EE‘Prj.‘ - .(‘m A. \‘l/ m,“ 0".— i-‘4 - ~1f-«3— : H _.. , J, l/ , 1;: J, C a I“. w .-- -..._ n .- - *1» re; - -‘ . -_ , ”."W‘ramf :‘W'mmtum‘ “yaw-mm! «.3! 3 14 ll I-l I I I .5.‘lo"-l n|1|..‘lul\‘ln'kn . . r v. . ., .. .I . . . . . !|...|- . Iil‘ I11... 3!; . .. il-sai .ui‘lgw \s 105.1»...7“ u.lnl, . . u ' ...ls.%‘dlhfhv.r iii :1 .. . .- .. . .A 222:1... 33.2.1 _ {3:17.11qu 7 . 3.---.. ‘5 u. . .r .3... I ll . .14? IWIIV‘I « ...\.‘\A§J..“.<| Q. .. r V . n » u ’33.. ., . I. . v. .. c .,-l..al‘4..1a‘.Hn..uuq :. v .Lfa... .3. co C .- I-c-galr 2...; . . .. z. (e... . . p n.....nnnhwn...in. trunking... Ar. aunt-III' .. $14.”. _ :3... ‘.__.n............... :2 . 05.1.]! a ..a....: u...‘. . . 0. ‘0 :V . u v" . u... t l.‘hm.3.IM.-an0 .unilnu‘ u.litl oh. ,fllx.lbII thou"..- . o-IJI‘ I . . 1.: )I. . . DP. 7 0 . m“ o n .. K , . , . . 5|}, {.0 .‘. ‘ t- . '1‘“, r ‘ .I F. la a I: A l . . .l . ... ‘ . 4., .. J, .l‘ . afiF‘: 3‘...‘ Lb? «Ina! I. b «I. A. . a , . . #. I. ... Ihtu,.t¢ law . . I , t .fsl “luff a: .. "VI 93:"! State mvcrsi ty Mich U I— “2 ”M L—On , H w‘é‘ \- lafi "Hams NET}: 9611! i ...«.11.o...t§\.l‘tl¢llu.ctn|\a§u I. BaHZYLATION O? 0- CRESJL A,Thesis Submitted t3 the Faculty of the Michigan State College of Agriculture and Applied Science in part fulfilment of the requirements for the Degree of master of Science Hdbart Oscar Swartout May, 1927 fT-‘-"""‘“‘l' LO”! ’7” 5’4 7 .5; ,1 =1 v /1 ~- ~ 1" II. mmmmmr I most gladly acknowledge my indebtedness, and exprele my grat- itude, to Dr. R. G. Huston, under Whose direction the following page- were written. To his unfailing good nature. his well-timed advice, and his sympathetic encouragement, is largely due whatever measure of auc- ceae that has been achieved in the work upon which this thesis is based. H. O. awartout I i 3 18:6 9 III. F O Fifi}? T The eXperiencea chemist will doubtless find many statemente in the following pages that betray the work and the type of obeervation of a childieh dabbler, rather than a mature scientific investigator. I have written, however, more for the sake of those who will follow in my footsteps; and it may be that some Of them will appreciate the details of technique and observation that would be tiresome to a person of more eXperienoe. If I have told too much, let me plead in extenuation the fresh interest of a man with a new tadk,- his first essay in the line of research. H. o. Swartout CONTENTS PREVIOUS WORK AND GE‘IERAL PRINCIPLES Introduction Alkylation or Phenol: Benzylation of Phenols Olainen' a Work and Method The Problem Stated EXPERIMENTAL WORK ‘ Benzylation of 0- creeol by Glaioen'e .o Method Benzylation or 0- creeol by A1013 Condensation Comparative Tests and Further Data Plate 1. and Plate II. Plato III. Plate IV. Plato V. CONCLUSION «Immw 12 13 16 19 20 22 23 24 28 IV. 2 snizmnon or o-cassoi . : PREVIOUS WORK AND GENERAL PRINCIPLES Introduction ' Among the thousands of organic syntheses which have been effected during the past few decades, not the least interesting are those in which hydrogen atoms belonging to simple or substituted benzene rings have been rcplaced by aliphatic or aromatic radicles. Compounds that in a gen- eral way may be classed as alcohols or esters theoretically could be, and in fact have been, used to bring about these reactions; and water or acids have consequently been elim- inated in the process. It is easy, therefore, to see wrw dehydrating agents or catalysts have usually proved help- ful,- often necessary,- in effecting such syntheses, the simplest scheme of which is as follows: E H H H R-OH + H< Fa {K >H + 320 H H H H 3 RH R-Cl . H< I>H H H 4. HCl H - H R If one or more substituents are already present, it is .II but reasonable to suppose that they will exert an influence upon further substitution. Thus reactions concerning 0- cresol, the results of certain investigations of which com- pound have occasioned this report, depend partly upon the OH and OH} groups originally attached to the ring. Perhaps Holleman (Chem. Rev., July, 192A, 187 seq.) has best summed up present views which bear cn.this point. He quotes Boil- steinFs rule: "If a substitusnt c enters a compound 0534AB, both A and B exert an influence; but the group whose influ- ence predominates directs 0 to the place it will occupy.” Then he lists the common groups which direct to p- or 0-- positions, in the