) and E" = 0.05 V (or AGO = 4.8 kJ mol-') and for Fe3+ in HC104 solutions with added substrate, Amax x 350 nm (or 309 kJ mol-l) and E" = 0.772 V (or AGO = 74.5 kJ mol-l).Both as regards the photochemical and electrochemical terms, then, Fe"' might be expected to show behaviour intermediate between Ce'" and Uvl in its photochemical interaction with organic substrates. In its interaction with the two simplest alcohols, excited iron(1rr) ion, denoted FeI1I4:, behaves both as CeIV* in abstracting a hydrogen atom from the hydroxylic carbon atom to give RCHOH (R = H or Me).This takes place with both FeCI, and Fe(ClO,),, indicating the process to be due to a genuine attack by FelI1* rather than by intermediate formation of a chlorine atom,6 followed by a secondary attack of the latter upon the matrix. While both of these mechanisms are conceivable for FeCl,, the CI. atom mechanism cannot operate in the case of Fe(C104),, nor can it explain the formation of readily detectable alkyl radicals as a principal path- way in the Fe"'* oxidation of tertiary alcohols which is found with both Fe(C104), and FeCl,. These originate from a C-C cleavage photoprocess previously found with CeIV*, but not with Uvl*, viz., and Uv'* R1R2R3COH + FelI1* + R1R2R3&H + Fel1 (1) (2) + - R1R2R3COH 4 H+ + R1* + R2R3C=0 Towards carboxylic acids, FelI1* shows ambivalent behaviour.[steps (1) and (2) may be concerted). The radical derived from RC02H, where R = Me, Et, (Me),CHCH, and CH2=CHCH2, is predomi- nantly or exclusively R e , which comes from a process of oxidative decarboxylation found with CeIV* and PbtV* ; l4 with isobutyric acid, however, the seven-line spectrum must be due to (Me),CC02H, i.e., the route is one of hydrogen-atom abstraction [the possibility that the primary isopropyl radical is highly reactive towards the labile tertiary hydrogen atom at 77K is ruled out by the observation of Me,CH under these conditions during CeIV* oxidation], Clearly FelI1* behaves very much like CeIV* [and unlike Uvr';] in these particular oxidations, with the exception of isobutyric acid towards which it behaves like Uvl*.2484 E.S.R.OF PHOTO-OXIDATION PRODUCTS FelI1* attacks formamide and dimethylformamide by an H-abstraction process to give CONH2 and HCON(Me)cH2, resembling both CerV* and Uv1*.4 It is un- necessary, therefore, to invoke the intermediacy of C1 atoms in the FeC1, photo- reaction with dimethylformamide, although such intermediacy cannot be excluded. The significance of the anion has been demonstrated l6 in the FeCT3 photo-oxidation of succinic, glutaric and adipic acids, when the o-chloro-derivative of the correspund- ing mono-acid is produced, suggesting attack of H02C[CH2],CH2* upon unreacted FeGl3. Incorporation of the chlorine into certain of the products also occurs during the FeC13 photo-oxidation of diethyl ether.I1 With diethyl ketone, the MeCIHCO- CH,Me radical was produced, as with Uvl* and Ce'v*,3 although the latter gave additional peaks of C,IES.The acetyl radical given by CeIV* attack on acetalde- hyde was not found with FelI1*, which gave a spectrum yet to be assigned. To summarise, FelI1 is an effective photo-oxidant for a wide range of organic substrates, behaving more like CeIV than Uvl although differing from the former in points of detail. Its effectiveness does not depend on the presence of chloride ligands to give reactive Cb intermediates, as iron(zI1) perchlorate yields the same radicals as does FeCi, and moreover in certain oxidations the product radicals are inconsistent with attack by CI-. We thank the S,R.C. for a grant to purchase the spectrometer, and Mr.S. H. Jenkins for both recording the spectra and maintenance of the spectrometer. I D. Greatorex and T. J. Kemp, Trms. Faraday