PART I DIF‘OSITIVE CARBONIUN IONS PART II PEROXQYTRIFLUOROACETIC ACIDeBORON FLUORIDE AS A SOURCE OF POSITIVE HYDROXYL Thesis Ior Hm Degree OI DI'I. D. MICHIGAN STATE UNIVERSITY Charles Allen Buehler 1963 .-_ LIBRA R Y Michigan State niversity w... IVERS ICHIC ‘- 3 CT .. . '\V-- '| x: | ,ATF ‘Jr‘.’ 1*.”"7’ "5",: ,-:’, .v- I; ITY .. .x-‘ -~ ABSTRACT PART I DIPOSITIVE CARBONIUM IONS PART II PEROXYTRIFLUOROACETIC ACID-BORON FLUORIDE AS A SOURCE OF POSITIVE HYDROXYL by Charles Allen Buehler The primary purpose of the first part of this thesis was to further investigate the scope of dipositive carbonium ionnchemistry. The aluminum chloride-catalyzed reaction of carbon tetrachloride with pentamethylbenzene (1-), was extended to the preparation of 3-halo (bromo land- chloro)-Z, 4, 5,6-tetramethylbenzotrichlorides and Z-chloro- 3, 4,5,6-tetramethy1benzotrichloride. « Each of these benzotrichloride derivatives dissolved in 100% sulfuric acid to form intensely red solutions and evolve hydrogen chloride. Hydrolysis of the colored solutions gave nearly quantitative yields of the corresponding 3- and 2-halotetramethylbenzenecarboxylic acids. By analogy with earlier work, the best explanation for these data is the formation of stable dicarbonium ions. -+ ' X The structure of these dipositive ions was investigated by ultraviolet, visible and proton magnetic resonance spectroscopy. [Charle s Allen Buehler Also investigated were the solutions of various benzophenone dichlorides (2,3, 4, 5-tetramethy1-, 3,‘5-dimethyl-, z, 4-dimethy1-, 4,4'-dime-thyl- and benzophenone dichloride itself) in sulfuric acid, with the hope of observing dicarbonium ions of the following type: c=© The benzophenone dichlorides were prepared by refluxing a solution of the corresponding ketone with phosphorus pentachloride in carbon tetrachloride. The visible, ultraviolet and proton-magnetic resonance spectra of these compounds in 100% sulfuric acid were similar regardless of the substitution. However, their spectra also compared favorably with the spectra of similarly substituted diphenyl- methylcarboniurn ions in the same solvent. The mono-ionization of these dichlorides in sulfuric acid was substantiated by their spectra and the production of only one mole of hydrogen chloride when a solu- tion of 2, 3,4, 5-tetramethylbenzophenone dichloride in sulfuric acid was swept with dry nitrogen, the phenyl apparently stabilizing the mono-ion more than the dicarbonium ion. The second part of this thesis .is concerned with the use of peroxytrifluoroacetic acid-boron fluoride as a potent source of positive hydroxyl. Synthetically, this reagent proved useful in preparing mesitol and .isodurenol from mesitylene and isodurene in excellent yields (88 and. 65%). The-reagent is, however, quite severe , and was found to bring about new and interesting side reactions in addition to electrOphilic replacement. Prehnitene gave, besides the expected prehnitol, Z, 2';.,3,' 3', 4, 4', 5, 5'-octamethyldiphenylmethane, 4, 5, 6, 6-tetramethyl- cyclohexa-Z, 4-dienone, 2, 3, 5-trimethylphenol, Z, 3, 6-trimethylphenol Charle s Allen Buehler and isodurenol as products of the reaction with peroxytrifluoroacetic acid-boron fluoride. Similarly, chloro- and nitromesitylene underwent hydride abstraction to form coupled materials of the following structure: X = C1, N02 The mechanism of these reactions is discussed in light of the known ability of peroxytrifluoroacetic acid to function ionically- (Z). The reactions of less substituted aromatics (such as benzene) with this reagent were extremely complex and gave largeamounts of intractable tar. - REFERENCES 1) H. Hart and‘R. w. Fish, J.-Am.-Chem. Soc., 333,. 5419 (1960) Z) R. D..Chambers,- P. .Goggin, and W.-K. -R.~ Musgrave, J. Chem. Soc. , 1804 (1959). PART I DIPOSIT IVE CAR BON IUM =IONS PART II PEROXYTRIFLUOROACETIC ACID-BORON FLUORIDE .AS A SOURCE OF POSITIVE HYDROXYL BY . Charles Allen- Buehler A THESIS ' Submitted to Michigan. State Univer sity in partial fulfillment of the requirements forrthe: degree of DOCT OR OF PHILOSOPHY Department of Chemistry 1963 ACKNOWL’ED GMEN T . The author~wishes to expresshis sincere appreciation and gratitude toProfessor' Harold Hart for-his inspiration, guidance and understanding. duringthe-course of this investigation. - Appreciation is extended to S. ’Meyerson for‘determining the mass spectra. - Apprec‘iation'is extended to the Petroleum Research‘Fund of theAmerican-Chemical Society for financial support from-September, 1961 through‘June, 1962 and to the National Science Foundation for financial support and ~‘aCoopera-tive Fellowship, June, 1962 through June, 1963. . Appreciation'is also-extended to my friends, colleagues and loved-ones for their inspiration and guidance throughout my tenure at MichiganState University. Grateful appreciation is also extended tomy wife, . Millie, for- her patience and encouragement at all times. .*****s******* ii TABLE OF CONTENTS 'PART- I ((DIPOSIT'IVE CARBONIUM IONS ‘INTRODUCTION.I............. RESULTSANDDISCUSSION.. . . . . . . . . . . . . . . I. Trichloromethylbenzene Derivatives. . . . . . . A.Preparation................. ~B. Dicati‘on Formation'in 100% Sulfuric Acid II. BenzophenoneDichlorides . ‘AoPreparation'oo0000.00.ooooooo B..Solutions of Benzophenone Dichlorides ~Su1furic Acid . . . . . . . . . III. Miscellaneous Experiments EXPERIMENTAL........... I. Trichloromethyl Compounds . . . . . . A.Synthesis ............ ‘Preparation of Pentamethylbenzene ‘Preparation of Isodurene. . . .. . . ’Preparation of Prehnitene . . . . . . . . 'Attempted Preparation of Trichloromethyl- -isodurene Attempted - Pr eparationi of Tr ichlor om ethyl- in O O O O O O‘O O O O O O O O O O ‘Preparation of Bromoprehnitene. . . . . . bromoprehnitene. . . . . . . . . . . . . Preparation of Bromoisodurene . . . . . Preparation of Trichloromethylbromoi-z isodurene..-............ 'Prepara ion of 3-Bromo-Z, 4, 5, 6-tetra- methylbenzoic acid . . . . . . . . . . Hydrolysis of Trichloromethylbromo- isodurene iii Page 12 21 21 .32 49 56 56 56 56 57 58 58 60 61 63 63 66 66 TABLE OF CONTENTS - Continued Preparation of Chloroisodurene . . . . . . 'Preparation of Trichloromethylchloro- isodurene. . . . . . . ._._.,.. . . . . Preparation of 3-Chloro-2, 4, 5, 6-tetra- methylbenzoic acid . . . . . . . . . . . Hydrolysis of Trichloromethylchloro- isodurene ........... Preparation of Chloroprehnitene. .A. . . . Preparation of Tric hloromethylchloro- prehnitene.......... ...... - Hydrolysis of Trichloromethylchloro- prehnitene................ ’Preparation of Trichloromethylprehnitene ‘Preparation of Methoxyisodurene . . . . . Attempted 'Preparation of Z, 4, 5, 6-‘I‘etra- methyl-3-methoxybenzotrichloride. . . . Page . 68 . 69 O 70 . 73 . 73 . 76 . 76 79 . 80 80 Attempted-Preparation of Z, 4, 6-Trimethoxy- benzotrichloride. . . . . . . . . . . . . . 82 II. Benzophenone. Dichlorides .". .‘ . . . . . . . . . . . . 83 A.Synthesis.................... 83 Attempts to Prepare 2, 3, 4, 5, 6-Penta- methylbenzophenone Dichloride. . . . . . 83 Preparation of 2, 3, 4, S-Tetramethylbenzo- phenone.................. 85 *Preparation of Z, 3, 4, 5-Tetramethy1benzo- phenone Dichloride . . . . . . . . . . ... 87 Hydrolysis of Z, 3, 4, 5-Tetramethy1benzo- phenone Dichloride . . . . . . . . . . . . 88 Preparation of 1-(2, 3, 4, 5-tetramethylphenyl)- l-phenylethylene . . . . . . . . . . . . . 90 ’Preparation of 2,4-Dimethy1benzophenone . 91 - Preparation of 2, 4-Dimethylbenzophenone Dichloride................. 94 'Preparation of 4, 4'-Dimethy1benzophenone \Dichloride................. 94 'Preparation of 1, 1-Bis(p-tolyl)-ethylene . . 94 ‘Preparation of 3, S-Dimethylbenzophenone . 98 %Preparation of 3, 5-Dimethy1benzophenone Dichloride................. 98 iv TABLE OF CONTENTS -- Continued Page Preparation of Benzophenone Dichloride. . . 'Preparation of 2, 4, 6-Trimethy1benzo- phenone.................... Attempted Preparation of 2, 4, 6-Trimethyl- benzophenone Dichloride . . . . . . . . . Hydrolysis of a Solution of 2, 4, 6-Trimethy1- benzophenone , , Phosphorus ' Pentachloride and Carbon Tetrachloride after Seventeen ‘HoursReflux................ B. Quantitative Determination of Hydrogen‘Chlor- ide From the Reaction of 2, 3,4, 5-Tetra- .methyl benzophenone Dichloride with Sulfuric Acid...................... 111. Miscellaneous Experiments . . . . . . . . . . . . . . Preparation of 2, 3, 4, 5, 6-Pentamethylaceto- phenone............ ...... 'Preparation of 1-Chloro-1-(2, 3, 4, 5, 6- pentamethylphenyl)ethylene . . . . . . . . Hydrolysis of a 101% Sulfuric Acid-Solution of Chloro-(Z, 3, 4, 5, 6-pentamethy1pheny1)- ethylene.................. Preparation of 1-(2, 3, 4, 5, 6-Pentamethy1- phenyl)acetylene . . . . . . . . . . . . Hydrolysis of Sulfuric Acid Solutions of 1-(2, 3, 4, 5, 6-Pentamethylpheny1)ac etylene Hydrolysis of Sulfuric Acid Solutions of Phenylacetylene . . . . . . . . . . . . . . SulfuricAcid................... . Spectra...................... . SUMMARY O O O O O O O O O O O O O O O O O O O O O O O O O O O O PART II PEROXYTRIFDUOROACETIC ACID-BORON FLUORIDE AS ASOURCE OF POSITIVE HYDROXYL INTRODUCTION O O O O O O O O O O O O O O O O O O O O O O O O O RESULTSANDDISCUSSION. . . . . . . . . . . . . . . . . .. 100 100 103 103 107 108 108 108 112 112 113 115 115 116 117 120 123 TABLE OF CONTENTS - Continued EXPERIMENTAL . . . . . I. Oxidatiens . . . 'A. Mesitylene . Isodurene . . ~ Pr ehnitene mmoow II. -Vapor-Phase Chromatography. . SUMMARY ‘ LITERATURE CITED . . Chloromesitylene. . .-Nitromesity1ene. . . . Benzene. . . . vi Page 138 138 138 140' 141 156 163 164 168 169 170 TABLE III. -IV.. 'VI.. VII. VIII. IX.. XI. LIST OF TABLES Ultraviolet and visible absorption maxima of some trichloromethylbenzenes in-100% sulfuric acid . . . . Proton magnetic resonance spectra of some halotetra- methylbenzotrichlorides and their corresponding carboxylic acids in 100% sulfuric acid . . . . . . . . . .Proton‘magnetic resonance spectra of substituted benzophenone dichlorides in carbon tetrachloride before and after reflux.with phosphorus.pen.ta.chloride Proton magnetic resonance spectrum of 2,4, 6-tri- methylbenzophenone in carbon tetrachloride before and after reflux with phosphorus pentachloride . . . . Visible absorptionmaxima for some benzophenone di- chlorides in sulfuric acid. . . . . . . . . . . . . . . . Ultraviolet and visible maxima for some substituted benzophenones in 100% sulfuric acid. . . . . . . . . . Visible absorption maxima for some diphenylmethyl- carbonium ions in sulfuric acid . . .Proton magnetic resonance spectra of various benzo- phenone dichlorides in 101% sulfuric acid. . . . . . . .Proton.magnetic resonance spectra of some substituted benzophenones in 100% sulfuric acid . . . . . . . . . vii Page .Proton magnetic resonance spectra of some halo- tetramethylbenzotrichlorides in. carbon tetrachloride. 8 .Ultraviolet absorption maxima of some trichloro- methyl compounds in cyclohwexane and n-hexane. . . . 15 17 19 28 31 36 38 41 ,42 42 LIST OF TABLES - Continued TABLE XII.. XIII. XIV. XV. Page Proton magnetic resonance spectra of some substi- tuted 1,1-diphenylethylenes in 101% sulfuric acid. . . .Stoichiometry with respect to. hydrogen chloride of 2, 3, 4, 5-tetramethylbenzophenone dichloride in. sul- furic aCidO O O O O O O O O O O O O O O O O O O O O O O O .Proton'magnetic resonance spectrum of some substi- tuted phenols invcarbon tetrachloride from the oxi- dation of prehnitene with peroxytrifluoroacetic acid- boron fl‘lorideO O O O O O O O O O O O O O O O O O O O O O Products from the oxidation of prehnitene with peroxy- trifluoroac etic acid with and without boron fluoride. . viii 45 48 127 130 ‘ FIGURE 1. 9.. 10. 11. 12. 13.» LIST OF FIGURES Ultraviolet spectra of various trichloromethyl benzenes in n-h'exane and cyclohexane. . . . . . . . . . . . . . . Ultraviolet and visible spectra of various trichloro- methylbenzenes in 100% sulfuric acid. . . . . . . . . . Proton magnetic resonance spectra of 2, 3, 4, 5-tetra- methylbenzoPhenone in carbon tetrachloride before and after reaction with phosphorus pentachloride. . . . . . Ultraviolet and visible spectrum of 2, 3,4,5-tetra- .methylbenzophenonedichloride in 101% sulfuric acid asafunCtIODOftime. o o o o o o o o o e o o o o o e o o .Visible spectra of various benzophenone dichlorides in 100% sulfuric acid. O O O O O O O O O O O O O O O O O O O Ultraviolet and visible spectra of some substituted benzOphenones in 100% sulfuric acid . . . . . . . . . . -Ultraviolet spectra of XLII and XLIII in n-hexane . . . Visible spectrum of XLIII in 101% sulfuric acid . .. .. .. Ultraviolet and visible spectra of XLII in 98~and 101% sulfuric aCid. O O O O O O O O O O O O O O O O O O O O O O Infrared spectrum of isodurene (Cstolution) . . . . . Infrared spectrum of bromoprehnitene (CS; solution) . Infrared spectrum of bromoisodurene (CS; solution) . Infrared spectrum of trichloromethylbromoisodurene (CS; andCCh salution8)0 o o o o o o o o o o o o o o o 0 ix ’Page 14 16 26 33 35 37 51 5-3 54 59 64 65 LIST OF FIGURES -»-Continued FIGURE 14. ISO 16. 17. 18. 19. 20. '21. 22. 23... 24.- 25.. 26.- 27 . Infrared spectrum of 3-bromo-2, 4, 5, 6-tetramethy1- benzoic acid-(CHC13 solution). . . . . . . . . . . . . . Infrared spectrum of Chloroisodurene (CS; solution) . Infrared spectrum of trichloromethylchloroisodurene (CSzandCCI‘SOIUIIOn).o.........o..... 0 Infrared spectrum of bromochloroisodurene (CS; solu- tion) O O O O O O O O O O O O O O O O O O O O O O O O O O O Infrared spectrum of 3-chloro-2, 4, 5, 6-tetramethy1- benzoic acid (CHC13 solution). . . . . . . . . . . . . . Infrared spectrum of chloroprehnitene (CS; solution) . Infrared spectrum of trichloromethylchloroprehnitene (CSZ andCCI‘ 801ntion). o o o o o o o o o e o e o o o 0 Infrared spectrum of 2-chloro-3,4, 5, 6-tetramethyl- benzoic acid (CHC13 solution). . . . . . . . . . . . . . Infrared spectrum of methoxyisodurene(CCl¢ solution) Infrared spectrum of 2, 3, 4, 5-tetramethylbenzophenone ~(CCh salution) O O O O O O O O O O O O O O O O O O O O O O Infrared spectrum of 2, 3, 4, 5-tetramethylbenzophenone dichloride~(CSz and CC14 solution). . . . . . . . . . . . Infrared spectrum of 1-(2, 3, 4, 5-tetramethylphenyl)-1- phenylethylene (CCI‘ solution). . . . . . . . . . . . . . Infrared spectrum of 2,4-dimethy1benzophenone‘(CC14 salution)0 O O O O O O O O O O O O O O O O O O O O O O O O Proton magnetic resonance spectra of 2,4-dimethy1- benzophenone and phosphorus pentachloride in carbon tetrachloride before and after reflux. . . . . . . . . . Page 67 70 71 72 74 76 77 78 81 86 89 92 93 95 LISTOF FIGURES - Continued FIGURE 28. 29. 30.. 31. 32.. 33.. 34.. 35. ‘36. 37. '38. 39. 40. 41... .Protonmagnetic resonance spectra of 4,4"-dimethy1- benzophenone and phosphorus pentachloride in carbon tetrachloride before andafter reflux. . . . . . . . . . . Infrared spectrum of 1,11 -Bis(p-tolyl)ethy1ene -(CC14 801111231011)...o.......ooo...v...-.....ooo Infrared spectrum of 3, 5-dimethy1benzophenone (CC14 salution) O O O O O O O O O O O O O O O O O O O O O ‘ O O O O O .Proton magnetic resonance spectra of 3, 5-dimethyl- benzophenone and phosphorus pentachloride in carbon tetrachloride before and after reflux. . . . . . . . . . . Infrared spectrum of benzophenone dichloride (neat) . . Infrared spectrum of 2, 4, 6-tri1nethy1benzophenone . . . . Proton magnetic resonance spectrum of 2,4, 6-tri- ~methy1benzophenone-in carbon tetrachloride. . .. . . . . Infrared spectrum of pentamethylac etophenone- (CCI‘ salution) O O O O O O O O O O O O O O O O O O O O O ‘ O O O O O Infrared spectrum of 1-chloro-1-(2, 3,4, 5, 6-penta- .methylpheny1)ethy1ene(CCL, solution) . . . . . . . . Infrared spectrum of pentamethylphenylacetylene“ (CC14 salution) O O O O O O O O O O O O ‘ O O O O . O O O O O O ‘ O O O O Infrared spectrum of isodurenol (CCL, solution) . . . . . .Vapor- phase chromatogram of the alkali-soluble fraction-from the oxidation of prehnitene with peroxy- trifluoroacetic acid-boron fluoride . . . . . . . . . Infrared-spectrum of prehnitol (CC-14 solution). . . . . . Infrared spectrum of 4, 5, 6, 6-tetramethy1-Z, 4-cyclo- hexadienone (CC14 solution). . . . . . . . . . . . . . . . xi Page 96 97 99 101 102 104 105 109 111 114 142 144 145 147 ‘ LIST OF 'FIGURES - Continued FIGURE ' v Page 42. Infrared spectrum of 2, 2", 3, 3',4,4',5, 5’-octamethy1- diphenylmethane (CCI, solution). . . . . . . . . . . . . 149 ’43.-Infrared-spectrum of intractable tar-found in the neutral fractionfrom the oxidation'of prehnitene with peroxytrifluoroacetic acid-boron fluoride (CCl. solution)......................... 151 44.-Vapor phase chromatogram of the alkali-soluble fraction from the oxidation of prehnitene with peroxy- trifluoroacetic acid. . . . . . . . . . . . . . . . . . . 153 45.« Infrared spectrum of fraction 3 from Figure 44. . . . 154 46..Proton magnetic resonance spectrum of fraction 3 from‘Figure44eeveeeeeeoeeeeooeeeeee1.55 47.- Infrared spectrum of LI" (CCl, solution) . . . . . ,. . . 158 48.— Infrared spectrum of the petroleum- ether 'elutable fraction from the oxidation of Chloromesitylene with peroxytrifluoroacetic acid-lboron bluoride (CCl‘ solu- tion)............................. 160 49.7- Infrared-spectrum of LVIII from the oxidation of nitromesitylene with peroxytrifluoroacetic acid-boron fluoride-‘(CHC13 solution). . . . . . . . . . . . . . . . 165 . xii PART I DIPOSITIVE CARBONIUMIONS INTRODUCTION , Several years ago it was shown:(1) that trichloromethylpenta- methylbenzeneionized insulfuric acidto produce a stable dicarbonium ' ion-(II). . Support for the formation of 11 came from cryoscopic, '\ . + _ - -<:c13 + szso. --..——->- c-c1 + ZHCI- + :szo. I II stoichiometric and conductivity measurements on the acid solution, .and- alsefrom the visible, ultraviolet and nuclear magnetic resonance spectra of this solution. This was substantiated more recently by. the isolationof acrystalline tetrafluoroborate and tetrachloroborate of . theadicarbonium ion (2). It seem-ed important to determine what factors contribute to the. formation and-stability of "extraordinary" dicarbonium . ions of this .typei. e_. , ions produced formallyby the loss of two anions from a single carbon atom, and this part of the-thesis is designed to ’shed‘ light on this problem. ‘ -Relief of steric strain in the trichloromethylpentamethylb'enzene is- certainly one important factor which contributes to dicarbonium .ion formation. .Examinationaof a‘Fisher-Herschfeldermodel of trichloro- methylpentamethylbenzene shows thatthe-ilarge trichloromethyl group cannot rotate freelywith respect to the neighboring orth'ovmethyl groups. Ionization of onevchlorine from the trichloromethyl group does not lead to a stabilized monocation (III). because in. order for the positive C1 C1 +/ . / O a C c C1 -<-—-->- \ C1 111 IIIa charge to be delocalized into the ring :(IIIa), the remaining-“two chlorines ~must be coplanar with thearom-atic ring; such'a conformation results in unfavorable interactions between the chlorines and the or-tho methyls. ~ Loss of two) chlorines relieves steric interactions markedly because the C=C-C1 group'(see‘II) is probably linear. -Also»important to the formation of dicarbonium ions is the type of s-ubstituent on the aromatic ring. -Althoughperchlorotoluene has the necessary steric hindrance for ionization, this compound hydrolyzes slowly. to the carboxylic acid, even in hot concentrated sulfuric acid (3). . Since chlorine is inductively electron-withdrawing, it appears that electron-donating substituents, A such- as methyl groups, are necessary to stabilize the dicarboniumvion. ~ Incontrast, . selective replacement of . various methyl groups‘on the original trichloromethylpentameth-ylbenzene does not prevent ionization to dicarbonium ions. . Compounds ‘IV, V,- VI and' VII give stable dicarbonium ions in sulfuricacid- (1,.2, 4-6). . col, _ cc13 .cc1, col, 0 7 .F . H ' x X . -IVa X- = C1 V VIa 'Xa H VII IVb‘X= Br VIIb ‘X=F -"IVC X-= F ‘VIb X = C1 VId X.= Br The rates of hydrolysis of these substituted-phenylchlorodicarbonium ions-were determined spectrophotometricallyin various concentrations of 90 a100% aqueous sulfuric acid (4). Thehydrolysis was pseudo first- order “with-respect to dipositive ions.- The kinetics revealed that the ability of substituents to. stabilize dicarbonium-ionswas in the order (methyl > hydrogen > halogen, as might be expected. . One objective of this work was to-synthesize compounds VIII and’IX. These compounds CC13 CC13 X X VIIIa X = Cl , IXa X = C1 VIIIb X = Br IXb X.= Br ‘VIIIc X = H ' ' retainthe basic spatial requirements of trichloromethylpentamethyl- benzene andzalso permit a clear'comparison of. substituents'on the ortho (Ix, VII), para. (IV) and meta (V, VIII) positions of the dicarbonium ion. - Another objective of this work was to investigate the effect of structural. variations inthe a -substituent Y (X) on the formation of Y C1+ X- XI dicarbonium .ions. . Chlorine in this position presumably donates electrons Via valence bond structures of the type XI. Other. groups which might alsodo this are fluorine ,[F > 61 (7)], methoxy- (~OCH3) or phenyl substituents. It is well-known that phenyl groups stabilize carbonium ions by the delocalization ‘of charge (although inductively they destabilize as seen from the work of Batiste (8) who found that heptaphenyltropylium cationvis less stable than the unsubstituted ion). Introduction of a phenyl substituent (Y) into dicarbonium .ion X would stabilize the charge in the following manner (XII). - However, the 1:60 a are monocation‘(X-III) should also be stabilized by the phenyl substituent, ~ ('31 C1 Cl]. .c : é . c XIII X1118. X1111) maximum) stabilization occurring when the entire system is coplanar '"(XIIIa and XIIIb). -Examination of Fisher-Her schfelder models of XIII . shows that. the ortho. methyls prevent such coplanarity; consequently, .maximum. stabilization is unlikely. The degree towhich XIII is stabilizedrelative to the dicarbonium ion should determine whether the monocation ionizes further. The same maybe said for other Y substituents (X) such-as fluorine, methyl or) methoxy; i: E: , the substituent contributes tothe stability of the mono-ion as well as the dicarbonium lion, 7 and which is formed will depend ona balance of factors. » RESULTS AND DISCUSSION I. Trichloromethylbenzene Derivatives A. Preparation The 3-halo(chloro, bromo)g-2, 4, 5, 6-tetramethylbenzotrichlorides (VIIIa and VIIIb) were synthesized from the corresponding haloiso- durenes using aluminum chloride and excess carbon tetrachloride. col3 . + col. £31571? VIIIa x = CI " VIIIb x = Br \x x The same reaction with chloroprehnitene gave 2-chloro-3, 4, 5, 6-tetra- methylbenzotrichloride (IXa). Incontrast, the condensation of carbon cc13 - C1 C1 AlCl . + CC ——%> 14 35-42 m - IXa tetrachloride with less substituted-aromatic compounds, such~as benzene, gives a mixture -of- di- and triphenylmethane derivatives (10, 11,112). The use of thisFriedel-Crafts reaction to add (a trichloromethyl groupto a poly-methyl substituted (benzene nucleuswas originally re- portedby Hart and Fish (1) who used aimethod communicated to them by- Rolih‘and» Peters (9) to .make pentamethylbenzotrichloride (I) from pentamethylbenzene. The method employed a seven-fold excess of carbon tetrachloride, . probably to prevent the possible formation of di- and. triphenylmethane derivatives. - Subsequent preparations of- other substitutedvtrichloromethylbenzenes '(IV, V, VI and VII) showed the method to be quite general and the products were usually obtained in good yield- The-same was true for the analogous preparations of - VIIIa, VIIIb and IXa (78-89% yield). - However, there .is good reason to suspect alternative paths for this reaction, for. durene treated with the same reagents (6) gave tri- chloromethylprehnitene (VII). The methyl migration which occurred whensthe bulky trichloromethyl group was introducedeparallels the rearrangement which takes place during the‘Jacobsen reaction. occ1, AlCl: VII Durene, isodurene and prehnitene all give prehnitene on sulfonation andhydrolysis (13). There is additional precedent for sucha rearrange- ment in the work of Baddely and Pendleton (14) who have shown. that acetyldurene-when heated to 1000 with an excess of aluminum chloride yielded a mixture of products, , 80% of which was acetylprehnitene. - Therewas no reaction between acetylprehnitene and excess aluminum chloride when refluxed at 1500 for three hours. Therefore acetyl- prehnitene was assumed to be the most stable isomer. Although such rearrangements seem unlikely for completely subs-titutedaromatics (VIIIa, V'IITb and IX), it was necessaryto unequivo- cally establish their structure. zThe infrared and ultraviolet spectra of these compounds were entirely consistent with their. structures. I Their nuclear“ magnetic resonance spectra gave only aromatic methyl absorp- tion at ’1’ values of 7.8-8.0 (15) (see Table I). The compounds were readily hydrolyzed to thecorresponding carboxylic acids by refluxing Table I. .Proton Magnetic Resonance Spectra of Some Halotetramethyl- benzotrichlorides in Carbon Tetrachloride Numberaof ComLound ~ Positions ()units) Protons 2-Chloro-3,4, 5,6-tetra- 7.50, 7.69.7.81, 7.87 , 3 (apiece) methylbenzotrichloride 3-Bromo-2,4,5,6-tetra- 7.32,7.51,7.60,7.78 3 (apiece) methylbenzotrichloride . 3-Chloro-2,4,5,6-tetra- 7.35,7.48,7.65,7.78 3 (apiece) methylbenzotrichloride a Obtained by electronic integration. with aqueous acetone. m-Bromoisodurenecarboxylic acid and m-chloro- isodurenecarb‘oxylic acid were preparedfrom dibromoisodurene or chlorobromoisodurene via carbonation of the mono-Grignard reagent. The acids obtained by‘these methods were identical ineall respects (m.p., m.m.p.. and infrared spectrum) to the acids obtained by hydrolysis of the corresponding trichloromethyl compounds. The carboxylic acid from the hydrolysis of 2-chloro-3, 4, 5, 6- tetramethylbenzotrichloride (IX) was found to be different (m.p. , m.m.p. andnuclear magnetic resonance spectrum) from those acids obtained from VIIIa and lIVa; consequently, its structure must be as postulated. - A similar difference was found in the trichloromethyl compounds themselves (IXa, VIIIa and'IVa). .Attempts were made to synthesize VIIIc and IXb by the reaction d—e scribed previously. Trichloromethylation of bromoprehnitene gave, .ccn can Br " IXb VIIIC rEither surprisingly, trichlorom~ethylprehnitene,(VII). This result was Br CC13 A ,+ CC14 4&0)- 38-42 Verified by the identity of both the product of the reaction-and the Q arboxylic acid obtained by aqueous acetone hydrolysis (m.p. , m.m. p. a~1'Ic1 infrared spectru'm) with authentic samples of trichloromethyl- Drehniteneand prehnitenecarboxylic acid. -As mentioned previously, ' Q:l'Iloroprehnitene under the same conditions gave trichloromethylchloro- Drehnitene as expected. ‘Olah (16) has found that the order of reactivity 10 of halogens in displacements of this sort is Br >> Cl. Thisis in accordance with the relative stability of positively polarized halogen entities or, in the limiting case, halonium ions, Br+ > Cl+ . ' . In accordance with the relative basicity of halobenzenes' (bromobenzene more basic than chlorobenzene), it is to be expected that the bromo- benzene derivative will possess a higher degree of activity than the chlorobenzene compound in a mechanism involving protonation of the ring to a sigma complex and subsequent eliminatiOn of positively. polarized halogen from the sigma complex, __i_. _e_.: f— -)" con - con ccn H . Br Br + —fl—> ———-> + Br+ — - Thismechanismassumes that such‘a displacement occurs after the tri- cllloromethylation. There is no experimental evidence in support of 11115.8 assmnption but a significant driving force for this reaction must be the increased steric hindrance in the 2-bromo-3, 4, 5, 6-tetramethyl- benzotrichloride relative to the 2-methyl- and Z-chloro compound (Br) C1,. CH3). If displacement had occurred before trichloromethylation ‘ a éimilar~displacement should have been observed in the previous synthe- sisoi other bromotrichloromethylbenzenes-'(VIIIb, . VId (and VIIIb). >One c>‘Iilner possiblemechanism whichdeserves mention involves electrophilic di- splacement of the bromine by the incoming trichloromethyl complex. There is no experimental evidenceexcludingthis possibility; it would be 1 iliteresting to test this by use of appropriately: labelled compounds, or Q Qmpounds which have other alkyl groups in place of. methyl. Isodurene when treated with aluminum chlorideand excess carbon 1:E-itrachloride gave trichloromethylprehnitene.