order of the speed of the reactions which they produce, which of course determines their comparative directing influence: 03> N32> I >Br> 01>CH3. From this series it is clear that in o‘cresol the influence of the OH group is greatly predominant. and that it will direct fur- ther substituents to positions p- or o- to its own. These facts lead to the idea that the OH group, which is characteristic of phenols, should cause,§l; of them to react in an analogous manner, except in cases where further complications exist. Possible complications are, of course. not few; but to discuss them in detail would add.very little to the elucidation of our subject, and it would lead us too far afield. Because of the analogy among phenols, however, it will be worth while to note some of the work that has been done in alkylating or benzylating them, as.an introduc- tion to the problems connected with the benzylation of o- cresol. Alkylation of Phenols In 1881 Liebmann (Ber. 14, 1842) prepared a butyl phe- ' nol by condensing phenol and isobutyl alcohol with the aid of molten Zn012. He also prepared propyl phenol and amyl phenol by the same method. The next year*Mazzara (Gazz. 12, 505) condensed prepyl alcohol and.m-cresol with Mgclz, and a year later (Ber. 16, 242) he used the same catalyst in pre- paring methyl butyl phenol. In 1884 Auer (Ber. 17, 669) ob- tained an ethyl phenol by condensing absolute alcohol and phenol with ZnCle. Six years later Dennstedt (Ber. 23, 2569) prepared the same compound, but he used zinc dust in- stead of ZnCle, and he considered that his compound was a mixture of two of the three possible isomeric ethyl phenols. In 1894 Bauer (Ber. 27, 1614) condensed isobutyl alco- hol and o-cresol with the aid of Zn012 and heat, preparatory to the synthesis of a.nitrated isobutyl o-cresol. In 1895 Anschutz and Beckerhoff (Ber. 28, #08) prepared amylcppenol by condensing both isoamyl alOOhol and tertiary amyl alcohol with phenol. They concluded that the two products were the same, citing similarity of melting points, of boiling points, and of benzoyl derivatives as proof of the identity. In 1904 Clemmenson (Ber. 37, 54 seq.) prepared ethyl resorcinol, ethyl hydroquinone, ethyl pyrocatechol, and both mono-ethyl and di-ethyl pyrogallol, by reducing the appro- priate hetones with zinc amalgam and 1:1 or 1:2 861. In 1907 Herzig and Wenzel (Monatsch 27, 781) stated that they had.methylated phenols by treating them with CH3I in.alkap line solution. In 1913 Johnson and Hodges (J. Am. Ch. Soc. 35, 1014) used Clemmenson's general method to prepare a few alkyl phenols. They worked with others as well as ketcnes. The next year Johnson and Kohman (J. Am. Ch. Soc. 36, 1259) continued this line of research and succeeded in preparing alkyl phenols with long aliphatic chains. In 1889-‘90 Gattermann, Ehrhardt, and.Maisch (Ber. 22, llg9} 23, 1199) prepared condensation products of anisol, phenetol, etc. by treating them with various acyl chlorides and with benzoyl chloride in the presence of A1013. They expressed their belief that the constitution of the con- densation product or anisol and acetyl chloride is: 0 H3°'°‘<::::>'°<533 Their proof is the fact that the oxidation of this com- pound gives anisic acid, or p- methoxy bensoic acid. as it may as properly be called. The other condensation products which they prepared gave analogous oxidation products, and on these grounds they repeatedly stated (Ber. 23, 1203. 1204, 1205, 1208, 1210) that the acyl or benzoyl groups regularly entered the ring in the p- position. In 1891 Senkowski (Ber. 24, 2924) reported that he had succeeded in alkylating aniline and some other aromatic compounds. He stated_that in such reactions the new sub- stituent always takes the p- position, and that the same is true of the higher homologues of phenol, which are prepared by treating a mixture of phenol and the appropriate alcohol with ZnClg. He thus extended and confirmed the theory that was suggested by the work or Gattermann and his helpenp. Benzylation of Phenols Some of the earliest work which has a bearing on the benzylation of phenols was done by Kollaritz and.Merts from 187l-'7}. (Ztschr. Chem. 1871, 705; Ber. 5, 447: 6, #46) They succeeded in synthesizing diphenyl ketone, and they stated that they had worked according to principles govern, ing the condensation of