SOC., 1971,67, 56. D. Greatorex and T. J. Kemp, Trans, Favaday SOC., 1971, 67, 1576. D. Greatorex and T. J. Kemp, J.C.S. Fmadzy I, 1972, 68, 121. D. Greatorex, R. J. HiII, T. J. Kemp and T. J. Stone, J.C.S. Faraday f, 1972, 68,2059. V. Balzani and V. Carassiti, Photochemistry of Cmrdination Compounds (Academic Press, New York and London, 1974), chap. 10. A. I. Kryukov, L. V. Nazarova and B. Y. Dain, Ukrain. khim. Zhur., 1963,29, 812. P. B. Ayscough, K. Mach, J. P. Oversby and A. K. Roy, Truns. Fmaday Soc., 1971,67,360. P. B. Ayscough and C. Thomson, Trans. Faraday Soc., 1962,58, 1,477. S. R. Bosco, A Circillo and R. B.Timmons, J. Amer. Chem. SOC., 1969, 91, 3140. S. A. Ivanitskaya, A. N. Korol and A. I. Kryukov, Ukrdn. khim. Zhur.,, 1973,39,1248. 1260. lo V. 1. Trofimov and I. I. Chkeidze, High Energy Chem., 1967, 1, 282. l2 A. L. Poanyak, G. A. Shaghitanova and S. I. Arzhankov, Russ. J. lnorg. Chem., 1970, 15, l 3 W. M. Latimer, Oxidation Potentials (Prentice-Hall, New York, 2nd edn., 1952). l4 K. Heusler and H. Loeliger, Heto. Chim. Aci'a, 1969,52,1495. l 5 F. S. Dainton and R. G. Jones, Trans. Faraduy SOC., 1967,63, 1512. l6 J. A. Kuhnle, R. E. Lundin and A. C. Waiss, Jr., Chem. Comm., 1972, 287. Electron Spin Resonance Studies of Photo-oxidation by Metal Ions in Rigid Media at Low Temperatures Part 5.-Photo-oxidation by the Iron(m) Ion BY ALAN COX AND TERENCE J. KEMP* Department of Molecular Sciences, University of Warwick, Coventry CV4 7AL Received 21.d May, 1975 Iron(ur), both as its perchlorate and chloride salts, is an effective photo-oxidant for a wide range of organic compounds to give radical species readily trapped and detected by e.s.r.spectroscopy at 77 K. Its general behaviour parallels that of cerium(rv) photo-oxidation in that C-C fission processes are prominent with tertiary alcohols and carboxylic acids, although with simple alcohols, amides and certain other types of molecule, abstraction of a hydrogen atom from an activated C-H bond is found. The closely similar behaviour of the two salts towards the compounds we have examined suggests that Cl atoms are not essential intermediates in FeJII photo-oxidations, as has been previously suggested, although there are photo-oxidations by FeCl, in which chlorine is incorporated into the isolabte product.In Parts 1-3 1-3 we showed that photolysis of solutions of cerium(1v) salts in various organic media at 77 K using light absorbed only by CeIV species (usually a charge-transfer complex with the solvent) leads to the efficient production of solvent- derived radicals readily detectable by e.s.r. spectroscopy. Experiments with organic substrates dissolved in dilute perchloric acid and solutions of cerium(1v) salts gave essentially similar results. In Part 4 the behaviour of uranyl salts was examined under similar experimental conditions ; again, organic radicals were produced, but often of a type different from those encountered in cerium(1v) photosensitisation.In view of the considerable literature existing on the photo-induced reactions of iron@) salts with organic substrates, we report here the results of a brief survey of the e.s.r. spectra generated by photo-oxidation of a representative variety of organic substrates, mostly by iron(rn) perchlorate, making comparison with the behaviour of CeIV and Uvl, and with previous data on FelIr where this is available. EXPERIMENTAL The procedures described in Part 1 were used. The radiation from a 100 W point- source Xe arc was filtered through Pyrex. Solutions were made up in neat solvent where the organic molecule of interest was a liquid, otherwise saturated aqueous solutions were used. Where iron(m) perchlorate was comparatively insoluble, iron(m) chloride was used instead.