THE KENETFCS AND MECHANISM OF THE CYCLODEHYDRATION QF 2uPHENYLYRIPHENYLMEYHANOLS Thus}: far “he Dogma of D“. D. MICHIGAN STATE UNIVERSITY Edward A. Seder 1965 THESIS LIBRARY Michigan State University This is to certify that the L thesis entitled The Kinetics and Mechanism of the Cyclodehydration of 2-Phenyltriphcnylme tha nols presented by Edward A. Sedor ‘ has been accepted towards fulfillment of the requirements for Mdeqree in Chemistry Mfr/WM 4L?N( Major prolessor Date DK’(- ‘0' 1 ‘(fi'l ( ( I 0-169 "T r": aA' I’ll r ._ V , . l . "t '1. I L." 1 E“ u I 4 r 571 .190» —. k‘. D h'. l v ABSTRACT THE KINETICS AND MECHANISM OF THE CYCLODEHYDRATION OF 2-PHENYLTRIPHENYLMETHANOLS by Edward A. Sedor The purpose of this investigation was to synthesize several substituted 2—phenyltriphenylmethanols and to study the acid-catalyzed cyclization to 9,9-disubstituted fluorenes. The carbinols were prepared by the method first reported (1,2); that is, the addition of a Grignard reagent to a benzophenone, resulting in carbinols of the following general type: The rates of cyclodehydration of twelve substituted carbinols were determined Spectrophotometrically in 4,5 and 6 weight percent sulfuric acid in 80% aqueous acetic acid. The cyclizations were pseudo first order with respect to carbinol. The kinetics revealed that the rates were generally enhanced by electron—donating substituents and decreased by Edward A. Sedor electron-withdrawing substituents. However, the para-methoxy substituents were the exception. Here the reaction was favored by electron-withdrawal. A Hammett plot for most of the carbinols correlated best with 6 and had a negative p. The para-methoxy and the di-para— methyl substrates, however, differed showing a better corre- lation with 6+ with a positive p+. The reaction appeared to proceed by two different mechanisms. The activation parameters also divided the substrates into two groups, the bulk of the carbinols with small and positive A S* values and the para-methoxy substrates with large negative A 8* values. The activation energies generally were larger with the electron-withdrawing substituents. The effects of varying the initial acid concentration and using a deuterated reaction solvent (H2804-CH3COOD-D20) were investigated and helped to define the reaction path. Only one mechanism was shown to be Operative. In all cases, the first step was rapid and reversible protonation of the carbinol. The rate-determining step then depended upon the substituents. Most of the compounds reacted by an A-l process as shown best by deuterated solvent studies (kDao/ngo r»! 5 (5)); the ionization was slow and rate- determining followed by rapid aromatic electrophilic substi- tution. For the para-methoxy substrates ionization to the carbonium ion was rapid and complete, followed by rate— determining cyclization. Edward A. Sedor REFERENCES 1. F. Ullman and R. von Wurstemberger, Ber., §§, 4105 (1905). 2. E. Khotinsky and R. Patzewitch, Ber., $2, 5104 (1909). 5. K. Wiberg, Chem. Rev., 55, 715 (1955). THE KINETICS AND MECHANISM OF THE CYCLODEHYDRATION OF 2-PHENYLTRIPHENYLMETHANOLS BY Edward AT Sedor A THESIS ( Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1965 Dedication This thesis is dedicated to my wife, Jacqueline, for her patience and encouragement in completing the experimental work and for her invaluable assistance in perfecting this manuscript. It is also dedicated to my parents, not merely for their financial and moral support, but for their belief in the values of higher education. ii ACKNOWLEDGMENT The author wishes to express his sincere appreciation to Doctor Harold Hart for his invaluable guidance and understanding during this investigation. Appreciation is also extended to Michigan State Uni- versity for providing a teaching assistantship from September 1961 through May 1964, and to the United States Army Research Office--Durham, for providing financial assistance from June 1964 through November 1965. iii TABLE OF CONTENTS Page INTRODUCTION . . . . . . . . . . . . . . . . . . . . 1 RESULTS AND DISCUSSION . . . . . . . . . . . . . . . 8 I. Synthesis. . . . . . . . . . . . . . 9 A. The Starting Carbinols (2-—Pheny1tri- phenylmethanols). . . . . . . . . . 9 B. The Reaction Products (9, 9-Disubsti- tuted fluorenes). . . . . . . . . . . . 10 C. Product Study of Cyclodehydration . . . 10 II. Kinetics . . . . . . . . . . . . . . . . . . 15 A. Rates and Correlations. . . . . . . . . 15 B. Reaction Mechanism. . . . . . . . . . . 54 III. Miscellaneous. . . . . . . . . . . . . . . 48 A. Kinetics of Cyclodehydration of 9-(2- biphenylyl) -9- fluorenol . . . . . . . . 48 B. Attempted Cyclization of 1,9-Dipheny1- 9-fluorenol to 12—Pheny1indeno[1,2,5- ijfluorene . . . . . . . . . . . . . . 51 C. Sulfonation of Triarylcarbonium Ions. . 55 EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . 56 I. General. . . . . . . . . . . . . . . . . . . 57 II. Syntheses. . . . . . . . . . . . . . . . . . 58 A. Preparation of 2—Pheny1benzophenone . . 58 1. Preparation of 2-Phenylbenzoic Acid . . . . . . . . . . . . . . . 58 2. Preparation of 2-Carboxamido- biphenyl . . . . . . . . . . . . . 58 5. Preparation of 2-Cyanobiphenyl . . 59 4. Preparation of 2-Pheny1benzo- phenone (Method A) . . . . . . . . 59 B. Preparation of 2-Phenylbenzophenone (Method B). . . . . . . . . . 60 C. Preparation of 2- -Phenyltriphenyl— methanols . . . . . . . . . . . . . . . 61 iv TABLE OF CONTENTS - Continued Page 1. Preparation of 2-Phenyltriphenyl- methanol . . . . . . . . . . . . . 61 2. Preparation of 4-Chloro-2'-phenyl- triphenylmethanol. . . . . . . . . 62 5. Preparation of 4-Methoxy-2'-phenyl- triphenylmethanol. . . . . . . . . 65 4. Preparation of 4-Methyl-2'-phenyl- triphenylmethanol. . . . . . . . . 64 5. Preparation of 5-Chloro-2'bpheny1- triphenylmethanol. . . . . . . . . 66 6. Preparation of 5-Methoxy-2'-phenyl— triphenylmethanol. . . . . . . . . 67 7. Preparation of 5-Methyl-2'-phenyl- triphenylmethanol. . . . . . . . . 68 8. Preparation of 5-Trifluoromethyl- 2'-phenyltriphenylmethanol . . . . 7O 9. Preparation of 4,4'-Dichloro-29- phenyltriphenylmethanol. . . . . . 71 10. Preparation of 4,4'-Dimethoxy-2"- phenyltriphenylmethanol. . . . . . 72 11. Preparation of 4,4'-Dimethy1-2"- phenyltriphenylmethanol. . . . . . 74 12. Preparation of 4-Methoxy-4'-methyl- 2"-phenyltriphenylmethanol . . . . 75 D. Preparation of 9,9-Disubstituted fluorenes . . . . . . . . . . . . . . . 76 1. Preparation of 9,9-Diphenyl- fluorene . . . . . . . . . . . . . 76 2. Preparation of 9-(4-Chloropheny1)- 9-phenylfluorene . . . . . . . . . 77 5. Preparation of 9— (4-Methoxyphenyl)- 9-phenylfluorene . . . . . . 77 4. Preparation of 9- (4- -Methylphenyl)- 9-phenylfluorene . . . . . . . . 78 5. Preparation of 9- (5- Chlorophenyl)- 9— —phenylfluorene . . . . . . . . 79 6. Preparation of 9- (5- -Methoxyphenyl)- 9-phenylfluorene . . . . 80 7. Preparation of 9- (5-Methylphenyl)- 9-phenylfluorene . . . . . . . . . 80 8. Preparation of 9-(5-Trif1uoro- methylphenyl)-9-phenylfluorene . . 81 9. Preparation of 9,9—Di(4-Chloro- phenyl)fluorene. . . . . . . . . . 82 TABLE OF CONTENTS - Continued E. F. 10. Preparation of 9, 9-Di-(4-Methyl— phenyl)fluorene. . . . . . . . 11. Preparation of 9, 9-Di- (4-Methoxy- phenyl)fluorene. . . . . . . . . . 12. Preparation of 9-(4-Methoxyphenyl)- 9-(4-Methylphenyl)fluorene . . . . Preparation of Deuterated Reaction Med- ium for Kinetic Studies . . . . . . . . Evaluation of Products from Kinetic Runs. . . . . . . . . . . . . . . . . . III. Kinetics of Cyclodehydration . . . . . . . A. The Kinetic Method and Equations. . . . B. Examples of the Method. . . . . . . . . IV. Miscellaneous. . . . . . . . . . . . . . . . A. Preparation of 9-(2-Biphenylyl)-9- fluorenol . . . . . . . . . . . . . . . B. Preparation of 9,9-Spirobifluorene. . . C. Preparation of 1,9-Diphenyl—9-fluorenol 1. Preparation of 9-Fluorenone-1- Carboxylic Acid. . . . . . . . . . 2. Preparation of the Diacyl peroxide of 9-Fluorenone-l-carboxylic Acid. 5. Preparation of 1-Phenyl-9-fluor- enone. . . . . . . . . . . . . . 4. Preparation of 1, 9-Diphenyl- -9- fluorenol. . . . . . . . . . . . . D. Attempted Cyclization of 1,9-Diphenyl- 9-fluorenol . . . . . . . . . . . . . . SPECTRA - Infrared and Nuclear Magnetic Resonance. . LITERATURE CITED . . . . . . . . . . . . . . . . . vi Page 85 84 84 85 86 86 86 95 99 99 99 99 99 99 100 100 101 105 127 TABLE II. III. IV. VI. VII. VIII. IX. XI. XII. LIST OF TABLES Effect of Varying Initial Carbi ol Concen— tration on Reaction Rates at 25 and 4% H2804 o o o a o c o o o o o o o o o o o o o 0 Examples of Calculated Rate Constants for Cyclization at 25 . . . . . . . . . . . . . . Calculated Rate Constant for é-Methoxy-2'- phenyltriphenylmethanol at 25 . . . . . . . . Comparison of Rates of Formation of Fluorenes and Disappearance of Carbonium Ions for Com- pounds XXII, XXIII and XXIV . . . . . . . . . The Substituted Carbinols Used in the Cyclo- dehydration Studies . . . . . . . . . . . . . Rates of Cyclization of 2-Biphenylmethanols in 4% H2804 . . . . o . . . . . . . . . . . . Tabulation of Ea, AH: and ASi for the Cyclo- dehydration . . . . . . . . . . . . . . . . . Rate Constants for the Cyclodehydration in 4toe%Hgso4................ Rate Constants for the Cyclization of 4,4'- Dimethyl-, 4-Methoxy- and 4-Methoxy-4'-methyl- 2"-phenyltriphenylmethanols in Various Per Cent H2504 (sec’ ). . . . . . . . . . . . . . Ratesoof Cyclization in 4% H2804-CH3C00D-D20 at 25 . . . . . . . . . . . . . . . . . . . . Rate Constants for the Cyclodehydration of 9- (2-Biphenylyl)-9-fluorenol Compared with those for 2-Phenyltripheny1methanol . . . . . Comparison of the Activation Parameters for 2-Phenyltriphenylmethanol and 9-(2-Biphenylyl)- 9-fluorenol . . . . . . . . . . . . . . . . . vii Page 14 17 18 20 20 21 51 41 42 44 49 50 LIST OF TABLES — Continued TABLE Page XIII. Calculations of Rate Data by Computer forO 5'-Methyl- 2-phenyltriphenylmethanol at 20 and 4% H2804. o o o o o o o o o o o o o o o o 96 XIV. Calculations of Rate Data by Computer forO 4-Methyl-2'-phenyltriphenylmethanol at 150 and 4% H2804. . . . . . . . . . . . . . . . 97 viii FIGURE 1. 10. 11. 12. 15. 14. LI ST OF FIGURES ACD Page PlOt of 2.505 log [ h 1 versus Time a) (Ah - Ah) for 4—Chloro-2'-pheny1tripheny1methanol at 250 . . . .... . . . . . .... . . . . . . . 15 Plot of log ( a} ) versus Time for Ah ' Ah Cyclization of 4-Methoxy-2'-phenyltriphenyl- methanol at 200 . . . . . . . . . . . . . . . 18 Plot of log kr versus 5 at 200 and 4% H2804 . 25 Plot of log kr versus 6 at 250 and 4% H2804 . 24 + Plot of log kr versus 6 at 250 and 4% H2804. 25 Plot of log kr versus 5 at 500 and 4% H2804 . 26 + Plot of log kr versus 6 at 500 and 4% H2804. 27 Arrhenius Plot for the Cyclodehydration . . . 50 Plot of AH* versus As* for the Cyclodehydra- tion. . . . . . . . . . . . . . . . . . . . . 52 Plot of,AHi versus 6 for the Cyclodehydration 55 Hammett Plot for Hydrolysis of Benzoate Esters in 99.9% H2504 . . . . . . . . . . . . 57 The Hammett Plot for the Formation of Semi- carbazones from Substituted Benzaldehydes at pH 5.9. O O O O O O O O O O O O O I O O O O O 58 Comparison of Powell Plot for 4—Chloro—2'- phenyltriphenylmethanol with Standard Curve for First and Second Order Reaction . . . . . 95 Plot of Absorbance versus Time for 5-Methyl- 2'-pheny1triphenylmethanol at 200 and 4% H2504 . . . . . . . . . . . . . . . . . . . . 95 ix LIST OF FIGURES - Continued FIGURE 15. 16. 17. 18. 19. 20. 21. 22. 25. 24. 25. 26. Plot of Absorbance versus Time for 4-Methyl- 2'-phenyltriphenylmethanol at 150 and 4% H2804 o o o o o o o o o o o o o o o o o o o 0 Comparison of Ultraviolet Spectra of 5-Chloro- 2'-pheny1triphenylmethanol and 9-(5-Chloro- phenyl)-9-phenylfluorene. . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 4-Chloro-2'-Phenyltriphenyl- methanol. . . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 4-Methoxy-2'-phenyltriphenylh methanol. . . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 4-Methy1-2'-phenyltriphenylmethanol Infrared'and Nuclear Magnetic Resonance Spectra of 5-Chloro-2'-phenyltriphenylmethanol Infrared and Nuclear Magnetic Resonance Spectra of 5-Methoxy-2'-pheny1triphenyl- methanol. . . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 5-Methy1-2'-phenyltriphenylmethanol Infrared and Nuclear Magnetic Resonance Spectra of 5-Trifluoromethyl-2'-phenyltri- phenylmethanol. . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 4,4'-Dichloro-2"-phenyltriphenyl- methanol. . . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 4,4'-Dimethoxy-2"-phenyltriphenyl- methanol. . . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 4,4'-Dimethy1-2"-phenyltriphenyl- methanol. . . . . . . . . . . . . . . . . . . Page 95 98 104 105 106 107 108 109 110 111 112 115 LIST OF FIGURES — Continued FIGURE 27. 28. 29. 50. 51. 52. 55. 54. 55. 56. 57. 58. 59. Infrared and Nuclear Magnetic Resonance Spectra of 4-Methoxy-4'-methyl-2"-phenyltri- phenylmethanol. . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9,9-Diphenylfluorene . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9-(4-Chlorophenyl)-9-phenyle fluorene. . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9-(4-Methoxyphenyl)-9-phenyl- fluorene. . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9-(4-Methylphenyl)-9-phenyl- fluorene. . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9-(5-Chloropheny1)-9-phenyl- fluorene. . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9-(5-Methoxyphenyl)-9-phenyl— fluorene. . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9-(5-Methylphenyl)~9-phenyl- fluorene. . . . . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9-(5-Trifluoromethylphenyl)-9- phenylfluorene. . . . . . . . . . . . . . . Infrared and Nuclear Magnetic Resonance Spectra of 9,9-Di-(4-Chlorophenyl)fluorene. Infrared and Nuclear Magnetic Resonance Spectra of 9,9-Di-(4-Methylphenyl)fluorene. Infrared and Nuclear Magnetic Resonance Spectra of 9,9-Di-(4—Methoxyphenyl)fluorene Infrared and Nuclear Magnetic Resonance Spectra of 9-(4-Methoxyphenyl)-9-(4-methyl- phenyl)fluorene . . . . . . . . . . . . . . xi Page 114 115 116 117 118 119 120 121 122 125 124 125 126 INTRODUCTION The first acid-catalyzed cyclodehydration of Z-bi- phenylylmethanols to fluorenes was reported in 1905 by Ullman and von Wurstemberger (1), who obtained 9,9-dipheny1- fluorene by heating crude 2—phenyltriphenylmethanol with concentrated sulfuric acid. Khotinsky and Patzewitch (2) obtained a similar result by refluxing the crude carbinol in glacial acetic acid. Much of the early research on 2-biphenylylmethanols was concerned with the attempt to convert the carbinols to the corresponding chlorides with HCl; the result was always the formation of the cyclic product. For example, Sergeev (5) in 1929 found that his attempt to convert the carbinol (I) to the chloride resulted only in the formation of 4-benzhydryl-9,9-dipheny1f1uorene (II). This was also one of the first investigations to show the general nature of the cyclization. © oo ©9© > . ROOH O w© R = -COOH, -COC5H5 Clarkson and Gomberg (4) in 1950 first reported the isolation of pure cystalline 2-pheny1tripheny1methano1. Again, any attempt to form the chloride always led to a quantitative yield of 9,9-diphenylfluorene. Their work showed that acid caused cyclization even when present only in catalytic amounts. In their attempt to produce a stable 2-biphenylyl- carbinyl chloride, Clarkson and Gomberg synthesized 9-(2- biphenylyl)-Q—fluorenol (III) which was inert to refluxing acetic acid. But upon the addition of just one drop of HCl, the carbinol quantitatively cyclized to the 9,9-spiro- fluorene (IV). Go. 00 HCl >\ g ©‘© HOAC / 0.0 III IV This was an easy route to many Spiro compounds, for example spiro-9-f1uorene—9'—xanthene (VI). Qoo o O o O > O 0 C) V VI In the following years Bergmann and Bondi (5) reported a variety of cyclizations such as HCl \\ H3 HOAC / ©.