610 J.C.S. Perkin IElectrochemical Syntheses. Part V1I.l Isomer Distribution in the PartiafOxidation of n-AlkanesBy Heinz P. Fritz" and Thomas Wurminghausen, Institute for Inorganic Chemistry, Technical University ofMunich, 8 Munchen 2, ArcisstraRe 21, West GermanyThe electrochemical oxidation of alkanes in weak acidic solution (CH,CI,-CF,*CO,H) was investigated. Thenon-statistical isomer distribution of n-decane oxidation products seems to parallel slight differences in the reac-tivity of the CH2 groups. A mechanism involving free carbocations is suggested.ALKANES can be oxidized both chemically, e.g. insolutions of cobalt(n1) acetate in acetic acid withaddition of a strong acid or of lead(1v) or cobalt(II1)trifluoroacetate in trifluoroacetic acid, arid electro-chemically, e.g.in some aprotic in fluoro-sulphuric acidJ8 and in trifluoroacetic acid.gIn the course of synthetic work on the electro-chemical oxidation of alkanes,1deg; we recently demon-strated that the oxidation potentials of certain alkanesare lowered (to more cathodic potentials) not only influorosulphuric acid (a ' super acid ') as compared withaprotic media, but also in the more weakly acidicsolutions of, for example, fluorosulphuric acid, methane-sulphonic acid, or trifluoroacetic acid in acetic acid.On the basis of electrochemical measurements weassumed the previously postulated * rapid protonationreaction to occur prior to the electrochemical reactionin the oxidation of alkanes. For our synthetic purposesit was of interest to know whether the decrease inalkane oxidation potential occurs only after protonationby very strong acids or if it is a general result of thepresence of protic substances.A preliminary study ofour reaction system was carried out by voltammetry atstationary platinum electrodes. Mixtures of trifluoro-acetic acid, acetic acid, and dichloromethane were usedas solvents.RESULTS AND DISCUSSIONVoltammetry.-Cyclic voltammetry at ' inert ' tstationary platinum electrodes was applied. InCH,CI,-O. ~M-(Bu~N)PF~ as well as in cH,Cl,--6.5~-MeC0,H-O. 1M-(Bu4N)PF6 no oxidation peak for cyclo-hexane was observed. Only in the extreme potentialregion of solvent decomposition was it possible tomeasure an additional weak current due to substrateoxidation. In MeCO2H4.O~-CF3-CO2H (t rifluoroaceticacid is weak in acetic acid 12) an additional current duet The half-wave oxidation potentials for cyclohexane a t Au,Part VI, G.Bockmair and H. P. Fritz, Electrochim. Actu, inJ. Hanotier, J.C.S. Perkin 11, 1972, 2247.R. Tang and J. K. Kochi, J . Inorg. Nuclear Chem., 1976,D. B. Clark, M. Fleischmann, and D. Pletcher. J . Electro-M. Fleischmann and D. Pletcher, Tcfralzedron Letters. 1968,7 T. M. Siegel, L. L. Miller, and J. Y . Becker J.C.S. Chem.8 J. Bertram, J. P. Coleman, M. Fleischmann, and D. Pletcher,Rh, and I r are similar to that a t Pt anodes.11the press.3 R. E. Patch, J . Anzer. Chem. Soc., 1967, 89, 3662.35, 3845.analyt. Chern., 1973, 42, 133.6266.Comm., 1974, 34.J.C.S.Perkin 11, 1973, 374.to cyclohexane oxidation again could only be detectedin the extreme potential region of solvent decomposition.Different results were obtained with CH2C1,-CF3*C0,Hmixtures. With increasing CF,*CO,H concentration weobserved a cathodic shift in the oxidation potential forcyclohexane (Figure). This shift might be due to aCF3.C0,H I / mol 1-'Epl2 vs. CF,CO,H(~M) in CH2C12--0.1~-Bu,NPF,-3.7 xM-C~H,,; Scan rate ( v ) = 60 mV s-lchange in alkane activation energyl3 and/or to achemical reaction following electron transfer (Le. anucleophilic reaction