[Fe1Ir] was typically 0.05 mol dm-3. RESULTS ALCOHOLS METHANOL A 1 : 2 : 1 triplet was produced with QH = 1.98 mT and g = 2.002 4 which, by comparison with our previous data and those in the literature. is assigned to CH20H radical. This result agrees with that of Kryukov et aLY6 who used both perchlorate and chloride salts. 2490A . COX AND T . J . KEMP 249 1 ETHANOL A basic five-line spectrum of binomial intensity distribution was produced in agreement with the result of Kryukov et u Z . , ~ which is identical to that found in both CeIV and Uvl photo-oxidations,'* and which is assigned to CH,tHOH. As with Cexy, however, additional small peaks appeared near the maxima of the second and third absorption lines (numbering from low field) indicative of the presence of a second, minor radical.PROPAN-~-OL Photoreaction was much slower in this instance and the spectrum was far less intense than that found with either of the lower alcohols or using CeIV as oxidant1 It consisted of a central singlet flanked by a number of weaker, but sharp and repro- ducible peaks. This substrate gave complex behaviour with CeIV and (in the liquid state) yielded only a very weak spectrum with Uv* as photo-oxidant. t-BUTYL ALCOHOL A relatively sharp 1 : 3 : 3 : 1 quartet was obtained with aH - 2.2 mT. runs traces were apparent of an additional 1 : 2 I 1 triplet with aH 1.89 mT. species are assigned to CH3 and -CH,C(CH,),OH respectively. In some These 2 - MET HY L B U T A N - 2 - 0 L( t - A M Y L A 1, C OH 0 L) Photoreaction was again rather slower than with CeIV.At 77 K a very clear spectrum of ethyl radical gradually emerged ; on warming the samples to - 150 K the spectrum changed to a 1 : 2 : 1 triplet suggestive of attack of C2H5 upon substrate to yield a CH2X species. BUTAN-2-OL A radical mixture was obtained comprising a central singlet and other peaks typical of C2H5 radical (major coupling 2.6 mT). 3 -METHY LHEXAN- 3 -0L Quite rapid photolysis occurred to yield a pure spectrum of ethyl radical. BENZYL ALCOHOL Quite rapid photolysis gave a broad singlet showing hyperfine coupling suggestive of a benzyl-type radical (possibly C6H&HOH). CARBOXYLIC ACIDS ACETIC ACID At 77 K a radical mixture (of approximately equal concentrations) was obtained assigned to .CH3 and CH2C02H from the couplings and g-factors.On warming, the methyl radical disappeared leaving -CH,C02H which displayed the anisotropy apparent in the spectrum of Ayscough et a1.' PROPIONIC ACID The principal feature was a five-line spectrum of binomial intensity distribution together with two quite sharp additional peaks with al, 2.6 mT, suggestive of ethyl2492 E . S . R . OF PHOTO-OXIDATION PRODUCTS radical. The quintet coupling of 2.35 mT suggests assignment to MeCHC0,H (the remaining peaks of the ethyl radical are submerged under those of the quintet). ISOBUTYRIC ACID An intense, very well-defined seven-line spectrum was produced with a binomial intensity distribution and a, 2.06 mT, due clearly to MezeCO2H (and not the isoprop yI radical).ISOVALERIC ACID An intense five-line spectrum resulted (a, - 1.89 mT) with the appearance of that for y-irradiated isobutyl halides suggestive described by Ayscough and Thomson of the formation of isobutyl radicals. VINYLACETIC ACID A very intense five-line spectrum of binomial intensity distribution and with aH = 1.51 mT was given, suggesting the formation of ally1 radical. MISCELLANEOUS SUBSTRATES FORMAMIDE An intense five-line spectrum was obtained, identical in appearance to that obtained by Bosco et aL9 for matrix-isolated CONH2 radical. (One proton coupling is too small to be resolved and one line of the six expected is submerged in the broad central peak.) Our couplings are slightly larger, aN = 2.3 