© This work was confirmed by Dice et a1. (6) in 1950. Among the more interesting examples of this type of reaction which have been reported are the following: R R R ”R OH R R R . ‘ n R R R © Q R , R = Phenyl Kohler and Blanchard (7) in 1955 9 e t 9 :25) , C...) Koelsch (8) in 1954 and Hatt (9) in 1941 to \ ©0163 / CH3 Wittig and Fuhrman (10) in 1940 : 9H HCl \ R R , R HOAc7 . 0 ‘R and CH3 H3 CH3 HCl \\ . HOAC/ ©‘O 0H Anchel and Blatt (11) in 1941 R = COOEt Ray et al. in 1950 (12) HCl \\ HOAc 7 Suzuki (15) in 1951 x0 :3ij C99 ©© > ©fi© Gilman and Gorsich (14) in 1956 X II Cl,Br The cyclization is not limited to the formation of fluorene ring systems, but can also lead to six-membered rings as in Q, .Q o‘fl on where X = -CH2-(15,16,17,18), -0-(4) and -S-(19). Although this cyclodehydration reaction has been known for sixty years, all of the reported work was syn- thetic in nature. No kinetic study of the reaction has been reported, nor has there been any systematic study of the effect of structural variations in the carbinol, or reaction conditions (i.e., acidity) on the ease with which cyclization occurs. It was the purpose of this investigation to examine the kinetics and mechanism of this cyclization. The reaction can be looked upon as a two—stage process. The first involves formation of a carbonium ion, by protonation and water-loss from the carbinol. This stage should be influenced by the same kinds of effects which are characteristic of 8N1 solvolytic processes. The second stage of the reaction is presumably an intramolecular electrophilic aromatic substitution. One of the goals of this research was to differentiate these two processes, and to determine whether substituents would affect their relative importance. Accordingly, compounds of the type © C) were synthesized. Suitable conditions were found for carrying Y out the cyclization at measurable rates. The reaction kinetics were followed Spectrophotometrically, and solvent deuterium isotope effects were also studied. These results are incorporated in an overall reaction mechanism which fits the data. RESULTS AND DI SCUS SI ON I. Synthesis A. The Starting Carbinols (2-Phenyltriphenylmethanols) The general method for preparation of the 2-phenyltri- phenylmethanols was the same as first reported (1,2), that is, by the addition of an aryl Grignard reagent to a benzophenone: © 9 + Mgsr O PH zoo Rina? © g ‘ O EHQX :FgBr + {g KEY—90 © The carbinols were obtained as oils, mixed with starting material and side products, and were difficult to purify. Clarkson and Gomberg (4) reported that steam distillation of crude 2-phenyltriphenylmethanol, using a trace of sodium carbonate to neutralize any acid, removed unreacted starting material and contaminants. This technique was not successful with the substituted carbinols. Instead, liquid-solid column elution chromatography on silica gel or alumina was employed. To facilitate the separation, the reaction was run in such a manner as to keep the amount of unreacted ketone at a mini- mum, because the carbinol and ketone had similar Rf values. 10 I More complete conversion of ketone to carbinol was accomplished by using a two molar excess of Grignard reagent, tetrahydro- furan as solvent, and a reflux time of between twenty-four and forty-eight hours. The amount of unreacted ketone was reduced to 5% or less, and the purification was simplified. The separation and isolation of pure carbinol in most cases required more than one pass through the column. The pure carbinols were difficult to crystallize. However, all except 5-trifluoromethyl- and 5—methy1-2'—phenyltriphenyl— methanol were obtained as crystalline materials. B. The Reaction Products (9,9-Disubstituted Fluorenes) In order to follow the reaction kinetics, a spectrum of pure fluorene hydrocarbon was needed to establish the position of Amax and the extinction coefficient. The pro— cedure for preparing the fluorenes from the carbinols was simple. A refluxing solution of carbinol in acetic acid was treated with a few drops of sulfuric or hydrochloric acids and refluxed for two minutes. Upon cooling a quantitative yield of crystalline hydrocarbon was obtained. The hydro- carbon contained no detectable contaminants (t.l.c.). C. Product Study of the Cyclodehydration It was necessary to establish with certainty that the products of cyclodehydration were in fact the 9,9-diaryl- fluorenes, as anticipated. It is conceivable that fluorenes might be formed by a route different from the expected 11 cyclodehydration. For example, Patai and Dayagi (20) showed that the triarylcarbonium ion VII formed the fluorene VIII almost quantitatively after one hour in refluxing acetic acid. VII VIII This is a general reaction of triarylcarbonium ions and could provide an alternate path for the cyclization. The following scheme indicates the possible products from this type of reaction, using the para-disubstituted carbinols IX as an example. X is the expected product and XI and XII are possible products from cyclizations of the type described by Patai (20). :5st do o© IX x X oo __> 0o o \ one XI 12 By hydride abstraction from intermediate carbonium ions, it is also possible that spiro products of the type XIII could arise. Structures XI and XII were eliminated in favor of X on the following spectral data. The infrared spectrum of the product showed a. the absence of an 890 cm‘1 band characteristic for an isolated aromatic hydrogen (21) b. the absence of bands at 700 and 750 cm'1 character- istic for a monosubstituted phenyl ring (21). The n.m.r. Spectra also eliminated structures XI and XII since a resonance for the 9-hydrogen was not observed (9-phenylfluorene, 9-hydrogen at T = 5). Structure XIII was also eliminated in favor of X since a. the n.m.r. proton integration showed two more hydrogens than would be expected of XIII b. the carbon-hydrogen analysis consistently fit X better than XIII for a variety of X's and Y's. Similar arguments can be brought to bear for the products from monosubstituted carbinols, again eliminating structures corresponding to XI, XII and XIII. It should also 15 be pointed out that the reaction time for formation of such products is generally appreciably longer than is required for the normal cyclodehydrations possible when a 2-phenyl sub- stituent is present in the carbonium ion. II. Kinetics A. Rates and Correlations The apparatus, techniques and equations used to de- termine the kinetics are described in detail in the Experi- mental section of this work. The kinetics were run in solvent consisting of 4% by weight of sulfuric acid in 80% aqueous acetic acid. This solvent was similar to those used previously to affect cycli— zation, but the use of aqueous rather than glacial acetic acid decreased the reaction rates enough to allow them to be measured accurately. Progress of the reaction was followed by recording the appearance of a band at about 508 mu due to product. The starting carbinol had no absorption in this region and cause no interference. Pseudo first order rate constants were obtained for the appearance of product with time. The following expression was used to calculate the rate constants: (1) Ah kt = 2.505 log W h h 00 where Ah and Ah were the absorbances of the product at infinite time and any time t, respectively. 14 The reactions were shown to go to completion without formation of side products by comparison of the products ob— tained from kinetic runs with authentic samples of the pure hydrocarbons. The residue from kinetic studies of compounds XIV, XIX and XXII, after removal of solvent, showed only one compound by t.l.c. analysis. The Rf values and infrared Spectra of the residues were identical to those of the corres— ponding pure hydrocarbons. In all cases, the extinction coefficient of the kinetic product at 508 mu was compared to the extinction coefficient of the corresponding pure hydro- carbon. The values checked within 5%. All the reactions were shown to obey first order kinetics by a straight line plot of log [ASD/(Agé — Ah)] versus time, for at least 88% reaction (see Figure 1 for an example). The reaction rates were also independent of the initial carbinol concentration (see Table I for examples). Table I. Effect of Varying Initial Carbinol Concentration on Reaction Rates at 25 and 4% H2804 Concentration x 105 Rate Constant x Carbinol (moles/liter) 104 (sec-l) 4-Chloro-2'-phenyl- triphenylmethanol (XVII) 7.05 7.96 i 0.04 14.10 8.06 i 0.05 25.00 7.78 i 0.02 4-Methyl-2'-phenyl- 5.50 87.6 . triphenylmethanol (XXI) 5.71 86.8 9.55 86.6 l+|+l+ 000 WWW 15 oo A Figure 1. Plot of 2.505 log -E?——-— versus Time for Ah ‘ Ah 4—Chloro-2'-phenyltriphenylmethanol at 250. 0'8 " k = slope = 7.86 x 10-éseC-l 120 600 1200 1900 2400 Time (secs.) ] ) ..Ah oo h /(A 00 h 2.505 log [A 16 Agreement of rate constants calculated at various times dur- ing the reaction (Table II) and the similarity of the Powell treatment plot (22) with standard plots also substantiate first order kinetics. When a paramethoxy substituent was present in the car- binol, the solutions were colored red immediately upon adding the acidic reaction medium to the carbinol. This color re- sulted in a slight tailing from a maximum in the visible region which interferred with the absorption maximum at 508 mu, and the following revised kinetic expression was used to calculate the rate constants: - O C+ (eh Qt) (I) ~(Ah " Ah) kt = 2.505 log where c: is the initial concentration of carbonium ion, eh and 6t are molar extinction coefficients at 508 mu of the product and the carbonium ion, respectively; and AEJand Ah are the absorbances of product at infinity and at any time t, respectively. First order kinetics are shown to be followed, because straight line plots of log [1/(AEO- Ah)] versus time were obtained for greater than 80% reaction (see Figure 2 for an example). The reaction rate was independent of varying concentrations of starting carbinol. The good agreement of the calculated rate constants at various times during the reaction (see Table III for an example) also supported the first order kinetics. Table II. zation at 25 17 Examples of galculated Rate Constants for Cycli- 2-Phenyltriphenylmethanol 4,4'-Dichloro-2"—phenyltriphenyl- methanol kr (x 103) Time (secs.) kr (x 104) Time (secs.) 2.58 500 2.60 1020 2.59 560 2.56 1620 2.61 420 2.55 2220 2.59 480 2.54 2820 2.58 540 2.54 5120 2.52 600 2.54 5720 2.58 660 2.55 4520 2.60 720 2.55 4920 2.62 780 2.54 5520 2.62 840 2.55 6120 [2.58 900 2.57 6720 2.62 960 2.58 1080 2.64 1200 Average kr = 2.59.: 0.05 Average kr = 2.54 i.0.01 x 10"4 x 10"3 sec‘1 sec‘l Calculated from least squares 2.55 i 0.04 x 10"4 sec":L Calculated from least squares 2.60.1 0.04 -3 - x 10 sec 1 18 Figure 2. Plot of log oo 1 .] versus Time for (Ah Ah) Cyclization ofO4-Methoxy-2‘-pheny1triphenyl- methanol at 20 C. CD h-Ah)] L09 [ 1/(A l l l I l l l l l l l 240 480 720 960 1200 Time (secs.) Table III. Calculated Rate Constant for 4-Methoxy-2'-phenyl- triphenylmethanol at 25 Time (secs.) k25(sec‘l) x 103 60 1 . 48‘“ 180 1.47 560 1.49 480 1.50 600 1.51 780 1.47 Average kr = 1.49 $.0.01 x 10‘3 sec-1 Calculated by least squares 1.47 i.0.01 x 10‘3 sec‘1 19 "The para-methoxy and di-para-methyl substrates formed red solutions attributed to carbonium ions; the red color faded as the cyclization reaction proceeded. The reaction rates could now be followed at two wavelengths: at 508 mu for product formation and in the visible region, near 470 mu, for the disappearance of carbonium ion. The results are tabulated in Table IV. Within experimental error, the rates of product formation and carbonium ion disappearance are identical. The wavelengths and molar extinction coefficients oftflmacarbonium ion solutions compare very well (see Table IV) to the values reported for similarly substituted tri— phenylmethanols (25). Table V lists the compounds whose rates of cyclization were measured, in order of increasing rate constant with the exception of those with a para-methoxy substituent. The rate constants, with their average deviations, for all of the compounds at all of the temperatures investi- gated are tabulated in Table VI; the rate constants are the average values of at least three experiments unless other- wise noted. 1 As Table VI illustrates, the rates of cyclization are influenced markedly by substituents. With the exception of the para-methoxycarbinols, the rates are generally increased by electron-donating groups and are decreased by electron- withdrawing groups. When the log of these rate:constants is . + plotted aga1nst Hammett's 0 values or Brown's 6 values (24), 20 Table IV. Comparison of Rates of Formation of Fluorenes and Disappearance of Carbonium ions for Compounds XXII. XXIII. and XXIV XXII XXIII XXIV Rate Constant for Disap—V 1.59 x 10'2 1.45 x 10‘3 4.72 x 10-4 pearance of red peak, (25°, sec‘l) Rate Constant for Appear- 1.55 x 10"2 1.42 X 10"3 4.67 x 10"4 ancg of Product (25 , sec'l) kmax (é) of red peak 458 mu 474 mu 487 mu (5.5 x 104) (5.5 x 104) (7.9 x 104) xmax (e) for similar 456 mp 476 mu 482 mu 4 4 4 substituted triphenyl- (5°50 X 10 ) (5.85 x 10 ) (8.51 x 10 ) carbinol (25) Table V. The Substituted Carbinols Used in the Cyclo- dehydration Studies Compounds x Y Z XIV H CF3 H XV Cl H Cl XVI H Cl H XVII Cl H H OH XVIII H OCH3 H | Z XIX H H H C XX H CH 3 H XXI CH3 H H XXII CH3 H CH3 XXIII OCH3 H H X XXIV CH3 H OCH3 XXV OCH3 H OCH3 21 uoHQ B\d .m> .H many 03¢ maco mo mmmuw>m I * x 00H Eoum 09Hm> pmumHommuuxm A 0 >.0 HHO>.0 Amd¢.00 *m00.0 Hmmm.0 AONH.00 I >xx Aom.mdv *m0.0 H 00.0 0N.0 H $0.d H0.0 H mm.m A¢0.NV >Hxx >.0 H 0.0m 0.0 H «.wm d.0 H >.¢H «0.0 H N0.m m0.0 H 00.m HHHXX Ammmv Ammmv m H mmd 5.0 H N.mm «.0 H $.mm HHXN Ammmv *H H HMH *m.0 H n.mm N.0 H m.¢¢ m.0 H n.mm Hxx Amwav m.H H m.em 0.0 H m.mm H.o H m.Hm H.o H n.ma xx 0.d H N.0> 0.H H $.0m 0.0 H $.0m H.0 H m.mfi mm.0 H 00.0 NHN Am.>m0 m.0 H 0.mm N.0 H m.ma No.0 H 00.0 ma.0 H mm.0 HHH>X 0.0 H m.mm *0.0 H N.Nd >H.0 H mm.> 00.0 H om.m mH.0 H hm.a HH>K >m.0 H $0.0 00.0 H 00.m *N0.0 H #0.N No.0 H em.d A0¢0.00 H>x mH.0 H 00.0 m0.0 H mm.