of the elFctrochemically generatedspecies with CF3C0,H). Comparing our cyclovolt-ammetric data with the diagnostic criteria of Nicholson,Shain, and Wopschall 14-16 we only found qualitativecorrespondence to a reversible electrochemical reactionfollowed by a rapid chemical reaction.Having con-sidered the criticisms levelled at interpretations basedon electroanalytical measurements l7 we favour a directoxidation without change of activation energy (e.g. byproton assistance) and a trapping reaction by the solventfor alkane oxidation in these weakly acidic media.Preparative EZectrolyses.-The preparative electro-lysis of cyclohexane in CH,C~~-~.~M-CF,*CO,H yieldedcyclohexyl trifluoroacetate (91 yo) and cyclohexyl chloride(4). The reaction in this nucleophilic medium did notlead to solvolysis products of the methylcyclopentylD. B. Clark, M. Fleischmann, and D. Pletcher, J.C.S. Perkinlo H. P. Fritz and Th.Wiirminghausen, f. Electroanalyt.11 H. P. Fritz and Th. Wiirminghausen, unpublished results.l2 B. M. Rode, A. Engelbrecht, and J. Schantl, 2. Phys. Chern.,l3 K. J . Vetter, ' Elektrochemische Kinetik,' Springer-Verlag,l4 R. H. Wopschall and I. Shain, Analyt. Chem., 1967, 39,l5 R. S. Nicholson and I. Shain, AnuZyt.,Chem., 1964, 36, 707.lE S. Piekarski and R. N. Adams, Physical Methods ofChemistry,' Part IIa, eds. Weissberger and Rossiter, Wiley-Interscience. New York, 1971, p. 566.17 R. N. Adams, ' Electrochemistry a t Solid Electrodes,'Dekker, New York, 1969, p. 255.11, 1973, 1578.Chem., 1974, 54, 18.1973, 253, 112, 17.Berlin, 1961.16141976 61 1ion, which according to Olah et at?. and Fleischmannet aZ.,8 is more stable in ' super acids ' than the cyclohexylion.Preparative electrolyses of n-decane (which is moredifficult to oxidise than cyclohexane) were performed inthe systems CH2C12-6.5~-CF,*C0,H-0.05M- ( Bu4N) PF,and MeC02H4~-CF,*C02H-1~-MeC02Na.In the firstsystem we obtained isomeric trifluoroacetoxy-n-decanes(68) with a current yield of 36 yo, bistrifluoroacetoxy-n-decanes (1 1 yo), and non-distillable, highly viscousproducts (21) after reaction of 90 of the n-decane.Neither olefins nor rearrangement products (secondaryto tertiary) were detected. The trifluoroacetoxy-n-decane isomer distribution is given in the Table (a).The isomer ratios varied by less than 2 during theelectrolysis. In the second system Table (b) weand found an isomer distribution Table (d)) similar toour other results. A change from divided to undividedcells did not affect the oxidation products.The change from n-octane to n-decane should not beresponsible for the observed drastic change in isomerdistribution.From the work of Asinger et all8 onalkane radical reactions (showing statistical isomerdistributions) it is known that the finding of a non-statistical isomer ratio can be due to systematic errorsin the analytical work-up of products. We thereforetook special care not to introduce any ' artificial'alteration of the experimental isomer distribution. Theelectrolyte-product mixtures were analysed by g.1.c.either directly or after extraction of the esters withether, and the results were checked against a mixture ofauthentic trifluoroacetoxy-n-decanes. The results didIsomer distribution of n-decane oxidation products (mono-acetoxy- or -trifluoroacetoxy-n-decanes)Isomer : 1- 2- 3- 4- 5-(a) Trifluoroacetates 14 22 28 36Standard d e n .