aldehydes, ketones, and phenols, us- ing phosphoric anhydride as a dehydrating agent. In 1872 Paterno (Gazz. 2, 20) prepared benzyl phenol by treating phenol with benzyl alcohol or with benzyl chloride in the presence of zinc turnings. He also bensylated anisol. This work was repeated in 1875 by Paterno and Filetti, (Gazz. 5, 381) using a mixture of acetic and sulfuric acids instead of zinc turnings. Three years later Paterno and.Haasara, (Gaza. 8, 303) aeain using zinc as a catalyst, condensed cresol with benzyl chloride. In 1880 0. Fischer (Ann. 206, 113) condensed.bensyl al- cohol and dimethyl aniline with ZnCle, also benzhydrol and dimethyl aniline. He found that he could use either 211012 or phosphoric anhydride to assist in the reaction. While this was not a case of’benzylating a phgggl, yet it is of interest because of the constitutional analogy between ani- line and phenol. As already noted, Liebmann prepared.several alkyl phe- nols, using Zn012 as a catalyst; but he also prepared a ben- zyl phenol. (Ber. 14, 1844; 15, 152) Incidentally he raised the question as to whether the zinc used by Paterno was the active catalyst, or whether it was Zn012 formed from the zinc and the free H01 always found in benzyl chloride. He referred to some of his own eXperimental observations as an evidence that a very small quantity of Zn012 could eat- alyse the reaction. In 1881 Merz and Weith (Ber. 14, 187 seq.) tried the street of both ZnCle and A1013 on phenol. The result was not a benzyl phenol, but a diphenyl ether. Their work is of interest in connection with our subject, however, for it marks one of the first uses of A1013 as a catalyst or dehydrating agent in reactions concerning phenol, being eight years in advance of the work of Gattermann.and.his associates, who used this catalyst in effecting the syn- theses already mentioned in connection with their names. In 1909 Khotinsky and Patsewitch (Ber. 42, 3104) called attention to the fact that aromatic tertiary car- binols may easily be condensed with many substances, in- cluding phenol, by the aid of acetic acid to which a little H2804 or Zn012 has been added. But the condensation of aliphatic, or aromatic, or mixed aliphatic and aromatic alcohols with aromatic com- pounds, using A1013 as a catalyst or dehydrating agent,- its exact office is not yet fully known,- has, so far as I can learn, not been reported by any investigators except Dr. Huston and his helpers at Michigan state College. Reé ports of their work have appeared from time to time since 1916. (J. Am. Ch. Soc. 38, 2527; 40, 785; 46, 2775; 48, 1955) Only the 1924 report, however, dealt with phenols. Condensations of benzyl alcohol with phenol, with anisol, and with phenetol, also of benzyl chloride with phenol, were effected, A1013 being used to aid in the reactions. The opinion was expressed that the benzyl phenols thus pre- Pared were p- compounds, in agreement with the theory of Gattermann, Senkowski, and others, as previously noted. Claisen's Work and.Method The work of Claisen deserves particular notice. He has prepared both alkylated and benzylated phenols by a method which is fundamentally different from any that have thus far been mentioned, and his discussions throw much light upon the question as to whether his products are identical with, or isomeric with, those prepared by ordi- nary condensation methods. I shall refer only to his 1924 report, (Ann. 442, 212 seq.) as that is sufficient for our present purpose. . Claisen generally uses an alkali metal salt of a phe- nol, dissolved or suspended in a suitable medium, and treats it with an alkyl or benzyl halide. The result is usually a mixture, containing a mono- alkyl or benzyl phenol, an other which may be removed by extracting with ligroin, and a small proportion of a di- alkyl or benzyl phenol. When suitable precautions are taken the chief product is the mono- alkyl or benzyl compound. A brief consideration of s’ ,t x -s E . ~- a t I s — l . r A 1 t \ .94 I n. 7 I I . \ A I .‘ > . . , , . , 1 ,. ' . . { ‘ \/ . . ' . ‘ . .1 ‘~ " -- t, * ° . g ‘ . O . . . . » .‘ v . . . - . d N a . ' ~ ‘ ~. - .‘ ‘: ~ g. .. . | u 4 - - . .s . V . 