¢ ma.0 H mm.N no.0 H 0N.H Ammm.00 >N *H.0 H. m.md No.0 H mm.m d0.0 H mm.d AHM0.00 Aemm.00 >Hx own on 0mm oom omH essomfioo COHHMH>00 mmmum>< paw AvOH x .atommv ucmumcoo mumm ammum>¢ vommm SH CH maoamnumfiamamcmsmHmlm mo GOHHMNHHUMU mo mmumm .H> manna 22 using the sums of the 6 or 6+ values when there is more than one substituent, linear plots are obtained. These plots, using rate data at several temperatures, are shown in Figures 3-7. The rate data at 350 or 150 are not used for the 6 p or 6+ p+ plots since many of the values for the rate constants were extrapolated values. The discussion of the Hammetto p relationships is based on values obtained at 250, but the discussion applies equally well to the plots at 200 or 500. A glance at Figures 5-7 shows that substrates with a para—methoxy substituent require separate treatment; the following comments are restricted to the other carbinols. At any one temperature, the criterion for determining whether the best substituent constant to use was 6 or 6+ was the comparison of average deviations of the least squares plots of log k versus 6 or 6+. The value of the smallest average deviation was subtracted from the largest deviation and the difference was divided by the smallest deviation. The result was the percentage by which the rate data best correlated with the substituent parameter 6. At the three temperatures the rate data showed a better correlation with 6 than with 6+. At 200 the p value was —2.54 and p+ value was -2.46 with the fit 30% better for 6. At 500 the p value was —2.67, the p+ value was -2.17 and the fit was 520% better for the 6 plot. At 250 the p value was -2.58 and the p+ value was 25 Figure 5. Plot of log kr versus 6 at 200 and 4% H2804. Log k 24 Figure 4. Plot of kr*versus 6 at 250 and 4% H2804. -1.4 4- -1'8 "di-p-CH3 o ‘ p-CH3 -2.2 o r m-CH3 H '* H p =-2.58 -206 r Om‘OCHg Gp-OCHa -5.0 - p—Cl 0 -CH - '-OCH _5.4 “ P 3 P 3 m-Cl p = +6 ' . di-p— -3.8 u .m-C4§\\ -4.2 o o di‘p‘OCHa -.6 -.'5 -.'4. -.'5 -.'2 -.'1 6 I1 :2 :5 I4 .5 25 Figure 5. Plot of krversus 6+ at 250 and 4% H2804. -1.4 . P = —1.92 ‘1°8‘ di-p-CH3* ‘p‘CHs -2.2 ' m-CH3 ‘ m-OCH3 LI O p-OCHa M 3 0 4+ § . o p-Cl P = +2.85 . p-CH -p -OCH -5.4- 3 3 di-p-Cl ' 1 m-Cl —308 1b TIP-CPS ‘ ‘402 IL ‘4'6 “ ‘0 di-p-OCHS -1.8-116—1Z4-122-1Io Log k 26 Figure 6. Plot of log kr versus 6 at 500 and 4% H2304. --l.6 4- ' di‘p-CHg p = -2.67 P‘CHs —2.0 + m-CH3’ f m-OCH3 H -204 «L ‘ p-OCH3 ~208 Jr a p-Cl ‘ p-CHg‘p I ‘0CH3 -5'2 ' m-Cl ' 'di-p-Cl -506 J» III-CH3 ‘ -.5 -04 ‘05 -02 “.1 6 6 Figure 7. Log k 27 Plot of log kr versus 6+ at 500 and 4% H2504. -2.l7 di-p-CHa ! 0’ p-CH3 m-CH3° .m-OCH3 'H p = +2.92 . p-OCH3 + o-Cl o P'CHS‘P"OCH3 oITl-Cl di-p-Cl' m-CF3 i l A l A L I . . l . . . - . . . . I . . -l.4-1.2-1.0-.8 -.6 -.4 -.2 O .2 .3 .4 6+ 28 -1.92. The fit with 6 was 75% better. The correlation coefficient (25) of the fit of points to the line at 250 was 0.995, indicating an excellent correlation of the rate data with the 6 substituent parameter. The three substrates with a para-methoxy substituent did not even come close to the correlation line for the other compounds. Since the reaction had a negative p, a large rate enhancement might have been expected for carbinols with para-methoxy substituents. These substituents, however, actually retarded the rate of product formation. This re— tardation might have been the result of protonation of the ether oxygen. A protonated methoxy substituent would be expected to be electron-withdrawing, since the oxygen electrons would be tied up by the proton. The oxygen would now be electron deficient and would not be able to stabilize the carbonium ion intermediate. However, the meta-methoxy compound, XVIII, which should show the same effects of protonation, was found to lie on the 6 p plot in the correct position. This was evidence that protonation of the oxygen was not the cause of rate retardation. The Hammett plots, however, do show that the para— methoxy compounds lie on a separate line with a lepe (p) opposite in sign from the other compounds. NOW‘the corre- lation of rate data was found to be best when 6+ rather than 6 was used as the substituent parameter. At 250 the fit was 240% better using 6+ rather than 6 with a p+ value of 29 +2.85 and a p value of +5.85. At 50° the fit was 120% better, again for 6+ compared to 6, with a p+ value of +2.92 versus a p value of +4.65. Thermodynamic parameters were calculated from the rate data using an Arrhenius plot of log kr versus l/T (see Figure 8), the resulting slope being equal to Ea/2.505R (26). The enthalpies (AH*) and entropies (A81) of activation were calcu— lated from a plot of log kr/T versus 1/T using the equation k t i r _ -AH AS .5 ln(T ) ' RT + R + 1“ (h)] where K = Boltzmannb constant and h = Planck's constant. The slope was -AH¢/2.503 R and the intercept was ¢ AS K 2.503R + log (h) (27). $ ¢ The Ea’ AH and AS values are listed in Table VII. The activation energies show that cyclodehydration, regard- less of rate-determining step, proceeds a little easier for substrates with electron-donating substituents than for those with electron-withdrawing substituents. The AS: values divide the substrates into two groups: most of the compounds have small positive values (-1 to +5 e.u.), but the di-para- methyl and para-methoxy substituted carbinols showed large negative values of AS: (—5 to -17 e.u.). This can be inter- preted as showing that the rate-determining step for most of the carbinols has a transition state in which there are approximately the same, or perhaps slightly fewer restrictions to molecular motion than there are in the ground state. 30 Figure 8. Arrhenius Plot for the Cyclodehydration. XXII XXI Log kr XIX XVIII ‘XXIII XVII 'XXIV XVI XV XIV L I 3.25 3.30 3.35 3.40 3.45 3.50 l/T x 1000 31 Table VII. Tabulation of Ea' AH*, and AS: for the Cyclo- dehydration Compound Ea(Kcal/mole) AH*(Kcal/mole) AS*(e.u.) XIV 21.25 22.75 4.37 XV 23.67 22.83 1.60 XVI 22.79 21.72 2.68 XVII 21.27 20.11 -0.81 XVIII 25.27 21.09 4.30 XIX 19.24 18.88 -2.64 XX 23.97 19.80 1.40 XXI 21.73 18.62 -1.05 XXII 15.39 14.90 -12.26 XXIII 18.30 14.96 -16.70 XXIV 18.30 18.30 -7.78 XXV 21.80 20.7 -4.9 However, reactions of carbinols with para-methoxy sub- stituents involve a rate-determining step requiring a large amount of restriction of motion or orientation in the transition state. ¢ ¢ # Plots of AH versus AS and AH (see Figure 9) versus 6 (see Figure 10) were made. Figure 9 shows that AH: and AS* vary fairly linearly with each other; therefore the 6 p correlations should also be fairly linear. The best approxi- mation of a straight line drawn through the points gave an approximate isokinetic temperature (5) of 4000K. Since 8 was considerably greater than the reaction temperatures, the isokinetic relationships should hold and the 6 p relationships are in the correct order (28). ‘ ::.,~.; 32 AoSomvfimq b 1 fll m- m- 0H: NH- HH- wan ma- A» u m a m e m- a mm00um . mH ma .coHumufiwzmpoHuwo How H m4 momum> #34 mo HoHa «N musmHm (SION/I 93X ) iHv 33 .mmoou . glamoum . 3 .0H tam coHumupmcapoHomu How 0 msmum> H 34 Ho uoHa flm .OH magmam (SI 0111/]: ED){) iHV 34 In Figure 10 the rough proportionality between AH: and 6 indicates that the Hammett relationship is valid (28). B. Reaction Mechanism Before discussing the mechanism of the cyclodehydration reaction, it will be well to summarize the pertinent experi- mental results which have just been presented in some detail. These results must be accommodated by any mechanism which would be proposed. It was established for all the substituted 2-phenyl- triphenylmethanols, that the acid-catalyzed cyclodehydration obeyed pseudo first order kinetics. In general, cyclization was favored by electron-donating substituents. A Hammett plot showed a good correlation of the reaction rates with 6 and had a negative p. The para-methoxy substituted car- binols, however, were an exception. Here the reaction was favored by electron withdrawal. The Hammett correlation fit best using 6+ values, with a resulting positive p+. It ap- peared that the cyclization was proceeding by two different mechanisms. I The para-methoxy carbinols formed red solutions im- mediately, attributed to rapid and complete ionization to the carbonium ion. Product formation and disappearance of the red color were found to proceed at identical rates. One compound, the di-para—methyl carbinol, may belong to either family of reactions. With acid, it immediately formed a red solution, behavior similar to that of the 35 para-methoxy carbinols. Yet its rate fell in an ambiguous position on the Hammett plots, fitting reasonably well with either 6 (1p) or the 6+ (+p+) correlation. It reacted like the para-methoxy carbinols in that the rate of product formation and the disappearance of color (carbonium ion) were identical. The rate of reaction was unfortunately so rapid that initial points were taken after the reaction was 45% complete. Thus it was not clear whether all of the carbinol was instantly converted to carbonium ion, or whether this had happened more slowly, but was complete by the time the first kinetic point was taken. It is likely that this carbinol reacted mainly by the same mechanism as the para-methoxy carbinols, but it is possible that the rate of ring closure was not much slower (perhaps one power of ten) than the rate of carbonium ion formation. As a rule, the activation energies were larger for carbinols with electron-withdrawing substituents, indicat- ing inhibition to cyclodehydration due to electron with- drawal from the reaction site. The ASat values were small (—1 to +5 e.u.) for carbinols on the (-p) correlation line, but moderately large and negative (~ESto -17 e.u.) for carbinols with di-para-methyl and para-methoxy substituents. The results of the kinetic study presented thus far indicate that cyclodehydration of 2-phenyltriphenylmethanols proceeds by two different reaction paths, depending upon substitution. The Hammett relationship showed that the 36 process of product formation changed with substituents. Leffler and Grunwald (29) have given examples of Hammett plots with abrupt breaks in linearity. If the break in the curve is concave up, a change in mechanism is always the cause. However, if the break is concave down, a change in rate-determining step with an otherwise constant mechanism is the cause. An excellent example of a reaction which shows an abrupt change in the slope of the 6 p plot caused by an abrupt mechanism change is the hydrolysis of benzoate esters in 99.9% sulfuric acid (30). The hydrolysis of the methyl esters obeys a 6 p relationship with the rate de- creasing as the substituents become more electron—attracting. This is what is expected for the acyl-oxygen fission. For ethyl esters, the same relationship is observed for electron- attracting substituents, but the strongly electron—withdraw- ing groups now cause an abrupt shift in mechanism. Alkyl— oxygen fission now predominates and the rate decreases as the substituents become more electron-withdrawing. The 6 p plot in Figure 11 shows this abrupt change. An example of an abrupt break in a 6 p plot with a constant mechanism is the formation of semicarbazones from substituted benzaldehydes (31). The reaction has a two- step mechanism, either step of which can be rate-limiting: O H p + Ar-'C'-H+R—NH2—kl>Ar-c-NHRH—>Ar-c=N-R+H;o E1- H RB 1'4 The rate-limiting step is greatly dependent upon the acidity. 37 Figure 11. Hammett Plot for Hydrolysis of Benzoate Esters in 99.9% H2304 (29). o Ethyl Esters 0 Methyl Esters 0.7 — _ .54 01-0-7 '— 1 0 +4 0 -2.1 h - l I 1 0.0 0.7 1.4 6 In acid solutions (pH 1.75), kl is rate-determining and in neutral solutions k2 is rate-determining. In both cases substituents did not have a pronounced effect on rates. However, at pH 3.9, a change in the nature of the sub- stituents causes a change from k; rate-determining to k2 rate-determining. The plot (see Figure 12) shows the abrupt break in the Hammett plot due to change in rate- determining step. Since the observed break in the 6 p plots for the cyclodehydration was concave down (see Figure 4), a change in the rate-determining step is indicated. The following general mechanistic scheme seemed plausible, as a working hypothesis: 38 Figure 12. The Hammett Plot for the Formation of Semi- carbazones from Substituted Benzaldehydes at pH 3.9 (29). I T l 0.0- m B-O.5- '— o ,M m 0 O A _1.0 I l l -0.4 0.0 0.4 6 + ROH + H30 fit ..R0H2+ + H20 . (1) k2 ROH2+ <—:i> R+ + H20 (2) 4 + R —£5> Product + H+ (3) The carbinol is first protonated in a reversible step. The protonated carbinol can now reversibly lose water to form a carbonium ion, which then undergoes ring closure and proton loss to give the product in an irreversible step. Using a steady state treatment (see Experimental), the following expression for appearance of product was obtained: d (Product) = k; [H+] lROH] dt 1 + Jig [1 + k5 [H20]J k3 k5 ~ 39 If the mechanism is correct, then the rate, although dependent upon initial acid concentration, should not exhibit such dependence during the course of a given run, since the proton consumed in step (1) is regenerated in step (3). Also since the concentration of water is essentially constant (solvent 20% H20), the equation reduces to the pseudo first order equation I _ + d (Product) = kl [ROH] k1 ' kl [H 1 dt k k; . 1 + ESL (1 + k5 k4 = k4 [H20] Since with most of the carbinols studied no color due to build-up in the concentration of carbonium ion was seen, it may be assumed that for these cases k5 is large and not rate-determining. If k5 and k3 > > k1, the protonation step is rate-determining. On the other hand, if the equilibrium in step (1) is rapidly established and if kg and k4 are much larger than k3, then the ionization step (i.e. k3) may be rate-determining. In either case the simplified rate ex- pression d (PrSEUCt) = kr[ROH] k = conglomoration of r constants results and either assumption would fit the kinetic data. For para-methoxy carbinols, the formation of R+ was rapid and complete. This conclusion is justified by the observation of immediate color formation and by agreement of the extinction coefficients [when the data are extrapolated back to zero time] with literature values for similar 40 carbonium ions. In these cases the rate-limiting step is step (30 and the rate of product formation is defined by the expression d (Product) = + + and if [R ] [ROH] d (Product) This result was previously deduced from the data in Table IV, where the rates of product formation and disappearance of carbonium ion were shown to be equal. For the para-methoxy compounds, therefore, the rate-determining step must be the electrophilic aromatic substitution reaction (k5). The rate-limiting step (k; or kg) for the majority of compounds could not be obtained from the knowledge of the kinetic order of the reaction, and further variables were studied. A change in initial acid strength was found to affect the reaction rates. For the compounds which fit the nega- tive p correlation line, a rate enhancement of nearly two- fold was found for each 1% increase in sulfuric acid strength in going from 4% to 6% (see Table VIII). These data however, still did not allow an unambiguous assignment of the rate- limiting step, although it did support some of the mechanistic conclusions already reached. If proton transfer were limiting, 41 Table VIII. Rate Constants for the Cyclohydration in 4 to 6% H2504 kr (x 104) in % H2304 Compound 4% 5% 6% XIV 1.16 i 0.01 1.85 i 0.03 2.90 i 0.05 xv 2.55 1 0.12 4.16 i 0.05 6.58 i 0.02 XVI 2.66 i 0.02 4.55 i 0.06 6.75 i 0.17 XVII 7.91 i 0.14 14.40 i 0.26 21.88 I 0.56 XVIII 19.51~: 0.2 51.5 i 0.6 - xxx 26.4 i 0.5 47.6 i 0.8 75.5 i 0.4 the rate should be increased due to the larger concentration of hydronium ion. But, if k3 were rate-limiting, one might still expect an increase in the amount of ROH2+ present due to the mass effect. The observed rate increase would, there— fore, result in either case. However, if our mechanistic conclusions for the para- methoxy compounds are correct (k5 rate-limiting) an increased acid strength should have little effect on the rate for these substituents. Since ionization is complete at low acid strength (4%) a small increase in acid strength could not increase the concentration of R+ or the rate of product formation. The results are shown in Table IX. Although the rate constants at any given acid strength vary by a factor of over 30, it is seen that a change in acid strength has only a slight effect on the rates. The results are in agree- ment with the proposed mechanism for the para-methoxy compounds. 