(VII). This wasverified 11 . cc:13 AlCl: CCl‘ 38-42o VII by the .identity of both the product of the reaction and the carboxylic acid obtained on hydrolysis with authentic samples of trichloromethyl- prehnitene and prehnitenecarboxylic acid (m.p. , m.m. p.. and infrared spectra). This rearrangement parallels that. found by‘Rafos (5) who obtained-tricyloromethylprehnitene from durene, under the same con- ditions. The particular stability of VII was substantiated by the failure of prehnitene to rearrange when treated in a similar manner, the product isolated being trichloromethylprehnitene. . Ineach case the methyls migrate so that only one methyl group remains ortho to the bulky trichloromethyl substituent. This is strong evidence-for steric Strain which can be relieved by migration of one of the methyls. However, bromomesitylene which has steric requirements similar to is odurene,, gave trichloromethylbromomesitylene without rearrange- ment (4). This result maybe explained by the destabilizing influence of bromine on~a sigma complexed intermediate such as XIV (undoubtedly an oversirnplification, for the aluminum chloride would most likely col3 . CCI3 Br - XIV XIVa c c>11'nplex with the substrate in some manner). ~ Of course, a mechanism Involving: XIV as anintermediate implies rearrangement after trichloro- 111 ethylation, an assumption which has not yet been verified experimentally. 12 However, if isom-erization occurred priorto trichloromethylation, disproportionation products similar to those found when durene is (treated withexcess aluminum chloride ought to be produced “(17). . Since such products were not isolated, it seems reasonable to. conclude that rearrangement occurs after trichloromethylation. . Further support for the argument that migration occurs to relieve steric strain is the observation of similar rearrangements in the J acobsen reaction andthe 'reaction of acetodurene with aluminum- chloride (14). - Unfortunately, a detailed-study of these migrations islacking, thus preventing a more c oncrete assignment of mechanism. . Attempts to ‘pr epare 3-methoxy-2, 4, 5, 6-tetramethylbenzotri- Chloride and-2, 4, 6-trimethoxybenzotrichloride were not successful. 5— T richloromethylationof methoxyisodurene and. 1, 3, 5-trimethoxybenzene under the same-conditionsfound successful in .earlier- preparations gave largely intractable tars. - Methoxyisodur ene,-when treated‘with excess ~ carbon tetrachloride andealuminum chloride, yielded in addition to. 1'ecovered» startingmaterial a p-rpduct which gave a. deepred- solutionwhen ‘11 ssolved- in 100% sulfuric acid. -- However, hydrolysis ofthis solution on .ice gave no-crystalline material. B... Dicatioanormation in 100% Sulfuric Acid The previouswork of Hart and co-workers (lab) has firmly estab- 1iShed the formationof dications in.100'% sulfuric-acid solutions of Dentamethylbenzotrichloride(I), 2, 3, 4, 5-tetramethylbenzotrichloride (VII), 4-halotetr-amethylbenzotrichlorides(IV) andvariously substituted 2'. 4, 6-trimeth-ylbenzotrichlorides(VI). -It was felt that the rigorous proof 9f dicarbonium ion formation by these workers made it unnecessary to 1‘ epeat- such experiments there. .13 The steric requirements-for dicationeformation' are present in tall the compounds. reported~(VIII and IX). The ultraviolet spectra of these compounds in n-hexane amply. demonstrates this strain. . Benzene normallyxabsorbs inrtwo-main regions of, the ultraviolet, one of high - intensity near*198 my. (6 ca. 8, 000‘) called therK-band, and the other of 'much lower intensity between ‘230 impland 27.0 mu. (6 ca... 230) called the‘B- band. . Ballester and Castaner-‘(3) have shownthat increasedsteric hindrance produces bathochromic shifts in both. these bands with the pre- dominant. shift occurring in'the Bivband. - Similar results were found by - Hart and co-workers for variously substituted trichloromethylbenzenes. Figure 1. shows a comparison of VIIa, VIIIb andIX withr4-bromo-2, 3, 5, 6- tetramethylbenzotrichloride (5.). Substitution of chlorine ortho, meta or " para does not seem to‘substantially. affect the gross features of the Spectrum. -Likewise,. substitution of bromine para ormeta causes no ‘ Serious changes (see *Table‘II). This is to ,be expected due to very Similar resonance and steric contributions of thesesubstituents. . Figure 2. and Table III showva. comparison of the ultraviolet and vis- ible spectra of these same trichloromethyl derivatiVesin' 100% sulfuric acid. The greatdifference between the spectra of the parent compounds and. the corresponding dicarbonium ions is due totheeffect ofintro- ducing‘apermanent charge into, the chromophoric systemgiving rise to ‘Chargetransfer-resonance spectra (18). -The most obvious difference O3|:Iserved is the hypochromic effect in the 2-.chloro—3,.4, 5, 6-tetra- ‘ methylchloirodicarbonium ion versus the -3-chloro derivative. . Sufficient information (is not available toaccurately discuss this decreased. intens- i12y, . but it is to be noted that the same effect isnot observed inthe Spectra of the corresponding carboxylic‘acids in 101% sulfuric ”acid; A. 3., Z-chloro-3,.4, 5, 6-tetramethylbenzoic acid (lmax 293 mu,- e = 1 9, 950) and 3-chloro-2, 4, 5,.6-tetramethylbenzoic acid ‘(xmax 294mg, 6 = 19. 300). These spectra-are presumably those of the corresponding 14 #\ \. \ .c-csl3 \. \ \ \ \O \, 1 \O \ '. O O 4. O —a \ .3. CC13 3.0 ~ logE 2.0 ~ 2 O O L Figure 1. Ultraviolet spectra of various ,trichloro- 0 methylbenzenes in n-hexane and . .. cyclohexane. ‘ e . O 10 O l\ \\ . . O O \ . . O L , \ a 2‘00 250 Wavelength (mu) 350 15 Table II. UltraviOIet Absorption‘Maxima of Some Trichloromethyl - Compounds in Cyclohexane and n-Hexane Compound ‘Wavelength Molar Absorbancy (mp) _ Index"(€) 4-Chloro—2, 3, 5, 6-tetra- a 301 l, 503 methylbenzotrichloride 251 8, 980 4--Bromo-2, 3,5, 6-tetra- .303 1,611 methylbenzotrichloridea 254 10, 320 3 —Chloro-2, 4, 5, 6-tetra- b 305 1, 669 methylbenzotrichloride '248 7, 950 218 37, 090 3-Bromo-2,4, 5, 6-tetra- b 305 1,676 methylbenzot .riichloride 248 7 ,_ 8 1 0 217 33, 590 2-— Chloro-3, 4, 5, 6-tetra- 305 1,040 -methylbenzotrichloride 250 4,470 217 .21,200 16 38v, Aumsofiocrmb? one coo com com ‘ ems. 8e omm. , com 1 emu Be _ J _ _ _ . . _ _ _ ,_ lac; a ..W .33. oassnom s2: 5 nsoa y? gflaonnmofln— mflowpo> mo «.36on 033C, one «oaogmnfiD. .N oudmfim o.N a . - . a .27 Ho 6 s 'O-’ _ .V/I ...// H010 HUtUflUI L min .../I .+ + o I dag/l . .V ./ \ wmoH . 0/, 9 n./// ,OOeOWHII'II ,O [Gem 0 so u ”local: “/6 s. -§\,u /e.. O // \ , O Hm O O 9 \ ’ C . 0/ / \ o t o ’ // o \\O / O 0/ ) I o o‘lle'o e 56. ..// P. .. .\ / ,.\ in m .+ o I e \o. I .1\ o. // x I. \l I too/I II- . . .0 I 0 III ‘O .. 0 I . .0 / \\o- *0 sale auto...cc. OO \ OOO e o / \ o. i I... o v H010. HO 0 I .o 1.. co .0. s Q . J me¢ 17 Table III. -U1traviolet and Visible Absorption Maxima of Some Trichloro- methylbenzenes in 100% Sulfuric Acid W L Wavelength Molar Absorbancy Compound (mu) Index (6) 4-Chloro-2, 3, 5, 6-tetramethy1- 520 (shoulder) 2, 095 benzotrichloride 406 24, 550 399 22, 100 299 5, 730 260 7, 500 226 11,000 3-Chloro-2,4, 5, 6-tetramethy1- 550 1,493 benzotrichloride 402 20, 200 392. (shoulder) 18, 200 295 3, 190 264 4,980 2 ~Chloro-3, 4, 5, 6-tetramethyl- 550 1, 242 benzotrichloride 402 12, 600 388 (shoulder) 10, 550 296 7,750 268 5,900 4 - Brom‘o-Z, 3, 5, 6-tetramethyl- 530‘ (shoulder) 2, 130 benzotrichloride 419 27, 150 405 (shoulder) 21, 300 310 4,800 262 7,820 243 8,150 3-Bromo-2,4, 5, 6-tetramethyl- 549 1,550 benzotrichloride 406 20,200 396 (shoulder) 17, 600 300 5,580 265 5,190 18 acylium ions (19), XV and XVI. . 0 0 II II =C+ .C_+ Cl _ C1 . XV XVI The nuclear magnetic resonance spectra of these compounds (VIIIa, VIIIb and IX) in 100% sulfuric acid are consistent with such highly charged species. In Table IV are recorded the spectra of the di- ions in 100% sulfuric acid along with the corresponding acylium ions produced by dissolving the carboxylic acidsein 100% sulfuric acid. ' Although it is dangerous tocompare ions which differ basically in , Structure at the position of ionization, _i_. 6_. ,1 acylium ions, versus dicarbonium ions, it seems important to-record here those differences 'Which doaexist in the proton magnetic resonance spectra. A point which deserves mention here is the value of 6. 46 Y given to .the internal reference, methanesulfonic acid. This value-was based 0n- the observed position. of this material in 100% sulfuric acid relative to an external reference (contained in a capillary tube inserted into the NMR tube) of tetramethylsilane to which the ‘1‘ value of 10. 00 is ‘ assigned (20). ' Since all the spectra are taken 'in' approximately 100% s"~1-1fu‘ric acid there is internal. consistency. Previous spectra reported by Hart and,co-workers used an internal reference of tetramethylsilane ' but the insolubility and ultimate reaction of this material with 100% S“-llfuric acid prompted the use of methanesulfonic acid. Consequently. the T values reported here differ from those reported previously (3. 4,5,6). 19 .cofimnwofi: 3:03on >3 poawEuoquo F3 .0 .znOmnmo . 6838qu 6?»:me and mo masogmfimmm unomoumvu mwmocficonmmnfi muvnEsZm me. .> Amvmwnh oguoanowuuouconanuoe Amomamv m Nieigso.» Amrmgfivfiggaoos -mhfi-o.m.w.m-ouofio.~ Q 3. .30w .0 wm .h . u t .. ognoEofiuuounonfififioE somamv m 8.5.8.586 Amrm§gsfid -mfi3$.m.w.~-ofionm-m. o co .N. . . Amy H w . v upwaozowhuouaonflwfiog A3293 m «mg dog. Emoéévmodézwd -mb3$,.m.v.~-3ofio.m omcououm. Amfig F .sofifimonfla 335.. L, 50330an pcdomaoo, mo noncgz. a. £3 Haifa? , .m coH tgmconumofifl . .634 3.3.615 $2: GM wages. oflkwxoovumnu .magdommounonu MESH. was mupwhofifiomuwouconfiifio8.933363 080% No mhuomqm mocmfismox owuofimwsfigohm. SN mEmrm 2.0 ‘A tentative assignment of bands in the sulfuric acid solutions of the dicarboniumions is made in Table IV. The 3-halo derivatives A(Br,w- C1) have in common the consistent assignment of the ortho methyl group adjacent to the halogen at lowerfield. This is based first on the observation of the proton magnetic resonance spectrum of 2,4, 6-tri- methylphenylchlorodicarbonium ion‘(XVII) which exhibits two bands in the aryl-methyl region in the relative areas of 2 to 1, the latter C1 l (3+ XVII appearing at higher field (6). The symmetry inherent in this molecule demands the assignment of the ortho methyls to lower field in agree- Inent with the relative areas. Second, it is to be expected that methyls adjacent toelectron-withdrawing halogens should appear at lower field than those adjacent to. methyl groups. The proton ~magnetic resonance Spectrum of chloromesitylene exhibits twobands in the aryl-methyl region at T values 7. 75 and 7.84 with relative areas of 2 to 1 in agree— -1'I1ent with this postulate. The assignment of the meta methyl to highest ‘ 15:1 eld is in-accordance with‘a similar assignment in previous dicarbonium ions"(l-6).‘ The same arguments were used-to assign the bands found in ‘ 2-chloro-3, 4, 5, 6-tetramethy1chlorophenyldicarbonium ion. A The 'c orn‘plexity of those factors to be considered in the other spectra re- perted forces us to forego assignments until sufficient model compounds 3. 1‘ e available . 21 II. Benzophenone Dichlorides 'A. Preparation Benzophenone dichlorides can be prepared by two general methods: first, the reaction of diaryl ketones with phosphoruspentachloride and secondly, the Friedel-Crafts condensation of an excess of the aromatic compound with carbon tetrachloride and aluminum chloride. The stabil- ity of these dichlorides varies widely; some are stable to .cold water‘ (XX), others hydrolyze rapidly in moist air (XVIIIb). Those compounds which .have been reported may be separated into two classes, those for’which CCIZ CCIZ X X X xvm xxx 'a) x = (cagzu (21, 22) a) x = no, (29-31) b) x = H (23.25) b) x = CI '(32) c) x = 00H, (26, 27) c) x = e5 (33) d) X = 95 (28) cm, : col, col, .<:1, (:12 xx (34, 35) XXI (36) accurate analytical data are'available (XVIII, XIX, XX and XXI) and those, ' a‘Pparently unstable, . reported as oils or solids containing varying amounts of ketone-and other‘impurities (XXII, «XXIII). There appears to be no (2 erelation between the type of substituent and the stability of the dichloride. 22 cm2 . cc12 X X X ‘XXII . XXIII a) X. = CH; (37) a) X = Cl (39) b).X =Ft~rbutyl (38) b) X = CH; (40) . Our concern was the preparation of 2, 3, 4, 5, 6-pentamethylbenzo- phenone dichloride (XXIV). It was found by'Fish"'(41) that pentamethyl- benzophenone does not react with phosphorus pentachloride to give XXIV. 1:1“1‘0 ’ . XXIV If carbon tetrachloride or benzene were used as solvents, the reaction ‘did not go; the use of tetrachloroethane caused decomposition. Fish Suggested that-the dichloride might be preparedfrom benzotrichloride and pentamethylbenzene in the presence of stannic chloride.'-Severa1 ~ cm, (:012 atttempts to prepare the dichloride (XXIV) by this route failed. {Only tar land polymericmaterial were isolated regardless of the solvent used or ' the length and temperature of the reaction. -- There was evidence of a diaryl ketone in the infrared spectrum (6.00 H) of the tarry portion of the product but the low yield prompted nofurther investigation. ' Another attempted method which proved unsuccessful, was the condensa- ti On of trichloromethylpentamethylbenzene with benzene and aluminum 23 chloride. There was noreaction at 250-, but when the temperature was raised polymeric material was obtained. One interesting result which was not further‘investigated was the formation of diphenylmethaneas one of the products of this reaction. - When the reaction of benzotrichloride and pentamethylbenzene wasrepeated with aluminum chloride as the catalyst, two identifiable materials were isolated. ' One of these proved to be 2, 3,4, 5-tetramethyl- benzophenone (XXV) and the otherhexamethylbenzene. The structure Iii 9f the ketone was verified by independent synthesis from prehnitene and benzoy’l chloride. Its proton magnetic resonance and infrared spectra Were also consistent with its structure. Hexamethylbenzene was verified by comparison with anauthentic sample (m. p. , m.m. p. and infrared Spectrum). The migrationof the‘ortho methyl group to unreacted penta- methylbenzene suggests that the failure to-synthesize XXIV is caused by steric hindrance. Examination of Fisher-Hirschfelder models shows that, as with trichloromethylpentamethylbenzene, the ortho methyl groups prevent rotation of the chlorinecand ‘phenyl substituents. A180,.thephenyl substituent can only be accommodated if it lies ortho— “ gonal to the‘pentamethylbenzenering. The removal of an orthomethyl group permits a much greater degree of freedom and certainly is one I"P—ason for thedisproportionationproducts observed. This same sterichindrance may explain the failure of the ketone to react with phosphoruspentachloride. . Fuson'and co-workers (42) haVe given‘numerous examples of the anomalous behavior of sterically hindered ketones on reaction with nucleophiles, as, for example: 24 9-95 I O OCH3- ~ Icl) ¢ ' . c . . +- p *- '¢-'¢Mg‘lBr‘—> The preferential displacement of the -methoxy substituent is presumably O: ~due to the highenergy necessary for attack at the carbonyl carbon because of steric hindrance provided by the ortho methyl groups. In our case, introduction of twochlorine atoms into pentamethylbenzo- -phenone, 3.3. , formation of pentamethylbenzophenOne dichloride, would lead to a very rigid and sterically hindered structure. The ketone because of its trigonal hydridization at the carbonyl carbon has reason- ably-free rotation-about the carbonyl carbon-carbon bonds, although the molecule is not completely free of strain. ‘An examination of Fisher- Hirschfelder rmodels of pentamethylbenzoph-enone indicates that the phenyl rings cannot both lie in the plane of the carbon-oxygen bond simultaneously, therefore preventing maximum overlap of r-bonds throughout the molecule. However, attack'at the carbonyl carbon by an incoming nucleophile remains a high energy process because of S-teric hindrance, certainly the main reason reaction did not occur. It is important: to considerfurther the isolation of :.3,4, 5, 6-tetra- “methylbenzophenone from the reaction of benzotrichloride and penta- methylbenzene. {Undoubtedly, . the loss of the ortho methyl group was. ' sufficient to impart somestability to the presumed intermediate di- -Chloride; otherwise other-disproportionation products would b'e-expected. ‘ C Onsequently, it might also be possible to prepareZ, 3,4, S-tetramethyl- b€>Inzophenone dichloride from the ketone and phosphorus pentachloride. Refluxing a solution of 2, 3,4, 5-tetramethy1benzophenone and a slight ‘ 33‘0- e53 of phosphorus pentachloride in carbon tetrachloride (gave a light, . gr'lien liquid when the excess phosphorus pentachloride, phosphorus o3"‘3rchloride and carbon tetrachloride were removed by distillation under 25 vacuum. The infrared spectrum of this liquid showed that little un- reacted ketone- remained. - The observed susceptibility of this material tohydrolysis or decomposition prompted us to forego its purification ’and examine the carbon tetrachloride solution after reflux. The proton magnetic resonance spectrum of this solution strongly suggests com- plete conversion of the ketoneto the dichloride (see Figure 3). Thearyl hydrogen on the methylated benzene ring shifts from .3. 19 to 2. 227‘ . this low value being similar to the position of the aryl hydrogen in tri- chloromethy’lprehnitene (2. 27‘1“). The aryl methyl hydrogens also shift, from bands at 7. 88 and 7. 78‘r' (corresponding to three and nine protons) in the ketone to bands at 8.18, 7.79, 7.88 and 7.67%, each withequal Hydrolysis of this solution with aqueous acetone gave quanti- intensity. These data, in tative recovery of 2, 3, 4, 5-tetramethylbenzophenone. conjunction with the results found when this solution is extracted with “100% sulfuric acid (see next section), offer strong support for the formation of 2,63, 4, 5-tetramethylbenzophenone dichloride (XXVI). :éjrwr: XXVI O ' In a- similar manner, . 4, 4'-dimethylbenzophenone dichloride " (XXVI), 2, 4-dimethy1benzophenone dichloride (XXVII) and 3, S-dimethyl- benzOphenone dichloride (XXVIII) were prepared from the ketones; their XXIX XXVII . XXVIII . 26 '9 1D 0% Rxn. m TMS r l 1 L 2.19 ‘Ho 2.73 +3. 19 7.78 7.88 __.) 1" Figure .3. -Proto‘n magnetic resonance spectra of 2, 3, 4, 5-tetramethy1benzophenone and phosphorus pentachloride in carbon tetrachloride before and-after reaction. 100%-Ram. 7. MS . . . .9). N 2.222.38 'H 2.88 32.67 7.797.38 8.18 10.00 0 3 T 27 protonmagnetic resonance spectraare consistent with their'structures (see Table 'V). ‘No attempt was made to isolate these dichlorides be- cause .for'our purposes impurities of ketone (which would undoubtedly persist on purification) would only conflict with their study in sulfuric acid. 4,4"-Dimethylbenzoph—enone dichloride-(XXVII) is reportedly formed (40) from the reaction of 4,- 4"-dimethylbenzophenoneandphos- phorus pentachloride, but attempts to further purify the thin green oil obtained-ledto decomposition and/or hydrolysis. The assignment of those remaining bands appearing in the proton magnetic resonance spectrum of 2, 3, 4,»5-tetramethy1benzophenone di- chloride is now possible with the information provided by the spectra of the other dichlorides, XXVII, XXVIII and XXIX. The band appearing ,at‘8. 18 T is apparently due to the Z-methy‘l hydrogens for-the following reasons; examination of Fisher-Hirschfelder models shows that this dichloride has similar steric requirements as 'amodel of pentamethyl- benzophenonedichloride, althoughthere seems to be more freedom of rotation withonly one ortho .methyl group present. -Consequently, . the phenyl ring probablyrlies orthogonal to the tetramethylbenzene ring'in its equilibrium position, .placing the ortho .methyl hydrogens in the shielding region of the benzene ring current (XXX). C1 \\\b/C1 CH3 ‘ CH3 8 CH3 CH3 . XXX Such anisotroPic effects with aromatic rings have been observed on murderous: occasions (43). The dimensions of the shielding (+7) and deshielding’(-) regions in benzene are known quite accurately (44) and 'XXXI, serves toillustrate their general shape. 28 .muofinsop m< m0 3mm .m we souconvc. $333.58 a mo moEoHuXo 0.25 may oudooonmom U down—summon“ 030.3030 .3 configuouofl n— .cofioom Housofiwuomxm 0:... 5 @050“ 03 Geo 83mm.“ mega—masque w oom.~ u :5 oomd n cad e S .s Np c. «cocoreouauflsfiufises .m w pow N co» N e 3 . s 3 . s «coconaouaoflsfiufis .. a .s w omoé uo~.m oomd .. om.m . 38:5 m 3;. .88 fig .34. maoaunmouceflsfionseés wagon“ Ann—Mad Lx 503.303 Amid: #503303 wonuongz m upwuogoflfi econ—0M padomaoo oguofiaomunona manogmmognfi :33 03.3th32 can 0.8on oguoasoduuofi GOQHMU. ad 0063030?“ oaocoamoudom coedfiumafim m0 ouuoomm ooGMGOmoM .oflodmmz 030nm .> 3an 2:9; , + XXXI + C D This assignment of the ortho methyl hydrogens to higher field is substantiated by the proton magnetic resonance spectrum of Z, 4- dimethyl benzophenone dichloride. The ortho methyl group is found up field (8.07% ) relative to the ortho methyl of the ketone (7. 69 or 7.72‘1‘). “Since this band position is unusually high it seems only reasonable to ascribe this shift to a shielding effect of the phenyl ringcaused by the steric hindrance of the ortho methyl group. .Several factors contribute to the failure to observe similar. effects in either of the ketones, _i. e_. , 2,4-dimethy1benzophenone or 2.3.14.5- tetramethylbenzophenone,V the major contributor being the trigonal hydridization at the carbonyl carbon. -As was .mentioned earlier; this bonding permits a greater freedom of rotation about the carbonyl carbon- carbon bonds; consequently, the ortho methyls experience an averaging effect of conformations such as XXXII and XXXIII. ‘Many other conform- ations are possible but these serve to illustrate the shielding (XXXII): and deshielding (XXXIII) positions which the phenyl ring may take relative to the ortho methyl group. II II ‘ CH3 1 C - CH3 ‘ __3 , I <— . CH3 . CH3 XXXII . XX XIII 30 Also to be considered is the electron-withdrawing resonance effect of the carbonyl which-lowers the positions of the methyls in the ketones relative to the dichlorides. r The demand for overlap also plays an important role indeciding the best possible conformation. . Certainly the proton magnetic resonance spectra of these ketones rule out the exclusive contribution of conformations similar to those found in the corresponding dichlorides. The large number of factors to be con- sidered prevents any definite assignment of a major contributor to the overall conformations of the molecules. The proton magnetic resonance spectrum of 2, 4, 6-trimethyl- benzophenone illustrates the situation where two ortho .methyl groups are present. The ortho methyl hydrogens appear much higher (8.00% ) than the single para methyl group (7. 761‘ ) indicating that the phenyl ring is again forced to lie orthogonal to the trimethylbenzene ring in its equilibrium position. Consequently, there remains'little doubt that the anisotropic effect of the benzene ring is the cause of these rather high values for methyl groups bonded toan‘aromatic ring. ~Although one might argue that'in 2, 4, 6-trimethylbenzophenone, the carbonyl group is causing the ortho methyl group toappear at higher :field‘ (see'Jackman (45) for theanalogous anisotropic effect of the carbonyl), we prefertoeascribe the greater portion of this effect to be due almost entirely to the benzene ring because of similar effects observed in. the dichlorides, where no carbonyl is present. The remaining bands in 2, 3,44, 5-tetramethy1ben20phenone di- chloride are more difficult. to assign but those appearing at 7. 79‘and 7.88 1‘ are presmnably the meta methyl hydrogens since the amethyls in 3, 5-dimethylbenzophenone dichloride appearat higher field (7. 81"") than the para methyls found in 4, 4'-dimethylbenzophenone dichloride (7.69'1"). 31 .Benzophenone dichloride was prepared for comparison with the other dichlorides. Inlthis case, the dichloride was isolated and was found to have properties-in agreement with those reported (b.p. and infrared spectrum) (45). . Several attempts to prepare 2, 4, 6-trimethylbenzophenone di- chloride by reacting phosphorus pentachloride with 2,4,6-trimethyl- benzophenone were not successful. In all instances there was evidence of considerable unreacted ketone in the product mixture. In a typical experiment, 2,4,6-trimethylbenzophenone and excess phosphorus penta- chloride were refluxed in carbon tetrachloride for seventeen hours. The protonmagnetic resonance spectrum of this solution after reflux showed bands virtually identical to 2, 4, 6-trimethy1benzophenone (see Table VI), althougha small (0.1m ) but real downfield shift was found in the entire spectrum. The visible and proton magnetic resonance spectra of an acid solution, obtained by extracting the same-carbon tetrachloride solution with‘101% sulfuric acid, was identical inlall respects to a spectrum obtained by dissolving 2, 4, 6-trimethylbenzo- phenone in 101% sulfuric acid. Table VI. Proton Magnetic Resonance Spectra of 2, 4, 6—Trimethyl- benzophenone in. Carbon Tetrachloride Before 'and‘After ' Reflux with ‘-Phosphorus ~ Pentachloride Before Reflux » After Reflux Number of ' (position, ‘1‘ units) (position, ‘I‘. units) Protons 8. 00 7. 97 6 7 . 7 3 7 . 70 3 ;- 3. 30 3. 20 2 2.90-2.20a 2.81-2.12a 5 aRepresents the range of a multiplet. 32 B. Solutions of the'Benzophenone Dichlorides in1Sulfuric‘Acid f A 2. 