(m = 14) 0.3 0.3 0.3 0.4Rel. tfi bne 4.66 3.10 2.72 3.41 2.28(b) Trifluoroacetates ,JAcetatesRel. t R of acetates17 22 28 3316 18 29 376.38 4.81 4.16 3.76 3.68(c) Trifluoroacetates 0 15 22 27 36(d) Trifluororoacetates a 16 23 27 34(e) Calc. rel. yo 8 2 8 20 32 38(f) RH+ - RH charge difference 0.7 13.4 20.8 29.7 35.4Kel. yields (g.1.c.) (). ' b With respect to n-decane (rel. tB 1.00). Column of di-n-decyl phthalate. Column of ApiezonFrom ref. 23. f From ref. 25.obtained acetoxy- and trifluoroacetoxy-n-decanes in theratio 1.8 : 1 (the molar ratio of the acids was 3.3 : 1) andwith a similar isomer distribution.The result of the electrochemical oxidation of n-octaneat platinum electrodes in CF3*C02H-0.4~-(Bu,N)BF4is in contrast to the isomer distribution found by uswith n-decane: from the previous experiment a strictlystatistical ratio of the isomeric 2-, 3-, and 4-trifluoro-acetoxy-n-octanes is reported, whereas we obtained thenon-statistical isomer distribution shown in the Table (c)after controlled potential electrolysis of n-decane inCF,gC0,H-0.05~-(B~4N)PF6.We considered that thisdiscrepancy might be ascribed to the ca. ten-foldsupporting electrolyte concentration used by Fleisch-mann et al.; in that experiment the Helmholtz doublelayer could mainly be built up by aprotic alkyl groupsand so a different reaction path might be induced.Wetherefore tested a solution in CF3*C02H-0.4~-(Bu,N) PF,* Studies of rearrangement products in the anodic oxidationof branched alkanes strongly hint a t ' free * carbocation inter-mediates.**F. Asinger and K. Hdcour, Clzem. Ber., 1961, 94, 83, andreferences cited there.l* J. T. Keating and P. S. Skell, ' Carbonium Ions,' vol. 11,eds. G. A. Olah and P. von R. Schleyer, Wiley-Interscience,New York, 1970.not change during g.1.c. analysis, nor was the isomerdistribution altered in pure trifluoroacetic acid, in whichacid-catalysed solvolysis might occur. We thereforepostulate that this non-statistical isomer distribution isnot due to a radical reaction path and that the numberof electrons transferred (12) isIn view of the general observation that the weakestbond in a molecule is most readily oxidised, it can beunderstood why no l-isomer is obtained.As it is un-likely that a given methylene group of an alkane chainis preferentially orientated towards the electrode, theisomer distribution should be directly representative ofthe electronic structure of the intermediate carbocation,which should be ' hot ' or ' free ' Is** rather than ' en-cumbered ' 21 in view of its mode of formation. Fromthe concept of inductive effects 22 one would expect theprobability of finding a positively charged carbon atomto increase towards the centre of the molecule. Thenet positive charge distribution in the carbon skeleton ofan n-decane radical cation has been calculated by2o H.P. Fritz and Th. Wiirminghausen, unpublished results.21 H. Maskill, R. M. Southam, and M. C. Whiting, Chem.See e.g. calculations of S. Fliszar et nl., J. Amer. Chem. SOG.,Comm., 1966, 20, 496.1972,94, 1068; 1974, 96, 4352, 4358612 J.C.S. Perkin ILorquet.= Assuming a similar relative charge distri-bution for R+, we estimate the relative isomer distri-bution of the products of reactions of such ions with anucleophile to be as in the Table (e). The C-H bondcharge density was calculated to be significant only atposition 2. This would presumably lead to an increasein the percentage of the isomer. In the light of thesenumerical values the similarity of the experimental andthe estimated isomer distributions is reasonable.A recent INDO calculation on n-decane 25 yielded thevalues given in the Table (f) for the differences in chargedensity between the ion and the neutral molecule.Anisomer distribution similar to our