42 Table IX. Rate Constants for the Cyclization of 4,4'- Dimethyl, 4-Methoxy- and 4-Methoxy-4'-Methyl- 2"-phenyltriphenylmethanols in Various % H2804 (sec‘l) % H2S04 Compound 4% 5%' 5% XXII 1.55 x 10'2 1.59 x 10-2 — XXIII 1.42 x 10‘3 1.57 x 10-3 1.48 x 10-3 XXIV 4.67 x 10‘4 4H88 x 10'4 - Another attempt was made to distinguish step (1) from step (2) as the rate-degermining step; this involved a measurement of the solvent deuterium isotope effect. If the proton transfer from the hydronium ion (kl) were rate-limit- ing, an A-SE2 reaction (32), the ratio of kH should 20/kD20 be much greater than 1 as postulated (32,33,34). An unam- biguous example of the A-SEZ reaction where the protonation occurs on an oxygen is not easy to find. It has been reported that the mutarotation of glucose is faster in H20 than D20 (30% faster) (35). However, this effect was attributed to a complicated A-1 mechanism. Later work (34) states that the rate enhancement is actually due to rate-determining protonation by H30+ or D30+, but the interpretation is still questionable. An A-SE2 mechanism is suggested for other cases (32), but not unambiguously proved. Paul and Long (32) state that where nitrogen, sulfur or oxygen act as a base, its protonation should be very rapid and not rate— determining. An A-SE2 reaction should occur almost 43 exclusively where proton transfer to a carbon atom is re— quired. For the cyclodehydration, if the proton transfer were rapid and equilibrium established, followed by slow ioni- zation (k3 rate-determining; an A-1 mechanism) a rate enhancement in the deuterated solvent would be observed, by a factor of two or greater (34,36). An example is the acid; catalyzed hydrolysis of ethylene oxide, an established A-1 mechanism, where kDgO/ngo = 2.20 (37). Another reported example is the hydrolysis of ethyl orthoformate where kDgO/kHZO = 2.35 was found (38). In both examples, the reactions had been previously established as proceeding by A-1 mechanisms, and in both cases a rate enhancement in D20 was observed. To study the effect of deuterated solvent, the solvent system Dgo-CH3COOD-H2804 was chosen. The kinetics were run in 4% H2804 in 80% CH3COOD-D20 where the deuterium enrich- ment of acidic hydrogen was 97.3%. The results are given in Table X. The results show that for the three carbinols which fit the (-)p Hammett plots, a two- to three-fold rate en- hancement resulted. Therefore, the rate-limiting step must be an A-1 process (i.e., kg). The cyclization occurs by rapid protonation followed by slow ionization of the carbon~ oxygen bond of the protonated carbinol, thus eliminating kl as rate-limiting. 44 Table X. Ratesoof Cyclization in 4% H2804 - CH3COOD-D20 at 25 Compound kDEO/kH20 4-Methoxy-2'-pheny1triphenylmethanol (XXIII) 0.94 4,4'-Dimethyl-2"-phenyltriphenylmethanol (XXII) 1.55 4-Methyl-2'-pheny1triphenylmethanol (XXI) 2.88 2-Phenyltriphenylmethanol (XIX) 3.10 4-Chloro-2'—phenyltriphenylmethanol (XVII) 2.91 For the para-methoxy carbinol hardly any change in rate was found when deuterium was substituted for protium in the solvent. This is in agreement with the mechanism in which k5 was rate-limiting. The overall effect when D20 replaces H20 in an acid-catalyzed reaction is the same as if a stronger acid were present (29). Since k3 is essen— tially independent of acid strength the kDgO/ngo ratio of 1 (i.e., no effect) is in agreement with the proposed rate- limiting step. For the di-para—methyl carbinol, the rates were faster in deuterated solvent, but less than predicted if only an A-1 mechanism were operative. The intermediate value of the kDgO/ngo ratio suggests that kg and k5 may have com— parable magnitudes with this carbinol. The Hammett correlations can now be explained in terms of the mechanism. The negative sign of p showed that for substituents which fit this correlation, the rate-limiting transition state is deficient in electrons. If carbonium 45 ion formation were nearly complete in the transition state for step (2) p might be expected to have a value approach- ing that which was observed in the ionization of triphenyl- carbinols, where it was reported to be -3.64 (39). This correlation was for an equilibrium where ionization was complete; this resulted in a large negative p value, showing maximum need for electronic stabilization. Therefore, the 7 somewhat smaller (-) p value for the cyclodehydration (-2.58) supports the postulate that k3 is rate-determining, with only partial development of positive charge in the transition state. The better agreement with 6 rather than 6+ showed that direct resonance interaction with the positive carbon atom was small. This is quite understandable since, of the carbinols on the (-)p plot, only those with a para-methyl group had substituents which might interact directly by resonance. Thus, the plot would not be highly sensitive to the difference between 6 and 6+. It may be that because of the 2-phenyl substituent, the transition state is sufficiently non-planar as to minimize the importance of resonance stabili- zation. For the para-methoxy and the di-para-methyl substituted carbinols, the correlation with 6+ and the positive p are now meaningful. Since k5 is rate-determining the better correlation with 6+ rather than 6 is expected, due to the direct resonance interaction of the substituents with the 46 positive carbon. This same effect has been reported for many systems (40). The ground state is the triarylcarbonium ion, resonance stabilized by para-methoxy or para-methyl substituents. In the transition state, where electrophilic substitution in the 2-phenyl ring has already begun,there must be less posi— tive charge on the carbonium carbon atom. Product + + C H+ Y Thus substituents which stabilize a positive charge should stabilize the ground state more than the transition state. The reaction will therefore be retarded by electron—donating substituents, and the rather large positive p which was ob- served is expected. The activation parameters, particularly the activation entropies, fit well with the proposed mechanisms. Compounds for which step (2) is rate-determining show a small posi— tive AS*. As the protonated carbinol ionizes, the central carbon changes from sp3 to sp2 hybridization. The tran- sition state is thus less congested than the ground state, there is greater freedom for molecular motion, and the re- action has a small favorable entropy change. 47 On the other hand, when step (3) is rate-determining, the transition state must be much more severely restricted in number of degrees of freedom than the ground state. Electrophilic substitution demands that the 2-phenyl ring and the carbonium carbon atom be rather specifically oriented with respect to one another. In particular, the carbonium carbon must attack the 2-phenyl ring in a direction roughly perpendicular to the ring plane, as shown. This must result in a large negative ASi for the reaction, as observed. The activation energies for the carbinols whose rate- determining step is carbonium ion formation increase with increasing electron-withdrawing ability of the substituents, as would be expected. The activation energy for ring closure must be considerably less than that for ionization, since for these substituents k5 > > k3. Even when k5 is rate-determining, the activation energy for this process seems to be lower than that for ionization. As expected, however, it does seem to increase with increasing stability of the carbonium ion. 1 48 To summarize, it appears that only one mechanism is Operative, as outlined on page 38. In all cases, the first step is the rapid and reversible protonation of the carbinol. The rate-limiting step then depends on the nature of the substituent. If the substituent can greatly stabilize the intermediate carbonium ion, then ionization is also rapid and the aromatic electrophilic substitution is rate— determining. But in most cases, ionization is slow and rate- determining, and the ring closure is fast. III. Miscellaneous A. Kinetics of Cyclodehydration of 9-(2-Biphenyl)- 9-fluorenol It has been known for some time (see page 2) that 9-(2-biphenylyl)-9-fluorenol (III) undergoes cyclodehydration to form spirobifluorene (IV). QC 0 0 0-0 0.0 III IV It seemed of interest to determine the effect of "tying back” the two phenyl groups of the 2-phenyltriphenylmethanol (XIX) on the rate and activation parameters of cyclization. 49 9.9. _.. 00 XIX The kinetics for the cyclodehydration of 9-(2-bi- phenyly1)-9-fluorenol were carried out under the same conditions as for the 2-phenyltriphenylmethanols. Analysis of the rate data was slightly different, because the alcohol had an absorption at 308 mu due to the fluorene ring present in the starting material. The absorbance at the time of 00 h absorbance minus the zero time absorbance. The rate con~ mixing was taken as zero absorbance and A was the final stant calculations were then identical with those for the triphenylmethanols. The rates are given in Table XI. Table XI. Rate Constants for the Cyclodehydration of 9-(2-Biphenylyl)-9-fluorenol Compared with Those for 2-Phenyltriphenylmethanol m.—_. 9-(2-biphenyl-9- 2-phenyltriphenyl- Temperature fluorenol rate con- methanol rate con- stant (x 103) sec-1 stant (x 103) sec-1 21° 0.291 i 0.05 1.55 4 0.01 (20°) 25° 0.474 : 0.18 2.64 i 0.06 50° 0.855 t 0.45 5.04 i 0.16 55° 1.45 1 0.40 7.62 i 0.16 50 The rates at each temperature are slower by a factor of about 5 than those for 2-phenyltriphenylmethanol. A comparison of activation parameters is shown in Table XII . Table XII. Comparison of the Activation Parameters for 2-Phenyltriphenylmethanol and 9-(2-Biphenylyl)- 9-fluorenol Ea(Kcal/mole) AH¢(Kcal/mole) AS*(e.u.) 2-Phenyltriphenyl- methanol 19.24 18.88 -2.64 9-(2-Biphenylyl)-9- fluorenol 20.88 20.79 -4.06 There was no evidence (as for example, the appearance of a color) during the kinetic runs with III for the for- mation of significant concentrations of the intermediate carbonium ion. It seems reasonable to assume, therefore, that the rate-determining step, as with XIX, is the formation of the carbonium ion, and that this is followed by a rapid cyclization step. The slower rate and higher activation energy, then, must be a reflection of the greater difficulty in forming the carbonium ion. This is reasonable because in the carbonium from III, one places a positive charge on a cyclo- pentadiene ring; this is not in accord with HUckel's rule, which indicates that a negative charge in such a position is 51 very much better than a positive charge. For comparison, one may note the differences in pKR+'s of 9-phenyl-9- fluorenol pKR+ = -10.27 (23) and triphenylcarbinol pKR+ = —6.65 (23). B. Attempted Cyclization of 1,9-Diphenyl-9-fluorenol to 12-Phenylindeno [1,2,3 - jk] fluorene Rapoport and Smolinsky (41) reported the preparation of fluoradene(indeno[1,2,3 - jk] fluorene; XXVI) from 9-(2-aminophenyl) fluorene (XXVII). XXVI XXVII The hydrocarbon (XXVI) is unusual in that it is a rather strong acid for a hydrocarbon, having a pKa of 11.1 0.5. There are no reports of a good synthetic route for preparing 12-substituted fluoradenes, and in fact the above synthesis is a multi-step process which has discouraged the investi- gation of the chemistry of this interesting ring system. It was found that the cyclization of 2-phenyltriphenyl— methanols occur with great ease. It seemed desirable to try to use the cyclodehydration reaction in an attempt to pre- pare 12-phenylfluoradene (XXIX) starting with 1,9-diphenyl- 9-fluorenol (XXVIII). 52 XXVIII XXIX The substrate (XXVIII) had the same feature as the 2-phenyl- triphenylmethanols, being an ortho-phenyl triarylcarbinol. Although the carbinol (XXVIII) was not reported in the literature, it was easily obtained from a rather straight- forward reaction, the addition of a phenyl Grignard reagent to 1-phenyl-9-fluorenone. The fluorenone was a known com- pound (42) and was obtained in a low yield from fluoranthene in a three step synthesis (see Experimental): CrO , Na 0 CH3008H> 820% Q‘O 2 Benzene 1> Reflux " When the cyclization was attempted in 85% H2804, a green-brown solution was formed; work-up led to recovery 53 of starting material or sulfonated product, depending on the time of contact with the acid. When the cyclization was attempted in acetic acid using 4% H2804 as the catalyst for two days or for six months, only starting alcohol was re- covered. No evidence of the cyclized product was found. It appears that the steric requirement in this cyclization is too demanding for the conditions under which the reaction was carried out. C. Sulfonation of Triarylcarbonium Ions It was reported by Hart and Sulzberg (43) that when 4—phenyltriphenylmethanol was dissolved in concentrated sulfuric acid a red solution was obtained, the color of which was attributed to the corresponding carbonium ion. However, this solution turned from red to orange in less than 1 minute. The orange color was ascribed to the for- mation of a sulfonated carbonium ion. When the orange solu- tion was decomposed with ice—water, only a water-soluble gummy red mass was isolated; no starting carbinol was recovered. The red gum was shown to be sulfonated by its infrared spectrum. Identical results were obtained when (9-(4-biphenylyl)-9-f1uorenol was treated under the same conditions. Preliminary kinetic data were obtained for the sulfonation in 93.9% H2804, but the position of the sul- fonation was not determined. Sulfonation was presumed to occur in the biphenyl ring, most likely in the 4'-position. 