00ml. aliquot of 2, 3,74,-5-tetramethylbenzophenone dichloride (0.403 M) in carbon tetrachloride (_i_. e_l, the solution of phosphorus pentachloride and 2,3,4, 5-tetramethylbenzophenone after twelve hour reflux), when extracted with 101% sulfuric acid, gave a deep red colored sulfuric acid solution, the color persisting for long periods of time. .Copious evolution of hydrogen chloride was also observed partly due to the presence of excess phosphorus pentachloride ~andphosphoru-s oxychloride. After hydrogen chlorideceased to be evolved (10 min.),' the acid solution was separated and poured onto ice, giving a quantitative recovery of Z, 3, 4, 5-tetramethylbenzophenone (verified by m.p. ,‘ m.m.p. and infrared-spectrum with an authentic sample). These results showed that complete extraction of the dichloride was accomplished and that no gross changes in thebenzenoid structure had occurred. . The visible spectrum of this acid solution had bands at 468 and 355 mp. which changed with time to 429 and-345 mp. (see Figure 4). This change of spectrum is attributed to sulfonation for-the following reasons. -When the acid-solution was allowed to stand at room temperature for long periods of time (2 hr.),‘ hydrolysis on ice gave lower yields of ~ recovered-ketone (41%), the remaining material being water-soluble. The rate of change of this spectrum varied with‘acid concentration, faster at high-er concentrations (103%) and slower at lower concentration (100%). .Similar observations were emade by Hart and Sulzberg (46) in the-visible spectrum of 4-biphenylyldiphenylcarboniumlion in sulfuric acid. At high acid concentration the spectrum changed from bands at 510 and 420 my. to bands at 476 and432 mu. Hydrolysis experiments confirmed this change to be due to sulfonation. .Similar observations have been'made-in the solutions of other molecules in sulfuric acid ' (47,48). .33 35v 83:30.53 owe on; owe cums an:s owm com . 00m cos owm com osm o~m com _. _ fi __fi_ a J _ a q 4 a a . b mafia; vwom. vN.oN . 00.3.. . flax; .. mafia?” n0um0 15 00¢ 00 A525 mafip00m 0&5 HNMV‘M ..'. 0:20.003 \. .053 m0 don—05¢ 0 mm .300 0303.90 $2: I. a a Q 5 03.3303 0noc0nmoun0£>50amnu0uum as m N no guo0nm 0333/ find uofiowkmsab .w 0.2%«h gOIx-9 om 34 ‘Carbon tetrachloride solutions of benzophenone dichloride‘(XVIIIb), 2.,4-dimethylbenz0phenone dichloride(XXVIII), 4, 4'-dimethylbenzo- phenone dichloride (XXVII) and 3,5-dimethylbenzophenone dichloride , ”(XXIX) also give orange to red sulfuric acid solutions when extracted with 101% sulfuric acid. Their visible spectra are completelyanalogous to the spectrum of the tetramethylbenzophenone dichloride (XXVI) in sulfuric acid‘(see Figure 5). In contrast to Z, 3, 4, 5-tetramethy1benzo- phenone dichloride (XXVI), these compounds gave stable spectra with instability observed only after long periods of time (6 hours). .When the dichlorides were extracted with 96% sulfuric acid, the spectra changed with time to the absorption characteristic of the respective ketone. A comparison of this spectrum after hydrolysis with that of the ketone, not only verified ketone formation, but acted as a check on the concen- tration of dichloride initially present. In‘Figure 6 can be found the visible and ultraviolet spectra of 2, 3, 4, 5-tetramethylbenzophenone,3 . Z,4-dimethylbenzophenone, 3, 5-dimethylbenzophenone, 4,4'-dimethyl- benzophenone and benzophenone in 100% sulfuric acid. Included for [comparison is the spectrum, of 2,4, 6-trimethylbenzophenone in 100% sulfuric acid. The spectra of this ketone andthat of benZOphenone have been previously reported in sulfuric acid (xmax 340 mg for benzo- phenone and xmax 300 mp. for 2, 4, 6-trimethy1benzophenone) by Fischer . and co-workers '(49) and are in-agreement with our findings. However, - these workers found that 2, 4, 6-trimethylbenzophenone was unstable above 75% sulfuric acid. The spectrum found by us remained stable in 100% and 85% sulfuric acid for 15 minutes after mixing. .This stability was verified by examination of the proton magnetic resonance spectrum which also remained stable and‘by hydrolysis of the acid solution,tgiving quantitative recovery of ketone. If the sulfuric acid concentration was raised to 101% however, sulfonation did occur, the visible spectrum now exhibiting bands at 309.5 (6: 22,000) and 360 mp. (e = 6, 000). 35 .pflom QCSESm $03 5 mopflnogofip ocoqosmoncon— "303.9,. mo 9300mm “038$, .m ondwfim oom A165 Aumamfioxrmg omw _ _ on; o—mm oom , _ .. .. a . a I. in...» x x. \\ “Mama... . a I .\..|./ \\+ ..... \ . .. .0 Cy. \- I| /o ‘HJVJ. . .... 5 \. \\H§¢afl0r \ +.... .3 . \\+.7$.oo . .fln \ .7 . My \ \ ’1’ o ”0* .9 to I \. \ £4.19 ... V . ... 4. w. . )4... .. c v . g a. \ 1‘ ’ 4’ Q 0‘ x x . . Kw .. . y + \. a» . I no .\ fig 0 o /4/ \ .9 I. $~ {$4. / (I... .\ 1.1“ .. 50c \oco . o 4 \ .o «I o fie/40’. f\¥4U\\ .0 .400 O m . o o, M o \ o. no w OH I I’lfiofl \ .o 0’ .\.o o no . . \ I... c o o. . x. \. \I r. o o o O o O 0 9:00 . i m M £00 QINJUU Ill 3300 3 4.4.5.5. x. t I 0. 3-300 36 Table VII. .Visible Absorption Maxima of Some Benzophenone Dichlorides in Sulfuric Acid Wavelength Molar Absorbancy Compound (mp) Index (6) z, 3, 4, 5-Tetramethylben20phenone 4683' 28, 900 Dichloride .. , ‘ v 355 13, 000 2, 4-Dimethylbenzophenone 454 23, 320 Dichloride 344 14, 480 3 , 5-Dimethy1benzophenone 444 40,810 Dichloride 342 16, 880 4, 4'-Dimethy1benzophenone 47 1 68,800 'Dichloride 343 18, 000 Benzophenone Dichloride 435 36,456 335 18,790 a . vExtrapolated to A zero. time. A E Q ¢ ___- \¢ mluaxxii. Figure 6. Ultraviolet and visible spectra of some benzo- Q phenones in 100% sulfuric acid. {\ 3. 0 — 1 I l 2 50 300 350 400 Wavelength (mp) 38 Table VIII. Ultraviolet and Visible Maxima of Some Benzophenones in 100% Sulfuric Acid --Wavelength » MolarVAbsorbancy Compound (mp) Index (6) z, 4-Dimethylbenzophenone 298 14,490 351 .21,500 3, 5-Dimethy1benzophenone 297a 12, 100 352 30,450 4, 4' ~Dimethylbenzophenone 302 ll , 920 373 37,400 2,33, 4, 5-Tetramethy1benzophenone 300 13,:975 37 1 15, 528 Benzophenone 345 24,064 2,‘4,.6-Trimethylbenzophenone 305 22,952 4, 136 aTaken in 96% sulfuric acid. 39 Sulfonation was verified by hydrolysis experiments which gave poor recovery of ketone (50%) and, by examination of the proton magnetic resonance spectrum which changed with time to a spectrum that had three distinct methyl groups, a result inconsistent with simple ketone in sulfuric acid (see Table IV). Instability was also observed in the visible spectrum of 2, 4-dimethy1benzophenone, but only after standing for long periods of time. Very littlework has been done with various benz0phenone di- chlorides in sulfuric acid.‘ Hantzch (25) noted that benz0phenone di- chloride itself gave an intense yellow solution in sulfuric acid which quickly decolorized, while losing hydrogenchloride. These workers also found that deep orange crystals could be isolated whenthe di- chloride was treated with stannic chloride in chloroform. To these crystals they assigned the structure shown below. Unfortunately, this C1 T c': + SnCls- _. _l material was unstable even in a desiccator containing phosphorus pentoxide, leaving the question of structure incompletely answered. 4,4'-Dipheny1benzophenone dichloride, when dissolved in sulfuric acid (28) gave a deep blue solution, but no attempt was made to determine the species present. Our original reason for synthesizing 2, 3, 4, 5-tetramethylbenzo- phenone dichloride(XXVI) was to see whether ionization in sulfuric acid would give a stable dicarbonium ion. 'Although'the choice of this dichloride rather than 2,3,4, 5, 6—pentamethylbenzophenone dichloride was initially based onits ready synthesis, there remained sound reasons to suspect dicarbonium ion formation. Examination of Fisher—Hirsch- felder models shows that complete coplanarity of the initially formed 40 mono-ion (XXXV) is impossible and only one phenyl ring at a time can 3:1\ XXXV +. lie coplanar with the -C-Cl group. The loss of a second chlorine atom would relieve this steric hindrance because of the linearity of the result- ing dicarbonium ion in all its resonance forms. The successful ionization of trichloromethylprehnitene, which has similar steric requirements as 00~©s 2,3 4, 5- -tetramethylbenzophenone dichloride, to a dicarbonium ion, gave credence to these predictions. -However, the visible spectra of the benzophenone dichlorides in sulfuric acid reported here bear. a striking resemblance to the reported spectra of various diphenylmethylcarboniumions. Table IX records the visible absorption maxima of some typical mono-positive carbonium ions ~- in sulfuric acid (50). .0 It is also difficult to reconcile the similar visible spectraingsulfuric acid for all the dichlorides regardless of steric hindrance unless, of course, steric hindrance is less important when the resulting di-ions have unusual stability, i_. e: , diphenylmethyldicarbonium ions. The formation of mono-positive carbonium ions for these dichlorides in sulfuric acid, rather than dicarbonium ions, 7 was supported by exami- nation of the proton magnetic resonance spectra of these solutions (see Tabl-e‘X). The proton magnetic resonance spectra of the respective ketones in sulfuric acid are reported in Table XI for comparisonJPage 42). 41 Table IX. Visible Absorption‘Maxima for Some Diphenylmethyl Carbonium Ions in Sulfuric Acid «:Wavelength ‘Molar'Absorbancy Compound (mil) . Index (6) 1, l-Diphenylethylcarbonium ion 430 ‘22, 300 1, 1-[4, 4"-Dich~lorodiphenyl]ethyl- 466 45, 700 carbonium ion 1, 1-Diphenylpropy1carbonium ion 432 54, 700 1, 1-Bis[p-toly1]ethylcarbonium ion 458 63, 100 The species present in the sulfuric acid solutions of these benzo- phenones may be represented-as a series of resonance forms. + OH OH OH In .2, 4-dimethylbenzophenone one methyl group appears at'lowerfield than the other. -An assignment of this peak is possible by comparing this, spectrum with 4,4'-dimethylbenzophenone in sulfuric acid. .Here the para‘meth-yls‘ appear 'at‘ 7. 21 'l’ , almost identical with the highfield band present in 2,4-dimethylbenzophenone (7. 24 'T). . Thus the band appearing at’ low field must be due to the ‘orthormethyl group. Incon- trast, a similarspectrum of 2, 4, 6-trimethylbenzophenone in sulfuric acid has the orthomethyls appearing at higher. field (7. 60 '1’). ' However, the high value of the para methyl (7. 35 '1'), may explain this discrepancy. . The same ketone in carbon tetrachloride also has the ortho ‘methyl groups at higher field and this was ascribed to nonplanarity caused by steric hindrance. ~ Therefore, the ortho methyls lie in the shielding 42 Table X. .Proton‘MagneticResonance Spectra of Various Benzophenone Dichlorides in 101% Sulfuric Acid-dz Position: Number of Compound ' ‘ . ‘. . ~ (Tunits) Protonsd 2, 3, 4, 5-Tetramethylbenzophenone 7. 26, 7. 38, ‘3 (apiece) Dichloride 7.42, 7.48, 1.34-2.21b 6 2, 4-Dimethy’1benzophenone 7 . 14 6 Dichloride 1 . 56-2. 56" 8 4, 4'-Dimethylbenz0phenone 7 . 11 6 Dichloride 1 . 9 1c 8 '3, 5-Dimethylbenzophenone 7 . 27 6 Dichloride 1 . 29-2. 16 8 Table XI. Proton Magnetic‘Resonance Spectra of Various Benzophenones in 100% Sulfuric Acid =- Positionsa Number of Compound (’1' units) Protonsd 2,3,14,5-Tetramethy1- 7. 38, 7. 42 12 benzophenone 2. 53 b 1 l . 54-2. 24 5 2, 4-Dimethy1benzophenone 7. l4, 7. 24 3 (apiece) w 1 . 59-2. 74b 3 4, 4'-Dimethylbenzophenone 7 . 21 6 2. 0°C 8 2, 4, 6-Trimethy1benzophenone 7 . 60 6 7 . 35 3 2. 63 1 1.36-2.24b -5 aReference: 1 Methanesulfonic acid - 6.46 T. cIndicates ’T' values on either, side of a multiplet. lCenter of an AB quartet. Obtained by electronic integration. 43 region of the benzene ring current. ~Although the demand for 'planarity is greater» for the ketone in sulfuric acid,~ because of a need to disperse the positive charge, models indicate that planarity is impossible. This non-planarity causes the para methyl group ,to-appear 'at‘high field. * This same effect was not. found with 2,4-dimethylbenzophenone nor ~with 2, 3, 4, 5-tetramethylbenzophenone, either inicarbontetrachloride or‘sulfuric acid. One orthomethyl group is insufficient toforce the molecule to be non-planar; therefore, the orthomethyl experiences both shielding and deshielding regions of the ring current. ~One can ascertain the relative amount of positive charge on‘an atom in a given compound by studying the hydrogen bonded directly to that atom XE proton 'magnetic resonance spectra. - By-comparingthe chemical shift of such hydrogenswith various references, the degree of positive charge is‘discernable. This technique has proved useful in the determination of the relative charge distribution of triphenyl- carbonium ion (51,52). «A comparison of the proton magnetic resonance spectra of the dichlorides 'in sulfuric acid with their corresponding ketones certainly, shows a similarity of charge density. For example, the position of the para methyl in4, 4"-dimethylbenzophenone~ (7. 211') appearsat only slightly higher field than the-same methyl inthe corres- ponding dichloride (7,. 114*). If we are dealing with monopositive carbonium ions, ,this is in agreement with the greater ability of hydroxyl ' to. stabilize ‘a positive charge than'a chlorine, .i_._e_. ,, XXXVI we XXXVII. .Since the visible spectra of protonated ketones differ substantially ch 030 XXXVI XXXVII 44 from normal carbonium ions, it is questionable whether they can be compareddirectly with molecules which developa full positive charge. ~ A more likely candidate for comparison with 4 ,n4'-dimethy1benzo- phenone dichloride might be A 1, 1-(p-tolyl) ethylene, which-reportedly gives 1,11-~(p-tolyl)methy1carbonium ion (XXXVIII) in sulfuric acid. ‘;b XX XVIII The visible spectrum of a sulfuric acid solution of this olefin-(Table IX) bears 'a striking resemblance to 4, 4'-benzophenone dichloride in sulfuric acid (xmax 458 my. (6 3.63, 100) versus xmax 471 mp. (e = 68,800), - The proton magnetic resonance spectrum of this bis(p-tolyl)methylcar- bonium ion‘(see Table XII) has a band'at]. 11 ’T‘for the para methyls,- the identical positionfound forethe para methyls in 4, 4'-dimethylbenzo- phenone dichloride in sulfuric acid. .Consequently, there remainslittle doubt that thesedichlorides, when diss‘olved'in 100% sulfuric acid, give mono- ratherithan dicarbonium ions, withlloss of one chlorine atom. I Still further proof was sought for the formation of mono-ions, particularly in~the case of 2, 3, 4, 5-tetramethylbenzophenone dichloride, . since it was our earlier feeling that dis-ion formation would mostilikely . occur‘with this (compound. -It‘is interesting to compare in detail the proton magnetic resonance spectrum of 2., 3, 4, 5-tetramethylbenzophenone -- insulfuric acid with that of the corresponding dichloride. - In the. ketone, the four aryl methyls appear'at 7. 38 and 7. 42 '1‘ , whereas in the dichloride the -methyls appear as four separate bands at 7.48, A. 7.42., 7. 36 and 7. 26’T. - The-low field band inthe dichloride is in agreement with the shift observed in the spectrum of 4, 4'-benz0phenone dichloride 45 "Table XII. . Proton'Magnetic Resonance Spectra of Some Substituted 1, l-Diphenylethylenes in 101% Sulfuric Acid. L ‘__-‘ "Positions Number of Compound (’1' units)a Protons 1", 1~Bis(p-tolyl) ethylene 7. 11 6 6. 26 3 (approx.) 1. 72b .8 1-(2, 3, 4, 5-tetramethylphenyl)- 7. 58 '3 (approx.) 1-phenylethylene 7. 41 6 (approx.) 7. 26 3 (approx.) 6. 26 3 (approx.) 2. 57 1 2.24—1.54C 6 afReference CH3SO3H, 6. 46 ’T' bCenter‘ of a pair of AB doublets CLimits of a multiplet relative to the ketone (7. 11 w 7. 211'). Quite-surprising is the appearance of one of the methyl groups of 2, 3,4, 5-tetramethylbenzo- [phenone dichloride 'at higher field (7.48 ’1‘"), higher than any band appearing inthe corresponding‘ketone. 1A. clue as to which methyl this -might be is provided by the proton magnetic resonance spectrum of 2,4-dimethylbenz0phenone dichloride in sulfuric acid. -Here the two methyl groups appear at the same band position (7. 141’) a surprising result since (the ketone in the same solvent has the ortho methyl at lower field (7. 14T) than the para methyl (7. 241'). It appears that the ortho methyl group in both of these dichlorides (i_. e_. , XXVIII and XXVI) in sulfuric acid, has shiftedupfield from the position‘one would normally Ifi‘ ’ ["1 46 expect. -Previous data with these dichlorides in carbon tetrachloride gave an identical result, explained by the shielding effect of the phenyl ring forced to lie orthogonal to the methylated ring because of steric hindrance. The observation of a similar effect here substantiates the postulate that mono- rather than dicarbonium ions were formed, for if the latter were present, there would be little reason to expect the - ortho methyl groups to lie in the shielding region of the benzene ring, , since the molecule would most likely be linear. In contrast, the con- formation of the mono-ions would be such that the phenyl rings would .=. lie over the orthomethyls in their preferred conformation'(XXXV). It was felt that l-(2, 3, 4, 5-tetramethylphenyl)-l-phenylethylene would be extremely useful as-a Amodel compound since in sulfuric acid it would presumably give 2, 3, 4, 5-tetramethylphenylphenylmethyl- carbonium ion' (XXXIX) :13; simple protonation, as found by Deno and co-workers ’(50) in similarly substituted ethylenes. This olefin was 5H3 C +\ XXXIX prepared by the reaction of 2, 3, 4, 5-tetramethylbenzophenone with methylmagnesium bromide. Its structure was verified by infrared, ultraviolet and proton-magnetic resonance spectra'along with micro- analysis. Insulfuric acid, the similarity between it and 2, 3, 4, 5- tetramethylbenzophenone dichloride is remarkable. 1 Its visible spectrum 47 in 96% sulfuric acid (imax 454 (e = 23,100)~ and 330 m... (6: 6,180» was quite stable, but instability was observed at higher acid concen- trations. The proton magnetic resonance spectrum in 101% sulfuric acid‘(see Table‘IX) exhibited bands at 7. 54, 7.41 and 7. 26 ’T corres- ponding to three, six and three aryl methyl hydrogens respectively, . a band‘at 6. 26 ’7' for three aliphatic methyl protons, a single band at 2. 37 ’T integrating for one hydrogen, and a multiplet from 1. 54 to 1.44 ’1" corresponding to five aromatic hydrogens. The similarity of this carbonium ion‘(XXXIX) to the species formed when‘2, 3,4, 5- tetramethylbenzophenone dichloride is dissolved in sulfuric acid sub- stantiates the formation of (2, 3, 4, 5-tetramethy1phenyl)phenylchloro- carbonium ion from the latter compound. These differences observed can be attributed to a methyl versus a chloro substituent onrthe e-carbon atom. The stoichiometry with respect to hydrogen chloride appeared particularly promising as a means to establish, unequivocally, the formation of 2, 3, 4,.5-tetramethylphenylphenylchlorocarbonium ion when the corresponding dichloride is dissolved in sulfuric acid. -One mole of hydrogen chloride should be produced if the monocarbonium ion is formed, whereas two moles would be expected if the ion‘were the dicarbonium ion. To accurately measure the moles of hydrogen chloride evolved, all chlorine containing impurities must be absent. 1 Consequently, pure, crystalline 2,, 3, 4, 5-tetramethy1benzophenone dichloride (m.p. 116-1170) was isolated by cautiously removing, _under vacuum, the solvent, phosphorus oxychloride and excess phosphorus pentachloride from the carbon tetrachloride solution of 2, 3, 4, 5-tetra- methylbenzophenone and phosphorus pentachloride after reflux. This compound proved-surprisingly. stable and was identical in carbon tetra- chloride to the solution obtained prior to isolation (proton magnetic resonance spectrum). Its infrared spectrum and elemental analysis 48 were in- agreement with its structure. ~ The amount of hydrogen chloride produced when 2,3, 4, 5-tetramethylbenzophenone dichloride is dissolved in sulfuric acid was determined quantitatively by passing a stream of dry nitrogen through‘thered sulfuric acid solution; hydrogen chloride was trapped in a 2% sodium hydroxide solution. The results are sum- ,marizedin Table VIII. -Each mole of 2., 3, 4, 5-tetramethylbenzophenone dichloride dissolved inI100% sulfuric acid produced one mole of hydrogen chloride. This quantity of hydrogen chloride was rapidly flushed from the solution, and prolonged sweeping (10-12 hours) gave negligible additional hydrogen chloride. - The visible spectrum of this solution after sweeping was unchanged and identical tothe spectrum reported earlier, 1. e. , the spectrum of the sulfonated ketone (see Figure 4). Table XIII. Stoichiometry with Respect to Hydrogen Chloride of 2,3, 4,5- Tetramethylbenzophenone Dichloride inSulfuric Acid a , b Moles of Cl-/ Wt- (g.) Time (hrs.) Ml. of AgNO3 Mole of Dichloride 0.22178 6 12.11C 0.81 0.7383 4 A 43.55d 1.02, 0.6392 4 36.75d 0.99b aln 25 ml. of 99.75% Sulfuric'Acid. ‘b Length-of time that nitrogen swept through the solution; the rates of flow varied. C0.05046 N, d0.05908 'N. eSample of dichloride was impure. . Whenithe hydrogen chloride-free solutionwas poured on cracked ice, there was obtained a deep-brown-aqueous solution, since by this 49 time all of the organic material would be completely sulfonated. The presence of sulfonated ketone prevented the determination of the remaining chloride. - These results, in conjunction with the earlier spectral data, con- firm the formation of mono-positive carboniumions when benzophenone dichlorides such as, XXVI, , VIIIb, . XXVIII, XXVII and'XXIX, are dis- solved in concentrated sulfuric acid. ~As was mentioned earlier, Whether one gets a mono- or a dicarbonium ion will depend on the effect which the substituent will exert on each ion; _i_. g. , . phenyl apparently , stabilizes the mono-ion more than the dicarbonium ion, since the latter is necessarily a rather high energy species anyway. .Consequently,.any attempts to further stabilize dicarbonium ions of the extraordinary . type must pay particularheed to the energetics of the initially formed mono-ion. , III. Miscellaneous Experiments Peach and Waddington (53) reported that phenylac etylene when dissolved in liquid hydrogen chloride gave phenylmethyldicarbonium ion‘(XL), its structure based solely on the determination of the moles of HClz P present. This diprotonation of a phenyl substitutedac etylene éCH CH / + / 3 .c 0+ _ -+ 4HC1 -———9 [:::T -+ 4ch ’ XL derivative suggested a similar approach to dicarbonium ion, XLI. + c-CH3 XLI 50 Consequently, the corresponding pentamethylphenylacetylene (XLIII) was prepared from the a-chloro olefin (XLII) by reaction with sodium [Cl CH2 EtOH + N aOEt —-—-———> X LII XLIII ethoxide in ethanol. The a-chloro compound (XLII) was prepared from pentamethylacetophenone, phosphorus pentachloride, phosphorus oxychloride and phosphorus trichloride. The structure of these com- pounds were verified by their spectra and elemental analysis. It is interesting to compare the ultraviolet spectrum of the two, . XLII and XLIII, in n-hexane, for the olefin has similar steric requirements as pentamethylbenzotrichloride , while XLIII, because of the linear aCetylene grouping, has considerably less steric hindrance (see Figure 7). As expected, the latter compound has much more detail absorption from 270 to 290 mu while the former chloro-olefin has absorbance quite similar to pentamethylbenzotrichloride. Unfortunately, the visible spectrum of the acetylene derivative in sulfuric acid (see Figure 8) is extremely difficult to interpret, particularly inview of the isolation of only tarry materials when these acid solutions were poured on ice. Time did not permit a further investigation of this compound. However, it is interesting to note that the unsubstituted homolog, phenylacetylene, in 100% sulfuric acid, reacts violently to give a red-brown sulfuric acid solution. Hydrolysis of a similar solution (99% sulfuric acid) on ice gives besides water—soluble materials, a low yield of acetophenone (55%). The yield of acetophenone ‘was improved by dissolving the phenylac etylene in 98% sulfuric acid, then pouring this solution on ice (68%). The proton magnetic resonance log 51 Figure 7. methylphenylacetylene in n-hexane 1 ‘3 \ \\ /\ , "s I \ \ \ I, \\ \ I \ \ .I \ \ I 1 \ \ \ I \ \ \ I \ \ I 1 \‘ \ \ r \ « 1 1 - 1 \ 1 \ 1 l 1 250 300 7 Wavelength , (mu) Ultraviolet Spectra of 1-chloro-l-(2, 3, 4, - 5, 6-pentamethylphenyl)ethylene and 2, 3, 4, 5, 6-penta- 52 spectrum of these latter-solutions had bands at 6. 19 and 4. 25 "T‘ in the relative areas of two to one and a multiplet from 2.00 to 0. 90 ’T' inte- grating for an area of seven. -Acetophenone in the same acid concen- tration has a band at 6. 19 ’7' (three protons) and a multiplet from 2. 00 to 1. 00 ’1' corresponding to five protons. Therefore, it appears that acetophenone is partially formed in the sulfuric acid solutions along with other materials which were not identified. . Similarly the visible spectrum of phenylacetylene in sulfuric acid (98%) has bands at 432 (E: 503) and 295.8 mu (6 =20, 950) whereas acetophenone in 95.7% sulfuric acid has Xmax 295 mu (6 = 19, 550). The visible spectrum of the chloro olefin, XLII, in sulfuric acid is shown in Figure 9 and shows remarkable differences when taken in 98% sulfuric acid or in 100% sulfuric acid. 53 00> 35V 33:30.45? ooc com oov com CON ‘-'”n—-—-- _- _ . — _ — .Bom 03316 «mos 9." 0:330um~>ao£mfi>£uo~bmucomso .m .44 .m .N mo 55.3023 03fi3> _ .w 0.3th I 801 54 35V Sumnmamxfimg om m oLom omv o3 omm com J a _ _ _ 7 I / / \mfi $3 I I l. l/ // I, l I! I .1 I I z / // / \ . I ll.\ «Omam $2: , // \ .Bom 035345 $2: can 3 E max mo 938% «33> 98 3338:: .o 3&3 \ // \. / \ 301 55 The proton magnetic resonance spectrum of this chloro olefin in 101% sulfuric acid has bands at 7.33, 7.04 and 6.88 1‘ correspond- ing'to approximately six, three and three protons, a band at 5. 88 1, for two protonsand a broad band at 4. 18 7- corresponding to one proton. The latter band was not present when the spectrum was taken in 101% deutero- sulfuric acid. There is some uncertainty in the electronic integration because the bands were unusually broad. Hydrolysis of these acid solutions on ice gave l-chloro-l-(Z, 3,4, 5, 6-pentamethy'l- phenyl) ethylene in 80% yield. In contrast, the proton magnetic resonance spectrum in 99% sulfuric acid has bands at 7.01, 7. 25 and 7. 53 ‘1‘ corresponding to two, three and twelve protons. At thistime, there seems to be no simple explanation of these results and any interpretation must wait until further work can be done. E XPERIMENTAL I. Trichloromethyl Compounds A. Synthesis Preparation of Pentamethylbenzene. This compound was prepared by a modification of the methodvdescribed by Smith‘and Shac‘klett‘ (55). 7 Improved yields of pentamethylbenzenewere obtained by using lithium aluminumhydride in place of aluminum hydride as the reducing agent and by chloromethylating durene using a procedure suggested by‘ Aitken and co-workers (56). (a) Chloromethylation of Durene. - In a two-liter, three-necked, round-bottomedrflask equipped with a Tru-bore stirrer, a thermometer and a reflux condenser was placed 154.0 g... (1.148 moles) of durene, 350 m1. of acetic acid, A 0. 770 m1. of concentrated hydrochloric acid, .and 65- g. of formalin. .The mixture was stirred vigorously for. four hours at 70°. -An additional 38. 1 g. of formalin was then added and the mixture stirred and heated for four more hours. The flask was then cooled with stirring and 250 ml. of benzene added to dissolve the pre- cipitated solid. .An additional 100 ml. of benzene was used to wash the reactidn flask. -The combined benzene layers were washed with two -50—m1. , portions of water, then neutralized with a 5% sodium bicarbon- ate solution, , dried over anhydrous potassium carbonate, and the solvent removed on a Rinco rotary evaporator. The residue was distilled through a twenty-inch glass-helices packed column using a heat lamp to prevent solidification in the distillationhead. The boiling point was 3:: All analysis by Spang Microanalytical Laboratory, Ann‘Arbor, .Michigan. Melting points are uncorrected unless otherwise indicated. 56 57 147-14801at 22 mm. The yield of chloromethyldurene was 160 g... (78%), .m.p. 63-650. The residue weighed 15 g. - (b) Reduction of Chloromethy’ldurene. To a suspension of 14. 21 g..- (0.375 mole) of lithium aluminum hydride in.350 m1. of dry,. stirred, refluxing tetrahydrofuran was added'over twohours, a solution of 137. 