experimental onewould be expected from this set of data.' Free carbocation ' reaction products obtained afterdeamination of n-octylamine or l-propylpentylamine 21or after decarboxylation of lauric acid l9 were to a greatextent the isomer with the functional group at the samecarbon atom. In direct hydrocarbon oxidation, as here,the isomer distribution is not influenced by a leavinggroup. In addition it is not necessary that the moststable carbocation be formed via rearrangement before* trapping,' * although the possibility of hydride shiftscannot be excluded.The isomer distribution of trifluoroacetoxy-n-decanesafter oxidation of n-decane with Et,NO(Et,NOH)FeII inCF3*C0,H found by Den0 and Poh126 differs from ourresults, although these authors also suggest a reactionpath via carbocations (' encumbered ' in this case).Kdditional experiments are in progress to answer theremaining questions.EXPERIMENTALThe cyclic voltaminetry cell and the apparatus havebeen described elsewhere.1deg; For electroanalyses a silver-10-2M-silver trifluoroacetate reference electrode in dichloro-rnethane-6.5~~trifluoroacetic acid-0.1M-tetrabutylammon-ium hexafluorophosphate separated by a D4 sinter wasused.During measurements with dichloromethane-tri-fluoroacetic acid the reference electrode was separated fromthe bulk solution by a dichlorornethane-6.5~-trifluoro-acetic acid-lM-tetrabutylammonium hexafluorophosphatebridge (D4) .a7 The potential of two identical referenceelectrodes in the various electrolytes was found to differby less than 3 mV.The potential drop due to resistancewas shown to be negligible by measuring with a TacusselCDCO instrument. For the cyclohexane oxidation thestandard deviation of 10 Ep, a measurements in dichloro-methane-'7. 8~-trifluoroacetic acid-0.h-tetrabutylammon-ium hexafluorophosphate was calculated to be 13 mV.For preparative work a Teflon cell with two 98cm2 platinum electrodes 1.7 mm apart and with a Normagglass-Teflon pump for forcing convection was employed.* krearr./kh.ca. 3 for the deoxidation of butan-2-01 (ref. 19,p. 580).23 J. Lorquet, ' Theory of Mass Spectrometry,' vol. 3, ed. W. L.Mead, Elsevier, Amsterdam, 1966.24 See, however, R. Hoffmann, J. Chem. Phys., 1964, 40,2480.z6 L. Eberson, personal communication.2o N. C. Deno and D. G. Pohl, J. Amer. Chem. SOC., 1974, 95,6680.The potential of the working electrode was measured fromthe rear of the electrode with saturated calomel electrode(s.c.e.), separated from the electrolyte by a D4 glass frit.Products were analysed by i.r. spectroscopy (BeckmannIRlO instrument), quantitative C and H determin-ations, and comparison of g.1.c. retention times with thoseof authentic samples on two different columns.The g.1.c.analyses were carried out on a Perkin-Elmer 116 E instru-ment with a thermal conductivity detector and columns ofdi-n-decyl phthalate on Chromosorb B or Apiezon M onCelite 545 at 160 "C.Reagents and G.2.c. Standards.-Trifluoroacetic acid(Roth) was distilled prior to use. Methylene chloride(Riedel-De Haen) was dried (P,O,,), distilled, and purifiedchromatographically on alumina. Tetra-n-butylammoniumhexafluorophosphate was precipitated from water by mix-ing aqueous tetra-n-butylammonium hydroxide (Fluka)and hexafluorophosphoric acid (Merck), filtered off, re-crystallised from ethyl alcohol, and dried in high vacuum at140 "C for 24 h. The trifluoroacetate standards wereprepared by mixing equimolar amounts of the correspond-ing alcohols and trifluoroacetic acid anhydride.29 Afterdistillation the trifluoroacetates were gas chromato-graphically