54 The object of the extension of these experiments was first to determine the positidn of the sulfonation and secondly to study the kinetics more extensively. The sul- fonation was carried out as reported and the red gummy mass was obtained. Attempts to isolate the sulfonic acid were fruitless. Liquid-liquid extractions of the water solutions of the acid with ether or benzene yielded the same red gum. An attempt to isolate the sulfonic acid as the barium salt was attempted, but the separation was complicated by the presence of a large amount of barium sulfate, presumably from sulfuric acid trapped in the gummy mass. The experiment was repeated, using concentrated nitric acid instead of sulfuric acid. Again a red solution formed first and gradually became orange in about 10 hours. When the orange solution was poured onto ice, a yellow precipitate was quantitatively formed. The melting point of the yellow nitro compound (1710) agreed with the reported value for (4'-nitro-4-biphenylyl)diphenylmethanol (44). In nitration as in sulfonation, the carbinol first forms the carbonium ion. Both acid solutions were dark red, with an absorbance maximum at 510 mu. The orange solutions both showed absorption maxima at 476 mu with almost identical extinction coefficients. There seems to be little doubt that the red solutions contained the carbonium ion, and the orange solutions contained the nitrated or sulfonated car- bonium ion. 55 The position of electrophilic attack in nitration of the carbonium ion appeared to be the 4'-position. Sulfona- tion is probably similar, but this will require further experimentation to verify. EXPERIMENTAL 56 57 I. General Spectra The infrared spectra were obtained on a Unicam SP200 spectrophotometer or a Perkin-Elmer 237B Grating Spectro— photometer. The ultraviolet Spectra were obtained on a Beckman DB Spectrophotometer equipped with a Sargent Model SRL Recorder. The n.m.r. spectra were taken on a Varian Model A-60 instrument. All n.m.r. spectra were obtained at 60 MC in carbon tetrachloride, using tetramethylsilane as an internal standard. The resonance positions are reported in T units and the relative peak areas were obtained by electronic integration. Microanalysis All of the microanalytical data were obtained from the Spang Microanalytical Laboratory, Ann Arbor, Michigan. Melting Points All melting points were obtained on a Eimer and Amend melting point block and are uncorrected. Kinetics The kinetics were obtained on a Beckman DB Spectro- photometer with a thermostated cell compartment (see page 86 for details). Thin Layer Chromatography The chromatograms were obtained on 1" x 3" glass 58 microscope slides, coated with silica gel H (E. Merck Ag. Darmstadt, Germany; Brinkman Instrument Inc., Westbury, Long Island, New York) applied by dipping the slides in a chloroform slurry of silica gel (15 g. in 50 ml. of chloro- form). The visualization was carried out by using a Spray solution made from 1 ml. of a 5% solution of p-anisaldehyde in 95% ethanol diluted to 10 ml. with 95% ethanol followed by the addition of 10 drops of concentrated sulfuric acid. The plates were then heated at 1200 to produce the colored spots. Computer Analysis of Rate Data Rate constants were obtained by linear least square analysis. The linear least squares equation and the rate equations (pages 91 and 94) were programmed, using Fortran notation, and the data were processed on a Control Data 3600 computer. II. Syntheses A. Preparation of 2-Phenylbenzophenone 1. Preparation of 2-Phenylbenzoic Acid The carboxylic acid was prepared as reported (45) using 100 g. (0.565 mole) of fluorenone, 250 g. of powdered potassium hydroxide, and 700 ml. of toluene. The yield was 110 g. (98.5%), m.p. 114-115°. 2. Preparation of 2-Carboxamidobiphenyl The method of Hénigschmid (46) was used to prepare 59 the crude acid chloride from 110 g. (0.555 mole) of biphenyl- 2-carboxylic acid and 250 ml. of thionyl chloride. After stripping off the excess thionyl chloride under reduced pressure, the crude oil was added to 200 ml. of ice-cold concentrated ammonium hydroxide. The yield of crude amide was 55 g. (51%), m.p. 1750. 3. Preparation of 2-Cyanobiphenyl The nitrile was prepared according to the procedure of Goldschmidt and Veer (47), using 55 g. (0.279 mole) of crude dry 2-carboxamidobiphenyl and 75 ml. of thionyl chloride. Distillation of the crude product, b.p. 1140 at 0.03 mm., yielded 35 g. (70%). m.p. 410. 4. Preparation of 2-Phenylbenzophenone (Method A) (48) In a three—necked one-liter round-bottomed flask fitted with true-bore stirrer, reflux condenser, gas inlet tube and a dropping funnel was prepared the Grignard Reagent from 3.5 g. (0.146 mole) of magnesium turnings and 23 g. (0.146 mole) of bromobenzene in 200 ml. of anhydrous ether. To the solution was added dropwise a solution of 23.0 g. (0.128 mole) of 2-cyanobiphenyl in 100 ml. of anhydrous ether. The mixture was then heated at reflux with stirring for four hours. The Grignard complex was treated with 100 ml. of 5% sulfuric acid. To the mixture was then added 100 ml. of glacial acetic acid, 200 ml. of toluene and 25 ml. of 60 concentrated sulfuric acid in 100 ml. of water. The mixture was refluxed for five hours and allowed to cool. The toluene layer was separated, washed with dilute hydrochloric acid, followed by saturated sodium carbonate solution and water. It was then dried over anhydrous magnes- ium sulfate. The toluene was removed, leaving 25 g. of crude ketone. Recrystallization from methanol yielded 20.0 g. (61%), m.p. 87-89° (49). B. Preparation of 2-Phenylbenzophenone (Method B) The preparation was similar to that of Bradsher (49). In a 1-1. round-bottomed three-necked flask fitted with a tru-bore stirrer, dropping funnel, reflux condenser and gas inlet tube was prepared, under a flow of nitrogen, the Grignard reagent from 5.0 g. (0.281 mole) of high purity magnesium (Dominion Magnesium Ltd.) and 47.7 g. (0.205 mole) of 2-bromobiphenyl (b.p. 1000 at 0.5 mm.) in 400 ml. of anhydrous ether, using two drops of 1,2-dibromoethane to initiate the reaction. To the Grignard reagent was added dropwise a solution of 25 ml. of freshly distilled benzal- dehyde (b.p. 1750 at 752 mm.) in 50 ml. of anhydrous benzene. Ether (250 ml.) was removed by distillation, an additional solution of 50 ml. of benzaldehyde in 150 ml. of anhydrous benzene was added and the solution was refluxed for 21 hours. The solution was cooled and treated with 100 ml. of a saturated solution of ammonium chloride. The organic layer was separated, washed with water, dried over anhydrous 61 magnesium sulfate and the ether-benzene removed with a rota- tory evaporator. The dark oily material yielded benzaldehyde and 25 ml. of benzyl benzoate upon distillation (100-1500 at 3 mm.), and the residue was dissolved in hot methanol. Brown crystals (20.0 g.) were isolated, redissolved in hot methanol and treated with Norite. The solution yielded 16.2 g. (30.6%) of 2-phenylbenzophenone, m.p. 88-89.5O (reported m.p. 88-87°). A second reaction with 2.6 g. (0.107 mole) magnesium, 23.3 g. (0.1 mole) of 2-bromobiphenyl and 100 ml. benzalde- hyde yield 4.8 g. (20%) of the ketone, m.p. 87.5-89°). C. Preparation of 2-Phenyltriphenylmethanols 1. Preparation of 2-Phenyltriphenylmethanol In a 300-ml. round-bottomed flask fitted with a tru- bore stirrer, dropping funnel, condenser and nitrogen inlet was prepared the Grignard reagent from 0.6 g. (0.025 mole) of magnesium turnings and 5.0 g. (0.022 mole) of 2-bromobi- phenyl dissolved in 100 ml. of anhydrous ether. To the solution was added 3.9 g. (0.022 mole) of benzophenone in 100 ml. of anhydrous ether and the mixture heated at reflux for 4 hours. The complex was decomposed with 50 ml. of a saturated solution of ammonium chloride. The solvent was removed and the residue steam distilled with a trace of Na2C03. The pot residue was extracted with benzene, dried over anhydrous magnesium sulfate, and the benzene removed on a rotatory 62 evaporator. The residue was dissolved in pentane, yielding 2.1 g. (28%) of 2-phenyltriphenylmethanol, m.p. 91-92O (reported m.p. 87-880) (4). 2. Preparation of 4-Chloro-2'-Pheny1triphenylmethanol In a ZOO-ml, round-bottomed three-necked flask fitted with a reflux condenser, tru-bore stirrer, dropping funnel and gas inlet tube was prepared, in a nitrogen atmosphere, the Grignard reagent from 0.6 g. (0.025 mole) of high purity magnesium and 5.0 g. (0.026 mole) of p-bromochlorobenzene in 50 m1. of anhydrous ether. .To the Grignard reagent was added dropwise 5.0 9. (0.0194 mole) of 2-phenylbenzophenone in 20 ml. of anhydrous ether and 10 ml. of anhydrous benzene. After distillation of 25 ml. of ether, 25 m1. of anhydrous benzene was added, and the red solution was refluxed with stirring for two hours. The solution was hydrolyzed with 50 ml. of a saturated ammonium chloride solution, washed with water, dried over anhydrous magnesium sulfate, and the solvent removed on a rotatory evaporator. Boiling methanol was added to the 4.0 g. of oil remaining and a white precipitate (0.6 g.; m.p. 173-1750 from benzene-petroleum ether) was collected. It was probably 2-phenylbenzopinacol (9). The filtrate yielded 3.0 g. of a clear oil, 1.0 g. of which was chromatographed on 70 g. of silica gel using 3:1 CC14: benzene as the eluant. The first five 50-ml. fractions yielded 0.37 g. of 4-chloro-2'-phenyltriphenylmethanol, 63 m.p. 105.0-106.5O after recrystallization from benzene- petroleum ether. The remaining ten 50-ml. fractions yielded 0.5 g. of a mixture of carbinol, ketone and two unknown components. The infrared Spectrum (Figure 17), the ultra- 95% EtOH max 275 mu (6 = 800). and violet Spectrum, with A absorption 220 mu, and the n.m.r. spectrum (Figure 17) with T 7.28 (1H), T 3.5-2.58 complex multiplet (18H) were con— sistent with the assigned structure. Anal. Calcd. for C25H19C10; C, 81.00; H, 5.16; Cl, 9.56. Found: C, 81.02; H, 5.15; Cl, 9.54. 3. Preparation of 4-Methoxy-2l—phenyltriphenylmethanol In a 250-ml. round-bottomed three-necked flask fitted with a reflux condenser, tru-bore stirrer, dropping funnel and gas inlet tube was prepared, at 00, a solution of n-butyllithium from 0.21 g. (0.03 mole) of lithium, 4.11 g. (0.03’mo1e) of n-butyl bromide and 50 ml. of anhydrous ether. To the solution, maintained at 00, was added 5.61 g. (0.03 mole) of 4-bromoanisole in 35 ml. of anhydrous ether. The solution was stirred at 00 for 20 minutes to affect the lithium exchange (50). To the solution of 4-lithioanisole was added quickly a solution of 4.0 g. (0.016 mole) of 2-phenylbenzophenone in 50 ml. of anhydrous ether, the solution was stirred at 00 for an additional 30 minutes and the reaction was hydro- lyzed with 50 ml. of water. The ether layer was separated 64 and dried over anhydrous magnesium sulfate. The ether was removed with a rotatory evaporator, leaving a clear oil. The clear oil was chromatographed on a silica—gel column using benzene as the eluant. yielding 1.0 g. of a clear oil containing mostly carbinol, but a trace of ketone (infrared) and 2.0 g. of a complex mixture. The 1.0 g. of clear oil was rechromatographed on 30 g. of Beckman Analyzed basic alumina using benzene as the eluant. A clear oil was isolated, free of carbonyl. This was dissolved in 15 ml. of hexane, and stored in the refrigerator. White pellet-like crystals, 0.7 g. (16%), m.p. 87.5-89O were isolated and recrystallized from hexane, m.p. 87.5-890. A check of purity on t.l.c. plates using either carbon tetrachloride or benzene eluants Showed a Single spot. The infrared Spectrum is Shown in Figure 18 and the n.m.r. spectrum in Figure 18; T 7.43 (1H), 1 6.27 (3H), T 3.42-2.72 multiplet (18H). The ultraviolet Spectrum Showed end absorption at 220 mu, and absorbance maxima at 280 and 260 mu (6 700). Anal. Calcd. for C25H2202 : C, 85.22; H, 6.05. Found: C. 85.02; H, 6.04. 4. Preparation of 4-Methyl-2'-phenyltriphenylmethanol In a 250-ml. round-bottomed three-necked flask fitted with a reflux condenser, tru-bore stirrer, dropping funnel and gas inlet tube was prepared para-tolyllithium from 0.40 g. (0.057 mole) of lithium and 10.0 9. (0.0585 mole) of 65 p-bromotoluene in 75 ml. of anhydrous ether, using a nitrogen atmosphere. To the gray solution was added a solution of 4.0 9. (0.0155 mole) of phenylbenzophenone in 100 ml. of anhydrous ether dropwise over a period of fifteen minutes. The solution was refluxed for an additional hour and was hydrolyzed with 50 ml. of water. The ether layer was separated and dried over anhydrous magnesium sulfate, and the ether removed with a rotatory evaporator. The light yellow oil (6.0 g.) which remained was chromatographed on a silica gel column (Baker Analyzed Grade) using carbon tetrachloride as the eluant. The first three fractions (150 ml.) yielded 3.0 g. of alcohol and impurities, with the remaining fractions (350 ml.) yielding 2.5 g. of a clear oil which became a glass when the final traces of solvent were removed. The glass resisted crystal- lization from pentane or benzene-pentane. The glassy solid stood at room temperature for two months, resulting in the formation of a small amount of crystalline material, m.p. 60-750, which was used to seed a hexane solution of the glassy material. White crystals, 2.0 g. (37%) m.p. 72-740 were isolated and recrystallized from hexane, m.p. 72.5—74.50. A check of purity on t.l.c. plates using either carbon tetrachloride or benzene as the eluant showed only one spot. The infrared 95% EtOH spectrum (Figure 19), the ultraviolet Spectrum kmax 280 mu (6 790), end absorption at 240 mu; and the n.m.r. 66 Spectrum (Figure 19) T 7.35 (1H), T 7.70 (3H), T 3.38-2.66 (complex multiplet) (18H), were consistent with the assigned structure. Anal..Calcd. for C26H220: C, 89.11; H, 6.33. Found: C, 89.27; H, 6.25. 5. Preparation of 3-Chlor042'-phenyltriphenylmethanol In a 100-ml. round-bottomed three-necked flask equipped with a tru-bore stirrer, dropping funnel, condenser and gas inlet was prepared the Grignard reagent from 1.0 g. (0.042, mole) of high purity magnesium and 8.0 g. (0.042 mole) of 3-chlorobromobenzene in 30 ml. of anhydrous ether. One drop of 1,2-dibromoethane was used to initiate the Grignard formation and a stream of nitrogen was slowly passed through the flask. To the yellow solution was added dropwise with stirring 3.5 9. (0.0135 mole) of 2-phenylbenzophenone in 15 ml. of anhydrous benzene and 10 ml. of anhydrous ether. The red solution was refluxed for 5 hours, after which the solution was colorless. The complex was decomposed by the dropwise addition of 25 ml. of water, the ether layer was separated, and the water layer was extracted with four 10-ml. portions of ether. The combined ether solutions were dried for 20 minutes over anhydrous magnesium sulfate and the solvent removed with a rotatory evaporator, leaving 5.5 g. of a yellow oil. 67 The oil was chromatographed on 100 g. of Woelm neutral alumina. The carbon tetrachloride and benzene fractions yielded 2.32 g. (14.5%) of pure carbinol as a clear viscous oil which yielded only one spot on t.l.c. plates using benzene as the eluant. The oil eventually crystallized and was recrystallized from pentane, m.p. 101.5-1020. The forerun fractions and column yielded 2.0 g. of a mixture of compounds. The infrared spectrum (Figure 20) and the n.m.r. spectrum (Figure 20), T 7.13 (1H), T 3.38-2.66 (18.3H) were consistent with the assigned structure. Anal. Calcd. for C25H19ClO: C, 81.00; H, 5.16; Cl, 9.56. Found: C, 81.07; H, 5.14; Cl, 9.44. 6. Preparation of 3-Methoxy-2'-phenyltriphenylmethanol In a 500~m1. round-bottomed three-necked flask equipped with tru-bore stirrer, condenser, dropping funnel and gas inlet was prepared the Grignard reagent from 0.643 g. (0.0267 mole) of high purity magnesium and 5.0 9. (0.0267 mole) of 3rbromoanisole dissolved in 90 ml. of anhydrous tetrahydro- furan, while flushing the apparatus with a slow stream of nitrogen. To the gray solution was added dropwise with stirring a solution of 3.0 g. (0.0166 mole) of 2-phenylbenzophenone in 50 ml. of anhydrous tetrahydrofuran. The solution im- mediately turned blue, but changed to yellow after approxi- mately 30 hours of reflux. Reflux was continued for a total of 48 hours, and the reaction mixture was decomposed by the 68 dropwise addition of 100 m1. of distilled water. The mix- ture was extracted with three 100-ml. portions of ether, the ether dried over anhydrous magnesium sulfate for 15 minutes and the solvent removed on a rotatory evaporator. A yellow viscous oil, 6.1 g., remained. A sample, 3.5 g., was chromatographed on a column of 100 g. of Woelm neutral alumina, activity 1, using 50% by volume carbon tetrachloride and benzene as the eluant. The major portion was 1.1 g. of oily carbinol (44.5% based on 1.89 g. in 6.1 g. of crude oil), showing one major spot on t.l.c. plates. The remaining fractions yielded 1.8 g. of a mixture of compounds. The carbinol crystallized on standing and was recrystallized from pentane, m.p. 101-105.50. The infrared Spectrum (Figure 21) and the n.m.r. spectrum (Figure 21) T 7.28 (1H), T 6.42 (3H), T 3.50-2.68 (18H), were consistent with the assigned structure. Agal. Calcd. for C25H2202: C, 85.22; H, 6.05. Found: C, 85.07; H, 6.09. 