3 1g... (0. 750 -mole) of chloromethyldurene in 750 .m1.. of dry tetrahydrofuran. iThe mixture was stirred at reflux for an additional one-half hourethen cooled in an ice bathand small pieces of ice added cautiously‘to hydrolyze the excess lithium aluminum hydride. To this mixture was added .100 ml. of 10% hydrochloric acid and‘an equal volume of water. ~The result- ing mixture was extracted with 500 m1. of ether and the separated ether layer dried over anhydrous potassium carbonate. ~ The solvent was - removed ona Rinco rotary evaporator and the residual solid recrystal- lized fr'om- acetone, ‘ m.p. 51—520 (55). The yield of pentamethylbenzene ~was 198 g.. (85%). . Preparation of Isodurene. This substance was prepared by a method similar to that used in the preparation of pentamethylbenzene," starting with the readily available -mesitylene. . In a two-liter, three-necked, round-bottomed flask equipped with -a Tru-bore stirrer, a thermometer and» a reflux condenser wasplaced 117 g.., (0.984 mole) of mesitylene, 300 m1. of acetic acid, 600 m1. of concentrated hydrochloric acid,. and 57.0 g.. of formalin. This mixture was stirred vigorously at 400 for» four hours, after which-an additional - 26.4 g. , of formalin was added and the mixture stirred for four-more hours at 40°. -Workup was identical to that used for pentamethylbenzene. .Disvtillation througha‘twenty-inch glass-helices packed column gave chlorometh-ylmesitylene, , b.p. 119-1200 at 15 mm. The yield was 119. 7 g. . (72. 2%) . 58 ‘Reduction was accomplished in exactly the same way as with Chloromethyldurene. Isodurene was obtained in 82. 1% yield, b. p. 62-630 at 5mm. (55). The infrared spectrum of this compound is shown ianigure 10. .Preparation of Prehnitene. ‘ - This compound was preparedfrom hemimellitene '(1, 2, 3-trimethy1benzene) following essentially the method used in the synthesis of the other polymethyl aromatic hydrocarbons. A mixture of 59.04 g.. (0.492 mole) of hemimellitene.-. 300 m1. of concentrated hydrochloric acid, 28. 5 g.. of formalin and 150 m1. of acetic acid was stirred and heated to 400- for. four hours after‘which' 13.2 g. of additional formalin was added and the. stirring andiheating continued ‘ for'four -more hours. . Workup gave 60. 5 g- of Chloromethylhemimellitenei,‘ . b.p.. 127-1280 at 15mm) (73%). .Reduction by the same method used-for the pentamethylbenzene and isodurene preparations, gave prehnitene in 83.5% yield, b.p. 86-870 at 16mm. The infrared spectrum of this material was identical to a known sample (41). Attempted'Preparation of Trichloromethylisodurene. -An attempt was made to'prepare this compound by'the Friedel-Crafts reaction of isodurene-with carbon tetrachloride according to the procedure used in the synthesis of trichloromethylpentamethylbenzene (57) with slight modific ations . > .In a dry,. BOO-ml. , three-necked, round-bottomed flask equipped with -a ‘Tru-bore stirrer, .a thermometer, a reflux. condenser and an addition funnel was placed 26. 7 g.. (0. 2 mole) of anhydrous, . powdered aluminumchloride and-33 m1. of carbon tetrachloride. To this stirred suspension, a. solutionof 13.4 g... (0.1 mole). of isodurene in 33ml. of carbon tetrachloride-was slowly added. - The slurry became deep purple and hydrogenychloride evolved immediately. The stirred re- . . . . o . . action mixture was maintained at 37—42 throughout the addition 59 Amconowgvcuwcofioermg Ma Ma S A: . o. m N. o m, «s _. a _ _ ._ _ _ _ _ \ Anon—53m ~mUV oooudvofl no 2330on “schema. .oH ohnmwh. ,, 60 ‘(two hours), andfor two more hours. . The cooled, purple, aluminum chloride complex was poured into a vigorously. stirred mixture of 33 ml. of carbon tetrachloride, 100 g. of ice,_ 50 ml. of waters and 16 m1. .of concentrated hydrochloric acid. rThe orange organic layer was separated,— . washed with three 16-ml. portions of warm (400) water and dried over anhydrous sodium sulfate. The solvent was removed on a Rincolrotary evaporator and the oily residue distilled through a six-inch Vigreux column. ,-It was distilled rapidly in order to minimize decomposition. The distillate collected at b. p.. 14:0«144o at 1 mm. weighed 12. 5 g.. The residual tar weighed 11.0 g. . The distillate solidified, m.p. 87-900, and recrystallization from anhydrous acetone, using a dry ice-isopropyl alcohol bath, gave white crystals, m.p. 89-900 which had an identical infrared spectrum withan authentic sample of trichloromethylprehnitene (41). A mixture m.p. was not depressed. I A small sample of the trichloromethyl compound, . when dissolved in 15 ml. of 100% sulfuric acid, gave a deep red solution with evolution of hydrogenchloride. - After gas ceased to be evolved (15 min.),‘ the solu- tion was poured onto alargeexcess. of ice. - The resulting precipitate was filtered and recrystallized from anhydrous ethanol. -A 92% yield of white crystals, m.p.. 166-1670, was found tobe 2, 3,4, 5-tetramethy1- benzoic acid, identical with an authentic sample (57). -A mixture m.p. was not depressed. The same acid was obtained by refluxing a sample of this trichloro- .methy'l compound in 50% ac etone-water mixture for five hours. . Preparation of Brom0prehnitene. This compoundwas prepared by brominating prehnitene using the procedure of Smith and‘Moyle (57). In a300-m1, three—necked, round-bottomed flask equipped with a Tru-bore stirrer, addition funnel and condenser was placed 12.5 g. (0. 094- mole) of prehnitene,. 25‘ m1. of glacial acetic acid and a crystal of iodine. Light was excluded from the reaction mixture by wrapping 61 the flask with black friction tape. To this stirred reaction mixture was added 15.7 g. (0.098 mole) of bromine at 00 over a period of one and one-half hours. - The reaction products were then poured into water, and the water decanted from the semi-solid mass. The solid residue was washed thoroughly with 50 m1. of 5% sodium hydroxide followed by distilled water, then dissolved in ethanol and recrystallized at dry ice temperatures. These crystals'were dissolved inpetroleurn ether'and dried over anhydrous magnesium sulfate. 'After removal of the solvent on a'Rinco rotary evaporator, the residual solid was dis- tilled under vacuum through a short Vigreux column. Bromoprehnitene was obtained in 65%, yield, b.p. 154-1550 at 31 mm. , m.p.. 29-2300 (57). Its infrared spectrum is shown in Figure 11. . Attempted Preparation of Trichloromethylbromoprehnitene. _ An attempt was >made to prepare this compound from bromoprehnitene, carbon tetrachloride and aluminum chloride by a method essentially the same as that used in the attempt to prepare trichloromethylisodurene. In'a dry, 100-ml. , three-neckedflask equipped with a Tru-bore stirrer, thermometer, reflux condenser and addition funnel was placed 8. 0 g- (0. 06 mole) of powdered anhydrous aluminum chloride and 10 m1. of carbon tetrachloride. ~ To this stirred and heated (3.7-420) suspension was added asolution of 6. 37 g. (0. 0304 mole) of bromo- prehnitene and 10 ml. of carbon tetrachloride. Addition took two hours after which the purple complex was stirred and heated for three more hours.~ After the same process of hydrolysis and workup as described before, 5. 7 g. of tan crystals were obtained. m.p._84-887O. e Recrystallization from dry acetone, by cooling to dry ice temperature, gave a white crystalline material, m.p. 88-890. This material was found to be identical to an authentic sample of trichloromethylprehnitene by infrared spectrum and mixed melting point. a: Amcongfiv cohesive/m? ma NH 3 0H m w _ _ _ _ a _ . _ 62 .Acofisfioo umUv odouficgopmogouo. mo £5.30on popmflafi . .: oudwflh , i 63 Hydrolysis of this material with 50% aqueous acetone gave 2, 3,4, 5-.-~ tetramethylbenzoic acid, m.p. 168-1690. Identity with a known sample was verified by comparison of their infrared spectrum and by mixed melting point. . Preparation of Bromoisodurene. This compound was prepared following the procedure used in the preparation of bromoprehnitene. From 25 g.~ (0.187 mole) of isodurene dissolved in 50 ml. of acetic acid and 31.4 g. (0. 196 mole) of bromine, there was obtained 32.3 g.. of bromoisodurene, b.p. 108--110o at 5 mm. , r125 ="~l.5614; reported; n? = 1.5614 (57). Its infrared spectrum is shown in Figure 12. .Preparation of Trichloromethylbromoisodurene. This compound was synthesized from bromoisodurene,, carbon tetrachloride and an- hydrous aluminum chloride using similar conditions to those described inthe preparation of trichloromethylpentamethylbenzene (3). In a dry, 300-ml. , three-necked flask equipped with aTru-bore stirrer,thermometer, reflux condenser and addition funnel wasplaced 24.0 g... (0.18 mole) of powdered, anhydrous aluminum chloride and 30 m1. of carbon tetrachloride. i To this stirred suspension was added a solution of 19.11 g.. (0. 0912 mole) of bromoisodurenein 30ml. of carbon tetrachloride. The reaction mixture was stirred and kept at 37-420 during the addition and for two more hours. ~ After the usual hydrolysis and'workup, 24. 0 g- of light tan crystals were obtained, _ m. p. 86-880, in 80% yield. -Recrystallization from pentane after treat- ment of the solution with Norite, gave a pure sample of trichloromethyl- bromoisodurene, m.p. , 87-88. 50. Its infrared spectrum is shown in Figure 13. [Anal. Calcd. for C11H12C13Br: C, 39.37; H, 3.66; CI, 32.19; Br, 24. 18. Found: C, 40.14; H, 3.72;: C1, 32.10; Br, 24.00. 64 a; Amoongav gumooaoefim? . MA MA : 0H Lo . w m. o m _ _ _ _ _ _ _ _ _ .AGOCBHOm mmUv oaonfipoflogofiy mo gupoomm penanwcH .NA oudmfm pm 65 a: Amoou 015 fiwaoaocfio? .2 NH AH a; o m N. w m _ a _ d _ _ _ _ _ 3 .xsoraom :00 our No.0. 0G0u5p0mfioEounanuoEonoHnuwuu mo Efiuuoomm penduwaa . .mH oufimfm um nH00 66 Preparation of 3-Bromo-2, 4, 5, 6-tetramethylbenzoic Acid. This acid was prepared by a modification of the method described by Newman ‘and'Lloyd'(58). - Ina dry, 300-m1, three-necked flask fitted with adropping funnel, -Tru-bore stirrer and a reflux condenser with a drying tube was placed 21.9 g. (0.066 gram-atom) of magnesiumand 1.0 g. of dibromoisodurene {prepared by the bromination of isodurene(58)], .0.6 g. of ethyl bromide, and 10 ml. of dry ether under a nitrogenatmosphere. The reaction be- gan immediately and wasmaintained by the addition of 3. 0 g.. (total = 0.033 mole) of ethyl bromide and 6. 6 g.. (total = 0. 0262 mole)iof dibromo- isodurene in 250ml. of dry ether. After the addition was complete, the mixture was refluxedrfor an additional two hours. The ether was then removed by rapid evaporation witha nitrogen streamand replaced by dry benzene. -After reflux of this solution for two hours, the reaction mixture was cooled inanicegtbath’and carbon dioxide gas (from evapora- tion of dry ice) was bubbled into the mixture for20 minutes. Water was added andthe aqueous layer separated, extracted with 100 ml. of ether to remove any organic material (contained 2. 0 g. of star-tingmaterial and 2.0 g. of bromoisodurene, b.p. 110-1150 at 5 mm.), and acidified with dilute hydrochloric acid. The white precipitate was filtered, . washed with water and recrystallized from-ethanol-water, yielding 1. 95 g.. (48% based'on unrecovered-starting material) of 3-bromo- 2,‘4,‘5~,6-tetramethylbenzoic acid, . m.p. 199-2000.. - Its infrared spectrum can be found in Figure 14. ’fif‘il' Ca1cd..for-C11H13Br 02: ,C, 51. 39; H, . 5. 10; Br, _ 31. 08 NeutralizationEquivalent: 257.03 2 ' Found: ' C, 51. 21; H, . 5. 15; Br,. 30. 95. - Neutralization Equivalent: 256 ‘HLdIolysis of Trichloromethylbromoisodurene. Trichloromethyl- bromoisodurene was hydrolyzed by reflux in aqueous acetone andby pouring its solution in 100% sulfuric acid onto ice. 67 chouowncv flumcofioefim? a 2 2 S o m b M ._,M _ _ _ a . Acofioaom ”GTE. Bod oaossooafiofiosoooe .m .s .N uofionnum mo ghuoomm 009035 :3 ondmwh a Am 300 0 —IV* 68 ‘(a) Aqueous acetone. 'In a 2'5-m1. , pear-shaped flask was placed 0.50 g. of trichloromethylbromoisodurene and 10 ml. of 50% aqueous acetone. This solution was refluxed for ten hours, poured into .50 ml. of ice - water, and then extracted with. ether. The organic layer was i separated, stripped of solvent and the residue recrystallized from acetone. There was obtained 0. 38 g. - (95%), of 3-bromo-2, 4,5,6-tetra- methylbenzoic acid, . m.p. 197-199°. identical by infrared and mixed m. p. y to! an, authentic sample. . (b) Hydrolysis of 100% sulfuric acid-solution. Trichloromethyl- bromoisodurene (0. 5 g.) was dissolved in 10 ml. of cold, 100% sulfuric acid. eAfter standing for one hour, the dark red solution was poured on .approximately 30 g. of crushed ice and‘extracted with ether. The ether was then removedandthe residue recrystallized from acetone giving . 0. 38 g.. of 3-bromo-2, 4, 5, 6-tetramethylbenzoic acid, m.p. 197--l99o in 95% yield,— identical by infrared and mixed m. p. to an authentic sample. .‘Preparation of Chloroisodurene. This compound was prepared using the method of Illuminati and'Marino (59). In a 300-.ml.., three- necked flask was placed '25 g.. (0. 187 mole) of isodurene, 50 ml. of glacial acetic'ac‘id and a crystal of iodine. The flask was wrapped with frictiontape to exclude sunlight, cooled to 100 andchlorine gas (13.92 g....o. 196 mole) introduced slowly through-aninlet tube immersed-in the liquid. -The reaction mixture was stirred for one hour, then poured (on ice-water and extracted with ether. Theether-layer was washed with 10% sodium hydroxide, a saturatedsolution of sodium ,metabisulfite. and distilled water, then dried over anhydrous .magnesium sulfate. The solvent was removed and the residue distil-ledthrough a two—foot glass-helices packed column. The fraction collected as product boiled at‘130-132° at 26 mm., n3 = 1.5405 and weighed 18.7 g..-(64%). 69 Reported (59): b.p..108.5-108.9o at 12 mm., n3°5 = 1.5411. its infra- -red spectrumeis shown in Figure 15. ‘ Preparation of Trichloromethylchloroisodurene. This compound was prepared by the same method used to synthesize trichloromethyl- bromoisodurene. . In a 300-ml... three-necked flask provided with a TTru-bore stirrer, condenser and thermometer was placed 26.7 g. I (0. 2 mole) of aluminum chloride andl33. 3 ml. of carbon tetrachloride. To this stirred and heated (3'8-420) suspension was added-over a two hour period, 16. 8 g. (0. 1‘ mole) of Chloroisodurene dissolved inz33. 3tm1. of carbon tetrachloride. - After stirring and heating the reaction mixture for. two more hours, it was worked up as before to give 22. 3 g. (78%) of tancrystals, m.p. 75-790. Two recrystallizations from acetone at dry-ice temperature gave pure trichloromethylchloroisodurene, m.p. 83-84. 5°. Its infrared spectrum is shown ineFigure 16. .Anal. Calc'd-for cannon: c, 46.19;-H, 4.23;c1,49.58 Found: , c, 46. 16; H, 4.31; c1, 49.45. . Preparation of 3-Chloro-2, 4,5, 6-tetramethylbenzoic'Acid. This acid was prepared from bromochloroisodurene byvthe method of Newman and'Lloyd (58). (a) Preparation of bromochloroisodurene. .Following the procedure of Illuminati and Marina-(60),. 5.0 g.r (0. 0235 mole)of bromoisodurene was placed in a 300-ml. ,0 taped flask with‘20 ml. of acetic acid-and a crystal of iodine. .Chlorine gas (1. 76 g. , . 0.0248 'mole) was introduced slowly-and the addition was complete in ten'minutes. . Workup was identical to that used in the preparation of Chloroisodurene. The residue from the ether 'layer was recrystallized from ethanol—chloroform giving 4.5 g. (77%) of bromochloroisodurene, m.p. 187-189°. Its infrared spectrum is shown in-Figure l7. 70 NH Anson 02.3 5.3953583 : Ce C .AaoquOm «mov. 2.5.5603 10.330 mo guuuomm ponmnw5_ .3 0.33m _Mfljflw: fl\o H0. 71 «A Md 2 38a .fiumaofiocfim? o L '—I)N H a e n w m w m _ _ _ 4003300 #001000 «mUv onohdpogosoggswnuoa 10003030 0 gnu—000m 00.8.33 . .3 oufiwwh H0 :00. a 72 . Amconuwgv . numcvfivkrm? NH 3 A: o m N. _ _ _ _ A 1 . AGOSSHOm awn: 28.2603 nonoafioofioun mo 8.9.302? @manwcH_ .hd mydmfim #0 H U). 73 (b) Preparation of 3-chloro-2, 4, 5, 6-tetramethy1benzoic acid. In a dry 300-m1. , three-necked flask equipped with a Tru-bore stirrer, drop- ping funnel and a reflux condenser was placed 0. 84 g.. (0. 033 gram‘atom) of magnesium and 50 ml. of a solution of ethyl bromide (0. 0165 mole) . and bromochloroisodurene (0.0165 mole) in 200 ml. of ether‘(nitrogen atmosphere). The reaction commenced and was maintained by. the addition of the remaining solution. (After addition was complete, the solution was refluxed for two hours, then carbon dioxide was introduced at ice temper- tures. Usual workup gave white crystals of 3-chloro-2,4, 5, 6-tetra- methyl benzoic acid, m.p. 198-19150. Its infrared spectrum is shown inFigure 18. ’figfl. .Calc'd for CanCl 0;: C, 62. 12; H, 6.16; Cl, 16.67. Found: C, 61. 95;‘H, 6. 22;. C1, 16.71. .Hydrolysis of Trichloromethylchloroisodurene. (a) Aqueous acetone. A solution of trichloromethylchloroisodurene (0. 5 g.) in ‘25 m1. of 50% aqueous acetone was refluxed for twelve hours and gave, after the usual workup, 0. 37 g. (95%) of 3-chloro-2, 4, 5, 6-tetramethy1benzoic acid, .m.p.. 198-199.5°. This material was identical tothe compound prepared previously (infrared spectrum and mixture m.p. ). . (b) Hydrolysis of 100% sulfuric acid solution. When 0. 25 g. of trichloromethylchloroisodurene was dissolved in 100% sulfuric acid, a deep-redxcolor formed and hydrogen chloride evolved. - After the gas ceased to evolve, the solution was poured on-ice and the white precipitate was worked up as before. There was obtained a nearly quantitative yield '(0. 18 g..) of 3-chloro-2,4, 5, 6-tetramethy1benzoic acid, . m.p. 198-199. 5°, identical by infrared and mixture m.p. with an authentic sample. 1 Preparation of Chloroprehnitene. .Chloroprehnitene was prepared using the same method employed to prepare Chloroisodurene. -In a 300- m1., taped, three-necked flask was placed 12.5 g.- (0.094 mole) of 74 Ammo.“ 35v “imaged? 2 NH 3 cl. 0 w b m w A _ _ _ _ _ _ fl _ .AGOUSHOm £038 30m odouconafiofi inhumane .m J. .NuonoEoum mo 5.930on @9535. .2 oufimfm H0 3000 75 prehnitene, 75 ml. of glacial acetic acid and a crystal of iodine. ~Chlorine gas (6. 96 g. , O. 098 mole) was introducedat 00. Addition was complete in ten minutes and-the reaction mixture was stirred for one hour at room temperature, then worked up as before. Distillation through a two-foot column fitted with a spiral, tantalum wire gave afraction, b.p. 130-1320 at 22 mm. which was collected as product (9.0 g.; 56. 5%). Reportedfor chloroprehnitene; b.p. 1320 at 24 mm. (60). ~ Its infrared spectrum is shown in Figure 19. .Preparation of Trichloromethylchloroprehnitene. In a 100-ml. , three-necked-flask equipped with a Tru-bore stirrer, condenser,- thermometer and dropping funnel was placed 8. 9 g. (0. 066 mole) of aluminum chloride and 50 ml. of carbon tetrachloride. To this stirred and heated (38-420) suspension was added, over a two-hour period, a solution of 5. 2 g. - (0.031 mole) of chloroprehnitene in 10 m1. of carbon tetrachloride. The reaction mixture was then stirred and heated for two more hours. Workup was similar to previous trichloromethylations and gave 6. 0 g.. (68. 5%) of crude trichloromethylchlor0prehnitene, m.p. 88-9395. - Several recrystallizations from acetone at dry-ice temperatures gave white crystals of pure trichloroimethylchloroprehnitene, m. p. 95-970. Its infrared spectrum is shown in Figure 20. 53331. Calc'dfor eunuch: . c, 46. 19; H, 4.23;, c1, 49.58. Found: c, 46.35; H, 4. 34;. C1,.49.4o. Hydrolysis of Trichloromethylchloroprehnitene. - (a) Aqueous acetone. Trichloromethylchloroprehnitene (0. 5 g.) was dissolved in 50% aqueous acetone and refluxed for twelve hours. The usual workup gave a 95% yie1d(0. 37 g.) of 2-chloro-3, 4, 5, 6-tetramethy1benzoic acid, m.p. 195-1970. Its infrared spectrum is shown in‘Figure ’21. This compound was found to be different from 3-chloro-2, 4, 5, 6-tetramethy1benzoic acid 76 N Ammonoficv sumqoaoxfim? : OH 0 m N. D P I——"~1“ .—q F .2. _ e _ i _n .Lr... L 4.8338 :08 ocvuflagoumonogo no 8530on @9835 .mL mud—warm HO 77 Amcou 38v nuwnoaozmg i Q N: 2 E b m m e m M. _ 4 _ . _ _ q _ J _ 48:30... .8 was .88 odofidfioHmonoaaogauoEOHoEoCumo £3730on pondanH .ON 059th Mo moo 78 AmGOMUHEV ,zpwcoaocrm? me 2 2 S w w s w .m _ A _ _ 4 _ e _ _ .AGOUSHOm «HUEUV Rom. oflouconfffiognpou :0 .m .v .muouoHnouN mo gnpommm 60.2:de .HN onsmfih H0 3000 A 79 (m.p. 198-199. 5°) by mixed m.p.- (depressed) and proton magnetic resonance spectrum in 100% sulfuric acid (see Table IV). 311152. .Calc'd for CanCl 0;: C, 62.12; H, 6.16; Cl, 16.67. Found: C, 62.00; H, 6.16; Cl, 16.67. (b) Hydrolysis of 100% sulfuric acidwsolution. The deep-red solu- tion obtained when trichloromethylchloroprehnitene (0. 5 g.) was dis- solved in 15 m1. of 100% sulfuric acid gave, when poured on ice, a nearly quantitative yield of 2-chloro-3, 4, 5, 6-tetramethy1benzoic acid, m.p. l95-l97° after recrystallization from acetone. This acid is identical by infrared spectrum and m.m.p. to that obtained from aqueous acetone hydrolysis of trichloromethylchloroprehnitene. _ Preparation of Trichloromethylprehnitene. This compound was prepared by a similar method used in previous trichloromethylations. In a 100-m1. , three-necked, round-bottomed flask was placed 13.4 g. (0.10 mole) of aluminum chloride and 18 m1. of redistilled carbon tetra- chloride. To this stirred mixture was added dropwise 6.7 g. (0.05 mole) of prehnitene in 18 m1. of redistilled carbon tetrachloride. This mixture was heated during addition '(38-420’ 'two hours) and for four more hours. ‘The reaction mixture was hydrolyzed and worked up as in previous tri- chloromethylations giving 2. 5 g. (20%) of crude trichloromethylprehnitene, .m.p. 89-910. - Several recrystallizations from acetone gave pure tri- chloromethylprehnitene, m.p..91-920; reported: m.p. 91-920 (57). The product was found identical to an authentic sample by m.m. p. and infrared spectrum. Also isolated from the crude reaction product, after removal of the carbon tetrachloride, was an acetone-insoluble material, recrystal- lized from carbon tetrachloride, m.p. 299—300". This material was not further investigated. 80 'Preparation of Methoxyisodurene. This compound was prepared by allowing isodurenol. to react with an excess of sodium hydride, then addingmethyl sulfate. In a dry, nitrogen swept, 500-m1. , three-necked flask equipped with a condenser, dropping funnel and Tru-bore stirrer .was placed 2.41 g.- (0. 10 mole) of sodium hydride and 50 m1. of dry benzene. To this stirred suspension was added 20 ml. of isodurenol (10.0 g. , 0.067 mole, see'Part II in this thesis for its preparation, dis- solvedein. 100 »m1. of dry benzene). The mixture was brought to reflux and the remaining isodurenol added dropwise. -After addition was complete the greenmixture was refluxed for two hours, then cooled to room temperature. To this stirred solutionwas added dropwise 8.45 g.. (0.067 mole) of methyl sulfate while the contents of the flask were gradually heated to reflux. After addition was complete, the reaction mixture was ‘refluxedfor two-more hours, then cooled in‘an ice bath. ‘vWater (50 ml.) was added dropwise to the cooledand stirred solution, then the mixture allowed to warm to room temperature. The organic layer was separated, . washed with 20 m1. of Claisen's alkali (7 g. of potassium hydroxide,- 5 ml. of water and methanol to make the total volume 20 ml.),. 20 ml. of 10% hydrochloric acid, and 20 ml. of 10% sodium bicarbonate, then dried over anhydrous magnesium sulfate. . The. solvent was removed on a ~Rinco rotary evaporator'and the residue distilled througha seven-inch glass helices packed column giving two fractions; fractionI, b. p. 60--119o at 21 .mm. ,4, 4. 0 g.. ,. mainly methyl sulfate "(infrared spectrum), fraction *II,‘ b. p.. 119-1200-at 21 mm. . 8.0 g.. (78%). The latter fraction-was redistilled giving pure methoxyisodurene, b.p.. 119-1290 at 21 mm. , . reported; b.p. 119-1200 at 21 mm. (61). 'Its infraredspectrum is shown in Figure 22. Attempted Preparation of 2, 4, 5, 6-Tetramethyl-3-methoxybenzo- trichloride. , In-a 1004ml. , one-necked, round-bottomed flask was placed 12. 12 g. - (0. 091 mole) of aluminum chloride and 25 m1. of carbon tetrachloride. 81 m .4 RI Anson 38v gumcofioewmg. -D.——‘ 4.3.2.. .89 odou5@o§>unofioe mo Estevan @oudnmfi . .NN 0&9th 82 To this; stirred solutionwas added dropwise 7. 5 g.z(0. 046 mole) of methoxyis‘odurene in 25 ml. of carbon tetrachloride. The reaction mixture was heated (37-420) throughout addition, which-required two hours. -After addition was complete, the- solution was stirred and heated-for two and one-half hours, then worked up as in previous tri- chloromethylations. 'After removal of excess. carbon tetrachloride, the oily residue was distilled through a six-inch, . Vigreux column giving twofractions; fraction‘I, ‘ b.p. 45—500 at 2mm. , 1. 5 ml. which proved to berecovered methoxyisodurene by comparison with an authentic . sample (infrared spectra) and fraction II, b. p. 80-900-at 2mm. ,- 1. 2 ml. -A tarry residue remained, . 3.0 g.. . Fraction'II gave a redcolored sulfuric acid‘solution (100%) but did not give a crystalline material when the-acid solution was poured on ice. This material was not further investigated. - Attempted‘Preparation of 2, 4,16-Trimethoxybenzotrichloride. Several attempts to prepare this'compound from 1, 3,15-trimethoxy- benzene, aluminum chloride and excessvcarbon tetrachloride were not successful, givingin ,most instances intractable tars. . In typical . experiment, there was placed 6. 0 g. (0. 045 mole) of aluminum chloride vandv20ml. of carbon tetrachloride in a»300-.ml.,r three—necked flask. Tothis stirred suspension was added dropwise; 3. 0 g.-, (0. 0146 mole) of 1,-3,‘5-trimethoxybenzene [m.p. 48-510, . prepared-a6cording to the emethodof Benington-and co-workers (62). . Repo'rted;.m.p. . 519] dis- solved-in 20 ml. of carbon tetrachloride. . The reaction mixture was heated’(3‘7-429) duringadditionwhichrequired one hour. -After addition wasacomplete, 2. 0 g..- (total moles: 0. 0-6) of aluminum chloride was added and the mixture stirred for three ‘more hoursrat room temperature. The deep red complex was hydrolyzed by pouring onto ice and dilute hydrochloric acid, the aqueous layer separatedand the insoluble red oil extracted with. chloroform. The combined organiclayers were washed 83 with water, then dried over anhydrous magnesium sulfate. The solvent was removed by a Rinco rotary evaporator and the residue distilled directly througha still-head giving 0. 77 g.. b.p. 92° at 0.7 mm., of recovered starting material, , m.p. 48-520 and a black tarry residue, 0.96 g. II. . Benzophenone Dichloride s A. Synthesis Attempts to Prepare 2, 3, 4, 5, 6-Pentamethy1benzophenone Di- chloride. - It was found by'Fish (57) that pentamethylbenzophenone di- chloride could not be prepared by treating the ketonewith phosphorus pentachloride. - Unsuccessful attempts to prepare this compound by several different routes will be described here. (a) From trichloromethylpentamethylbenzene, benzene and aluminum chloride. In a three-necked, round-bottomed, 100-m1. , flask provided with a Tru-bore stirrer, condenser, and dropping funnel was placed 3. 51 g.- (0. 0263 mole) of aluminum chloride and 25 m1. of dry benzene. To this ice-cooled, 7 stirred solution was added dropwise, a solution of 7.0 g. (0. 