mT, aH = 3.5 mT. DIMETHYLFORMAMIDE A very intense 1 : 2 : 1 triplet was produced with a, = 1.89 mT due to the radical HCON(Me)CH2*, identical to that found on y-radiolysis of this material at 77 K," PENTAN-3-ONE A quintet of binomial intensity distribution was given with aH = 2.20 mT assigned to the radical MecHCOEt.ACETALDEHYDE The spectrum consisted of a basic singlet centred on g,, = 2.001 7 showing considerable hyperfine structure to both high and low field with coupling of - 1.0 mT. The nature of this radical is obscure ; MeCO has gav = 2.0005 while CH2CH0 has = 2.07 mT. TETRAHYDROFURAN AND 2-METHYLTETRAHYDROFURAN Weak spectra exhibiting fine structure were given. They did not correspond to those expected of the a-radical, produced during y-radiolysis,1° although we did find cerium(1v) ammonium nitrate photo-oxidation of tetrahydrofuran to produce the a-radical, with an intensity pattern 1 : 2 : 2 : 2 : 1 (aav 1.89 mT).The a-radical, MeCHOEt has been observed during FeC1, photo-oxidation of diethyl ether,l and we also found it during photo-oxidation by cerium(iv) ammonium nitrate at 77 K with a, = 1.98 mT.A. COX AND T . J . KEMP 2493 LACTIC ACID An intense absorption was obtained at 77 K which we regard as due to a radical mixture. Its basic pattern is a 1 : 4 : 6 : 4 : 1 quintet, but the coupling constant varied somewhat from peak to peak, averaging to 1.67 mT which seems too low a figure to be attributed to CH,eHOH. Curiously, Poznyak et aZ.12 found aqueous ethylene glycol solutions of FeIrl complexes of lactate and mandelate ions to photodecompose at 77 K (indicated by loss of the e.s.r. line of FelI1) but saw no production of RcHOH radicals until the matrix was warmed to 140 K ; this " delayed action '' in radical production was rationalised by the authors in terms of an [Fell-RcHOH] adduct which they believe to show no e.s.r.absorption; this thermally dissociates at 140 K into Fe" + RCHOH. DISCUSSION Thermodynamically Fe"' is intermediate in activity between Ce" and Uv' on which we have previously concentrated. l-' Whilst CeIV photo-oxidations of organic substrates involve light absorption in the charge-transfer band of the CeIv-substrate complex (A,,, - 300 nm or 361 kJ mol-l) together with net reduction of CeIv to Ce"', for which E" = 1.70 V in HC104 solution (1 mol dm-,), equivalent to a free energy change of 164 kJ mol-', the corresponding figures for UO$+ are Amax = 400 nm (or 271 kJmol-l) and E" = 0.05 V (or AGO = 4.8 kJ mol-') and for Fe3+ in HC104 solutions with added substrate, Amax x 350 nm (or 309 kJ mol-l) and E" = 0.772 V (or AGO = 74.5 kJ mol-l).Both as regards the photochemical and electrochemical terms, then, Fe"' might be expected to show behaviour intermediate between Ce'" and Uvl in its photochemical interaction with organic substrates. In its interaction with the two simplest alcohols, excited iron(1rr) ion, denoted FeI1I4:, behaves both as CeIV* in abstracting a hydrogen atom from the hydroxylic carbon atom to give RCHOH (R = H or Me). This takes place with both FeCI, and Fe(ClO,),, indicating the process to be due to a genuine attack by FelI1* rather than by intermediate formation of a chlorine atom,6 followed by a secondary attack of the latter upon the matrix.While both of these mechanisms are conceivable for FeCl,, the CI. atom mechanism cannot operate in the case of Fe(C104),, nor can it explain the formation of readily detectable alkyl radicals as a principal path- way in the Fe"'* oxidation of tertiary alcohols which is found with both Fe(C104), and FeCl,. These originate from a C-C cleavage photoprocess previously found with CeIV*, but not with Uvl*, viz., and Uv'* R1R2R3COH + FelI1* + R1R2R3&H + Fel1 (1) (2) + - R1R2R3COH 4 H+ + R1* + R2R3C=0 