7. Preparation of 3-Methyl—2'-phenyltriphenylmethanol In a 300—ml. round-bottomed three—necked flask fitted with a condenser, tru-bore stirrer, dropping funnel and gas inlet tube was prepared the Grignard reagent from 4.0 9. (0.0203 mole) of m—bromotoluene and 0.5 g. (0.0208 mole) of magnesium in 50 ml. of anhydrous tetrahydfofuran, using a nitrogen atmosphere. To the Grignard reagent was added dropwise with stirring a solution of 3.0 g. (0.0116 mole) 69 of 2-phenylbenzophenone dissolved in 65 ml. of anhydrous tetrahydrofuran. The reaction mixture became dark blue and, after complete addition of the ketone (about 20 minutes), was refluxed for an additional two hours. The reaction mixture was decomposed by the dropwise addition of 25 ml. of water and was extracted with three 50-ml. portions of ether. The ether layer was washed with 25 m1. of water and dried over anhydrous magnesium sulfate. The solvent was removed with a rotatory evaporator, leaving 6.2 g. of a yellow viscous oil containing alcohol and ketone (infrared). The oil was chromatographed on a column of 150 g. of Baker Analyzed basic alumina using benzene as the eluant. Ten 50-ml. fractions yielded 3.5 g. of a clear viscous oil containing mostly alcohol and a trace of ketone, detected by infrared and t.l.c. analysis. The oil was rechromatographed on a column of Woelm neutral alumina, activity grade 1, using Baker Analyzed carbon tetrachloride as the eluant. A clear viscous oil was obtained (2.44 g.; 55.8%) and 1.0 g. of a mixture. The oily carbinol would not crystallize, but the infrared spectrum (Figure 22), and the n.m.r. Spectrum (Figure 22), T 7.38 (1.1H), T 7.75 (2.7H) and T 3.50-2.58 (complex multiplet (18H), were consistent with the assigned structure. Chroma- tography on thin layer plates showed only one spot. Anal. Calcd. for C23H220: C, 89.11; H, 6.33. Found: C, 88.68; H, 6.41. 70 8. Preparation of 3-Trifluoromethyl-2‘ephenyltriphenyl- methanol In a 300-ml. round-bottomed three-necked flask fitted with a condenser, tru-bore stirrer, dropping funnel and gas inlet tube was prepared the Grignard reagent from 1.0 g. (0.0407 mole) of high purity magnesium and 50 ml. of a solu- tion of 9.0 9. (0.0407 mole) of 3-bromobenzotrifluoride (Aldrich Chemical Co.) in tetrahydrofuran, using a nitrogen atmosphere. To the gray solution was added dropwise with stirring a solution of 5.0 9. (0.0275 mole) of 2-phenyl- benzophenone in 50 ml. of anhydrous tetrahydrofuran. After addition, the solution was refluxed for 24 hours and allowed to stand for an additional 24 hours at room temperature. The solution was hydrolyzed by the addition of 100 ml. of water and extracted with three 50-ml. portions of ether. The ether solution was dried over anhydrous magnesium sulfate and the solvent removed with a rotatory evaporator, leaving a brown oil, 12.5 g. A portion of the oil, 5.0 g., was chromatographed on basic alumina (Baker Analyzed) using 50% by volume carbon tetrachloride-benzene as the eluant. The center fraction 2.5 9., containing mostly the desired alcohol, was rechroma- tographed on silica gel using carbon tetrachloride and benzene as eluants. The benzene fractions yielded 2.30 g. of a clear oil showing only one Spot on t.l.c. plates. The oil would not crystallize. Assuming that the 7.5 g. of crude product 71 would yield a proportional amount of alcohol, the yield was 52% . The infrared Spectrum (Figure 23), and the n.m.r. Spectrum (Figure 23), T 7.05 (1H), T 3.38-2.33 (complex multiplet) (18.6H) were consistent with the assigned struc- ture. Agal. Calcd. for C25H190F3: c, 77.25; H, 4.74. Found: C, 77.46; H, 4.99. 9. Preparation of 4,4'-Dichloro-2"-phenyltriphenyl- methanol In a 100-ml. round-bottomed three-necked flask equipped with tru-bore stirrer, condenSer, dropping funnel and gas inlet was prepared the Grignard reagent from 1.35 9. (0.0563 mole) of high purity magnesium and 13.1 9. (0.0563 mole) of 2-bromobiphenyl dissolved in 50 m1. of anhydrous ether. To the refluxing orange solution was added dropwise with stirring a solution of 5.5 g. (0.0219 mole) of 4,4'-di— chlorobenzophenone (Dow Chemical Co.) dissolved in 25 ml. of anhydrous benzene and 10 ml. of anhydrous ether. After the addition, the solution was refluxed for an additional 24 hours, sealed and allowed to stand at room temperature for an additional 48 hours. The white precipitated complex was then filtered through a glass wool plug and washed with two 20-m1. por- tions of anhydrous ether. The complex and glass wool were stirred with 75 ml. of ether and the complex decomposed by 72 the dropwise addition of 20 ml. of distilled water. The ether layer was separated, dried over anhydrous magnesium sulfate, and the ether removed on a rotatory evaporator. The heavy oily residue was stirred with 20 m1. of pentane which resulted in formation of white crystals of the desired carbinol, 2.48 g. (28%), m.p. 116.5-117.OO. Re- crystallization from pentane gave white crystals, m.p. 118-118.50. The pentane solution from the crude oil yielded 4.8 g. of a complex mixture containing some carbinol. The infrared Spectrum (Figure 24), and the n.m.r. spectrum (Figure 24), T 7.22 (1H), T 4.42-2.67 (complex multiplet) (16.9H) were consistent with the assigned structure. Anal. Calcd. for C25H180C12: C, 74.08; H, 4.48; Cl, 17.50. Found: C, 74.24; H, 4.52; Cl, 17.56. 10. Preparation of 4L4'-Dimethoxy—2"-phenyltriphenyl methanol In a 300-ml. round-bottomed three-necked flask equipped with a tru-bore stirrer, reflux condenser, dropping funnel and gas inlet was prepared the Grignard reagent from 0.62 g. (0.0258 mole) of high purity magnesium and 6.0 g. (0.0258 mole) of 2-bromobiphenyl in 100 m1. of anhydrous tetrahydro— furan, using 1 drop of 1,2-dibromoethane to initiate the reaction. To the gray solution was added dropwise with stirring, over a 20-minute period, a solution of 5.0 9. (0.0207 mole) of 4,4'-dimethoxybenzophenone (Dow Chemical Co.) 73 in 50 ml. of anhydrous tetrahydrofuran. The red solution was heated at reflux with stirring for 24 hours, stoppered and allowed to stand at room temperature for 24 additional hours. The complex was decomposed by the dropwise addition of 100 ml. of distilled water. The mixture was extracted 4 times with 80 ml. of ether, the organic layer dried over anhydrous magnesium sulfate for 10 minutes and the solvent was removed on a rotatory evaporator. The yellow oil crystallized on standing, yielding white crystals of starting ketone which, after washing with pentane, weighed 2.8 g. The oily residue, upon standing for 12 hours, yielded 0.65 g. (22%) of the desired carbinol, m.p. 123-1240. Recrystallization from benzenehpetroleum ether gave 0.60 g., m.p. 123-1240. The oily filtrate weighed 2.2 g. and contained at least five compounds (t.l.c.), one of which was the carbinol. The infrared Spectrum (Figure 25), and the n.m.r. Spectrum (Figure 25), T 7.46 (1.1H), T 6.27 (6H), T 3.47- 2.75 (complex multiplet) (17.5H) were consistent with the assigned structure. Anal. Calcd. for C27H2403: C, 81.79; H, 6.10. Found: C, 81.62; H, 6.01. 74 11. Preparation of 4,4'-Dimethyl:2:fphenyltriphenyl- methanol In a 300-ml. round-bottomed three-necked flask fitted with a condenser, tru-bore stirrer, dropping funnel, and gas inlet tube was prepared the Grignard reagent from 0.84 g. (0.0355 mole) of high purity magnesium and 8.1 9. (0.0335 mole) of 2-bromobiphenyl in 100 ml. of anhydrous tetrahydro- furan, under a nitrogen atmOSphere. To the solution was added drOpwise with stirring a solution of 5.0 9. (0.0223 mole) of 4,4'-dimethylbenzophenone (Dow Chemical Co.) in 75 ml. of anhydrous tetrahydrofuran over a 30-minute period. The reaction mixture was then refluxed with stirring for one hour, hydrolyzed by the addition of 100 ml. of water, and extracted three times with 50 ml. of ether. The ether solution was dried over anhydrous magnesium sulfate and the solvent removed on a rotatory evaporator, leaving a yellow viscous oil, 8.0 g., which crystallized upon standing. Filtration yielded 2.0 g. (25%) m.p. 1600 of white crystals showing one Spot on t.l.c. plates using either carbon tetra- chloride or benzene as eluants. Recrystallization from benzene-petroleum ether yielded white crystals of 4,4'-di— methyl-2"-phenyltriphenylmethanol, m.p. 162-1640. The infrared spectrum (Figure 26), and the n.m.r. Spectrum (Figure 26), T 7.45 (1H), 1 7.70 (6H), T 3.38-2.66 (complex multiplet) (16.5H) were consistent with the assigned structure. 75 Anal. Calcd. for C27H24O: C, 88.97; H, 6.64. Found: C, 88.97; H, 6.56. The oily filtrate yielded 0.5 g., m.p. 116°, of an unidentified carbinol and 5.4 g. of an oily mixture of at least four compounds (t.l.c.), one of which was the desired carbinol. 12. Preparation of 4-Methoxy-4'-methyl-2"-phenyltriphenyl methanol In a 300-ml. round-bottomed flask fitted with a tru— bore stirrer, condenser, dropping funnel and gas inlet was prepared the Grignard reagent from 0.62 g. (0.0258 mole) of high purity magnesium and 6.0 g. (0.0258 mole) of 2-bromobiphenyl in 100 ml. of anhydrous tetrahydrofuran. A Slow stream of nitrogen was passed through the system during the preparation. To the gray solution was added dropwise with stirring a solution of 3.0 g. (0.0142 mole) of 4-methyl-4'-methoxy- benzophenone (Aldrich Chemical Co.) dissolved in 50 ml. of anhydrous tetrahydrofuran. The addition took 30 minutes, then the solution was refluxed for 48 hours. The complex was decomposed by the dropwise addition of 150 ml. of water, and the aqueous layer was extracted with three 50-ml. portions of ether. The organic layer was dried over anhydrous magnesium sulfate and the solvent was removed on a rotatory evaporator, leaving 7.3 g. of a yellow oil. The oil crystallized on standing, yielding 76 6.15 g. of a white solid containing four compounds (t.l.c.). Chromatography of 3.0 g. on 75 g. of Woelm basic alumina yielded 0.6 g. of biphenyl, 0.59 g. of pure carbinol, m.p. 126-127.5O from pentane, and 1.6 g. of column residue con; taining four spots on a t.l.c. plate. The remaining 3.1 g. of mixture was dissolved in hot pentane and yielded 1 g. of alcohol and 2 g. of mixture. Based on 5.55 g. theoretical yield, the yield of pure carbinol was 28%. The infrared Spectrum (Figure 27) and the n.m.r. Spectrum (Figure 27) T 7.45 (1H), T 7.68 (3H), T 6.28 (3H), T 3.44-2.74 complex multiplet (16.5H) were consistent with the assigned structure. Anal. Calcd. for C27H2402: C, 85.23; H, 6.36. Found: C, 85.19; H, 6.36. D. Preparation of 9,9-disubstituted fluorenes 1. Preparation of 9,9—Diphenylfluorene . The procedure of Clarkson and Gomberg (4) was followed, using 0.5 g. of 2-phenyltriphenylcarbinol. The yield was quantitative. The infrared spectrum is shown in Figure 28 as is the n.m.r. Spectrum, T 3.0-2.16 (complex multiplet). The ultra— x95% EtOH max 507.8 mu (e 8.25 x 103). 296 (4.97 x 103), 285 (sh) (8.82 x violet Spectrum gave the following results: 103), 279 (sh) (1.15 x 104), 271.5 (1.50 x 104), 267.5 (1.46 x 104), 265 (sh) (1.57 x 104). 258 (2.61 x 104), 250 (5.40 x 104), 254 (sh) (5.55 x 104) . 77 2. Preparation of 9-(4-Ch1oropheny1)-9-phenylfluorene In a 25-ml. Erlenmeyer flask was placed 0.1 g. of 2-phenyl-4'-chlorotriphenylmethanol and 15 ml. of glacial acetic acid. The mixture was heated to reflux and 3 drops of concentrated sulfuric acid were added; the solution turned yellow, then clear. The solution was cooled and 0.1 g, of white crystals of 9-(4-chlorophenyl)-9-phenylfluorene were collected and recrystallized from methanol, m.p. 142.5- 144.00. The infrared spectrum (Figure 29), the ultraviolet EtOH spectrum, A 307.9 mu (7.47 x 103) 3 max 295-9 (7.07 X 10 ), 284.2 (sh) (8.71 x 103), 271.6 (1.51 x 104), 267.5 (sh) (2.114 x 104), 264.0 (sh) (1.568 x 104), 257.5 (2.84 x 104). 230.0 (3.47 x 104), 228 (sh) (3.37 x 104) and the n.m.r. spectrum (Figure 29) T 2.16-2.92 multiplet were consistent with the assigned structure. Anal. Calcd. for C25H17Cl: C, 85.09; H, 4.86; Cl, 10.05. Found: C, 84.78; H, 4.81; Cl, 10.13 (trace of ash). 3. Preparation of 9-(4-Methoxyphenyl)-9-phenylfluorene In a 25-ml. Erlenmeyer flask was placed 0.1 g. of 2-phenyl-4'-methoxytriphenylmethanol and 15 ml. of 80% by weight aqueous acetic acid; a red solution was formed. The red color remained for one-half hour at 250. Warming to about 800 caused‘the red color to change to yellow, then clear after heating for twenty minutes. The solution was cooled and a quantitative yield (0.1 g.) of white crystals 78 of the desired fluorene, m.p. 160?,was obtained. Recrystal- lization from methanol yielded white needles, m.p. 165.5- 166.0°. The infrared Spectrum (Figure 30), the n.m.r. spectrum (Figure 30) T 6.27 (3H), T 3.36-2.13 multiplet (17H), and 95% EtOH 508.6 mu (6 8.78 x 103), max the ultraviolet spectrum A 296.5 (5.15 x 103), 279.5 (sh) (1.42 x 104), 275.0 (1.76 x 104), 268.2 (sh) (1.75 x 104), 264.1 (sh) (1.647 x 104), 256.9 (sh) (1.26 x 104), 230.0 (1.77 x 104) were consistent with assigned structure. Anal. Calcd. for C26H2002 C, 89.62; H, 5.79. Found: C, 89.41; H. 5.84. 4. Preparation of 9-(4-Methylphenyl)-9-phenylfluorene In a 25-ml. Erlenmeyer flask was placed 0.1 g. of 2-phenyl-4'-methyltriphenylmethanol and 15 ml. of glacial acetic acid. The solution was heated to reflux, and two drops of concentrated hydrochloric acid were added. The solution turned red and became clear in a few seconds. After adding three drops of water and cooling to room temperature, a quantitative yield (0.1 g.) of white crystals of 9-(4-methylpheny1)-9-phenylfluorene, m.p. 134-134.50. was collected. A check on a t.l.c. plate using carbon tetrachloride as the eluant showed only one spot. The compound was recrystallized from methanol, m.p. 134-134.50. The infra- 95% EtOH red spectrum (Figure 31), the ultraviolet spectrum Nmax 79 508.1 mu (e 8.51 x 103), 296.2 (5.04 x 103), 284.1 (sh) (8.14 x 10°). 279.0 (sh) (1.150 x 104). 271.3 (1.512 x 104). 267.3 (1.496 x 104). 261.9 (Sh) (1.391 x 104), 237.5 (Sh) (2.74 x 104), 229.1 (5.605 x 104), 224.5 (sh) (5.59 x 104). and the n.m.r. spectrum (Figure 31) T 7.78 (3H), T 3.38-2.33 complex multiplet (17.4H) were consistent with the assigned structure. ‘Anal. Calcd. for CgeHgo: C, 93.94; H, 6.06. Found: C, 93.82; H, 6.18. 5. Preparation of 9-(3-Chlorophenyl)-9-phenylfluorene In a 25-ml. Erlenmeyer flask was placed 0.3 g. of 2-phenyl-3'-chlorotriphenylmethanol and 10 ml. of glacial acetic acid. The solution was heated to reflux, 4 drops of concentrated hydrochloric acid were added, the solution was refluxed for 1 minute and cooled to room temperature. A quantitative yield of white needles, m.p. 154-155O after recrystallization from methanol, was obtained. The infrared spectrum (Figure 32), the n.m.r. spectrum (Figure 32), T 3.08-2.08 (complex multiplet) and the ultra- 95% EtOH 307.2 mu (6 7.73 x 103). 