0263 mole) of trichloromethylpentamethylbenzene -in‘-25 ‘ml. of benzene. ~No hydrogen chloride was evolved‘after' 15 drops, . consequently the reaction flask was heated to reflux. -Addition was complete after 45 min. and the mixture then stirredand heated-for 24 hours until hydrogen chloride ceased to be evolved. The dark red complex was cooled, then poured onto-a mixture of ice, concentrated hydrochloric acid and carbon tetrachloride. The yellow organic layer was separated, washed with water and dried over anhydrous magnesium sulfate. The solvent was removed and the residue distilled through a six-inch, (vacuum-jacketed Vigreux column. Fraction 1, b.p. 60-650 at 2 mm. , 0. 5 g.. fraction II, b. p. 95-960 at 2 mm. , _ 3.9 g. , fraction 111, b.p. loo-110° at 2 mm. , o. 52 g. , residue,.4.09 g. of black tar. 84 Fractions II and III were found to be diphenylmethane by comparison of their infrared and proton magnetic resonance spectra with an authentic sample. -Fraction‘I contained three unidentified materials, as shown by vapor phase chromatography on a column of 20% silicon. oRunning the above reaction at lower temperature (40°) gave mainly recovered trichloromethylpentamethylbenzene. » The use of stannic chloride as. a catalyst or carbon disulfide as a solvent, gave only poly- .meric material or products similar to the above reaction, depending on the reaction temperature. (b) From benzotrichloride, pentamethylbenzene and aluminum chloride. Ina one-liter, three-necked, flask cOoled in an ice bath was placed 12.6 g.. (0. 099 mole) of aluminumchloride, 67. 5 m1- (0.495 mole)of benzotrichloride and500 ml. of carbon disulfide. - To this was added, over two and one-half hours, 15 g. (0. 099 mole) of pentamethyl- benzene in300 m1. of carbon disulfide. The purple complex was hydro- lyzed by pouring onto a mixture of concentrated hydrochloric acid and ice. The organic layer was separated, washed with water and dried over anhydrous magnesium sulfate. The solvent was removed by a Rinco rotary evaporator and the residue dissolved in ten-times its volume of petroleumvether (b. p. 60-900). ~The insoluble, high-melting material (m.p. > 300°) was filteredand the organic filtrate was stripped of solvent. The resulting residue was found to work up best in the following .manner; distillation at O. 08 mm. (T< 55°) removed most of the starting materials (65. 0 g.) and. the residue was transferred to a mole- cular still and“ "fractionated.- " The fraction collected at 90--110o pot temperature (0.08 mm.) proved to be mainly hexamethylbenzene (m.p. 159- 162°, 0. 8 g.) by comparison with an authentic sample (infrared spectrum). The residue was chromatographed on‘a mixture of Celite (Celite 535,- Johns-Manville) and charcoal (Darco G-60), 1:3 proportions by volume, using petroleum ether (60-900), benzene, ether and ethanol 85 as eluting solvents. .Several fractions were isolated, one of which,, elutable with ether, proved to be 2, 3, 4, 5-tetramethy1benzophenone . (1.5 g.) (m.p. 93-940, identical to a synthetic sample made from benzoyl chloride, aluminum chloride and prehnitene (the details of this synthesis are reported later). Two other main products were found and appeared while using petroleum ether-benzene as eluting solvents but could not be identified. . The first to appear had a m.p. 226-2270 (A) and the other a m.p.. 207-209° (B). The combined yield of A and B was 2. 55 g.. -No functionality was found in the infrared spectra of these compounds nor ~does the proton magnetic resonance spectra suggest anything but poly- methyl aromatic hydrocarbons. - The twolcould not be separated from one another, preventing adequate determination of their structure. Preparation of 2, 3, 4, 5-Tetramethylbenzophenone. This compound was prepared by the Friedel-Crafts reaction of benzoyl chloride,- prehnitene and aluminum chloride. In a 200-ml. ,0 three-necked, round- bottomed flask was placed 10.0 g. (0. 0747 mole) of prehnitene,- 11.5 g. (0.0866 mole) of aluminumchloride and- 38 ml. of carbon disulfide. To this stirred slurry was added dropwise, over twovhours, 10.15 g... (0. 0722 mole) of benzoyl chloride. The-complex was stirred for three hours,- then poured on. concentrated hydrochloric acid and ice. The organic ”layerwas separated and the carbon disulfide removed by air jet. ~ The residue was dissolved in ether, washed with 1096 sodium hydroxide and water, then dried over anhydrous magnesium sulfate. The solvent was removedand the residue, on recrystallization from ethanol, gave 15.0 g. (75%) of 2, 3,4, 5-tetramethylbenzophenone,,m.p. 86-890. - Two more recrystallizations from ethanol-water gave pure compound, m.p. .92-93.5°. Its infrared spectrum is shown inFigure 23. Its proton magnetic resonance spectrum in carbon tetrachloride was clearly consistent with its structure, 9 aryl methyl hydrogens at 7. 88‘!“ , 3 aryl methyl hydrogens at 7.78 ‘r , l aryl hydrogen at 3. 19 T and 86 Annoyed“: sewage/m3 Ma Ma HA OH 0 m. N. .P _ _ _ d 4 figs. m. _ \[lk .Acofigom JUUV osonvnm noncongfioambuutm .¢.m em. mo 8530on @oumnE .MN 0.":th 87 5 aryl hydrogens appearing as a multiplet from 2.19 to 2. 73 ‘I‘ (see ' Figure 3). Its-ultraviolet spectrum had ' kitch 278 (6 = 5, 029) and 249 mil-(6 = 13, 905). _ -._i§.£al. Calc'd for cuHmo: , c, 35.71; H, 7.62 Found: . C, 85.60; H, 7.70. Preparation of 2, 3, 4, 5-Tetramethy1benzophenonei Dichloride. This compound was prepared by refluxing 2, 3, 4, 5-tetramethylbenzophenone incarbon tetrachloride with a slight molar excess of phosphorus pentachloride. . In a 25-ml. , pear-shaped flask equipped with a con- denseriwas placed asolution of 1. 0355 g..- (0.00435 mole) of 2, 3,4,5-tetra- methylbenzophenone, 1.69 1'. 0. 01 g. - (0. 00811 mole) of phosphorus. penta- chloride and enough carbon tetrachloride to make the total volume 10. 0 ml. By. taking aliquots before refluxing began and as the reaction pro- -ceeded, one could determine the extent of reaction by examining the proton magnetic resonance spectrum (see Figure 3). The reaction was complete after 12 hours. A This‘carbon-tetrachloride solution,“ after reflux, was used in hydrolysis experiments reported below and for the determin- ation of various spectra. However, in order to determine the moles of hydrogen chloride evolved when the tdich'lbride, is dissolvedin 100% sulfuric acid, it wasnecessary to isolate the pure compound. In a dry 50-m1.,‘ one-necked, round-bottomedflaskeequipped with a small still-head and a side arm.having a stops-cock, was placed 1. 10 g... (0. 0047 mole) of 2, 3,4, 5-tetramethylbenzophenone,. 10 ml. of carbon tetrachloride and 1.85 g. (0.0096 mole) of phosphorus pentachloride. The solutionwas refluxed for twelve hours thenvacuum applied asthe solution cooled. , The vacuum was increased by slowly reducing a nitrogen bleed until the pressure reached 0.08 mm. The‘reaction vessel was warmed-slightly (60°) to remove the-last traces of solvent, phosphorus oxychloride and phosphorus pentachloride '(thelatter sublimed on the cold 88 finger of the still head). The yellow-white crystalline residue which remained was covered with a nitrogen blanket and the still head replaced quickly with a ground-glass stoppers while nitrogen was blown over the vessel Opening. - The flask was then evacuated through the side arm and kept at reduced pressure (0. 08 mm.) (for three hours. The vacuum was replaced with nitrogen and the residue taken up in n-pentane, treated ‘ with Norite then filtered and recrystallized repeatedly. (four times) giving'a white crystalline material, m. p.. 116-1170 (sealed capillary). . Care was taken not to expose the crystals to the atmosphere forilong periods of time. The proton magnetic resonance spectrum of pure, crystalline 2, 3, 4, 5-tetramethy1benzophenone dichloride in. carbon tetra- chloride was identical to the spectrum of the solution obtained after refluxing the ketone with phosphorus pentachloride (see Figure 3). Its infrared spectrum is shown in- Figure 24. .Anal. .Ca1C'd for c.7H,,c12: c, 69.63; H, 6. 18; c1, 24.19. I Found: C, 69.74; H, 6.07; Cl, 24.15. - Hydrollsis of 2, 3, 4, 5-Tetramethylbenzophenone Dichloride. (a) Aqueous acetone. A 2. 00 ml. aliquot was pipetted from the reaction mixture (phosphorus pentachloride, carbon tetrachloride and 2,3,4,05- tetramethylbenzophenone after twelve hours of reflux, 0.403‘M.) and placed in a 10 ml. pear—shaped flask with 25 m1. of 50% aqueous acetone. The solution was refluxed for twelve hours, cooled, then poured on water and extracted with carbon tetrachloride. The organic layer was washed with 10% sodium bicarbonate solutionand water, then dried over anhydrous magnesium sulfate. - The solvent was removedand the residue recrystallized fromaqueous ethanol giving a nearly quantitative yield (0. 18 g.) of 2, 3, 4, 5-tetramethy1benzophenone,. identical to an authentic sample by m.m.p. and: infrared spectrum. (b) Hydrolysis of a 100% sulfuric acid solution. .(1) After ten minutes. A 2.00 iml. aliquot of the reaction mixture-(phosphorus 89 Anson can: 5mdo~o>m3 .: a: m w ~. ‘ H T _ a _ _ _ A .AGOUSHOm nmov. 03.8303 odocoxmoucofi tgfioamfioptm Jim .N mo €530on @osmswsH .wm oudmfim MD.- JD 90 pentachloride, carbon tetrachloride and 2, 3, 4, 5-tetramethylbenzophenone after twelve hours reflux, 0.403 M.) was pipetted into 8 m1. of 100% sulfuric acid. The mixture was swirled and after ten minutes hydrogen chloride ceased to evolve. The deep red solution was poured on ice and the precipitate extracted with ether. The organic layer was washed with 10% sodium bicarbonate solution and water, then dried over anhydrous magnesium sulfate. The solvent was removed on a Rinco rotary evaporator and the residue recrystallized from aqueous ethanol giving 0. 19 g- (99%) of 2, 3, 4, 5-tetramethy1benzophenone, m. p. 91-930. (2) After two hours. The same experiment described above was repeated but the sulfuric acid solution was allowed to stand for two hours. 0 Workup gave a 41% yield of 2, 3,4, 5-tetramethylbenzophenone, m.p. 91-93 . The remaining product was water soluble. Preparation of 1-(2, 3, 4, 5-tetramethy1phtfly1)-l-phenylethylene. This compound was prepared from 2, 3, 4, 5~tetramethy1benzophenone and methyl magnesium iodide by a procedure analogous to that used by Fuson and co-workers (63) in the preparation of 1-(2, 3, 5, 6-tetramethy1- pheny1)- 1-phenylethylene . In a 100—ml. , round-bottomed, three-necked flask equipped with a condenser, stirrer and addition funnel was placed 0. 73 g.- (0. 030 gram atom) of magnesium and 30 m1. of dry ether. To this was added drop- wise, 4. 26 g. (0. 030 mole) of methyl iodide (nitrogen atmosphere) followed‘by refluxing to complete the reaction. Then 4. 0 g. (0. 0169 mole) of 2, 3,4, 5-tetramethy1benzOphenone in 30 m1- of dry ether was added dropwise at room temperature andthe mixture refluxed for one hour more. The ether was replaced with dry benzene and the solution refluxed for ten hours. The contents were cooled, poured into a saturated am- monium chloride solution andthe water layer extracted with benzene. The combined benzene layers were washed with 10% sodium bicarbonate and 10% sodium metabisulfite. A After drying over anhydrous magnesium 91 sulfate, the benzene was removed and the residue distilled through an eight-inch, vacuum-jacketed Vigreux column. The fraction collected at 145--150o at 5.2 mm. was redistilled, giving 3.8 g. (97.5%) of 1-(2, 3,4,5-tetramethy1pheny1)-1-phenylethylene, b.p. 133-1340 at 0.9 mm. Its infrared spectrum is shown in Figure 25. -The proton magnetic resonance spectrum of this compound had bands at 8. 05, 7.90 and 7.81 ‘P corresponding to three, six and three hydrogens respectively, two doub- lets centered at 4.90 and 4. 32 y (J = 9. 0 cps) corresponding to one hydrogen apiece, a singlet at 3. 13 ‘r corresponding to one hydrogen and a multiplet centered at 2. 82‘r integrating for five hydrogens, in agreement with its structure. -fin_a_l. Calc'd for CyeHzo: C, 91.44; H, 8.53. Found: C, 91.39; H, 8.66. Preparation of 2, 4-Dimethy1benzcn3henone. This compound was prepared by the Friedel-Crafts condensation of benzoyl chloride,- aluminum chloride and M-xylene (64). - In a 250-ml. , round-bottomed, three-necked flask was placed 30.0 g. (0.285 mole) of m-xylene, 43. 2 g. (0. 332 mole) of aluminum chloride and 100 m1. of carbon disulfide. . To this stirred suspension was added slowly 47 g.. (0. 332 mole) of benzoyl chloride. The solution was refluxed for five hours, then cooled and poured on a mixture of ice and concentrated hydrochloric acid. . The organic layer was separated and the carbon disulfide removed. The residue was dissolved in ether, .washed with 10% sodium bicarbonate and water, then dried over anhydrous magnesium sulfate. The ether was evaporated and the residue distilled through a twelve-inch, glass-helices packed column, giving 59. 8 g- (84%) of 2, 4-dimethy1benz0phenone, b.p. 125--130o at 0.7 mm.; reported (62); b.p.. 186-190 at 15 mm. Its infrared spectrum is shown in Figure 26. This compound was found identical to an authentic sample (Tennessee Products and. Chemical Corp.) by infrared and proton magnetic resonance 92 MM Anson 35v numzoaofm? NH 3 o. a m .. m _ .2. a- 4:03.39.“ JUUV osofwfiogsogmtatsfiaoam taguoamuuoutm J .m .3; mo 6.9.30on @osdswfi .mN shaman NTUno _ C“) 93 Annoy 38V newsman/m? 2 NH 7.. . a: c m. .c. c .m _ _ _ _ _ 1 _ m s .AGOSBOm JUUV econo£QONCoQ tH>£poEfi@te.N Ho 8530on @oumuwcH .om oudmfim J 3 J , g . . 94 spectra. . Its ultraviolet spectrum had bands at kitgj-IZSO (c: 15, 700) and 271 mu (6 = 8, 688). 'Preparation of 2, 4-Dimethylbenzgahenone Dichloride. This com- pound was prepared fromv2, 4-dimethylbenzophenone and phosphorus pentachloride under identical conditions used to prepare 2, 3, 4,5-tetra- methylbenz ophenone . ~A solution of 1. 0095 g. (0. 00474 mole) of 2,4-dimethy1benz0phenone, 0'. 97 g. (0. 00480 mole) of phosphorus pentachloride and enoughcarbon tetrachloride to make a 10. 00 m1. volume, was refluxed for twenty-four hours. The reaction was shown to be complete by comparison of the porton magnetic resonance spectra before and after reflux (see Figure 27). .Preparation of 4, 4'-Dimethy1benzophenone Dichloride. - A solution of l. 0027 g. (0.00474 mole) of 4,4'-dimethy1benz0phenone (Eastman Organic Chemicals, White Label), 0. 97 g. (0. 00480 mole) of phosphorus pentachloride and enough carbon tetrachloride to make a 10. 00 m1. volume, was refluxed for twenty-four hours. The reaction was shown to be complete by comparison of the proton magnetic resonance spectrum before and after reflux (see Figure 28). . Preparation of 1,.1-Bis(p-tolyl)-ethylene. This olefin-was prepared from 4,4'-dimethylbenzophenone andmethyl magnesium iodide, identical to the preparation of 1-(2,‘3, 4, 5-tetramethy1phenyl)-1-phenylethylene. A To 0. 04- mole of methyl magnesium iodide, prepared from 0. 97 g. (0.04 gram atom) of magnesium and 5.68 g.. (0.04mole) of methyl iodide in 30 m1. of dry ether, was added dropwise 7. 92 g- (0. 04 mole) of 4,4'- dimethylbenzophenone dissolved in 50 m1. of dry ether. After addition was complete the mixture was refluxed for one hour and worked up as before. There was obtained 6. 0 g. (76. 5%) of l, l-bis(p-';tc‘>1yl)-ethylene, b.p.. 1789at 15 mm. The compound was further purified by recrystalliza- tion from ethanol, m.p. 60-610. Reported (65); m.p. 610. Its infrared spectrum is shown in. Figure 29. 95 Figure 27. ’Proton magnetic resonance spectrum of 2,4-dimethy1— benzophenone and phosphorus pentachloride i carbonrtetrachlorid before and after reflux. )— 0% Rxn. TMS l l l)“ 2.30 3.30 T 7.69 7.77 10.00 HQ ——>- P” A: 100% Rxn. J TMS . !3/ I I .95 3.20 7.68 8.07 10.00 s __+ ’r 96 Figure 28. ~ Proton magnetic resonance 3 ectrurn of 4,4'-dimethy1- benzophenone» and phosphorus pentachloriin carbon tetrachloride before andafter reflux. ‘ \ 0% Rxn. . / . “mm—7’ 2. 70 ,‘v 7. 64 10. 00 Ho ; \H\ 100% Rxn. 1 I .Zo80 , n. 7.69 10.00 _ .‘ 11Wh~ 97 Anson 02.5 suwcoaoefim? mH NH 2 S m. m N. .w m w. m m. _ _ _ _ . d 4 _ A _ _ . .Aaofldaom JUUV 0:0350L3H073m3; J mo 5.9302; @053de .mm 0.“:th amuno. : e 98 Preparation of 3, S-Dimethylbenzophenone. This compound was prepared from3, S-dimethylbenzoyl chloride, benzene and aluminum chloride. (a) Preparation of 3, 5-dimethy1benzoy1 chloride. Twenty-five grams of 3, 5-dimethylbenzoic acid (0. 165 mole, Aldrich Chemical. Co. , Inc.) and 18. 9 ml. (0. 252 mole) of thionyl chloride were mixed in a 125 ml. , one-necked flask, refluxed for two and one-half hours, then dis- tilledat atmospheric pressure through‘a twelve-inch, glass-helices packed column to remove excess thionyl chloride. The residue, on distillation under reduced pressure, gave 21. 0 g.. (75. 5%) of 3, 5-dimethy1- benzoyl chloride, b. p.. llSo-at 15 mm. , reported (66); b.p. 109.5o at 10 mm. . (b) Preparation of 3, 5-dimethy1benzophenone. To a mixture of 17.65 g. (0.14 mole) of aluminum chloride, 21.4 g. (0.274 mole) of benzene, and 58 ml. of carbon disulfide was added dropwise over-two hours 21.0 g. (0. 125 mole) of 3, S-dimethylbenzoyl chloride. After re- fluxing and stirring for five more hours, the solutionwas cooled-and poured on a mixture of ice and concentrated hydrochloric acid. The organic layer was separated, stripped of carbon disulfide and the residue dissolved in-ether. The ether layer was washed with 110% sodium hydroxide and water, then dried over anhydrous magnesium sulfate. The ether solution was filtered, stripped of solvent and the residue recrystallizedfromethanol, giving 20.0 g.- (75%) of 3,.5-dimethylbenzo- phenone; m.p. 68-700; reported (67); m. p. .700. Its infrared spectrum is shown in- Figurew30. . Its ultraviolet spectrum had bands at kits: 255. 5 (e = 15,400) and 340 mp. (E = 500). Preparation of 3, 5-Dimethy1benzophenone Dichloride. This com- pound was prepared from 3, 5-dimethy1benzophenone and phosphorus pentachloride by a procedure identical to those used for the previously mentioned dichlorides . 99 Amcosflav nomaoaoiwg ma ma OH 0 w N. ('4 .--1 in V‘ .3383... :08 ecosoamonsonfafioggtm .m mo Eanuuomm @oumuHcH. .om unamfih 100 A solution'of 0. 9996 g.. (0.00475 mole) of 3, S-dimethylbenZOphenone and 1.00 g. (0.00485 mole) of phosphorus pentachloride was made to themark in a 10. 00 ml. volumetric with carbon tetrachloride then re- fluxed in a 100 ml. , three-necked, round-bottomed flask for twenty- four hours. The proton magnetic resonance spectrum of the solution after reflux was in agreement with the complete conversion of the ketone to the dichloride (see Figure 31). . Prepparation of Benzpphenone Dichloride. This compound was pre- pared by reaction of benzophenone with excess phosphorus pentachloride according to the method of Staudinger and Freudinberger’(45). r In a 2 00- ml. , one-necked flask was placed 50 g.. (0. 275 mole) of benZOphenone and 67. 3 g... (0. 323 mole) of phosphorus pentachloride. This mixture was heated for two hours at 145-1500 then fractionated through a ten-inch Vigreux column. Excess phosphorus pentachloride sublimed up on the column which was then removed and cleaned. - The residue was distilled giving 56.0 g. of benzophenone dichloride, b.p. 149-1520 at 5 mm.; reported: b.p. 201-2020 at 35 mm. Its infrared spectrum can be found in Figure 32. .Preparation of 2, 4, 6-Trimethylbenzophenone. This compound was prepared by the. Friedel-Crafts reaction of mesitylene and benzoyl chloride. In a 500-ml. ,0 three-necked flask equipped with a Tru-bore stirrer, addition-funnel and condenser'was placed 35.3 g. (0.29 mole) .of aluminum chloride, 30 g... (0. 25 mole) of mesitylene and 116 ml. of carbon disulfide. . To this stirred suspension was added dropwise (two hours) 34. 2 g.- (0. 242 mole) of benzoyl chloride. After addition was complete, the mixture was stirred for five more hours, poured onto a mixture of ice and concentrated hydrochloric acid, then the two layers separated. - The organic layer-was stripped of carbon disulfide and the * residue dissolved in ether. The ether solution was washed with 10% Sodium hydroxide and water, then dried over anhydrous magnesium sulfate. 101 Figure 31. . Proton magnetic resonance spectrum of 3, 5-dimethyl- benzophenone and phosphorus pentachloride in arbon tetrachloride before and after reflux. 7“ 0% Rxn. 1 - l ! z 2.20 2.90 7.72 10.00 (Ho ———>- S. N 100% Rxn. TMS J l 2.30 H E 10.00 102 (*3 Anson 0.5: 53550303 2 NH 3 2 .m m .N. w m v 3W _ _ a _ q _ 7 a _ _ _ _ t : 30-0.50, A m an _ i .3005 03.8303 0cos0smosn0n— mo 55300mm @9335. .Nm 0.3mm.” 103 The solvent was removed and the residue distilled through -a twelve- inch Vigreux column giving, besides recovered mesitylene, Z. 0 g.. ,9 b.p. 600 at 16 mm., .a second fraction, b.p. 1850 at 16 mm., ,_ 43.0 g. - (79.4%). Its infrared and proton magnetic resonance spectrum-(see ‘Figure 33 and 34 respectively) were consistent with 2,4,‘6-trimethy1- benzophenone. Reported‘(68): b.p. 1890 at 17 mm., m.p.. 350-. . Attempted 'Preparation of Z, 4, 6-Trimethylbenzophenone Dichloride. Several attempts were made to prepare Z,4,96-trimethylbenzophenone dichloride from 2,4, 6-trimethy1benzophenone and excess phosphorus pentachloride with and without solvent. .Solvent systems tried were tetrachloroethane and carbon tetrachloride. In both instances, unreacted ketone was obtained. . The same was true when no solvent was used even when special precautions were taken to prevent hydrolysis. -In a typical experiment, there was placed, in a ZS-ml. , 'three- neckedflask, a solutionrcontaining 1.0077 g. (0.00449 mole) of 2,4,6- trimethylbenzophenone, 1.65 g. (0.008 mole) of phosphorus pentachloride and enough carbon tetrachloride to make a 10. 00 m1. volume. r The mix- ture was then refluxedlfor seventeen hours. The proton magnetic resonance spectrum after reflux had bands at 7.97, 7. 71, and 3. 20$ - corresponding to nine, three and two hydrogens respectively, and a multiplet from 1.81 tol. 21$ integrating for five hydrogens, a sprctrurn virtually identical to unreacted ketone (see Figure 34). The visible spectrum of this product in.101% sulfuric acid' (obtained by extracting a Hal. aliquot of the carbon tetrachloride solution with 101% sulfuric acid) had bands at 309 (6‘ b 22, 370). and 355 mp. (e ‘= 6, 100), identical to the spectrum of 2,4, 6-trimethy1benzophenone in 101% sulfuric acid (actually sulfonated ketone). Hydrolysis of a Solutionof 2, 4, 6-Trimethy1benzoghenone, Phos- phorus Pentachloride and Carbon Tetrachloride after Seventeen Hours Reflux. » (a) Aqueous acetone. In a 25-ml. flask was placed 5. 0 m1. of a 104 m H Amconofiev £~m£o~o>m3 NH H OH «0 w h o M -r firm 4.5338 .88 ocoaonnoucofliflogwuuuc J. .N no 5300mm poudnwfi .mm onswwh 105 coda oo.w .mpfi. owlm omd 0N..N (P we)? .oEuoEodnuou confine 5 odonofimoudoagufioaiuno :v.N mo 8.930on 00:98.on oflofimmz dOuonnm. .wm ohdmmh a h_"'——l .x-n Eu; -1 106 solution containing 2, 4, 6-trimethy1benzophenone (0. 5472 M), excess phosphorus pentachloride and carbon tetrachloride after refluxing for seventeen hours. Aqueous acetone (10 m1., 50%) was added and the mixture refluxed for twelve hours, cooled, then extracted with additional carbon tetrachloride. The organic layer was separated, washed with 10% sodium bicarbonate and water, then dried over anhydrous magnesium sulfate. The solvent was removed by a Rinco rotary evaporator and the slightly. yellow, liquid residue (0. 58 g. , 95%) proved to be virtually pure 2, 4, 6-trimethy1benzophenone by comparison of its infrared and proton magnetic resonance spectrum with an authentic sample. (b) Hydrolysis of a 101% sulfuric acid solution. -A 5 m1. aliquot of a solution containing 2, 4, 6-trimethy1benzophenone (0. 5472 M), excess phosphorus pentachloride and carbon tetrachloride after reflux was extracted with 10 m1. of 101% sulfuric acid. The acid-layer turned a deep red color immediately and hydrogen chloride was evolved. . The proton magnetic resonance spectrum of this solution (separate experiment) had bands at 6. 99, 7.26,. 7. 36, 7. 54 and 7. 60 ‘1‘ and a multiplet from 2. 22 to 1. 24‘? (very unstable, spectrum of sulfonated and unsulfonated ketone). The spectrum appeared to change with time (one hour) to a spectrum having bands at 6. 99, 7. 24 and 7. 54 '1‘ corresponding to three hydrogens apiece and a multiplet from 2. 22 to l. 24 "h corresponding to approximately six hydrogens. (After the hydrogen chloride ceased to be evolved (4 minutes), the mixture was poured onto ice and worked up in the same manner as the aqueous acetone hydrolysis. .The residue (0. 32 g. , 52. 5%) proved to be pure 2, 4, 6-trimethy1benzophenone by comparison of its infrared and proton magnetic resonance spectrum with an authentic sample. When 2,4, 6-trimethy1benzophenone is dissolved in 101% sulfuric acid the proton magnetic spectrum of this acid solution is identical to the above, i_. e_. , also appears to sulfonate. Hydrolysis on ice likewise 107 gives low recovery of ketone (50%). A similar experiment in 99% sulfuric acid gave quantitative recovery of 2, 4, 6-trimethylbenzophenone. B. Quantitative Determination of Hydrogen Chloride From the Reaction of 2, 3, 4, 5-Tetramethylbenz0phenone Dichloride with Sulfuric Acid. Apparatus The apparatus consisted of two drying towers and two traps arranged in- series. The traps and towers had ground glass joints and were connected by Tygon tubing. Each trap had an inlet tube with a fritted glass tip protruding below the surface of the solution, and an exit tube connected to the next trap. The first tower contained reagent grade concentrated sulfuric acid and the second contained 99. 75% sulfuric acid (determined by titrating a weighed sample with standardized sodium hydroxide solution). The first trap, equipped with a ground glass inlet for a 50 ml. dropping funnel, contained a solution of 2, 3, 4, 5-tetra- methylbenzophenone dichloride in 99. 75% sulfuric acid and the last trap contained a 2 to 5% sodium hydroxide solution for collecting the hydrogen chloride. Procedure The apparatus was assembled as described. A known weight of 2, 3,4, 5-tetramethy1benzophenone dichloride was added to the first trap, the system closed, and 25 m1. of 99. 75% sulfuric acid added- The solu— tion was stirred magnetically and a rapid stream, of pre-purified nitrogen bubbled through the system. At various time intervals the aqueous sodium hydroxide trap was changed, the solution acidified with dilute nitric acid to pH 7-8 (pH was determined by phenolpthalein indicator), and the chloride ion determined by the Fajans' method (111). Towards thelatter stages of sweep, it was necessary to add a known quantity of 108 chloride to accurately determine the smaller amounts of chloride swept ‘ from the solution. - III. -Misc ellaneous Experiments Preparation of 2, 3, 4, 5, 6-Pentamethylacetophenone This ketone was prepared by reaction of pentamethylbenzene with acetic anhydride and aluminum chloride by the method of Smith and Guss (69). In a 500-ml. , three-necked, round-bottomed flask equipped with-a stirrer, dropping funnel and reflux condenser was placed 20. 0 g. (0.135 mole) of pentamethylbenzene, 100 m1. of carbon disulfide and» 39. 7 g.. (0.297 mole) of aluminum chloride. To this stirred suspension was added dropwise, 15. 2 g. (1.485 moles) of acetic anhydride, causing the solution to reflux. _After addition was complete (15 min.