pure.The absence of total rearrangement onesterification was tested by lH n.m.r. spectroscopy of the1- and 2-trifluoroacetoxy-n-decanes. They showed a tripletor a sextet, respectively, for the proton(s) adjacent toCF,CO, (6 4.3 and 5.1). The acetoxy-n-decanes wereprepared with sulphuric acid and acetic acid.,O The n-decane (analytical reagent) was used as supplied ; cyclo-hexane was treated with oleum (5) and dried withsodium prior to use. Cyclohexyl chloride standard wasused as supplied. For preparative electrolyses 110 mlof solvent containing supporting electrolyte and 0.1 moleof alkane were employed.Cyclohexane Oxidation.-The anode potential was pulsedfor 30 s a t 2.5 vs.s.c.e. and for 1 s at 0.15 V 71s. s.c.e.in dichloromethane-2.5~-trifluoroacetic acid-0.05~-tetra-butylammonium hexafluorophosphate. After 50 000 Chad passed and the cyclohexane had been consumed theelectrolysis was terminated and the electrolyte worked upby pouring it into water, neutralizing, extracting with ether,and evaporating the extract. The residue consisted of91 cyclohexyl trifluoroacetate and 4 cyclohexylchloride (by g.1.c.). The products were identified by g.1.c.retention times, separately and in mixtures. Three otherproducts (1, 1, and 3), were not identified.n-Decane Oxidation.-(i) The oxidation was carried outin dichloromethane-6.5~-trifluoroacetic acid-0.05w-tetra-butylammonium hexafluorophosphate galvanostatically at7 mA cm-, (2.6 V vs.s.c.e. average working potential) until90 of the n-decane had been consumed. (ii) The electro-lysis was done in acetic acid-4~-trifluoroacetic acid-lM-sodium acetate. The working potential of 2.6 V 'us. s.c.e.was applied for 30 s; this was followed by a 1 s pulse of0.3 V us. s.c.e. The oxidation was ended after the con-sumption of 47 n-decane. (iii) n-Decane was oxidised27 C. K. Mann and K. I. Barnes, Electrochemical Reactionsin Non-aqueous Systems, Dekker, New York, 1970, p. 17.25 R. Piontelli, G. Bianchi, U. Bertocci, C. Guerci, and B.Rivolta, 2. Elektrochenz., 1955, 59, 778.29 A. M. Lovelace, D. A. Rausch, and W. Postelnek, ' AliphaticFluorine Compounds,' Reinhold, New York, 1958.80 H.Henecka, 'Methoden der Organischen Chemie, Sauer-stoffverbindungen. 111, ed. Houben-Weyl, Thiemb Verlag,Stuttgart, 1952.Flow rates were 46 ml min-11976 613at 2.5 V us. s.c.e. for 10 s and at 0.4 V for 1 s in trifluoro-acetic acid-0.05~-tetrabutylammo~uni hexafluorophos-phate until 10 000 C had passed. (iv) The oxidation (iii)was repeated with 0.4~-tetrabutylammonium hexafluoro-phosphate instead of 0 . 0 5 ~ . Product identification wascarried out by direct injection of the electrolysis mixtureinto thc gas chromatograph and comparing g.1.c. retentiontimes separately and in mixtures. Relative yields werecalculated by comparing peak areas with those of g.1.c.standards. The standard deviations of run (i) are given inthe Table (a). The bistrifluoroacetoxy-n-decanes obtainedwere not resolved gas chromatographically. The corre-sponding fraction from vacuum distillation showed C ,45.85; H, 6.45 (Calc. for C,,H,,F,O,: C, 45.9; H, 5.45).The i.r. spectrum was typical of an alkyl trifluoroacetate(compared with authentic monotrifluoroacetoxydecane) :vmX. (film) 1 785s, 1 550w, 1 460m, 1 380m, 1 340m, 1 210vs,1 160vs, 970w, 860w, 770m, and 725m cm-l.The financial support of the Deutsche Forschungsgemein-schaft, Bad Godesberg, is gratefully acknowledged. Wethank Professor P. S. Skell for discussions and ProfessorL. Eberson and his co-workers for their INDO caIculations.6/1020 Received, 28th May, 1975
展开▼