295.6 max violet spectrum A (4.90 x 103), 285.2 (Sh) (1.02 x 104), 278.7 (sh) (1.24 x 104), 271.7 (1.62 x 104), 267.5 (1.50 x 104), 261.5 (sh) (1.44 x 104), 258.5 (sh) (2.78 x 104), 229.6 (5.41 x 104) were consistent with the assigned structure. Anal. Calcd. for C25H17Cl: C, 85.09; H, 4.86; Cl, 10.05. Found: C, 85.13; H, 4.85; Cl, 10.00. 80 6. Preparation of 9-(3-Methoxypheny1)-9-phenylfluorene In a 25—ml. Erlenmeyer flask was placed 0.2 g. of 2—phenyl-3'-methoxytriphenylmethanol and 10 ml. of glacial acetic acid. The solution was heated to reflux and 3 drops of concentrated hydrochloric acid were added forming a red solution which quickly turned colorless. Three drops of water were now added, and the solution was allowed to remain for 24 hours at room temperature. White crystals (0.19 g.; 95%) m.p. 1370 were isolated. Recrystallization from methanol yielded white crystals, m.p. 150.5-151.5°, showing only one spot on the t.l.c. plates. The infrared spectrum (Figure 33), the n.m.r. Spectrum (Figure 33),T 6.42 (3H), T 3.58-2.22 (complex multiplet) 95% EtOH max 307.8 mu (17H), and the ultraviolet Spectrum A (e 8.58 x 103), 295.7 (5.20 x 103), 285.2 (sh) (1.22 x 104). 270.0 (1.75 x 104), 267.5 (sh) (1.67 x 104), 263 (sh) (1.51 x 104), 257 (sh) (2.72 x 104), 228 (sh) (5.81 x 104) were consistent with the assigned structure. Anal. Calcd. for CgeHgoO; C, 89.62; H, 5.79. Found: C, 89.54; H, 5.69. 7. Preparation of 9-(3-Methylphenyl)-9-phenylfluorene In a 25-ml. Erlenmeyer flask was placed 0.5 g. of 2-phenyl-3'-methyltriphenylmethanol and 15 ml. of glacial acetic acid. The solution was heated to approximately 850, with the solution turning pink. To the pink solution was added 5 drops of concentrated hydrochloric acid, resulting 81 in a dark red solution which faded to a clear solution in about one minute. The Elear solution was cooled and 0.45 g. of white needles, m.p. 1500 were collected and recrystallized from methanol, m.p. 150.0-150.50. Chromatography on thin layer plates showed only one Spot when eluted with either benzene or carbon tetrachloride. The infrared spectrum (Figure 34) and the n.m.r. Spectrum (Figure 34) T 7.80 (3H), T 3.33-2.25 (complex multiplet) (17.6H) were consistent with the assigned structure. The ultraviolet Spectrum gave the following 95% EtOH max 283 (sh) (9.52 x 103), 278.5 (sh) (1.21 x 104). 271.6 (1.57 values: A 308 mu (E 8.34 x 103), 296 (5.20 x 103). x 104). 267.5 (1.54 x 104). 263.3 (sh) (1.45 x 104). 238.2 (2.71 x 104), 250 (5.58 x 104). Anal. Calcd. for CgeHgo: C, 93.94; H, 6.06. Found: C, 92.96; H, 5.95; ash 1.17. Corrected for ash: C, 94.05; H, 6.02. 8. Preparation of 9-(3-Trifluoromethylphenyl)-9-phenyl- fluorene In a 25-ml. Erlenmeyer flask was placed 0.1 g. of 2-phenyl-3'-trifluoromethyltriphenylmethanol and 20 ml. of a solution of 4% sulfuric acid by weight in 80% aqueous acetic acid, with an additional 4 drops of concentrated sulfuric acid. The mixture was swirled until the carbinol had dissolved and was allowed to stand at room temperature 230) for 3 days. About 5 ml. of the solution was evaporated 82 using an air jet and a white precipitate formed, m.p. 127.5-129.5O which was recrystallized from methanol, yield- ing white needles of 9-(3-trif1uoromethylphenyl)-9-phenyl- fluorene, 0.06 g. (60%). m.p. 127.5-129.5°. The infrared Spectrum (Figure 35), the n.m.r. Spectrum (Figure 35) T 3.05-2.13 (complex multiplet) and the ultra- 95% EtOH max (5.05 x 103). 282.5 (sh) (1.08 x 104), 278.0 (sh) (1.28 x 104). violet Spectrum A 307.2 mu (e 7.87 x 103). 295.4 271.0 (1.72 X 104). 267.0 (1.65 X 104). 265.0 (Sh) (1.56 X 104). 236.7 (2.66 x 104), 228.2 (3.34 x 104) are consistent with the assigned structure. Anal. Calcd. for C26H17F3: C, 80.81; H, 4.44. Found: C, 80.79; H, 4.43. 9. Preparation of 9,9-Di(4-chlorophenyl)fluorene In a 25-ml. Erlenmeyer flask was placed 0.25 g. of 2-phenyl-4',4"-dichlorotriphenylmethanol and 10 ml. of glacial acetic acid. The solution was heated to reflux and five drops of concentrated hydrochloric acid were added. After refluxing for 5 minutes, 5 drops of water were added and the solution was cooled to room temperature. The white crystals, 0.25 g. (100%), m.p. 165.5-167.0O were recrystal- lized from methanol, m.p. 165-166.50. Analysis on t.l.c. plates showed one spot. The infrared Spectrum (Figure 36), the n.m.r. Spectrum (Figure 36), T 2.97-2.22 (complex multiplet), and the ultra- 95% EtOH 3 max 508 mu (e 6.78 x 10 ), 296 violet spectrum, A 83 (4.41 x 103), 284 (sh) (9.16 x 103), 279.2 (sh) (1.150 x 104). 272.4 (1.50 x 104),'269.5 (sh) (1.45 x 104), 263.2 (sh) (1.34 x 104), 240 (3.15 x 104), were consistent with the assigned structure. Anal. Calcd. for C25H15C12; c, 77.55; H, 4.17; CI, 18.51. Found: C, 77.39; H, 4.10; Cl, 18.40. 10. Preparation of 9,9-Di(4-methylphenyl)fluorene In a 25-ml. Erlenmeyer flask was placed 0.1 g. of 2-phenyl-4',4"-dimethyltriphenylmethanol and 10 ml. of glacial acetic acid. The solution was heated to reflux, five drops of concentrated hydrOchloric acid were added, resulting in a dark red solution which rapidly lost its color. Following the addition of 25 drops of water, the solution was cooled to room temperature. White crystals, 0.07 g., m.p. 1600 were collected and recrystallized from methanol, m.p. 1600. Analysis on t.l.c. plates showed one spot with benzene or carbon tetrachloride as eluants. The infrared Spectrum (Figure 37), the n.m.r. Spectrum (Figure 37), T 7.74 (6H), T 3.33-2.16 (complex multiplet) 95% EtOH max 308.6 mu (16.6H), and the ultraviolet Spectrum A (e 8.5 x 103), 296 (5.09 x 103). 284 (sh) (8.16 x 103). 279.6 (sh) (1.152 x 104), 271.8 (1.45 x 104), 267.5 (1.45 x 104), 265.5 (sh) (4.48 x 104), 258.8 (2.80 x 104), 228.5 (4.03 x 104), were consistent with the assigned structure. 84 Anal. Calcd. for C27H223. C, 95.60; H, 6.40. Found: 0, 95.01; H, 6.45; 0.54% ash Recalculated correcting for ash Found: C, 93.51; H, 6.49. 11. Preparation of 9,9-Di14-methoxyphenyl)fluorene In a 50-ml. Erlenmeyer flask was placed 0.2 g. of 2-phenyl-4',4"-dimethoxytriphenylmethanol and 20 ml. of 4% sulfuric acid in 80% aqueous acetic acid. The red solution stood for 4 days until the red color had become light yellow. The solution was heated to about 700, 10 drops of water were added, and the solution cooled. After 2 days, 0.17 g. of white crystals were collected and recrystallized from methanol, m.p. 120-1220. The infrared spectrum (Figure 38) and the n.m.r. spectrum (Figure 38), T 6.40 (6H) and T 3.50-2.68 (16H) were consistent with the assigned structure. The ultraviolet 95% EtOH max 309.8 mu Spectrum gave the following results: A (e 1.08 x 104), 297.5 (5.62 x 104), 279.6 (sh) (1.90 x 104), 271 (2.44 x 104), 268 (2.48 x 104), 263 (2.48 x 104), 250 (7.98 x 104). Anal. Calcd. for C27H2202: C, 85.69; H, 5.86. Found: C, 85.42; H, 5.90. 12. Preparation of 9-(4-Methoxyphenyl)-9-(4—methyl- phenyl)fluorene In a 50-ml. Erlenmeyer flask was placed 0.3 g. of 2—phenyl-4'-methyl—4'-methoxytriphenylmethanol and 20 ml. 85 of glacial acetic acid. The solution turned pink and after the addition of 10 ml. of 4% sulfuric acid in 80% aqueous acetic acid, the solution turned dark red fading to yellow after 10 hours. The solution was heated to 800, 20 drops of water were added and the solution stood at room tempera- ture for 24 hours. White crystals, 0.29 g., m.p. 129-130O were collected. Recrystallization from methanol did not change the melting point. The infrared spectrum (Figure 39) and the n.m.r. spectrum (Figure 39) T 7.75 (3H), T 6.45 (3H), T 3.47-2.17 complex multiplet (16H) were consistent with the assigned structure. The ultraviolet spectrum gave the following data: 95% EtOH max (sh) (1.63 x 104) 271.5 (2.26 x 104), 268 (2.18 x 104). A 308.8 mu (6 1.04 x 104). 297 (5.97 x 103). 278.9 230 (5.03 x 104). Anal. Calcd. for C27H220: C, 89.47; H, 6.12. Found: C, 89.39; H, 5.94. E. Preparation of Deuterated Reaction Medium for Kinetic Studies To a 100-ml. pear-shaped flask was added 16.1908 g. (0.3202 mole) of D20 (Merck, Canada), followed by the drop- wise addition of 100.1 0.1% H2804 until 2.0367 9. had been added. The flask was now fitted with a condenser. To the solu- tion was added dropwise 32.6937 9. of acetic anhydride (b.p. 1360 at 720 mm. Hg). Initial additions caused warming and 86 slight Sputtering at the surface. After complete addition the system was closed, the entire apparatus was inverted several times to insure complete mixing, and the solution was allowed to remain at 250 for 20 hours. The final percentage was 4.00% H2304 in 80.00% DOAc and 20% DéO. The total deuteron enrichment was 97.3%. F. Evaluation of Products from Kinetic Runs The actual structure of the products from the kinetic runs had to be determined by comparison with the correspond- ing products from direct synthesis. For compounds XIV, XIX and XXIII the solutions from the kinetics were saved, most of the acetic acid removed on a rotatory evaporator, and the precipitate isolated and recrystallized from methanol. In all three cases the melting points and spectra agreed in every respect with the same hydrocarbon prepared directly on a larger scale. III. Kinetics of the Cyclodehydration A. The Kinetic Method and Equations The reactions were followed by recording the appear— ance of the absorption band at approximately 308 mu due to the product (a fluorene) and absent in the carbinol. A Beckman DB Spectrophotometer equipped with a thermostated cell compartment, using a Tecam thermostating circulating pump to control the cell compartment to.i 0.10 of the desired temperature, was used to follow the reactions. 87 A Sargent SRL recorder was used to record the absorbance and the reaction time in seconds. The reactions were carried out in 1 cm. three-ml. stoppered rectangular Silica cells using as the reaction solvent a 4.000% by weight solution of sulfuric acid in 80% by weight aqueous acetic acid. The sulfuric acid solution was prepared on a Mettler analytical balance by placing a 5 g. sample of 100.: 0.1% sulfuric acid in a stoppered weighing flask and diluting with 80.0% by weight of aqueous acetic acid until the desired percentage was obtained. This amount of solution was usually sufficient to run the kinetics of all the compounds at one temperature. Each time another sulfuric acid solution was prepared, care was taken to reproduce the percentage of H2504 to one part in four thousand (i.e. 4.000.: 0.001% H2804). To minimize any error due to deterioration of the acid solution (51). the solutions were used within five days of preparation. Stock solutions of the carbinols were prepared in 5- ml. stoppered volumetric flasks using anhydrous benzene as solvent. A 5-u 1. sample of such solutions, using a 10-u 1. Hamilton microsyringe for the measurment, was placed in 2-ml. stoppered volumetric flasks and thermostated until use. When the aliquots were diluted to 2 ml. with reaction solvent, the solutions were about 9 x 10"5 molar in carbinol. The procedure for a kinetic run was as follows: 88 A solution of each hydrocarbon was made up in acid reaction solvent by pipetting 5-u 1. of a benzene stock solution of hydrocarbon and diluting with acid reaction solvent. The solution was then used to set the exact posi- tion of the wavelength of the "308 mu" band on the instrument wavelength selector by using the Amax for each different hydrocarbon. The aliquots of carbinol solution were diluted to 2 ml. with thermostated sulfuric acid-acetic acid solution, mixed by six inversions, poured‘into the thermostated silica cells, and the recorder started. The time elapsed was at the most 30 seconds, and the time at which mixing was begun was taken as zero time of reaction. The recorder was then used to record the appearance of the absorption band versus time in seconds; the latter measurement was accurate to within 1 second in one hour. The value of the rate constant was obtained from the expression kt = 2.505 log [luff/(HEO - Ah)] where Ah is the absorbance at any time of the hydrocarbon and Ago the absorbance at infinite time (52) (see page 91). The kinetics were evaluated using steady state assump- tions as follows: ROH + H30+ —:.<__-L§ ROH: + H20 or for simplicity + + ROH + H —kJ-> ROHg + + HOH2 —k23—> R + H20 + R —h5%> Product + H+ 89 or by simplification then one can write dA + .5? = -klAH + kgc (l) dH+ + '6‘” = -klAH + kgc + k5E (2) t gig = klAH+- k2C "" k3C + k4ED (5) Assume steady state conditions on (2) and (3) where + dH _ .92 dt - 0 and dt 0 from (3) from (2) k AH+ + k ED + c = -l- 4 (4) k5E = klAH - k2C (5) k2 + k3 Substitution of (4) into (5) yields klkS + | = k2k5 + ksks + ksk.' AH k4 k4E since E (i.e. H2O) D: is essentially constant klkgks + d (Product) AH kgks + k3k5 + k2k4' dt = ksD = kl (kg/k3) + 1 + (k4'/k5) + AH 90 or d (Product) = k; [H+] [ROH] dt 1 + (kg/k;)(1 + k4'/k5) If k4 < < kg and [H+] remains constant d (Product) = 61' [ROH] dt 1 + (kg/k3) kr [ROH] a Simple first order expression. This must be true if k3 is the slow step and the steady state conditions on ROH2+ require that k2 > > k3. Since the appearance of product is stoichiometrically equal to the dis- appearance of alcohol d (Product) -d [ROH] dt = dt kr [ROH] .. +i. dD_ If we assume steady state conditions on R (l.e.-5E - 0) dB EE- = k3C ~k4ED-k5D=O = k3C=(k4'+k5)D=O _ ksC D — k4, + ks (6) from (4) and (6) E = kgklAH+ k4'k2 + (k5)(k2 + k3) d (Product) = kSD = kskgkl AH+ dt kaks + kng + k2k4' which is identical to the first approximation; again a first order relationship is obtained. 91 For the overall first order reaction A ———>-B, kt = 0 ln (cA /CA) where c A0 and CA are the concentrations of A initially and at any time t, reSpectively. If since kt = ln(CAO kt kt concentration of B at any time absorbance of B molar extinction coefficient of B 00 AB /€B m (A /eB) - (AB/e3) a) 00 ln [ AB / (AB - AB)] 2.505 log [ Ag°/(Ag’- AB)] which is the first order expression in terms of absorbance of the product. The rate data are sent to the computer for analysis where 2.303 log [Ago/(AED- Ah)] is plotted versus time (t) in seconds with a resultant slope equal to k, the first order rate constant. The rate data were shown to be pseudo first order by the plot of log [Ago/(Afip- Ah)] versus time (t) in seconds, resulting in a straight line for at least 88% reaction (see 92 Figure 1). The rate constants were calculated at times along the reaction with the constants obtained from the value of the slope (see Table II). Using the Powell method, the curve best fit the curve for a first order reaction (53) (see Figure 13). In all cases of the para-methoxy substituted carbinols, the expression was Simplified since the ionization appeared instantaneous and the rate-determining step was now the electrophilic attack on a benzene ring: ROH + H+ —El->» R0H2+ (rapid) E2 + ROH2+ éA R + H20 (rapid) 4 + Rf -E5€>’ Product + H (slow) + dP _ dR + dt ' dt k5 [R J + o kt = 2.303 log .1311 a Simple first order expression. [R 1 With the interference at 308 mu from the tail of the "red" band of the carbonium ion, the rate expression was modified: AObs = Ah + A+ (6) Aobs = observed absorbance Ah = hydrocarbon absorbance A+ = "carbonium ion” absorbance a) — C + C Ch ’ + h a) _ . Ch = Ch - C+ (7) Ch — concentration of hydrocarbon C+ = concentration of carbonium ion 93 A.mommv mEHH moq 00H om p p r n d d 1 did)! d A d. H!