4 u , w ‘I L . . . - . . , . , I ' » e ' I J u ' i .' k . .4 ._ '. A s ‘ . __ _ '7 O l . . a . _ the main reactions involved will make Claisen's discussions more intelligible: H H H H H< >0H 1» Na :3 H<:>0Na + H H H ' H H H -H H H H H H n<‘_‘:>o::a + org-(:33 z No.31 + HQ-o-g- H H H H s H a a H s H H H or HQ-fiQz-I Hog H Benzyl Phenol Phenyl Benzyl Ether A brief inspection of these equations shows why Claisen said that the‘gthgg is the product one would eXpect to be formed, and that the benzyl phenol is the result of "anoma- lous metal substitution", or, as he otherwise eXpresses it, the result of ring alkylation, when one would.naturally think that the benzyl group should take the place of the metal atom that is displaced. Why in the course of the re- action a hydrogen atom from the ring should give up its place to the benzyl group, and why the OH group should be re-formed in the process, are somewhat difficult questions to answer; but it is evident from the fact that a consider- able quantity of the mono- benzyl compound is regularly formed that this is exactly what occurs.- Claisen's discussion of this point, in the course of which he stated his belief that the substituent group in these cases always takes a position 0- to the 0H, was based on Michael's theory of the reaction between silver cyanide and methyl iodide. (J. pr. 37, 486; 46, 189) The following reaction scheme will help to make this theory clear: fiflks 1- CH3I 7-: fi<§8 a {'1'} + A51 N-CH; ' N-CH3 It can be seen that a shift in valence bonds occurs, that the CH3! separates, one part going to the carbon and one to the nitrogen, and that Agl is finally split off, leav- ing the CH3 attached to the nitrogen, instead of to the car- bon as we might expect. Claisen applies this principle to the reaction between an alkyl iodide and a metal derivative of an unsaturated organic compound, as follows: . /1 -c-0Na _1__R'I__fl___, '°\0Na - NaI -c=o -c(os) -éa .. HR -éna -t-R I In sodium phenolate, used above to illustrate Claisen's general theory, we find the group 28§ON33 and benzyl chloride is analogous to an alkyl iodide. We should expect, then, Just such a shift of hands, a splitting of benzyl chloride into parts which attach themselves to different carbon atoms, the elimination of NaCl, and the final formation of a com- Dound with the OH group restored and the benzyl radicle at— tached to a carbon atom in the ring, instead of to the oxygen. atom. Probably such a shift of bonds could come about only in the case of two adjacent carbon atoms, which would direct the benzyl group to a position 0- to the OH group. Thus, through Claisen's method of preparing alkyl or benzyl phenols, we are reasonably sure of obtaining prod- ucts with the OH group and the new substituent occupying 10 positions adjacent to each other in the ring. And, in this connection, we should give attention to his opinion that ordinary condensation methods are not sure to give pure p- products, as Gattermann and others seem to think, but are likely to produce mixtures of p- and 0- compounds, as might be predicted from Holleman's discussion of the general prin- ciples governing substitution in the benzene ring. Glaisen' s enumeration of the factors which affect the course of his ”anomalous" reactions and the preportions of the various products formed is also worthy of note. He says they are: - l. The kind of phenol. 2. The kind of alcohol,- whether saturated or unsaturated, aliphatic or aromatic. 3. The kind of halogen. 4. The kind of metal, in the metal phenolate. 5. The temperature. 6. The meditn in which the reaction occurs. It is not necessary to discuss these factors at any length; but they have a bearing upon the present problem, because Claisen' s method is one way of benzylating o- cre- sol, so a few comments will be made. Naturally a complex phenol would introduce more complications than a simple phenol. As to the second point, Claisen worked mostly with unsaturated alcohols, indicating that he found them best. suited to his purpose. He stated that the 2422.99: the bond between the halogen and the alkyl radicle the more noothly V I‘ll. illrllllil..'