-),6 the reaction mixture was stirred and heated for 30 minutes, then allowed to cool to room temperature. The contents of the flask were poured onto ice, the aqueous layer separated, then extracted with carbon tetrachloride. The solvent was remOVed from the combined organic layers by a‘Rinco rotary evaporator and the residue dissolved in ether, washed with50 'ml. of water, 50 ml. of 10% sodium hydroxide and water, then dried over anhydrous magnesium sulfate. The solvent was removed by a Rinco rotary evaporator and the residue distilled through a six-inch, vacuum- jacketed, Vigreux column giving 21.2 g. , b.p. 145-1550 at 8 mm. , of crude pentamethylacetophenone, m.p. 70-740.. . The crude product was recrystallized from pentane giving pure 2, 3,4, 5,6-pentamethy1aceto- phenone, m.p.. 82-83.50; reported: m.p. 840. .Its infrared spectrum is shown in Figure 35. - Preparation of l-Chloro- l—(2, 3, 4, 5, 6-pentamethylphenyl) ethylene This olefin was prepared from 2, 3, 4, 5, 6-pentamethy1acetophenone, phosphorus pentachloride, phosphorus oxychloride and phosphorus 109 AmcoH 02.5 33:39:35 .2 m: 3 2 e m m b. _ _ _ _ _ . .2538 400* ococosmobomaaaocfimucomuo .m J. .m .N no gnuoommpoumumfi .mm unfimfih . a 3 O“ I .20 110 trichloride by the method which Adams and Theobald (70) used to prepare l-chloro-l-(2,‘4, 5,6-tetramethy1pheny‘l)ethylene. In a 100-ml. , one-necked, round-bottomedflask cooled in an ice bath was placed 20. 0 g. (0. 105 mole) of 2, 3,4, 5, 6-pentamethylaceto- phenone, 17. 9 ml. of phosphorus oxychloride, 3. 6 ml. of phosphorus trichloride and 21. 9 g. (0. 105 mole) of phosphorus pentachloride. The mixture was allowed to warm to room temperature, then heated at 55C) for eight hours. The temperature was raised to 65-700 for fifteen more hours and the solution then allowed to cool to room tempera- ture. -The contents of the flask were poured onto ice with vigorous stirring and the aqueous mixture extracted with ether. m3 : E o w s _ w _ :i 4.2.2.. .80 200150Aadoamaaflofimusomuo .m 3v .m .3 sauouodaoua mo gnuoomm “000932;. .om 0.3%me . £0» 0 I H0 112 8. 02 ‘7‘ corresponding to approximately nine and, six hydrogens respectively and-a band at 5.90 ‘b corresponding to two hydrogens, suggesting the compound to be the a-chloro derivative of 2, 3,14,15,6- pentamethylac etophenone . The yield of olefin, based on unrecovered starting material, was 48%. ’HLdrolysis of a 101% Sulfuric Acid Solution of l-Chloro-l- (2, 3, 4, 5, 6-pentamethylphenj‘l) ethylene A carbon tetrachloride solution. containing 0. 30 g. of l-chloro-l- (2,3, 4, 5,6-pentamethy1pheny1) ethylene and 0. 5 ml. of solvent was extractedywith 1 ml. of 101% sulfuric acid. -The deep red acid layer stoodfor twenty minutes, then was poured onto ice, extracted with carbon tetrachloride and the organic layer separated. The carbon tetrachloride solution was washed with 10% sodiumhydroxide and water,- thendried over anhydrous magnesium sulfate. The solvent was dis- tilled and the crude residue, 0. 24 g. , recrystallized from aqueous ethanol giving pure crystals of l-chloro-l-(Z, 3,4,5, 6-pentamethyl- pheny1)ethy1ene, m.p. 73-740, 0. 23 g.- (80%),. identical to a known sample‘(m.m.p. , infrared and proton magnetic resonance spectra). . Preparationof l-‘(2, 3,4, 5, 6--Pentamethylphenyl)acetylene This compound was prepared by the reaction of l-chloro-l-(Z, 3, 4, 5,6- pentamethylphenyl) ethylene with sodium ethoxide in ethanol by themethod used by Adams and Theobald (70) to prepare l-(2, 4, 5, 6, -tetramethy1- pheny1)acety1ene. J In a 100-ml. , three-necked flask equipped with a still-head,- Tru- bore stirrer and dropping funnel was placed 1. 57 g. (0.067 gram atom) of sodium andi30 m1. of anhydrous ethanol. -To this mixture was added dropwise,_ 6. 6 g- (0. 0316 mole) of l-chloro-l-(Z, 3, 4, 5, 6-pentamethyl- phenyl) ethylene inr30 ml. of anhydrous ethanol. The reaction mixture "' 5- 113 was stirred and heated (110°) during addition (30 min.) and for five morehours. .Some of the excess ethanol was distilled during this - period of reflux, concentrating the solution to .30 ml. total volume. - The reaction mixture was heated (1100) for five additional hours, then 15 ml. more of ethanol was distilled. ~The residue was cooled, poured onto ice water, acidified with 10% hydrochloric'acid and washed with ether. -The ether‘ layer was separated, washed with water, 10% sodium carbonate and water, then dried over anhydrous potassium carbonate. rThe solvent was removed-and the brown-red residue (6. 25 g.) dissolved in-ethanol and treated withNorite. The solution was filtered and cooled to give 4. 55 g. of crude product, m.p. 69-759. The yellow solid was sublimed (850 at 0. 07 mm.) and the sublimate recrystallized from pentane giving 4.20 g.. (77. 5%) of pure l-(2, 3, 4,15,16-pentamethy1- phenyl)acetylene, m.p. 90-910. Its proton magnetic resonance spectrum shad bands 7.91, 7.68 and 6. 78 ‘1‘ corresponding to nine, six and-one hydrogen respectively, consistent with its structure. Its infrared spectrumis shown in Figure 37. . .5231. .Calc'd for C13H16: C, 90.62; H, 9.36. Found: -C, 90.50; H, 9.25. vadrolysis of Sulfuric Acid-Solutions of 1-(2, 3,4,l5,‘6-Penta- methylphenyl) ac etylene (a) 101% Sulfuric acid. -A carbon tetrachloride solutioncontaining 0. 30 ug. of l-(2, 3‘, 4, 5, 6-pentamethy1phen-y’l)ac etylene and 0. 5' m1. of solvent was extracted with 1 m1. of 101% sulfuric acid. . The brown- black solution was immediately poured onto ice, extracted with carbon tetrachloride and the organic layer separated. Theusual workup gave amorphous material (0. 10 g.) which. could not be purified (infrared spectrum had band at 5. 78 1.1.). .Some tar was also found (0. 10 g.) soluble in water and was isolated by salting the aqueous layer. ’22‘1151. 114 3:00 083 “Hawcoaoxwmg MA NH .3 0H m m N. A _ _ -_ _ d 43003300 JUUV oneatwaoomfirwconmgsuog ‘ {Sammie .m J. .m .3; mo Hahn—00mm pohmhwdm. .hm 0H5wrm ; 213-050 fl m flvw 115 (b) 95. 7% Sulfuric acid. -Using the method described above, a solu- tion of l-(2, 3, 4, 5, 6-pentamethy1pheny1)acetylene (0. 2 g.) in 95. 7% sulfuric acid’was poured on ice. The usual workup gave again only a tarry product (0. l g.~). -Hydrolysis'of Sulfuric Acid‘Solutions of Phenylaceglene (a) 99% Sulfuric acid. -A solution of phenylacetylene (1.0 g. , . Aldrich-Chemical Co. ,1 Inc.) in. 10 ml. of carbon tetrachloride-was ex- tracted with 20 ml. of 99% sulfuric acid, giving a deep red acidrxlayer which was swirled for five minutes then poured onto ice. The green aqueous mixture was washed with carbon tetrachloride, the aqueous ~ layervseparated, salted with sodium chloride and extracted with carbon tetrachloride. - The combined organic layers were washed-with 10% sodiumhydroxide and water, then dried over anhydrous magnesium . sulfate. . The solvent was removed on a’Rinco rotary evaporator and the yellowliquid residue (0. 65 g.) proved to be virtually pure aceto- phenone by comparison of its infrared and proton magnetic resonance spectra with those of an authentic sample. The yield was 55%. (b) 98% Sulfuric acid. The same procedure used in the hydrolysis of a 99% sulfuric acid solution of phenylacetylene, was repeatedhere withidentical amounts of material and 98% sulfuric acid. There was obtained,. after the usual workup, 0. 80 g. of acetophenone (68 %). Sulfuric Acid The sulfuric acid. used throughout this work was prepared by suitably, diluting J. T. Baker's reagent grade fuming(30-33% sulfur trioxide) sulfuric acid with reagent grade concentrated'(96-98%) sulfuric acid using a chart provided by'Beacon Chemical: Industries, .. Inc. , .Cambridge, - Mass. The percent acid reported should be withini’. 1% unless 7 otherwise stated. 13mm 116 Spectra The ultraviolet and visible spectra were obtained with the Beckman DK-Z‘Recording Spectrophotometer or a Cary 11. Recording Spectro- photometer using 1 cm. ground glass stoppered cells. The sulfuric acid spectra were obtained by placing 30-50 p. 1. aliquots (pipetted with a Hamilton 50 pl. syringe, Hamilton.Co., Whittier, Calif.) of the com- pound, dissolved in an appropriate solvent, in the requisite quantity of sulfuric acid. , The infrared spectra were scanned on a Perkin-Elmer Model 21 ‘ Recording InfraredSpectrophotometer with 0. 5 mm. thickness solution cells. Proton magnetic resonance spectra were determined with a Varian Associates Model A-60 analytical NMR Spectrophotometer. . Sulfuric acid spectra were obtained by extracting a carbon tetrachloride solution of the compound with the appropriate acid. _ . (211705“. SUMMAR Y 1.. The aluminum chloride-catalyzed reaction of carbon tetrau chloride with polymethylbenzene derivatives was extended to mono- substitutedisodurene and prehnitene derivatives to yield 3-bromo-, . 3-chloro-2,i4, 5, 6-tetramethy1benzotrichlorides and 2-chloro-3, 4, 5, 6- tetram-ethylbenzotrichloride. The structures of the latter were estab- lished by hydrolysis to the corresponding acids and by comparison with independently synthesized compounds or by observed differences of, the acid with its other possible isomers. 2.. Each of these halo substituted benzotrichlorides dissolved in 100% sulfuric acid to form dicarbonium ions, i_. e_.: + - CC13 + 2112504 C~Cl + mm. + 21150, X X Hydrolysis of the colored solutions gave more than 90% yield of the corresponding carboxylic acid. - The ultraviolet, visible and proton .magnetic: resonance spectrasupportedthe structure assignments for the dicarbonium ions. 3.. Several benzophenone-dichlorides were prepared (2, 3, 4, 5- tetramethy1-, 3, 5-dimethyl-, 2,4, -dimethyl-, 4,4'-dimethy1- and unsubstituted benzophenone dichloride) by refluxing a solutionof the respective ketonewith phosphorus pentachloride in carbon tetrachloride. The reactions were shown to be complete via their proton magnetic resonance spectra and only in the case of 2, 3, 4, 5-tetramethy1benzophenone dichloride was it found necessary to isolate the pure material. 117 118 4. These benzophenone dichlorides were found toformdiaryl— . chlorocarbonium ions when dissolved in101% sulfuric acid by com- parison of their-visible and proton magnetic resonance spectra with known diphenylmethylcarbonium ions in the same solvent. 5.. A solution of 2, 3, 4,95-tetramethy1benzophenone dichloride, when swept with dry nitrogen, gave one mole of hydrogen chloride,- substantiating the formation of diarylchlorocarbonium ions when this and the other dichlorides are dissolved in sulfuric acid. "1 «Ohm PART II _ PEROXYTRIFLUOROACETIC ACID-BORON FLUORIDE -AS A SOURCE OF POSITIVE HYDROXYL 119 ‘ r‘l‘m-u INTRODUCTION , Most direct hydroxylations of aromatic hydrocarbons have been either ascribed to attack by free radicals such as hydroxyl, or to additionof the so-called "double-bond reagents, '? followed by elimina- tion of water to give phenol (71-73). .In some cases even double-bond reagents may function simply as sources of hydroxyl radical (71). There are a few cases in which it has been suggested that an electro- philic reaction occurred. . In one of these the hydroxyl cation was pro- duced from hydrogen peroxide in acetic-sulfuric acid and used to pre- pare mesitol from mesitylene (74). Similarly, . peroxybenzoic acid reportedly oxidized cyclic aromatic hydrocarbons yiian electrophilic mechanism, but details are lacking (75). Two recent reports have used an 018,1abel to verify ionic mechanisms (76, 77). . Denny and Weiss found that benzoyl peroxide reacted with benzene in the presence of aluminum chloride to give phenyl benzoate and benzoic acid. - If the carbonyl oxygens were labeled with 018, less than 2% label was found in the ether oxygen, consistent with a mechanism involving displacement by benzene on the complexed peroxide. 018 ‘18 ’18 || \ E (I? .4101 O'C‘o 1,0 .+ ¢C—‘O-O-C¢ —_—3—>‘ + ¢C\OH / Of interest is the recent report that peroxytrifluoroacetic acid converted aromatic compounds directly to phenols and quinones by an electrophilic mechanism (78). This oxidant gave only 2,4-dimethy1phenol, 2, 6-dimethylphenol and m-xyloquinone from the oxidation of m-xylene. -However, peroxyacetic acid, which-like peroxybenzoic acid is used to prepare glycols from olefins by an electrophilic process (79), has been 120 121 . said to give no quinone from aromatic hydrocarbons (8'0), . and when it was‘used in a modified form (as hydrogenrperoxide and glacial acetic acid), reaction occurred under such; conditions that no decision on the mechanism was possible. The greater efficiency of peroxytrifluoro- acetic acid as an oxidizing agent for aromatic compounds and particularly in the Baeyer-Villiger reaction, can be attributed to the highly electro- negative trifluoroacetyl grouping, which allows facile heterolysis of . the 0-0 bond (81-84). Bourne and co-workers (85). showed that mixed anhydrides derived from trifluoroacetic acid are ionized slightly into .trifluoroac etate and acylium ions. A similar ionization of peroxytri- fluoroacetic acid should provide an excellent source of hydroxyl cations. In fact, peroxytrifluoroacetic acid has also been shown to convert aromatic ethers to phenolic ethers (86) and to produce quinones (78) or , cyclohexadienones (87) fromcertain phenols, all presumably by ionic mechanisms. 9 Our investigation of peroxytrifluoroacetic acid was prompted by the possibility that sucha potent source of positive hydroxyl would be extremely useful synthetically. However, in all previous cases no more than 50% yield of oxidized material (based onth—e peracid used) could be isolated. - The second part of this thesis reports the use of ‘ peroxytrifluoroacetic acid with boronfluoride'as a source of positive hydroxyl. - It was reasoned. that coordination of. a Lewis acid with an o rganic peracid might facilitate departure of ionic electrophilic hydroxyl ‘ from the latter, and furnish a potent oxidant under-mildconditions. - McClure and Williams used similar reasoning 7(88) and found that the combination of 90% hydrogen peroxide-boron fluoride etherate converted aromatic hydrocarbons to‘phenols and quinones but intlow- yield. This reagent was, however, . effective in the Baeyer-Villiger,»conversion of ketones to- esters. 1 *(ws. 122 'Our results indicate the combination of peroxytrifluoroacetic acid and boron fluoride provides an excellent means of synthesizing particular phenols in good yields. The reagent is, however, quite severe, andvcan bring about a number of new and interesting side reactions inaddition to electrophilic replacement of aromatic hydrogens. A 0*, .— RESULTS AND DISCUSSION The following description of the oxidation of mesitylene to mesitol. illustrates the general procedure which was'used in the present work. .Peroxytrifluoroacetic acid, prepared from trifluoroacetic anhydride and 90% hydrogen peroxide, was added to an excess of mesitylene in methylene chloride. .Excess'mesitylene serVed toureduce further oxida- tion of; the primary product. .Boron. fluoride was bubbled through the reaction mixture during addition. The reaction. was strongly-exothermic, . and. the temperature was kept below 70 by external cooling. .There was obtained on workup and distillation unchangedmesitylene and mesitol, -XLIV' (88% yield based on theamount of peracid used), the structure of which was, verified by comparison of its -m. p.. and. spectra with thoseof anvauthentic sample. -SeVeral variations of this experimental procedure failed to improve on these results. OH CH CH, CH3 CH3 0 O + cr,cf B—L—F CH3 ' OOH ~ "CH3 XLIV The earlier procedure of Musgrave' and co-workers‘ (78) with peroxytrifluoroacetic acid gave-mesitol inonly 17% yield basedonrthe amount. of peracid used, although the yield was 86% when based on unrecovered mesitylene. -Since»the peracidis expensive, its inefficient use rendered the method unsuitable for. synthesis. Theirrmethod was decidedly different from ours, involving equimolar'quantities of ; 123 . r"!‘%0flw 124 mesitylene andoxidizing agent. The peroxytrifluoroacetic acid was prepared.i_n.-_s_i_t_u. by adding 85%hydrogen peroxide to asolution of mesitylene, trifluoroacetic anhydride and methylene chloride. The reaction‘mixture was then stirred for 24 hours at 0°. - In contrast, omission of the boron fluoride from our procedure gave mesitol in45'% yield,- an improvement over the‘earlier method, but not as good as when boron fluoride is also used. - The significant increase in yield when boron-fluoride was‘used as a catalyst tends to substantiate ourtcontention-that a‘Lewis acid should facilitate the production of positive hydroxyl from peracids. -Complex- ation of the carbonyl oxygen with boron fluoride should not only promote heterolysis of the 0-0 bond, but also-stabilize the resulting trifluoro- acetate anion. 0 BF 0 T - //J‘ 3 // . .cr,c —->- CH3C t. so, + 0H \fi \‘0 O-OH L [It should be noted that the boronfluoride may catalyze the reaction by coordinationwith-either the carbonyl or‘ 'ether' oxygen; only‘one of thesepossibilities is shown in the equationJ .Although no detailed investi- gation of this reaction was undertaken, the knownability of peroxy- trifluoroac etic acid to function ionically. in-aromatic oxidations' (60, 68, 69,70), coupled with the catalytic activity of an ionic catalyst, suggests . that the reaction: involves-'0H+. ~A concerted mechanism can also be envisioned involVing displacement by mesitylene on the complexed peroxide. ~However, the-marked difference in reactivity between catalyzed and un-ca-talyzed reactionsuggests that a much larger concen- trationof hydroxyl cation is produced in the former case. -Other‘ Lewis acids may functionin the same-manneruas boron fluoride. -Aluminum chloride, a more acidic material than boron fluoride (89), could not be used since a vigorous reaction—occurred when 125 peroxytrifluoroacetic acid was added to a suspension of aluminum :chloride in methylene chloride. - A white precipitate was formed along with copious evolution of hydrogen chloride. The peroxytrifluoroacetic acid does contain trifluoroacetic acid,. since this acid is formedalong with the peracid when trifluoroacetic anhydride is treated with hydrogen peroxide (63). ,Consequently, the salt- formed is most'likelyaluminum /o ,o 0 OH 'OOH trifluoroac etate since an identical reaction occurred when trifluoroacetic acid alone was added. The amount of hydrogen chloride evolved was 3CF3C _(CF,COO),A1. + 3HC1 ‘r OH determined quantitatively by sweeping it into standard base, giving three equivalents of hydrogen chloride evolved, in agreement with the reaction shown. The precipitate formed proved to be a mild catalyst for oxidations with peroxytrifluoroacetic acid. . When an excess'of - mesitylene (three equivalents) was treated withperoxytrifluoroacetic acid - at room temperature in the presence of this salt, .mesitol was obtained in 65% yield with-excellent recovery of unreacted mesitylene. This» slight increase in yield over the uncatalyzed reaction-(65 versus 45%) could _be due to the lower temperature used in the latter (<70). No other . catalyst systems were investigated. '. Extension of this one-step: synthesis of mesitol torsimilarly substituted-phenols,. seemed of interest, in view of the reporteddiffi- culties in preparing such phenols (90). .Synthesis ofiisodurenol (2, 3, 4, 6-tetramethylphenol, .XLV) from isodur ene and peroxytrifluoro- acetic acid-boron fluoride appeared particularly promising since 126 . OH CH3 ' CH3 , CH3 CH3 XLV such a one-step synthesis should be a significant improvement over the earlier multi-step synthetic methods. As anticipated,. isodurenol was prepared in 65% yield by a procedure identical to that usedwith vmesitylene. The structure of the product wasverified by comparison (of its m. p. with published values. The infrared and proton magnetic resonance spectra (see Table IX) of this compound were also-consistent with its structure (XLV). In this instance, the phenol was isolated by ~elution chromatography on alumina, the results representing one experi- ment only. Unreacted isodurene was recovered in 80% yield. , Prehnitene (2, 3, 4, 5-tetramethylbenzene), when treated in- an identical manner, gave in addition to the expected 2, 3, 4, S-tetramethyl- phenol, .XIVI, .a number of other products (XLVII-LI). The phenols £306: <3! XLVI ‘ XLVII XLVIII XLIX L LI ,OH «Mhu 127 were isolated, after normal workup, by extraction of the methylene chloride solution with Claisen's alkali. ~On neutralization, the water- insoluble material was taken up in a minimum amount of carbon tetrachloride and subjected to vapor phase chromatography. The four phenols appeared in the following order; 2, 3, 6-trimethy1phenol (L),- 2., 3,5-trimethy1phenol (XLIX), isodurenol (XLVII) and finally prehnitol - (XLVI). All of the phenols are known and were found to compare favorably in m. p. , infrared and ultraviolet spectra with published-data. The proton -magnetic resonance spectra of these compounds were also consistent with their structures (see Table XIV). Table XIV. . Proton Magnetic Resonance Spectrum (CCL, Solutions) of Substituted Phenols from the Oxidation of Prehnitene with P eroxytrifluoroacetic Acid- Boron‘Fluoride f r Com ound ‘Position Number of p (‘l‘ units) Protons 2, 3, 5-Trimethy1phenol 7. 96 3 7. 86 6 5.47, 3.71, 3.58 1 (apiece) 2, 3, 6-Trimethy1phenol 7. 90, 7. 85, 7. 83 3 (apiece) 5. 66 1 3.4.323L 2 Isodurenol 7. 89 12 5. 78,, 3. 36 l (apiece) Prehnitol 7. 91 - 3 7. 88 9 5. 57, 3. 73 1. (apiece) a Center-of an AB quartet. 128 The neutral solution after extraction with base gave, on-vacuum distillation, recovered prehnitene and a second fraction boiling slightly , higher, containing besides prehnitene, 4, 5, 6, 6-tetramethy1-2,4-cyclo- hexadienone (XLVIII), 2, 3,.5-trimethy1phenol (XLIX), isodurenol. and prehnitol. The structure of the cyclohexadienone ((XLVIII) is based on a satisfactory analysis of its 2, 4-dinitropheny1hydrazone and on its spectra. - In carbon tetrachloride, the ketone (XLVIII) had infrared bands at 1663 and 1630 cm'vl and in ethanol it had xmax 327 mp. (logs-5 3.48) which compares favorably with a similarly substituted dienone' (LII) reported by Mandell and co-workers (91) (infrared bands at 1667 and 1634 cm“; xii): 328 mu, logy:= 3.55). . Likewise Cocker (92) reported that LIII has spectra in- agreement with the cyclohexadienone XLVIII (infrared bands at 1633 and 1663 cm'l,-XrEr1taS‘H 320 mu, . loge = 3.69).. The proton magnetic resonance spectrum of XLVIII showed singlets at 8. 85 and 8. 15‘» , corresponding to six protons O O ' CH 3 CH OH CH3 H3 H3 COZH LII LIII each~(aliphatic and allylic methyls respectively) and doublets at 4. 19 and3. 15 ‘I‘ ‘ (J = 19 cps) each corresponding to a single vinyl proton. .Also inkeeping with the structure was the observationthat the 2, 4-di- nitrophenylhydrazone failed, to show any syn-anti isomerismin dimethyl- sulfoxide-d6, carbon tetrachloride and acetone-d6 by examinationof its proton magnetic resonance spectrum-(93). The preferred geometrical isomer is most likely 'LIV, since the other isomer would havewserious steric interactions with the methyl hydrogens. 129 N02‘ CH3. CH3 N CH3 N02 . N CH3 LIV_ The residue, after distillation of the neutral fraction, . was chromatographed on Fluorosil using petroleum ether, benzene, and ether as eluting solvents. The main component, after recrystalliza- tion from ethanol, proved to be 2, 2', 3, 3', 4, 4', 5, 5'-octamethy1dipheny1- methane. Its structure was verified by comparison with a known 7 sample prepared from formaldehyde, prehnitene and sulfuric acid (94). Its proton magnetic resonance spectrum has bands at 7. 92, 7.87‘and 7.83 ‘1‘ corresponding to twenty-four hydrogens, at 6. 27 1- for two Imethylene hydrogens of the diphenylmethane type and at 3. 611- for two aryl hydrogens. The high value of the aryl hydrogens (3.61 7' ) compares favorably. with the position of the aryl hydrogen (3. 72 h ‘) in z, 233, 3',- 4, 4', 5, 5', 6-nonamethyldiphenylmethane (95), both values apparently, due to'shielding by the aryl rings. This same residue also gave quite a bit of intractable tar which showed bands in the aliphatic region (7.60 to 8. 70 ‘P ‘) of the proton magnetic resonance spectrum but no orverylittle aromatic protons and strong carbonyl and double bond-absorption in the infrared (bands at 1634, 1661,. 1695, 1736 and» 1750 cm‘l). In Table XV can, be found the yields of products iso-latedlbased on unrecovered prehnitene. r The material balanceaccountedfor 85% of the prehnitene consumed; it .is unlikely that any signifiCant products withlow molecular'weightsrmnder 200) were-missed. The number and nature of the products isolated reveal the potency of the reactive intermediate formedby this reagent. . Attack of OH+ at a vacant ring position can furnish prehnitol. .Similar attack at C-1, 130 ocofidaoua pono>ooouso mo Ewes? ofi 0“ 9,330.” um» mo 336.3653 .90 venom .. msoflowum Hmuudoc 0:... mo Gofiumfifimflp 50: 033mm.” 65 mo >£thmoumfiouaodofigm .. fl ooflomfl .oumfiflmflu Rondo: 9%. mo mo>udo .0 .m .> no cofimhwofim . hmuoefidmfinm e3 dodged: oESHOm 2.92m 05 mo mm>nso .U.nm .> mo coagumoufi .. '4wa m to .3 oo .2 323m >23. H ex: 0 .o emm ozofi; nosooom 682852: m vodka m .MN H1H. .0>«ud>ah0~u wadfiuvginfivgnmwm N N .H e; E>qx .oooeooooxoeooso N .fi 2 1m 3V .Hoooafisfiofiieé .m..~ N -g or: coda Hoqohfiiooficshnna N .J 8.: ed 3255 Hoeoooooha N .; m .2 i: ESE 3:5on 8:32 1.3 8355 m.mmfigz oeooofioo 3..; 338m 23» Eoohona , .204 ufioomouosfiiaxouom.5C? ocoficgoam mo _GofiumpflxO can Eoum muospounm .>N 3nt fact... 131 followed by a methyl shift, accounts for the cyclohexadienone and iso- durenol. The tarry residue probably contains dimer or polymer + _ OH " O OH a 1 .2 shift XLVIII I OH 1, 2 OH shift XLVII L. _J from the dienone, XLVIII and analogous cyclohexadienones from attack of OH+ at C-2. The remaining products (which account for a major fraction of the prehnitene consumed) either lack'(trimethy1phenols XLIX and L) or have an extra (LI) carbon atom. . Complete absence of trimethylbenzenes and of isomeric trimethylphenols suggests that the trimethylphenols XLIX and L arise from loss of the para methyl group of prehnitol (XLVI) or isodurenol (XLVII) respectively, according to the following scheme shown for prehnitol (XLVI): + H0 H +Hi in-Emf; prehnitene V6“ 26‘ + , CH OH 2 ~H+ . 9— CH2 prehnite LI + X LIX XLVI New»... 132 This mechanism requires that the sum of the moles of trimethylphenols (XLIX and'L) equal the yield of coupled material LI. In fact, from 0. 0332 mole of prehnitene consumed, there was obtained 0. 0048 mole of 2, 3, 5-trimethy1phenol (XLIX), 0. 0017 mole of 2, 3, 6-trimethyl- phenol (L) and 0. 0039 mole of coupled material (L1). The sum of the phenols XLIX and L is too large, but the yield of these products was determined from vapor phase chromatography curves, whereas that of the coupled material LI is of isolated, purified, crystalline product and is probably low. When the reaction was repeated with boron fluoride, but without peracid, prehnitene was recovered (94%) unchanged. . Study of the reverse conditions (peracid but no boron fluoride) gave essentially the same products as obtained when prehnitene was oxidized with peroxy- trifluoroac etic acid-boron fluoride, but- in low. yield based on the amount of peracid used. Particularly noticable was the low yield of trimethyl- phenols and coupled material. .In Table XV can be found the yields of those products isolated based on unrecovered prehnitene. One of the materials in the alkaline-soluble portion had not been observed previously. Its proton magnetic resonance, infrared and mass spectra are recorded in the experimental. section. Its structure has not been. determined. Although in both the catalyzedand uncatalyzed reaction similar products were isolated, the yield of recovered prehnitene (90 ‘70) in the - latter is considerably. higher than when boron fluoride was used (73%). This result againillustrates thecatalytic effect of a Friedel-Crafts catalyst on the ionic decomposition of the peracid. Chloromesitylene, when treated with peroxytrifluoroac etic acid- boron fluoride under the same conditions as previously used, gave LV as a major product. This material was obtained by elution chromatography HO .CHZ LV 1 c1 133 of the reaction mixture after workup. ' No attempt was made to extract phenolic material from the initial organic layer since previous attempts gave little or no alkali-soluble material. The pure coupled material (LV, m.p.. 142. 