- d- d. u- 4, O Hocmnume $.82on6156: . 78630.4 (Ill) ®>HSU HmmuHO UCOUGW Illll! 0>HSU HmmuHO umHHh ll...|Il| oflmmv WCOHHUmmH HmUHO USN Cam #mfi HOW ®>HSU UHMUCMUW : SHHB Hocmgumaamcmsmeuuahcmnml.NIOHOHEUIH How poam HHmBOA mo comHHMQEOO .MH musmHm 94 Aobs = EhCh + E+C+ (8) substitute (7) in (8) _ oo Aobs — (Ch - c+) eh + €+C+ — m - - Ch eh + (C+) (6+ eh) — m - Aobs — Ah + (C+) (€+ eh) 00 C = Ah ‘ Aobs (9) + eh _ €+ since kt = 2.305 log (c+°/C+) (10) substitute (9) in (10) o _C (e +,e ) kt = 2.505 log + h + AGD - A h obs Obs)] versus time in seconds A plot of log [ l/(A:)--A gave k = 2.305 slope. Again pseudo first order kinetics were observed (see Figure 2 and Table III). It is not known whether the observed precision in kr is the best that can be obtained with the present analytical methods. The largest source of error in the calculated rate constant was related to the analytical precision. Using the evaluation scheme as described by Benson (54), the per cent error in the calculation of the rate constant was estimated to be 4%. Absorbance Absorbance 95 B. Examples of the Method Figure 14. Plot of Absorbance versus time f8r S-Methyl- 2'-phenyltriphenylmethanol at 20 and 1% H2804. L 1 1 A A L A 4 1 A 1 1 L 1 0 100 220 540 460 580 700 820 Time (secs.) Figure 15. Plot of Absorbance versus time f8r 4-Methyl- 2'-phenyltriphenylmethanol at 15 and 4% H2804. 1 l 1 350 590 770 1010 1130 Time (secs.) 96 Table XIII. Calculations of Rate Data by Computer forO 5'-Methyl-Z-phenyltriphenylmethanol at 20 and 4% H2504 Absorbance Time (secs.) Rate Constant (x 103) 0.080 40 2.56 0.155 100 2.11 0.200 160 2.10 0.257 220 2.08 0.509 280 2.08 0.552 540 2.06 0.591 400 2.04 0.427 460 2.05 0.459 520 2.05 0.488 580 2.06 0.515 640 2.06 0.555 700 2.06 0.555 760 2.07 0.575 820 2.08 0.601 940 2.08 Rate Constant = 2.07 x 10-3 Intercept = 2.57 x 10‘3 Average Deviation = 0.00671 97 Table XIV. Calculations of Rate Data by Computer for 4-Meth 1-2'-phenyltriphenylmethanol at 15 and 4%YH2 SO 4 O Absorbance Time (secs.) Rate Constant (x 103) 0.065 50 2.27 0.152 110 2.24 0.192 170 2.25 0.248 250 2.29 0.295 290 2.51 0.555 550 2.28 0.566 410 2.27 0.598 470 2.28 0.425 550 2.27 0.448 590 2.29 0.468 650 2.28 0.485 710 2.28 0.500 770 2.27 0.515 850 2.22 0.525 890 2.27 0.556 950 2.29 0.552 1070 2.28 0.565 1190 2.28 0.575 1510 2.29 Average Deviation Rate Constant = 2.28 x 10‘3 0.00765 Intercept = -0.00278 98 Figure 16. Ultraviolet Spectra of 5-Chloro-2'-phenyl- triphenylmethanol and 9-(5-Chlorophenyl)-9- phenylfluorene. --‘-'5-Chloro-2'-phenyltriphenyl-“ 1 4 methanol ' ———— 9-(5-Chlorophenyl)-9-phenyl- fluorene I“ “- 1.2 1.0 +0.8. a) c m .Q n o .0 6 .3 ' <2 .O.4 .0.2 . f0.0 240 260 280 500 520 Wavelength 99 IV. Miscellaneous A. Preparation of 9—(2-Biphenylyl)-9-fluorenol The preparation was identical to that reported by Clarkson and Gomberg (4) with a yield of 51%, melting point 170°, reported 169-1700. B. Preparation of 9,9-Spirobifluorene A solution of 0.1 g. of 9-(2-biphenylyl)-9-fluorenol in S-ml. of glacial acetic acid was heated to reflux and one drop of concentrated sulfuric acid was added. The solution was refluxed for two minutes, 5 drops of water were added and the solution was cooled to room temperature. Filtration yielded 0.1 g., m.p. 2000 after recrystallization from methanol. .Reported melting point 198-1990 (4). C. Preparation of 1,9-Diphenyl-9-fluorenol 1. Preparation of 9-Fluorenone-l-carboxylic Acid The carboxylic acid was prepared from fluoranthene in refluxing acetic acid, as reported by Forrest and Tucker (55). Starting with 100 g. of fluoranthene 64 g. of acid was formed (45% yield). 2. Preparation of the Diacylyperoxide of 9-Fluorenone- 1-carboxy1ic Acid The peroxide was prepared as reported by Stiles and Libbey (42). The carboxylic acid, 25 g., was converted to the acid chloride with thionyl chloride (yield 18 g., m.p. 154-1550). The acid chloride (18.0 g., 0.072 mole) was 100 converted to the diacyl peroxide using Nagog. The yield was 14.7 g. of crude peroxide, m.p. 154-1570 with decomposition. 5. Preparation of 1-Phenyl-9-fluorenone (42) Crude diacyl peroxide of 9-fluorenone-l-carboxylic acid (14.7 g.) was suspended in 100 ml. of anhydrous benzene and refluxed for 60 hours. The crude product was chroma- tographed on alumina using benzene as the eluant. The yield of the desired ketone was 5.0 g. (17% based on carboxylic acid), m.p. 120-1210. 4. Preparation of 1,9-Diphenyl-9-fluorenol The Grignard reagent was prepared from 0.12 g. (0.005 mole) of magnesium turnings, and 0.8 g. (0.005 mole) of bromobenzene in 50 ml. of anhydrous ether. To the Grignard reagent was added dropwise a solution of 1.0 g. (0.004 mole) of 1-phenyl-9-fluorenone in 25 ml. of 50:50 ether-benzene. The mixture was stirred at reflux for 10 hours. The reaction mixture was decomposed with 20 ml. of a saturated ammonium chloride solution, washed with water and dried over magnesium sulfate. After removal of the solvent the yellow residue (1.5 g.) was chromatographed on 50 g. of alumina (Baker, acid washed) with benzene as eluant. The alcohol (0.5 g.; 0.001 mole; 25%) was isolated along with 0.1 g. of ketone and 0.7 g. of an oily mixture of undetermined composition. The carbinol was recrystallized from pentane, m.p. 105-1060. 101 Anal. Calcd. for C25H180: C, 89.79; H, 5.45. Found: C, 89.60; H, 5.50. The carbinol (0.05 9.) when dissolved in 10 ml. of 85% H2804, formed a green-brown solution which yielded only starting carbinol when poured on ice-water after 4 hours. It yielded only water-soluble products after 15 hours. The pKR+ value of the carbinol was determined using the absorbances at 656 mu and 504 mu. The values were pKR+ (656 mu) = -12.5 and pKR+ (504 mu) = -12.5. D. Attempted Cyclization of 1,9-Diphenyl-9-fluorenol The attempted cyclization of 1,9-diphenyl-9-fluorenol in acid media to phenylfluoradene (12-phenylindeno [1,2,5,-jk]- fluorene) was carried out by dissolving 0.1 g. of carbinol in 10 ml. of a solution of 85% H2804 or 80% aqueous acetic acid using 4% by weight H2804. In 85% H2804 after 4 hours starting material was recovered when the reaction mixture was poured onto ice-water. The infrared spectrum, melting point and R on t.l.c. plates were identical with the same f properties of the starting carbinol. The acetic acid - H2804 solution was allowed to stand two days. Ten drops of water were added, the solution warmed until turbidity disappeared and the solution allowed to cool to room temperature. Filtration yielded 0.09 g. of white crystals, m.p. 105°. The infrared spectrum was identical to that of the starting carbinol. A check for 102 purity on a t.l.c. plate showed only one compound with a Rf value identical to the starting material. When a solu- tion of the carbinol (0.1 g. in 10 ml.) was allowed to stand for six months followed by the identical work-up as above, only starting material was recovered. SPECTRA Infrared and Nuclear Magnetic Resonance 105 104 Figure 17. Infrared and Nuclear Magnetic Spectra of 4-Chloro-2'-phenyltriphenylmethanol in CCl4. “PM . J J l L l J l l l 4000 2000 1600 1200 800 Wavenumber (cm‘l) 105 Figure 18. Infrared and Nuclear Magnetic Spectra of 4-Methoxy-2'-phenyltriphenylmethanol in CC14. l 1 L, i I 1 I 1 4000 2000 1600 1200 800 Wavenumber (cm-l) I p I P I H :- 106 Figure 19. Infrared and Nuclear Magnetic Resonance -Spectra of 4-Methyl-2'-phenyltriphenyl- methanol in CC14. l I l l l l J 1 4000 2000 1600 1200 800 Wavenumber (cm-l) 107 Figure 20. Infrared and Nuclear Magnetic Resonance Spectra of 5-Chloro-2'-phenyltriphenyl- methanol in CC14. l l l l l I I l 4000 2000 1600 1200 800 Wavenumber (cm-1) 108 Figure 21. Infrared and Nuclear Magnetic Resonance Spectra of 5-Methoxy-2'-phenyltriphenyl- methanol.in CCl4. mmn 1 l l l i L l- l 4000 2000 1600 1200 800 Wavenumber (cm-l) 109 Figure 22. Infrared and Nuclear Magnetic Resonance ’ Spectra of 5-Methyl-2'-phenyltriphenyl- methanol in CCl4. l l ‘ l I 1 J t t 4000 2000 1600 1200 800 Wavenumber (cm-1) 110 Figure 25. Infrared and Nuclear Magnetic Spectra of 5-Trif1uoromethyl-2'-phenyltriphenylmethanol in CC14. C52 . III ! ' L I 11 I I L I 4000 2000 1600 1200 800 Wavenumber (cm") 111 Figure 24. Infrared and Nuclear Magnetic Resonance Spectra of 4,4'-Dichloro-2"-phenyltriphenylmethanol. in CC14. “WI/“7 I I L, I I I L 1 4L 4000 2000 1600 1200 800 Wavenumber (cm‘l) 1 I 4 i I I 1 1 2 4 6 8 10 112 Figure 25. Infrared and Nuclear Magnetic Resonance Spectra of 4,4'-Dimethoxy-2"-phenyltriphenylmethanol. in CC14 . MIN ’ ' II J I l I 1 IV ”I I 4600 2000 1600 1200 800 Wavenumber (cm‘l) F .- I— 4' I— p .— - 115 Figure 26. Infrared and Nuclear Magnetic Resonance Spectra of 4,4'-Dimethy1-2"-phenyltriphenylmethanol W“)! (M l l l l l l l I 4000 2000 1600 1200 800 Wavenumber (cm‘l) - — . 114 Figure 27. Infrared and Nuclear Magnetic Spectra of 4-Methoxy-4'4Methyl-2"-phenyltriphenyl- methanol in CCl4. ”WWW T— I) l I l l I I I I 4000 2000 1600 1200 800 Wavenumber (cm-l) 1 ) 1) ‘I ‘I 4 J? ( I) it 1 115 Figure 28. Infrared and Nuclear Magnetic Spectra of 9,9-Diphenylfluorene in CC14. I I I J I I I I 4000 2000 1600 1200 800 Wavenumber (cm-1) 116 Figure 29. Infrared and Nuclear Magnetic Resonance Spectra of 9-(4-Chlorophenyl)-9-phenylfluorene in CC14. ”7"“? I I I I I I I I 4000 2000 1600 1200 800 Wavenumber (cm-1) 117 Figure 50. Infrared and Nuclear Magnetic Resonance Spectra of 9-(4-Methoxyphenyl)-9-phenylfluorene in CCl4. cs2 L I 1 I I l I I I 4000 2000 1600 1200 800 Wavenumber (cm-1) 118 Figure 51. Infrared and Nuclear Magnetic Resonance Spectra of 9-(4-Methylphenyl)-9-phenylfluorene in CC14. €32 ' WW I on I l I I I I I I 4000 2000 1600 1200 800 Wavenumber (cm-1) -1 ‘~~ ‘_* _i -~* r l I I I I l | L l 2 4 6 8 10 119 Figure 52. Infrared and Nuclear Magnetic Resonance Spectra of 9-(5-Chlorophenyl)-9-phenylfluorene in CC14. cs2 I I I 'I I I I I 4000 2000 1600 1200 800 Wavenumber (cm-l) t I f 3 l I i I 2 4 6 8 10 120 Figure 55. Infrared and Nuclear Magnetic Resonance Spectra of 9-(5-Methoxyphenyl)-9-phenylfluorene in CC14. WW WW) I ' I I I I I I I 4000 2000 1600 1200 800 Wavenumber (cm-1) — 4H 121 Figure 54. Infrared and Nuclear Magnetic Spectra of 9-(5-Methylphenyl)-9-phenylfluorene in CCl4. cs2 r\/\~\/~J~—\ (v”"‘ . P5,}I L I I I I l I I - 4000 2000 1600 1200 800 Wavenumber (cm-l) 122 Figure 55. Infrared and Nuclear Magnetic Resonance spectra of 9-(5-Trifluoromethylphenyl)-9-phenylfluorene in CC14. cs;a I 1 I I I I I; I 4000 2000 1600 1200 800 Wavenumber (cm-1) - — — D - - F d. 125 Figure 56. Infrared and Nuclear Magnetic Resonance Spectra of 9,9-Di-(4-Chlorophenyl)fluorene in CC14. C82 “v” - . . 1.).. I 4000 2000 1600 1200 800 Wavenumber (cm-l) 124 Figure 57. Infrared and Nuclear Magnetic Spectra of 9,9-Di-(4-Methylphenyl)fluorene in CC14. CS2 WAN [ I I I I I I I L 4000 2000 1600 1200 800 Wavenumber (cm-1) - n I— - - - p — u-I 125 Figure 58. Infrared and Nuclear Magnetic Spectra of 9,9-Di-(4-Methoxyphenyl)fluorene in CCl4. MN) I) l I I I I I l I 4000 - 2000 1600 1200 800 Wavenumber (cm'l) Figure 59. 126 Infrared and Nuclear Magnetic Spectra of 9-(4-Methoxyphenyl)-9-(4-methylphenyl)- fluorene in CC14. WW (W L L I I L I I 4000 2000 1600 1200 800 Wavenumber (cm-l) l 1 l I l I : 2 4 6 10 LI TERATURE CI TED 127 (1) (2) (5) (4) (5) (6) (7) (8) (9) (10) (11) (12) (15) (14) (15) (16) (17) (18) (19) (20) 128 F. Ullman and R. von Wurstemberger, Ber., 58, 4105 (1905). E. Khotinsky and R. Patzewitch, Ber., 2, 5104 (1909). P. G. Sergeev, J. Russ, Phys. Chem. Soc., 61, 1421 (1929); Chem. Abstracts 24, 1565 (1950). R. G. Clarkson and M. Gomberg, J. Am. Chem. §2. 2881 (1950). Soc., E. Bergman and A. Bondi, Ber., 64, 1477 (1951). J. R. Dice, T. E. Watkins and H. L. Schuman, Chem. Soc., 12, 1758 (1950). J. Am. E. P. Kohler and L. W. Blanchard, J. Am. Chem. Soc., 51, 567 (1955). C. F. Koelsch, J. Am. Chem. Soc., _6, 480 (1954). H. H. Hatt, J. Chem. Soc., 478 (1941). G. Wittig and G. Fuhrman, Ber., 75B, 1197 (1940). M. Anchel and A. H. Blatt, J. Am. Chem. Soc., 65, 1948 (1941). J. H. Weisberger, E. K. Weisberger and F. B. Ray, J. Am. Chem. Soc., 12. 4250 (1950). H. Suzuki, J. Chem. Soc. Japan, 12, 825 (1951). H. Gilman and R. D. Gorsich, J. Am. Chem. Soc., 18, 2218 (1956). E. Barnett, J. W. Cook and I. G. Nixon, J. Chem. Soc., 504 (1927). F. F. Blicke and R. A. Patelski, J. Am. Chem. Soc., §§. 559 (1956). F. F. Blicke and R. J. Warzynski, J. Am. Chem. Soc., 6_2, 5191 (1940) . C. F. Koelsch, J. Org. Chem., 5, 456 (1958). C. C. Price, M. Hori, T. Parasaran, M. Polk, J. Am. Chem. Soc.,_§§, 2279 (1965). S. Patai and S. Dayagi, J. Chem. Soc., 72 (1962). (21) (22) (25) (24) (25) (26) (27) (28) (29) (50) (51) (52) (55) (54) (55) (56) (57) (58) 129 K. Nakanishi, "Infrared Absorption Spectroscopy," Holden-Day Inc., San Francisco and Nankodo Co. Ltd., Tokyo, 1962. A. A. Frost and R. G. Pearson, "Kinetics and Mechanism," John Wiley and Sons, Inc., New York, 1961, p. 14. W. N. White and C. A. Stout, J. Org. Chem., 21, 2915 (1962). H. C. Brown and Y. Okamoto, J. Am. Chem. Soc., 89, 4979 (1958). H. H. Jaffe, Chem. Revs., 55, 255 (1955). S. W. Benson, "The Foundations of Chemical Kinetics," McGraw-Hill Co., New York, New York, 1960, p. 67. S. Glasstone, K. J. of Rate Processes," 1941. p. 196. Laidler and H. Eyring, McGraw-Hill Co., "The Theory New York, New York, J. E. Leffler and E. Grunwald, "Rates and Equilibria of Organic Reactions," John Wiley and Sons, New York, New York, 1965, pp. 575-576. Ibid., pp. 188-190. D. N. Kershaw and J. A. Leisten, Proc. Chem. Soc., 84 (1960). B. M. Anderson and W. J. Am. Chem. 82, 1775 (1960). P. Jencks, Soc., F. A. Long and M. A. Paul, Chem. Rev., 51, 955 (1957). R. P. Bell, "Acid-Base Catalysis," Oxford Press, New York, New York, 1941, p. 150. E. L. Purlee, J. Amer. Chem. Soc., 81, 265 (1959). W. H. Hamill and V. K. LaMer, (1956). J. Chem. Phys., 4, 595 K. Wiberg, Chem. Rev., 55, 715 (1955). J. G. Pritchard and F. A. Long, J. Am. Chem. 6008 (1956). Soc., 18, J. N. Bronsted, M. Kilpatrick and M. Kilpatrick, J. Am. Chem. Soc., 51, 428 (1929). (59) (40) (41) (42) (45) (44) (45) (46) (47) (48) (49) (50) (51) (52) (53) (54) (55) 150 N. C. Deno and W. 5804 (1957). L. Evans, J. Am. Chem. Soc.,-12, J. E. Leffler and E. Grunwald, "Rates and Equilibria of Organic Reactions," John Wiley and Sons, New York, New York, 1965, p. 205. H. Rapoport and G. Smolinsky, 1245 (1957). J. Org. Chem., 22, M. Stiles and A. J. Libbey, J. Org. Chem., 22, 1245 (1957). H. Hart and T. Sulzberg, J. Org. Chem., 22, 1159 (1965). W. Therlacher and H. Muller, Chem. Ber., 22, 984 (1956). G. W. Kenner, B. R. Webster, M. J. T. Robinson, C. M. B. Tylor and J. Chem. Soc., 1756 (1962). O. H6nigschmid, Monatsh. Chem., 22, 568 (1901). St. Goldschmidt and W. L. C. Veer, .61. 489 (1948). Rec. Trav. Chim., R. Schlenk, Annalen, 464, 54 (1928). C. K. Bradsher, J. Am. Chem. Soc., 22, 45 (1944). H. Gilman, J. L. Towle and S. M. Spatz, J. Am. Chem. Soc.,.gg, 2017 (1946). B. M. Talbert and G. E. K. Branch, J. Am. Chem. 62, 1085 (1947). Soc., A. A. Frost and R. G. Pearson, "Kinetics and Mechanism," John Wiley and Sons Inc., New York, New York, 1961, pp. 29-50. Ibid., p. 15. S. W. Benson, "The Foundations of Chemical Kinetics," McGraw-Hill Co., New York, New York, 1941, p. 88. J. Forest and S. H. Tucker, J. Chem. Soc., 1141 (1948).