|\|z 11 the reaction progressed. We know that unsaturation in the alkyl radicle results in a comparatively loosely held halo- gen. The kind of halogen also has a bearing on the strength of such bonds. Points four and five hardly need comment, but not so with point six. Claisen found that the influp ence of the medium was uneXpectedly great, especially upon the proportions of the ether and the mono- alkyl or benzyl compound in the reastion.mixture. The use of a diggggiaplg medium gave over 90% of the ether, while a' o - sociabl medium, such as toluene, resulted in a yield of 60-70% of the ”anomalous" alkyl or benzyl phenol. While it is not easy to explain how it is so, it is probably true that such media as toluene cause a loosening of the valence bonds be- tween the alkyl or benzyl radicle and the halogen, thus fa- cilitating the "anomalous" substitution. Before making a definite statement of our problem, it is necessary to call attention to a reaction carried out according to Claisen's method by Schorigin in 1925. (Ber. 583, 2033) Be prepared sodium 0- cresolate, suspended in toluene as a medium, and treated it with benzyl chloride. He named the crystalline product which.he obtatned 2-methy1 6-benzy1 phenol, and stated that it melted at 51-52° and boiled at 187-188° (15 mm.) It will appear later that his benzyl phenol is identical with that which we prepared for ‘ the purpose of comparison with the condensation product of benzyl alcohol and o- cresol with A1013. Our work was begun before Schorigin's report came to our attention, however; and we record some additional data regarding the compound, which he evidently did.not study extensively. The Problem Stated There are, then, two general methods by which 0- cre- sol can be benzylated. Glaisen’s method should give an 0- product, which might preperly bear the name which ashor- 1gin gave it. The product of the A1013 condensation of benzyl alcohol and o- cresol should, according to Gatter- mann.and others, be a p- product, which.might be named 2~methyl 4abenzyl phenol: but if Claisen and.Holleman are right, it should be a mixture of the p- and 0- compounds, the p- predominating. Our problem, therefore, has been to benzylate o- cre- sol by the two methods; to purify the products; to look for any evidence that either of them is a.mixture3 to de- termine whether or not they are identical, and, if not, to distinguish them from each other by appropriate means, sudh as melting points, boiling points, solubilities, crystal forms, and the character and.behaviour of some of their simple derivatives: and to note the bearing of previous work and our own findings on the configuration of the mol- ecules of the two compounds. All these details, together with a discussion of certain.by-products, are given in the following sections. DERIMENTAL W0 RK Benzylation of 0- Cresol by Claisen‘s Method % mol of freshly chipped sodium was suspended in 130 g. of toluene in a l-liter Florence flask and treated in the cold with } mol of o- cresol. The reaction began at once. Heat was evolved and.hydrogen given off. The reaction seemed to be complete in two hours, but'the flask was fit- ted with a reflux condenser and gently heated for an hour longer. When the contents of the flask had cooled to room temperature, } mol of benzyl chloride was added through the condenser.[/£ gentle reaction began, the light gray, pasty mass of sodium cresolate dissolving smoothly to a reddish- brown liquid with a precipitate of Real. The mixture was let stand over night and then heated on an oil bath for 5 hours at 150-16GO.' After cooling, the NaCl precipitate was dissolved out by washing twice with water in a separatory funnel, and the trace of water and the toluene removed by distilling the mixture, stopping the distillation at 125°. The residue, consisting of a mixture of the reaction products, was dissolved in 250 cc. of Claisen's methyl al- coholic potash, (Ann. 442, 224) and shaken out with 200 cc. of ligroin in 50 cc. portions. The ligroin used.here and throughout the eXperimental work was Central Scientific Company's 40-500 petrolic ether. The purpose of this treatment was to form the ligroinpinsoluble potassium salt of the substituted phenol, so that the ether fraction and 14 other substances in the mixture could be separated from it by extraction. After the removal of the ligroin exp tract by means of a separatory funnel, the residual brown solution was made distinctly acid with 1:1 H01, thus pre- cipitating K01 and re—forming the free substituted phenol. This phenol was extracted three times with ether and the ether was removed by heating on a water bath, the residue being a reddish-brown oil. A further quantity of both the ligroin and the ether extracts was prepared by repeating the procedure described above, the corresponding extracts being united and frac- tionally distilled, with the following results: Ligroin Extract 250-280° (Atmospheric pressure) 16.0 g. No further study of this product, a yellow oil, was made, as it was probably 2-methyl phenyl benzyl ether, a compound in which we were not directly interested, except to get some idea of the proportion of it formed during the M W W, at £450,130?