5-143. 50), elutable from alumina with ether, accounted for 43. 5% of the total crude product. Its structure is based on satis- factory analysis, hydroxyl appearing at 3533 cm"1 in the infrared and its proton magnetic resonance spectrum which had bands at 7. 65, 7. 69,- 7.80, 7.89 and 7. 971- integr'ating for three hydrogens apiece, a single band at 6. 091' corresponding to twomethylene hydrogens ofthe di- phenylmethane type, a broad band at 5.45 1' for one hydroxyl hydrogen and two bands at 3. 97 and 3. 111- for one aryl hydrogen‘apiece. .The high-field aryl hydrogen (3. 97"» ') is ascribed to the ortho hydrogen on ‘LV, in agreement with previously observed shielding by aromatic rings in similarly substituted diarylmethanes. A The mass spectrum of this compound agreed with the proposed structure (96). Several points bear. mention before the mass spectrum of VL is considered in detail. .First, the isotopic distribution of a group of peaks for a chlorine-containing ion in the mass spectrum always shows unequivocally the number of chlorine atoms in the ion (97). .Second, ortho-methyldiphenylmethanes produce an intense peak of even mass number corresponding to .the pro- cess' (98): ' 7+ ._ .. + 214C ,CH; ‘——') ' or + _ Hz Ar' 1.. _. __ ‘ArCHz less H _J 'AR' plus H The mass spectrum shows that LV is a dichloro compound with M. W. 322-324-326. The ‘most abundant ion had a mass 166-168 (one chlorine atom), indicating an ortho-methyldiarylmethane, and establishing the formula: 134 C1 I ' .OH . CH, / CH3 3 ' CH3 ortho-Methyl substitution in the second ring would lead to'intense peaks at masses 168-170 (one chlorine atom). -Only a tiny peak-appeared ~ at 170ruling out such a structure. Peaks at 169-171 (one chlorine atom) correspond to loss, without rearrangement, of the first aryl group. . If the mesitylene carbon skeletons are assumed to remain .unchanged, the possible structures can be narrowed to: C1 ‘- CH3 CH3 . CH CH 3 2 C1 ‘CH3 *0H CH3 ; If the hydroxyl on the second ring ‘is in an ortho‘position it should behave like an ortho methyl; consequently only one structure (LI) remains. When this» compound was treated with-almninum-nickel alloy. under basic conditions (99) or lithium sand in n-butyl ether followed by hydrolysis, three products are isolated; recovered starting material,iLVI and a compound which appeared to contain one chlorine atom. The structure '- CH2 . Hog LVI of LVI was. verified by its mass spectrum where the parent peak falls at mass 254,. as expected. , The mass of the strongest peak, 132, indicated that one of the aryl groups is substituted ortho to the methylene. The spectrum showed essentially nothing at- mass 134, where‘a strong 135 peak would have been expected if the hydroxydimethylphenyl group were also‘ortho substituted. The proton magnetic resonance spectrum was also in agreement with the proposed structure,- LVI, having bands at 7. 76, 7. 88,. and 7. 89 1- corresponding to three, six and six aryl- methyl hydrogens respectively, a band at 6. 2.2 “h for two methylene hydrogens of the diarylmethane type and bands at 3. 60 and 3. 307‘ corresponding to two aryl hydrogens. apiece (the broad band of the hydroxyl was obscured in the background noise). No further work was done on the mono-chloro compound but its tentative assignment as LVII is based on the presence of three aryl hydrogens in its proton magnetic resonance spectrum in carbon tetrachloride, only one of which appears at 3. 97 '7‘ , the same position of the ortho aryl hydrogen in 'LV. Cl CH2 HO » LVII The remaining material from the oxidation was intractable tar (29%) along with three other crystalline materials (28%) which com- prised a minor portion of the total product. Two of these were elutable with benzene-ether and appeared to be coupled products lacking hydroxyl groups,'.as attested by the absence of hydroxyl bands in their infrared spectra. The third, elutable as an impurity with the coupled material ~ (LV) was extremely. difficult to purify; the proton magnetic resonance spectrum of the crude-material suggested it to be imainly 3-chloro- 2,4, 6-trimethy1phenol, having bandsxat 7. 76 and 7. 86‘? corresponding to six and three hydrogens respectively, at 5. 71 '1‘ corresponding to one hydrogen and a band at 3. 24 ‘h for one hydrogen. -No further work was done on- these materials. 136 The isolation of the coupled phenol, . LV, from the oxidationof chloromesitylene is consistent with the scheme suggested for the oxidation of prehnitene. The chlorine apparently does not prevent the hydride abstraction, but does inhibit the debenzylation. » Nitromesitylene, oxidized undersimilar conditions, gave a low yield of yellow crystals (m.p. 224-2250) which appeared to be similar in structure (LVIII) to the coupled material ,. LV, obtained in the oxi- dation of chloromesitylene. The assignment of structure LVIII is based Noz CH2 Noz H O LVIII on its satisfactory analysis, hydroxyl at 3535 cm"1 and its‘proton magnetic resonance spectrum which hadbands at 7. 75, 7. 82,. 7. 85' and 7.92‘!‘ corresponding six, three, three, and three protons, respectively, a band at. 6. 12 ‘l‘ for twomethylene hydrogens of the diarylmethane type and bands‘at 3.67 and 2. 83 ‘l‘ corresponding to one aryl hydrogen apiece. The yield (10. 6% based onxperacid) could be improved by running the reaction at reflux (21. 2%). The low recovery of pure, unreactedvnitro- mesitylene and considerable amounts of intractable tar, regardless of the temperature, suggest that the nitro-group, with its destabilizing influence on electrophilic substitution, makes side reactions much .more favorable. Several unsuccessful attempts were made to extend this oxidation with peroxytrifluoroacetic acid-boron fluoride toless substituted aromatic compounds. - In particular, benzene gave a black charcoal- like material insoluble in most organic solvents. Attempts to :moderrate the conditions by using a seven-fold excess of benzene gave similar results. ~Omission of boron fluoride from the latter oxidation gave a far ' “~15 137 less vigorous reaction, although the recovery of benzene was poor. "Aluminum trifluoroacetate, " which from earlier results appeared to be a milder catalyst than boron fluoride, gave in most instances large amounts of tar, containing traces of phenol. Again variation of the reaction conditions failed to give sufficient phenolic material to warrant closer examination. Similarly, Musgrave and co-workers (60) found the reaction of peroxytrifluoroacetic acid and benzene to be a complex one. The reactivity of initially formed phenol towards further attack certainly is one reason for the results observed, especially with peroxytrifluoroacetic acid-boron fluoride since this appears to be a particularly potent oxidant. Although a more detailed investigation of the tarry materials obtained in the oxidations of benzene might prove interesting, it was felt that in these initial investigations, more could be gained by careful examination of the less complex reactions. ‘ 4* I». EXPERIMENTAL . I. , Oxidations A. Mesitylene (a) Peroxytrifluoroacetic Acid-Boron Fluoride. In a 300-ml. , three-necked, round-bottomed flask equipped with-a condenser, jacketed dropping funnel, alcohol thermometer and gas inlet tube was placed 56. 1 g. (0.468’mole) of mesitylene and 100 ml. of redistilled methylene chloride. To this magnetically stirred solution was added peroxytri- fluoroacetic acid prepared by mixing in the cooled (0°) dropping funnel, 35 g. (0. 167 mole) of trifluoroacetic anhydride (Eastman. Kodak Co.), 50 m1. of methylene chloride and 4. 0 m1.. (0. 147 mole) of 90% hydrogen peroxide (Becco Chemical Division, F. M. C. Corporation), then allow- ing the mixture to warm to room temperature. * Boron fluoride was bubbled through the reaction mixture during addition, which required two and one-half hours. The reaction was strongly exothermic, and the temperature was kept below 70 by a salt-ice bath. After addition was complete, the boron fluoride addition was stopped and the solution allowed to warm to room temperature. Water (100 ml.) was added and the aqueous layer separated, then washed with three 25-ml. portions of methylene Chloride. The combined organic layers were washed with 10% sodium bisulfite (until the washings gave a negative potassium iodide test for peroxide) and 50 ml. of 10% sodium bicarbonate, then dried over *Adequate precautions should be taken when handling this material. ‘MC'Clure (88) states "the above mixture can be detonated by a modifi- cation of the drop weight test method of Bellinger (81). " See Bulletins No. 3 and'No- 46 of Becco‘ Chemical Division, ~ F.M. C. .Corporation, Buffalo 7, -N.Y. , on the handling of 90%h'ydrogeri-péro‘xide. 138 139 anhydrous magnesium sulfate. The solvent was distilled and the residue separated by distillation through a one—foot glass helices packed column giving 32. 0 g. of recovered mesitylene, b.p. 84-869 at 35 mm. and 17. 7 g. of mesitol (88. 5% yield based on peracid used), b.p. 98° at 10 mm., m.p. 69-700, m.m.p. 69-700 (100). The tarry residue from the distillation weighed 2. 0 g. The product was identical in all respects with-an authentic sample of mesitol (infrared and proton magnetic resonance spectrum). This same reaction when using only a slight excess of mesitylene (18. 7 g.), gave 6.0 g. (30% yield based on peracid used) of mesitol, 7.0 g. of recovered mesitylene and a considerable amount of tar (5.7 g.). (b) Beroxytrifluoroacetic Acid. In a 500-ml. , three-necked flask provided with a Tru-bore stirrer, alcohol thermometer, jacketed dropping funnel and a reflux condenser was placed 56. 1 g. (0.468 mole) of mesitylene and 100 m1. of redistilled methylene chloride. To this ice-cooled mixture was added dropwise, a solution containing 0.157 mole of peroxytrifluoroacetic acid prepared as before from‘35. 1 g. . (0. 167 mole) of trifluoroacetic anhydride, 50 ml. of methylene chloride and 4. 3 ml. of 90% hydrogen peroxide (0. 157 mole). Addition was complete in two and one-half hours and the reaction mixture was worked up in the same manner as when boron fluoride was used, giving on distillation 39.0 g. of recovered mesitylene and 9.67 g. of mesitol (45 C70). The tarry residue from the distillation weighed 4. 25 g. . (c) Peroxytrifluoroacetic Acid--‘"Aluminum Trifluoroacetate. " Ina 500-m1. , three-necked, round-bottomed flask provided with-a 'Tru-bore stirrer, jacketed dropping funnel: anda reflux condenser connected to a trap containing standard sodium hydroxide, was placed 22. 3 g. (0.167 mole) of aluminum chloride and 50 m1. of methylene chloride. To this was added dropwise, 57. 0 g. (0. 50 mole) of trifluoro- acetic acid (EastmanfiKodak Co. ). The vigorously evolved hydrogen 140 chloride was swept into standard base by a dry nitrogen stream. After addition was complete, the remaining traces of solvent were removed by slight warming of the solution (600), leaving a white-brown pre- cipitate. The unreacted base was determined using standard hydro- chloric acid and in this way 0.44 mole of hydrogen chloride was found liberated (89%). To this precipitate was added 56. 1 g. (0.468 mole) of mesitylene followed by dropwise addition (nitrogen atmosphere) of peroxytrifluoro- acetic acid prepared in the usual manner from 19. 1 g. (0. 167 mole) of trifluoroacetic acid, 50 ml. of methylene chloride and 4. 0 m1. of 90% hydrogen peroxide. The stirred reaction mixture was kept at 25-300 during addition by means of an ice-water bath. Addition was complete in two and one-half hours. The contents of the flask were then poured onto a minimum amount of ice and dilute hydrochloric acid, neutralized with 10% sodium bicarbonate and extracted thoroughly with additional methylene chloride. Workup of the organic layer was the same as used previously with boron fluoride as a catalyst, giving on distillation 37. 5 g. of unreacted mesitylene and 13.0 g. of crude mesitol (65%), m.p. 67-700. The tarry residue weighed 2.0 g. B. Isodurene This compound was oxidized in the same manner as mesitylene using peroxytrifluoroacetic acid and boron fluoride. No attempt was made to improve the following results which represent a single experi- ment. » In a 500-ml. , three-necked, round-bottomed flask was placed 54. 0 g.» (0.410 mole) of isodurene and 100 ml. of methylene chloride. To this stirred and cooled mixture (< 70) was added dropwise a solution of peroxytrifluoroacetic acid prepared in the cold (0°) from 30. 9 g. (0. 147 mole) of trifluoroacetic anhydride, 50 ml. of methylene chloride . Wihfi. 141 and 3.7 ml. of 90% hydrogen peroxide (0. 137 mole), then allowing the mixture to warm to room temperature. Boron fluoride was bubbled through the reaction mixture during addition, which required two and one-half hours. After addition was complete, the boron flouride addition was stopped and the solution allowed to warm to room tempera- ture. Water (100 ml.) was added and workup of the organic layer was the same as used in the similar oxidation of mesitylene. After removal of the solvent by distillation there remained a dark-brown liquid residue which was chromatographed on 200 g. of alumina using as eluting solvents, petroleum ether (60-900), benzene, ether and methanol. The benzene- petroleum ether elutable material gave on distillation 28. 5 g. of recovered isodurene, b.p. 76-780 at 10 mm. and a residue (7. 3 g.) which after repeated recrystallizations from pentane, gave 2. 3 g. of isodurenol, m.p. 74-760. Reported (101): m.p. 79-810). The remaining isodurenol was present in the ether elutable fractions and weighed 10.6 g. (total yield, 62%). Its infrared (see Figure 38) and proton magnetic resonance spectra (see Table XIV) after recrystallization, m.p. 79-810, were consistent with its structure. C . Pr ehnitene (a) Peroxytrifluoroacetic Acid- Boron Fluoride. In a 500 m1. , three-necked, round-bottomed flask was placed 16. 24 g. (0. 1212 mole) of prehnitene and 100 ml. of redistilled methylene chloride. To this cooled mixture (< 70) was added dropwise, peroxytrifluoroacetic acid prepared from 9. 0 g. (0. 0424 mole) of trifluoroacetic anhydride, 30 ml. of methylene chloride and 1.04 ml. (0.0383 mole) of 90% hydrogen peroxide in the usual manner. Boron fluoride was bubbled through the reaction mixture during addition, which required two and one-half hours. After addition was complete, the boron fluoride addition was stopped and the solution allowed to warm to room temperature. Water (100 ml.) was 142 Ma Enos 02.5 fiuwaoaocrm? NH 2 OH 0 w h o , m w ._ ._ .1 . a _ _ q _ _ ‘ $0 .AcofidHOm JDUV #985603 mo Efihuuomm Condoms .wm ohdmwh 143 added, the two layers separated and the aqueous layer salted thoroughly with sodium chloride, then extracted with methylene chloride. The com- bined organic layers were washed with 10 ml. of 10% sodium bisulfite, 10 m1. of 10% sodium bicarbonate, then 40 ml. of Claisen's alkali (14 g. of potassium hydroxide, 10 m1. of water then methanol to 40 m1. total volume). The basic extract was washed with methylene chloride, neutralized with dilute hydrochloric acid then extracted with ether. The water layer was separated and salted thoroughly with sodium chloride, then extracted with additional ether. The combined ether layers were dried over anhydrous magnesium sulfate, filtered and the solvent removed by distillation. The residue (1. 58 g.) was taken up in a minimum of carbon tetrachloride and subjected to vapor phase chromatography using a column of 20% silicon (SE-30) on Chromasorp W. (The conditions used can be found in Figure 39. Details on the column and instrument are reported in a separate section). The chromatogram (see Figure 39) showed five materials present: prehnitene (7. 6%), 2, 3, 6-trimethylphenol (15.8%), m.p. 59-610 [reported (102); m.p. 620], 2,3,5utrimethylphenol (40%),. m.p. 91-930, icyic’hexane 282, 278, and 273 mp. [reported (103); m.p. 95-960, ng’hexane 282, 278 and 274 mp (85)], isodurenol (10.2%), m.p. 810, m.m.p. 810 (104) and prehnitol (27.4%) m.p. 81-830, x33: 277.5, 282.5, and 286.5 mu [reported; m.p. 86-870 (105), REES: 277.5, 282. 5 and 286.5 mg (106)] in increasing retention times. The infrared spectra of the trimethylphenols were identical to published spectra (107). Isodurenol proved identical in all respects (m.p. , m.m.p. , infrared and proton magnetic resonance spectrum) to an authentic sample. The infrared spectrum of prehnitol can be found in Figure 40 and agrees with published spectra (107). The proton magnetic resonance spectra of these compounds were consistent with their structures and are recorded in Table IX. 144 .L—J— ! I' l I l I J 1 Figure 39. V.P.C. chromatogram of alkali soluble portion from oxidation of prehnitene with peroxytrifluoroacetic acid- boron fluoride. Sample size: 4 p1 Attenuator: 2 Column {[‘emp: 15200 Detector Temp.: 252 Sol ent Injector Temp: 259 ' V He. Press: 40 p. s.i. OH OH OH OH 1 11 5 /1o 15 20 25 3O 35 40 Time (min.) 145 Amsonofihv fiwcofioermg Ma Ma S 0H m m N. _ _ _ _ 2 _ _ _ .AcowudHOm JUUV Hofiafionm mo Enuuoomm possum": .ow oudmwh $0 146 The neutral fraction obtained after extraction with Claisen's alkali was washed with water, then dried over anhydrous magnesium sulfate. The solvent was removed by distillation and the residue distilled through a seven-inch, vacuum-jacketed Vigreux column giving 11. l g. of prehnitene, b.p. 85-900 at 16 mm. , (pure by vapor phase chromatography) and a second fraction (0.71 g.), b.p. 90-125o at 16 mm. which was further purified by vapor phase chromatography using a column of 20% silicon” (SE-30) absorbed on Chromasorb W. The column and conditions were identical to that used in the alkali-soluble fraction. This second fraction contained five materials; prehnitene (79.4%), 4, 5, 6, 6-tetramethyl- 2, 4-Cyclohexadienone (12. 2%), 2, 3, 5-trimethylphenol (3. 65%), isodurenol (2. 56 %) and prehnitol (Z. 24%) in increasing retention times. The structure of the ketone was based on its infrared spectrum (see Figure 41), ultra- violet spectrum [xEtOH max 327 my. (log 5 =- 3.48)] and proton magnetic resonance spectrum in carbon tetrachloride which showed singlets at 8.85 and 8. 15 ‘1‘ , corresponding to six protons each, and doublets at 4.19 and 3.15 ‘l‘ (J = 19 c.p. 8.), each corresponding to a single proton. The ketone was derivatized by allowing it to stand with 2, 4-dinitropheny1- hydrazine in ethanol for seven days, followed by refluxing for 30 min. The solution was cooled, water added and the precipitate taken up in chloroform. The chloroform extract was treated with magnesium sulfate and Bentonite, according to the method of Shine (108), then filtered and the solvent removed. Recrystallization from 95%aqueous ethanol gave deep red crystals,_m.p. 153°(after cooling, remelted at 156°). 33311. .Calc'd for C,,H,,N,o,;.c, 58.17; H, 5.49; N, 16.97. Found: C, 58.28; H, 5.53; N, 16.89. The proton magnetic resonance spectrum of this 2,4-dinitropheny1- hydrazone in carbon tetrachloride had bands at 8. 65 and 8. 12 ‘T‘ corres- ponding to six protons apiece, a single band at 3. 62 ‘1‘ for twohydrogens anda complex multiplet below 2. 00 7‘ for the hydrogens in the nitrated ring. 147 MA Annoyed»: 5383453 NH 3 0H m. m N. i . _ _ w . . 0852233333 .. $.N1H>Auo§uuouue .0 .m .v «o guuoomm pondpwfl . 3.. oufimfim 148 The single band at 3.62“? became an AB system with change in solvent but the remaining portion of the spectrum stayed virtually constant. The residue from the distillation of the neutral fraction (2. 55 g.) was chromatographed on Fluorosil using petroleum ether (60-900), benzene, ether and methanol as eluting solvents. The petroleum ether elutable material was recrystallized from ethanol giving a white crystalline material, 1.10 g.., m.p. 150-151°. Its ultraviolet spectrum suggested it contained two prehnitene moieties, kcyclohexane 271 mu CYC1°hexane 268-mp. (e = 652). The proton m ax magnetic resonance spectrum of this material in carbon tetrachloride (.5 F 290), prehnitene, )1 indicated it to be an octamethyldiphenylmethane. The high field position (3.61‘1~ ) of the aryl hydrogens is suggestive of a structure such as 2, 2', 3, 3', 4, 4', 5, 5'-octamethyldiphenylmethane [reported; m. p. 150--1510 (94)]. The infrared spectrum of this material is shown in Figure 42. The mass spectrum of this compound (96) indicated a di- ortho methyl substituted diphenylmethane and its molecular weight was consistent with C21H28- (Calc'd 280. 4,. Found; 280) final. Calc'd for C21H28= C, 89.94;" H, 10.06. -Found: .C,. 90.04; H, 9.86. The structure of this material was confirmed by independent synthesis from 8.98 g.- (0. 067 mole) of prehnitene, 0. 51 g. (0. 017 mole) of paraformal- dehyde and 23.8 ml. of an ethanol-concentrated sulfuric acid solution ' (1 to 2. 5 by volume) using the method of Smith and-Welch (94). The crude product, after workup, gave on distillation, prehnitene and 5. 0 g. . (83%) of crude 2, 2', 3, 3', 4, 4', 5, 5'-octamethyldiphenylmethane, m. p. 146-148°, b.p. 235° at 16 mm. Several recrystallizations from ethanol gave white crystals m.p. 150-1510, identical in all respects (m.m.p. , infrared and proton magnetic resonance spectrum) to the compound iso— lated from the oxidation of prehnitene. 149 Amaouowgv.. sumsoaoerm? ma NH .2 0H 0 m N. _ _ _ A 5 4 4:03.390. JOOV ocmauogcgoamwcanuoflnwuoo ...m .m 34.4 ._m .m ..N .N no Snooper possess .Ns onsmnh j £0 . — 150 The remaining material from this residue proved elutable with methanol-ether (1.20 g.). Its infrared spectrum (see Figure 43) and tarry nature suggested it to be a complex mixture. Its proton magnetic resonance spectrum had bands at 7.60 to 8. 70'1‘ but no detectable absorption in the aryl hydrogen region (2. 00 to 3. 80 ‘P ). , (b) Boron Fluoride. The same procedure was repeated with prehnitene but without peroxytrifluoroac etic acid. In a 300-m1. ,‘ three- necked flask was placed 8. 12 g. (0. 0606 mole) of prehnitene and 50 m1. of methylene chloride. To this cooled mixture ( < 70) was added drop- wise, 2.45 g. (0.0215 mole) of trifluoroacetic acid in 10 ml. of methylene chloride. Boron fluoride was bubbled through the reaction mixture during addition, -which required two and one-half hours. After addition was complete, the boron fluoride addition was stopped and the mixture allowed to warm to room temperature. Water was added, the layers separated and the aqueous layer extracted with additional methylene chloride. The combined organic layers were washed with 10% sodium bicarbonate, 10% sodium bisulfite, then dried over anhydrous magnesium sulfate. The solvent was removed by distillation and the residue dis- tilled through a seven- inch vacuum-jacketed, Vigreux column giving 7.50 g. (92.5%) of prehnitene, b.p. 86-870 at 16 mm., pure by vapor phase chromatography. No other products were isolated. (c) Peroxytrifluoroacetic acid. In a 300-ml. , three-necked flask ' was placed 16. 24 g. (0. 1212 mole) of prehnitene and 100 ml. of methylene chloride. To this cooled solution‘(< 70) was added dropwise, peroxytri- fluoroacetic acid prepared from 9. 0 g. (0. 0424 mole) of trifluoroacetic anhydride, 30 m1. of methylene chloride and 1.04 ml. (0. 0383 mole) of 90% hydrogen peroxide in the same manner'as employed in previous oxidations. After addition was complete (two and one-half hours), the mixture was allowed to'warm to room temperature. . Water (100 ml.) was added, the two layers separated and the aqueous layer salted thoroughly with sodium 151 MA chouficb fimc30>m>> .0Euodd “3.331300 030000uo3fint€nou0m 53, 022333099 Ho 0300ng got coo—33° .23 03000035 mo 80.50090 Cessna: .3. 0.33% -4 _ - 152 chloride, then extracted with methylene chloride. The combined organic layers were washed with 10 m1. of 10% sodium bisulfite, 10 m1. of 10% sodium bicarbonate, then 40 ml. of Claisen's alkali. The basic extract was washed with methylene chloride, neutralized with dilute hydrochloric acid, then extracted with ether. The water layer was separated, salted with sodium chloride and extracted with additional ether. The combined ether layers were dried over anhydrous magnesium sulfate, filtered and the solvent removed by distillation. The residue (1.05 g..) was taken up in a minimum of carbon tetrachloride and separated by vapor phase chromatography using a column of 20% silicon (SE-30) on Chromasorp W. The chromatogram‘ (see Figure 44) showed five .major materials present, prehnitene (3. 14%), a mixture of 2, 3, 6- and z, 3, 5-trimethy1phenol (6.7820), Fraction 3 (34. 5%), isodurenol (20.9%) and prehnitol (29.8%). All of those compounds found previously in the oxidation of prehnitene with peroxy- trifluoroac etic acid-boron fluoride were identified by comparison of retention times on the chromatography column under the same conditions and the identity of their respective proton magnetic resonance spectra with known compounds (see Table XV). Fraction 3 had not been observed previously. Its proton magnetic resonance spectrum‘(see Figure 46) had bands at 8.72, 8. 35 and 7.97‘1‘ corresponding to three protons apiece, a band at 7.87 7‘ corresponding to six protons and a single band at 7. 23 ’r integrating for approximately two hydrogens. The infrared spectrum'(see Figure 45) of this material had hydroxyl absorption at a 3. 00 p. with carbonyl and double bond absorption at 5. 80, 5. 90 and 6. 10 u. The mass spectrum gave a molecular weight of 182' and suggested the presence of an'acetyl group. Its ultraviolet spectrum had a broad band at 264 mp. ( €33 6, 350). It has not yet been identified. The neutral fraction obtained after extraction with Claisen's alkali was washed with water, then dried over anhydrous magnesium sulfate. The solvent was removed by distillation and the residue distilled through 153 Figure 44. Vapor phase chromatogram of the alkali- soluble fraction from oxidation of prehnitene with peroxy- trifluoroacetic acid. Sample size: 4 Pl Attentuator: 2 0 .Column temp: 152 Detector 259: Injector 259 H e Press: 40 p. s.i. fraction n3: 5 10 15 20 25 30 35 40 45 Time (min. ) 1.54 Amcou 35V aumnoaoxfim? 2 2 S cm a w e o _ _ _ _ w _ n 1 a 7 e. :3. 0.5"th _Eoum m Gowuumnm mo 53.30on pounds; .mv unamwh 155 Figure 46. -Pr‘oton magnetic. resonance spectrum of Fraction 3 from oxidation of prehnitene with peroxytrifluoroacetic acid (alkali soluble fraction). TMS I II I 7.23 7.87 7.97 8.35 10.00 156 a seven-inch, vacuum-jacketed Vigreux column giving 14. 28 g. of prehnitene, b.p. 88-900 at 17 mm. and a dark-brown liquid residue, 1.43 g. This material was separated into volatile (0.63 g.) and non- volatile (0. 80 g.) portions by distillation directly into a still-head at 0.07 mm. , the pot-temperature reaching a maximum of 1800. The volatile portion was further separated by vapor phase chromatography on a column of 20% silicone (SE-30) on Chromasorb W, separating the crude liquid into five materials, prehnitene (51. 2%), 4, 5, 6, 6-tetramethy1- cyclohexa-Z, 4-dienone (26. 6%), 2, 3, 5-trimethylphenol (7. 5%), isodurenol (12.4%) and prehnitol (2. 