Towards carboxylic acids, FelI1* shows ambivalent behaviour. [steps (1) and (2) may be concerted). The radical derived from RC02H, where R = Me, Et, (Me),CHCH, and CH2=CHCH2, is predomi- nantly or exclusively R e , which comes from a process of oxidative decarboxylation found with CeIV* and PbtV* ; l4 with isobutyric acid, however, the seven-line spectrum must be due to (Me),CC02H, i.e., the route is one of hydrogen-atom abstraction [the possibility that the primary isopropyl radical is highly reactive towards the labile tertiary hydrogen atom at 77K is ruled out by the observation of Me,CH under these conditions during CeIV* oxidation], Clearly FelI1* behaves very much like CeIV* [and unlike Uvr';] in these particular oxidations, with the exception of isobutyric acid towards which it behaves like Uvl*.2484 E.S.R. OF PHOTO-OXIDATION PRODUCTS FelI1* attacks formamide and dimethylformamide by an H-abstraction process to give CONH2 and HCON(Me)cH2, resembling both CerV* and Uv1*.4 It is un- necessary, therefore, to invoke the intermediacy of C1 atoms in the FeC1, photo- reaction with dimethylformamide, although such intermediacy cannot be excluded.The significance of the anion has been demonstrated l6 in the FeCT3 photo-oxidation of succinic, glutaric and adipic acids, when the o-chloro-derivative of the correspund- ing mono-acid is produced, suggesting attack of H02C[CH2],CH2* upon unreacted FeGl3. Incorporation of the chlorine into certain of the products also occurs during the FeC13 photo-oxidation of diethyl ether.I1 With diethyl ketone, the MeCIHCO- CH,Me radical was produced, as with Uvl* and Ce'v*,3 although the latter gave additional peaks of C,IES. The acetyl radical given by CeIV* attack on acetalde- hyde was not found with FelI1*, which gave a spectrum yet to be assigned.To summarise, FelI1 is an effective photo-oxidant for a wide range of organic substrates, behaving more like CeIV than Uvl although differing from the former in points of detail. Its effectiveness does not depend on the presence of chloride ligands to give reactive Cb intermediates, as iron(zI1) perchlorate yields the same radicals as does FeCi, and moreover in certain oxidations the product radicals are inconsistent with attack by CI-. We thank the S,R.C. for a grant to purchase the spectrometer, and Mr. S. H. Jenkins for both recording the spectra and maintenance of the spectrometer. I D. Greatorex and T. J. Kemp, Trms. Faraday SOC., 1971,67, 56. D. Greatorex and T. J. Kemp, Trans, Favaday SOC., 1971, 67, 1576. D. Greatorex and T. J. Kemp, J.C.S. Fmadzy I, 1972, 68, 121. D. Greatorex, R. J. HiII, T. J. Kemp and T. J. Stone, J.C.S. Faraday f, 1972, 68,2059. V. Balzani and V. Carassiti, Photochemistry of Cmrdination Compounds (Academic Press, New York and London, 1974), chap. 10. A. I. Kryukov, L. V. Nazarova and B. Y. Dain, Ukrain. khim. Zhur., 1963,29, 812. P. B. Ayscough, K. Mach, J. P. Oversby and A. K. Roy, Truns. Fmaday Soc., 1971,67,360. P. B. Ayscough and C. Thomson, Trans. Faraday Soc., 1962,58, 1,477. S. R. Bosco, A Circillo and R. B. Timmons, J. Amer. Chem. SOC., 1969, 91, 3140. S. A. Ivanitskaya, A. N. Korol and A. I. Kryukov, Ukrdn. khim. Zhur.,, 1973,39,1248. 1260. lo V. 1. Trofimov and I. I. Chkeidze, High Energy Chem., 1967, 1, 282. l2 A. L. Poanyak, G. A. Shaghitanova and S. I. Arzhankov, Russ. J. lnorg. Chem., 1970, 15, l 3 W. M. Latimer, Oxidation Potentials (Prentice-Hall, New York, 2nd edn., 1952). l4 K. Heusler and H. Loeliger, Heto. Chim. Aci'a, 1969,52,1495. l 5 F. S. Dainton and R. G. Jones, Trans. Faraduy SOC., 1967,63, 1512. l6 J. A. Kuhnle, R. E. Lundin and A. C. Waiss, Jr., Chem. Comm., 1972, 287.

ISSN:0300-9599    

DOI:10.1039/F19757102490

出版商:RSC    

年代:1975

数据来源: RSC