7%). All of these compounds were found previously in the oxidation of prehnitene with peroxytrifluoroacetic acid- boron fluoride and were identified by comparison of retention times on the chormatography column under the same conditions and the identity of their respective proton magnetic resonance spectra, with previously obtained spectra. The non—volatile portion of the residue, from the distillation of the neutral product (0. 80 g. ), showed 2, 2', 3, 3', 4, 4', 5,5'-tetramethyl- diphenylmethane present in trace amounts. as evidenced by its proton magnetic resonance spectrum (6.27‘1“ for the diarylmethylene protons and 3. 61 ‘l‘ for the aryl hydrogens). The small yield coupled with its tarry nature prompted no further investigation. Its infrared spectrum was almost identical to a similar fraction isolated in the oxidation of prehnitene with peroxytrifluoroacetic acid-boron fluoride (Figure 43). D. Chloromesitylene . - v” e_.... w..- (a) Peroxytrifluoroacetic Acid-Boron Fluoride. In a 300-ml. , three-necked flask equipeed with a condenser, jacketed dropping funnel, alcohol thermometer and gas inlet tube was placed 20. 0 g. (0. 130 mole) of chloromesitylene (prepared according to the procedure of McBee and Leech (109), b.p. 900 at 15 mm.) and 50 ml. of methylene chloride. 157 To this cooled (€ 70), magnetically stirred solution was added dropwise peroxytrifluoroacetic acid prepared from 9. 0 g. of trifluoroacetic anhydride, 15 m1. of methylene chloride and 1.04 ml. (0. 0383 mole) of 90% hydrogen peroxide in the usual manner. Boron fluoride was bubbled through the reaction mixture during addition, which required two and one-half hours. Water (100 ml.) was added, the two layers separated and the aqueous layer salted, then extracted with’additional methylene chloride. The combined organic layers were washed with 10% sodium bisulfite and 10% sodium bicarbonate, then dried over anhydrous magnesium sulfate. The solvent was removed by distillation and the residue distilled through a seven—inch, vacuum-jacketed Vigreux column giving 15.0 g. of recovered chloromesitylene, b.p. 90-930 at 15 mm. ,1 pure by vapor phase chromatography. The residue (4. 65 g.) from the distillation was chromatographed on neutral alumina, using as eluting solvents, petroleum ether, benzene, ether and methanol. The product separated into three distinct fractions, that elutable with petroleum ether-benzene, recrystallized from ethanol, 0.5 g., m.p. 108-1100, a material elutable with benzene-ether, 0. 5 g. , recrystallized from ethanol, m.p. 157-159°, and that elutable with ether, 3.65 g. , m.p. 139-143°. The latter material, obtained in the greatest yield (78. 5% of the total product weight) was further purified by re-chromatography on alumina, giving as a mafn fraction (2. 00 g. ), a white crystalline material, m.p. 142-1430 on recrystallization from acetic acid-water. The remaining material in the latter chromatography was accounted for by a tarry complex mixture (1. 00 g.) from which could be sublimed at reduced pressure, 80°, a white, cotton-like material, m.p. 83-840, yield 0. 3 g. The material found in the largest yield (2. 00 g. m.p. 142-1430, 39% yield based on unrecovered chloromesitylene) had a proton magnetic resonance and infrared spectra (see Figure 47) suggestive of a 158 Ammonoflbv fimdofioefim? NH HA OH 0 w * eel 3 — — .AGQSSHOm JUUV H4 mo 8:30on Usudnfifl. .5“ ouswflh -»\0 IO HO 4w all-Pm 159 diphenylmethyl structure containing a phenolic hydroxyl (2.80 H): The proton magnetic resonance spectrum had bands at 7.65, 7.69, 7.80, 7.89 and 7. 97 ‘1‘ corresponding to three protons apiece, a single band at 6. 09 ‘1‘ for two protons of the diphenylmethane type, a broad band at 5.45 7‘ for one hydroxyl hydrogen and singlets at 3. 97 and 3. 1111‘ for one aryl hydrogen apiece. The mass spectrum of this material further suggested an ortho methyl substituted diphenylmethane having structure 'LV (see page .132). M. W. calc'd for CmHzoOClz; 322, found; 322. Anal. Calc'd for C18H200C12: c, 66.87; H, 6.24; c1, 21.93. Found: C, 66.85; H, 6.25; C1, 22.02. The structure of this material was substantiated by examination of the product obtained on removal of the two chlorine atoms (see next section). The petroleum ether elutable material from the original chroma- tography (0.5 g. , m.p. 108-110°) was recrystallized again fromethanol giving white prisms, m.p. 118--119O (cooled and remelted, m.p. 110-1110). Its proton magnetic resonance spectrum had bands at 7.96, 7.85, 7.75, 5.97 and 3. 21 ‘1" in the relative areas of 3:3:3:1:1. Its infrared spectrum (see Figure 48) showed no distinct functionality. rA possible structure is shown below: C1 C1 CH3 , No attempt was made to distinguish this structure from other possible isomers. ~é_r_1_§_1_. Calc'd for ClgszCIZ: C, 71.02; H, 6.90; Cl, 22.07. Found: C, 71.14; H, 7.02; CI, 22. 10. The benzene-ether elutable material (0. 5 g. , m.p. 157-1590) proved to be a mixture, from examination of its proton magnetic resonance spectrum, and it could not be purified adequately since the yield was low. The same was true of the material found as an impurity 160 Amconufifiv numnofiorwm? Ma Ma .2 0H 0 w h _J .0300de Genoa-3,00 ufloomouogflngxouom 5E? ocogfimogonoflfi mo “53.32.88 05 80.3 dogwood 03.8.30 umauongmaouumm mo Eduuoomm ponmnde .ww 0.30th H0 H0 M 161 in the fraction elutable with ether. It could be separated partially by sublimation of the tarry portion obtained on rechromatography of the ether elutable material (m.p. 83-840,» 0. 3 g.). The proton magnetic resonance spectrum of the crude product had bands at 7. 76, 7. 86,: 5. 71 and 3. 24'1‘ corresponding to the approximate relative areas, 6:3:l:1 respectively, suggesting the material tobe mainly 3-chloro- 2,4, 6-trimethylphenol. ' Hydroxyl absorption (3. 00 p.) was found in the infrared spectrum in agreement with such a structure. - All attempts to further purify this sublimate resulted inconsiderable loss of the already small amounts, making complete identification impossible. (b) Removal of the chlorines from the coupled product, LV. In a 300-ml. , three-necked flask, flushed with dry argon and equipped with a high speed stirrer, sintered-glass tipped gas inlet tube, and condenser was placed 0. 069 gr. (0. 010 gram atom) of lump lithiumand 50 m1. of mineral oil. The flask was heated by means of a Fisher burner until the lithium melted, at which time the mixture was vigorously stirred. - When the lithium was thoroughly dispersed, the heating was discontinued, but the rapid stirring was maintained fora short period to pr event the lithium sand from fusing. After the mixture cooled to room temperature, most of the mineral oil was removed by the addition of pentane in portions, followed by suction on the inlet tube until most of the mineral oil was gone. ~The pentane was then replaced with anhydrous ether in the same manner until a suspension of lithium sand in 75 m1. of anhydrous ether remained. To this cooled (0°) mixture was added dropwise, l. 00 g. (0. 0031 mole) of 4-hydroxy-Z, 3-dichloro-3, 5, 2';- ' ‘4', 6'-pentamethyldiphenylmethane in 25 m1. of anhydrous ether. The mixture was refluxed for twelve hours, the ether replaced with anhydrous n-butyl ether, then refluxed for two more hours. The reaction mixture was cooled withan ice bath-and hydrolyzed by dropwise addition of 20 m1. of water. The crude product was extracted with ether, the two layers 162 separated and the organic layer washed successively with 5% hydro- chloric acid and water. The ether layer was dried and the solvent removed. The solid residue (1.00 g.) was separated by elution chroma- tography on alumina using petroleum ether, benzene, ether, acetone,- ethanol and methanol as eluting solvents. The ether-acetone elutable material (0.16 g. ,1 m. p. 113-114. 5°) contained impurities of a fraction appearing immediately after, 0. 50 g. and a third fraction elutable with acetone which proved to be starting material by comparison with an authentic sample, 0. 20 g. (m.m.p. ,6 infrared and proton magnetic resonance spectrum). The two earlier fractions obtained in this chroma- tography were then subjected to vapor phase chromatography using a column of 20% silicon (SE-30) on Chromasorb W. (The details on the instrument and column can be found in a separate section. Injector, 2680; detector, 252°; column temperature, 2300; helium pressure, 40 lbs. p. s. i.) The first fraction appearing off the column was collected (manner of collection described in a separate section) and recrystallized from acetic acid-water, m.p. 121-123°. Its mass spectrum was con- sistent with starting material minus two chlorine atoms, i_. e_. , 4-hydroxy-3, 5, 2', 4', 6'-pentamethyldiphenylmethane'(M. W. calc'd for (31stan 254, found: 254). The proton magnetic resonance had bands at 7. 76, 7. 88 and 7. 89’}- corresponding to three, six and six arylmethyl hydrogens respectively, a band at 6. 22 T for two methylene hydrogens and bands at 3.60 and 3. 30 ‘1‘ for two aryl hydrogens apiece (the broad band of the hydroxyl was obscured in the background noise). The second fraction off the column appeared to be starting material minus one chlorine atom, _'. e_. , .4-hydroxy-2-chloro-3, 5, 2', 4', 6'-penta- methyldiphenylmethane, as suggested by its proton magnetic resonance spectrum which had bands in the aryl-hydrogen region of 3. 30 and 3. 97‘]‘ . This material was not further investigated. A third fraction from the vapor phase chromatographic separation proved to be recovered starting- material. 163 This same mixture of products was obtained when the coupled material, ~LV, was treated with aluminum-nickel alloy under basic conditions, according to the method of Schwenk and Papa (99). E. Nitromesitylene (a) Peroxytrifluoroacetic acid-boron fluoride. In a 300-ml. , three- necked flask equipped with a reflux condenser, inlet tube, alcohol thermometer and jacketed addition funnel was placed 9.4 g.- (0. 057 mole) of nitromesitylene (prepared according to the method of Powell and Johnson (110), m.p. 42-44°, reported; m.p. 44°.) and 50 m1. of methylene chloride. To this cooled, magnetically stirred, solution (< 70) was added dropwise, peroxytrifluoroacetic acid prepared from 4. 5 g. (0.0212 mole) of trifluoroacetic anhydride, 0. 5 ml. (0. 0192 mole) of 90% hydrogen peroxide and 15 ml. of methylene chloride in the usual manner. Boron fluoride was bubbled through the reaction mixture dur- ing addition which required two and one-half hours. After addition was complete, the boron fluoride addition was stopped and the mixture allowed to warm to room temperature. Water (100 ml.) was added, the two layers separated and the aqueous layer salted thoroughly, then washed with additional methylene chloride. The aqueous layer gave a negative test for unreacted peroxide with potassium iodide. The com- bined organic layers were washed with 10% sodium bicarbonate and water, then extracted with 40 ml. of Claisen's alkali. The dark red basic extract was washed with methylene chloride, neutralized with 10% hydrochloric acid, then extracted with ether. The water layer was separated, . salted with sodium chloride and extracted with additional ether. The combined ether layers were dried over anhydrous magnesium sulfate, filtered and the solvent removed by distillation. The residue (0.7 g.) was dissolved in 95% ethanol, treated with Norite, filtered and precipitated in the cold after addition of water, giving yellow platelets, 164 m.p. 222-2250, 0. 35 g. This material was recrystallized again until the m.p. remained constant, 224-2250. A dark brown oil remained from these recrystallizations which could not be purified further. The proton magnetic resonance and infrared spectra (see Figure 49) of these yellow crystals suggested structure LVIII, analogous to the coupled material, LV, found in the oxidation of chloromesitylene. Anal. .Calc'd for CIBHZOOSNz: c, 62.77; H, 5.85; N, 8.14. Found: C, 62.69; H, 5.71; N, 8.02. The yield based on the peracid used (2 equivalents of peracid for every mole of coupled product) was 10.6% (0.35 g.). The neutral layer after extracting with base was washed with water, then dried over anhydrous magnesium sulfate. The solvent was re— moved with a Rinco rotary evaporator and the residue (7.9 g.) recrystal- lized from anhydrous ethanol giving 6. 2 g.. of pure recovered nitro- mesitylene. The same reaction using identical amounts of material was re- peated, but the peroxytrifluoroacetic acid was added to the methylene chloride solution of nitromesitylene at reflux. This method gave 0. 65 g. (21%) of coupled material, LVII, and 5. 3 g. of pure recovered nitro- mesitylene after the usual workup. The poor recovery of unreacted starting material in both reactions, 68.4% in the former and 60. 5% in the latter, can partially be ascribed to recrystallization losses along with unidentified impuritie s . F. Benzene (a) Peroxytrifluoroacetic acid-boron fluoride. In a BOO-m1. , three-necked flask equipped with a gas inlet tube, condenser, alcohol thermometer and jacketed dropping funnel was placed36. 7 g.. (0.468 mole) of benzene and 100 m1.. of methylene chloride. To this cooled magnetically stirred solution (< 70) was added dropwise, 165 .. _ Amcouofigv zuwcofioxrmg MA NH 2 OH o w w L ’ b 1 _ A _ a _ _ _ .AGOSDHOm «HUEOV 03.33“ couonugom ofloom quOdHfintflnonom so?» ocean—Mmogoufid mo coflmgxo ,, of 50.3 HHS/1H mo Sunbeam poudnfifi .3. oudmfm IO ~ 166 peroxytrifluoroacetic acid prepared in the usual manner from 35. 1 g. . (0.167 mole) of trifluoroacetic anhydride, 50 ml. of methylene chloride and-4. 3 ml. (0. 157 mole) of 90% hydrogen peroxide. Boron fluoride was bubbled through the reaction mixture during addition, which re- quired two and one-half hours. The solution turned dark blue almost immediately and darkened as addition proceeded. - After addition was complete, the boron fluoride addition was stopped and the mixture allowed to warm to room temperature. Water (100 ml.) was added and the two layers filtered to remove an insoluble black solid (4. 0 g.). This solid burned with difficulty and failed to melt below 250°. The water layer was separated from the filtrate, salted with sodium chloride and extracted with ether. The ether extract was dried over anhydrous magnesium sulfate and gave, on removal of the solvent, a black solid (1. 5 g.) which resembled the same material obtained above. The methylene chloride layer from the original filtrate was washed with water (the washings gave a negative potassium iodide test for peroxide), then='100 m1. of 20% sodium hydroxide. The base extract gave after normal workup, 1. 0 g. of a phenol smelling oil. The methylene chloride solution obtained after extraction with base was washed with 10 ml. of water then dried over anhydrous magnesium sulfate. Tye/.Solvent. was~rernioved by distillation and the residue dis- tilled, through a twelve-inch, glass-helices packed column giving 22. 0 g. of recovered benzene. The same reaction was repeated using a seven-fold excess of benzene (85. 7 g.., 1. 1 moles) and again gave besides recovered benzene an intractable black solid as the main product. (b) Peroxytrifluoroacetic acid-"aluminum trifluoroacetate. " In a 500-ml. , three-necked flask, swept with nitrogen and equipped witha Tru-bore stirrer, alcohol thermometer, condenser and jacketed dropping funnel was placed 22. 3 g. (0.167 mole) of aluminum chloride and 50 ml. 167 of methylene chloride. To this was added dropwise, 57.0 g. (0. 50 mole) of trifluoroacetic acid. After addition was complete, the remain- ing traces of solvent were removed by slight warming of the solution (60°) leaving a white-brown precipitate. To this solid was added, 85. 7 g. (l. 10 moles) of benzene followed by dropwise addition:(nitrogen atmosphere) of peroxytrifluoroacetic acid prepared in the usual manner from 35.1 g. (0. 167 mole) of trifluoroacetic anhydride, 520 m1. of methylene chloride and 4. 3 ml. (0. 157 mole) of 90% hydrogen peroxide. The reaction mixture was kept at 20-250 by an ice bath, and addition required two hours. The reaction mixture was stirred for one more hour, then poured onto a minimum amount of ice and hydrochloric acid. The aqueous layer was separated, salted with sodium chloride and extracted with methylene chloride. The organic layers were combined," and dried over anhydrous magnesium sulfate. The filtered solution was distilled through a six-inch Vigreux column giving after removal of the solvent, 106 g. of a mixture containing trifluoroacetic acid and benzene, b.p. 70-800. The remaining residue was distilled under reduced pressure giving 2. 5 g. of a red oil, b.p. 60-650 at 2 mm. This material had hydroxyl absorption (3. 00 u) in the infrared and was shown to contain phenol by comparison of its spectrum with that of a known sample. This same reaction was repeated several times with varying amounts of benzene (equimolar, three and five fold excesses) and gave in all cases essentially the same results. (c) Peroxytrifluoroacetic acid. This reaction was identical to the method used in the oxidation of benzene with peroxytrifluoroacetic acid-boron fluoride but without boron fluoride (same amounts, temperature and procedure). There was obtained after the usual workup, 23 g. of recovered benzene and 2. 1 g. of tar. In contrast to the boron fluoride catalyzed oxidation there was no‘ "black solid" melting above 250°. 168 ' II. Vapor-Phase Chromatography The separations by vapor phase chromatography reported in this portion of the thesis were accomplished on an Aerograph Model A-90—P Gas Chromatograph‘(Wilkens Instrument and‘Research, > Inc. ’- Walnut . Creek; California). The column used was that supplied by the manu- facturers and consisted of 20% silicon gum rubber (SE-30) coated on 60/80 mesh HMDS Chromasorb W (Johns-Mansville Product) and packed in a 5'x1/4" stainless steel column. Samples were collected by injecting 40 pl samples (any higher size sample causes flooding of the column) andcollecting the materials as they appeared (off the colmnn inta capillary tube bent at right angles. SUMMARY 1. - Peroxytrifluoroacetic acid catalyzed with boron fluoride oxidized excess mesitylene and isodurene to mesitol and isodurenol in excellent yields (88 and 65%) at low temperatures (< 70). The un- catalyzed reaction was far less exothermic and gave lower yields of phenol. 2. A similar oxidation of prehnitene gave, in addition to prehnitol,- z, 2', 3, 3', 4, 4',95,15'-octamethyldiphenylmethane, isodurenol, 4, 5,6, 6- tetramethylcyclohexa-Z, 4-dienone, 2, 3, 5- and 2, 3, 6-trimethylphenol. The results of this and earlier oxidations with peroxytrifluoroacetic acid-boron fluoride are best explained by the intermediacy of OH+, an extremely potent oxidant capable of not only. electrophilic replace- ment but electrophilic attack at already substituted positions, and hydride abstraction of aryl methyl hydrogens. 3. Oxidation of chloro- and nitromesitylene gave primarily coupled material of the following structure: X X HO CH2 (X = Cl, N02) The formation of this ‘material is analogous to the products obtained in the oxidation of prehnitene and were explained by a common mechanism. 4. Oxidation of less substituted compounds such as benzene were extremely complex and gave in most instances, regardless of conditions, intractable tars. 169 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) ll) 12) 13) 14) 15) 16) LITERATURE CITED H. Hart andR.-W. Fish, J. Am..Chem.-Soc., 82,5419 (1960). H. Hart and J. Fleming, TetrahedronLetters, No. 22, 983(1962). M. Ballester and C. Olinet,- Abstract of the D. So. Theses of Molinet, 1958, p- 18. M. Ballester,.C. Molinet and J.-Castanar,- J.- Am. .Chem..Soc. , _82, 4254 and 4259 (1960). N. Roobol, Ph. D. Thesis, MichiganState University, 1962. R. Rafos,-Ph. D. Thesis, MichiganState University, 1961.‘ H. Hart and R..W.~Fish, J. Am..Chem.~Soc., 82, 4460 (1961). E.~S. . Gould, "Mechanism and-Structure in Organic Chemistry," Henry Holt and Company, New‘York, 1959, p. 219. M. A. Batiste, J. Am. Chem.-Soc., 83, 4101. (1961). R. - J.- Rolih and H. . Peters, Standard Oil Co. (Ind.), -Whiting, Indiana. (Private Communication) . C. Friedel and J. M. Crafts, Ann. chim. phys., (6) l, 449 (1884). M. Gomberg, Ber.,._3_3_,.3144 (1900); J. Am..Chem. Soc., 2, 752 (1900). J. Boeseken, Rec. trav. chim., 3.4, l (1905). L. I..Smith and O. W.-Case, J.- Am.-Chem..Soc., 54, 1614 (1932). G- Baddeley and A. G. Pendleton, J. Chem. Soc., 52, 2392 (1917). L.~M. Jackman, "Applications of Nuclear'MagneticResonance Spectroscopy in'Organic Chemistry,- " Pergamon Press, New York, 1959, p. 58. G.~A. Olah,.W. S. Tolgyese andR. E. A. Deas, J. Org- Chem., _2_7_,.344l (1962). 170 17) 18) 19) 20) 21) 22) 23) 24) 25) 26) 27) 28) 29) 30) 31) 32) 33) 34) 35) 171 L..F. Fieser and M. Fieser, ~ "Advanced Organic Chemistry, " ReinhOld Publishing Corporation, New York, N. Y., 1961, p. 651. R. M. Silverstein and G. C. Bassler,-"Spectrometric Identification of Organic Compounds, " John Wiley and Sons, Inc. , . New York, N. Y. 1963, pp. 92-93. M. S. Newman and N. C. Deno, J. Am..Chem. Soc., 13, 3651 (1951). L. M. Jackman, 100. (23., p. 47. A. Schonberg, O..Schutz and S. Nickel, Ber., 61, 1384 (1928). H. Staudinger, Ber., 42, 3966 (1909); Ber., _4_3, 728(1910). A. H. Blatt, editor. "Organic Synthesis,.Collective Volume II, " John Wileyand'Sons, Inc., New‘York, N. Y., 1957, p. 573. M. Gomberg,. Ber., _3_3, 3144 (1900); J. Am. Chem. Soc., 32, 752 (1900). A. Hantzch, Ber., 552, 953 (1922). F. Strauss and A. Dietzmann, J. Prakt. Chem., _1_0_3_, 1 (1921). J. Boeseken, Rec. trav. chim., _2_4_, l (1905). W. Schlenk and G. Racky, Ann., 394, 215 (1912). W. W. Kaeding and L. J. Andrews, J.- Am. Chem. Soc., 12' 6189 (1952). G..Wittig and B..Faitman, .Ann. 554, 213 (1943). V.-A. Ismarl'skii and D. K. Surkov, J. Gen. Chem- (USSR), _1_3_, 848 (1943). W. W. Kaeding and L. J. Andrews, J. Am. Chem. Soc., 74, 6189 (1952). '— W..Schlenk _e_t .11.. Ann., 463, 98 (1928). J. F. Norris and E. H. Green,-Ann., 26, 492 (1901). S. D. Wilson and Y. Y. Cheng, J. Org.-Chem. _5_, 223 (1940). 36) 37) 38) 39) 40) 41) 42) 43) 44) 45) 46) 47) 48) 49) 50) 51) 52) 53) 172 A. Wohl and E. Mertyporoch, Ann., _4_8__1_, 30 (1930). F. C. Hahn, J. Am. Chem. Soc., _4__3_, 175 (1921). C. E. Coffey, J. Am. Chem. Soc., 83, 1623 (1961). J. F. Norris and E. H. Green, Am. Chem. J., _3_0, 392 (1903). M. Gomberg and J. D. Todd, J. Am. Chem..Soc., 3_9_, 2392(1917) D. Fish, Ph. D. Thesis, Michigan State University, 1960. R. c. Fuson and R. L. Albright, J. Am. Chem. Soc., 21». 488 '(1959). L. M. Jackman, 139: _c_it_., p. 125. F. A. Bovey and C. E. Johnson, Jr., J. Chem. Phys., 2_9_, 1012(1958). L. M. Jackman, 1_o_c_. 213., p. 124. T. Sulzberg and H. Hart, J. Org. Chem., _2_8_, 1159(1963). N. C. Deno,tP. T. Groves and G. Saines, J. Am. Chem. Soc.,) _8_1_,- 5790 (1959). R. J. Gillespie and J. A. Leisten, Quart. Rev., _8, 40 (1954). A. Fischer, B. A. Grigor, J. Pocker and J. Vaughn, J. Am. Chem. Soc., _8_3_, 4208 (1961). N. C. Deno, P. T. Groves and G. Saines, J. Am. Chem. Soc., 81, 5790 (1959). R. B. Moodie, T. M. Connor and R. Stewart, Can. J. Chem., _3_7_, 1402 (1959). R. Dehl, W. R. Vaughn, and R. S. Berry, J. Org. Chem., 33, 1616 (1959). M. E. Peach and T. C. Waddington, J. Chem. Soc., 600 (1962). 173 55) H. A. Smith and c. D. Shacklett, J. Am. Chem. Soc., 3;, 766 (1951). 56) R. R. Aitken, G. M. Badger and J. W. Cook, J. Chem. Soc., 331 (1950). 57) L. I. Smith and C. L. Moyle,~ J. Am. Chem. Soc., _5_5_, 1676 (1933). 58) M. S. Newman and H. Lloyd, J. Am. Chem. Soc., It, 2672 (1952). 59) G. Illuminati and G. Marino, J. Am. Chem. Soc., 18, 4975 (1956). 60) L. E. Smith and C.‘L. Moyle, J. Am. Chem. Soc., _5_8_, 1 (1936). 61) N. P. Biur-Hoi, G. Lejeune and M. Sy, Compt. rend., 240, 2241 (1955); C. A. 59, 11272. 62) F. Benington, R. D. Morin and L. C. Clark, Jr., J. Org. Chem., _l_9_, 11 (1954). 63) R. C. Fuson, M. D. Armstrong, W. Wallace and J. Kneisley, J. Am..Chem. Soc., 6_6_, 681 (1944). 64) C. Sollscher, Ber., _12, 1682 (1882). 65) M. S. Kharasch, R. Morrison and W. H. Uny, J. Am. Chem. Soc., 6_6_, 368 (1944). 66) M. Weiler, Ber., 32, 1910 (1899). 67) G. Baddeley, J. Chem. Soc., 332 (1942). 68) R..C. Fuson, J. Org. Chem., _1_1, 60 (1945). 69) L. 1. Smith and C. Guss, J. Am..Chem. Soc., 52, 804 (1937). 70) R. Adams and C. W. Theobald, J. Am. Chem. Soc., _6_5_, 2208 (1943). 71) W. A. Waters, Ann. Reports, 3, 130 (1945). 72) G. Stein and J. Weiss, Nature, _l-6_6_, 1104 (1950). 73) J. G. Wallace, "Hydrogen Peroxide in Organic Chemistry, " E. I. duPont de Nemours and Co. (Inc.), Electrochemicals Dept. 74) 75) 76) 77) 78) 79) 80) 81) 82) 83) 84) 85) 86) 87) 88) 89) 90) 91) 92) 174 D. H. Derbyshire and W. A. Waters,aNature, 165, 401 (1950). I. M. Roitt and W. A. Waters, J. Chem. Soc., 3060 (1949). D. B. Denny and H. M. Weiss, Chem. and Ind., 818 (1962). A. K. Brodskii and N. A. Vysotskya, Zhur. obsheher. khim., 25, 2273 (1962). R. D..Chambers, P. Goggin and W. K. R. Musgrave, J. Chem. Soc.,” 1804 (1959). D. sworn, "Organic Reactions, "Edited by R. Adams e_t 31., John Wiley, New York, N. Y., 1953,. Vol. VII, p. 318. R. T. Arnold and R. Larsen, J. Org. Chem., 5, 250 (1940). W. D. Emmons, J. Am. Chem. Soc., 16, 3468, 3470(1954). W. D,.Emmons,y A. S. Pagano and J. P. Freeman, J. Am. Chem. .soc., 16, 3472 (1954). W. D. Emmons and A. S. Pagano, J.1Am. Chem. Soc., 31, 89, 4557 (1955). W. D. Emmons and G. B. Lucas, J. Am. Chem. Soc., 11, 2287 (1955). E. J. Bourne, J. E. B. Randlis, M. Stacey, J. T. Tatlow and J. M. Teddler, J. Am. Chem. Soc., _7_6_, 3206 (1954). J. D. McClure and P. H. Williams, J. Org. Chem., _2_7, 627 (1962). J. D. McClure, J. Org. Chem., Z._8_, 69 (1963). J. D. McClure and P. H. Williams, J. Org. Chem., _21, 24 (1962). G. A. Russell, J. Am. Chem. Soc., 81, 4834 (1959). H. Pakshinamurty and M. Santappa, J. Org. Chem., _2_7, 1839 (1962). L. Mandell, D. Caine and G. E.-Kilpatrick, J. Am. Chem. Soc., _8_3, 4457 (1961). W. Cocker, Chem. and Ind. (London), 1041 (1955). 93) 94) 95) 96) 97) 98) 99) 100) 101) 102) 103) 104) 105) 106) 107) 108) 109) 110) 111) 175 G. J.,~Karabatsos,iR. A. Taller and F. M. Vane, J. Am. Chem. .Soc., _8_5_, 2326 (1963). H. A. Smith and C. M. Welch, J. Am. Chem. Soc., _7_3_, 4391 (1951). N. S. Bhacca, L.- F. Johnson and J. N. Shoolery, " NMR Spectra Catalog, " Varian Associates, Palo Alto, California, Spectra No. 357. S. Meyerson, American Oil Company (Ind.) (Private Communication). J. H. Beynon,- "Mass Spectrometry and its Applications toOrgam’c Chemistry, " D. Van Nostrand Co. , New York, 1960, p. 298. S. Meyerson, H.) J. Drews and E. -K. Fields; paper in preparation. E..Schwenk, B. Whitman, H. Ginsberg and D. Papa, J. Org. Chem. ,— 2, 1 (1944). F. Belinger (it 341., Ind. Eng. Chem. _3_8, 310 (1946). D. H. Hey, J.,Chem. Soc., 1590 (1931). G. T. Morgan and A. E. J. Pettit, J. Chem. Soc., 418 (1934). D. H. Key, J. Chem. Soc., 1590 (1931), R. A. Friedel and M. r Orchin,- "Ultraviolet Spectra of Aromatic Hydrocarbons, " John Wiley and Sons, Inc. , New York, N. Y. A. Tohl, Ber., _2_1, 904 (1888). "Organic Electronic Spectral Data, " M. J. .Kamlet, editor, Vol. I, 1946-1952, Interscience Publishers, Inc. ,1 New York. "Satdler Standard Spectra, " Sadtler'Research Laboratories, Philadelphia, . Penn. , 1962. H. Shine, J. Org- Chem. _2_4, 252 (1959). E. T. McBee and R. E. Leech, Ind. Eng. Chem.,}i, 393 (1947). E. Powell and F. R. Johnson,‘ "Organic Synthesis, " Collective Volume II, John Wiley and Sons, New York, 1957, p. 449. W. J. Blaedel and V. W. Meloche, - "Elementary Quantitative Analyses, " Row,.Peterson and Company, Evanston, Illinois, 1957, p. 288. m‘Al‘A g..-a...” Cuddle)“: uolwt)‘