“ Wfluwgz, yum/n5» Ether Extract ragotion. 16.0 g. is 8.1% of the t eoretical yield. Third fractionation,- 5 mm. 90- 1400 15.0 g. 140-155o (Mostly 150-1520) 62.0 g. 155-250° 25.0 g. Residue (Black tar) 29.0 g. Fraction 140-155o solidified in the receiver and was . Ill-Ill.“| lin‘lllll ill}! 15 recrystallized from ligroin to constant melting point. No change occurred after the fourth crystallization. Ten trials of the melting point were made. It was found to be 49.5-50.5°, which differs from that reported by Schorigin by l.5°. A test made with the purified crystals gave the boiling point of the compound as 150-152° (5 mm.), as com- Pared with sshorigin's reported 187-188° (15 mm.); but the differences are not great enough to raise any doubt as to the identity of the compounds, for the reacting substances were the same and the procedures much alike. The yield, 62.0 g., is 31.3% of theory. As a means of shortening the time needed to secure purified crystals the following plan was pursued: At each step in the recrystallizing process the mass of crystals was dissolved in the minimum possible quantity of boiling ligroin, the solution was set in the coldest possible avail- able place, and the mother liquor was pressed out of the thick felted mass of crystals which soon formed under such conditions. This mother liquor was allowed to evaporate, yielding comparatively impure crystals, while the felted crystal mass was dissolved in fresh ligroin and the solution evaporated, yielding much purer crystals. Before evapora- tion had proceeded to dryness, in all cases the last few cubic centimeters of mother liquor were poured off into a beaker containing a less pure product. In this way a few \ \,.‘ 16 days' time was sufficient for the production of crystals of constant melting point. Benzylation of 0- Cresol by A1013 Condensation A.mixture of 100 g. of redistilled o- cresol and 100 g. of pure benzyl alcohol was suspended in 200 g. of ligroin in a tall l-liter condensation Jar, Which was set in a water bath as a means of controlling the temperature. While cons stantly agitating the mixture by means of a motor stirrer, its temperature was gradually raised to 35° and maintained as nearly as possible at that level while 65 g. of anhydrous A1013 was slowly added in small portions over a period of about an hour. The whole Operation was carried out in the hood on account of the capious evolution of H01 gas which accompanied the reaction; and at times troublesome frothing occurred, which could be controlled only by the addition of further portions of cold ligroin. When the evolution of gas had nearly ceased, the mixture was let stand over night in the hood. The crude condensation product was a light grayish! brown, semi:gummy, semi-granular precipitate. It was de- composed by mixing it with shaved ice in a large beaker, adding a little 1:1 H01 toward the end to aid in the separ- ation. The mixture was extracted three times with ether, the combined ether extracts let stand several days over anhydrous K2003, and the dried solution freed from ether 17 on the water bath. Two more lots were made up in the same way, and the ether-free residues were combined and frac- tionally distilled. Fifth fractionation,- 5 mm. 90 -150° 36.7 s- 150-160° lie.) g. 160-180o (Mostly below 170°} 33.6 g. 180-2150 9.8 g. 215-235° (mostly at 225-23o°) ' 77.2 g. 235-280° 29.2 g. Residue (Brown tar) 11.2 g. Fraction 150-160o crystallized in 30 minutes, and fraction 160-1800 in four hours, in the ice-box. The crystals were pressed between filter papers and repeatedly recrystallized from ligroin, using the method that has been described in connection with the discussion of the Claisen’s method product. Six recrystallizations were necessary before a substance of constant melting point was obtained. The melting point was found to be the same as - that of the other compound,- 49.5-50.5°.- but When the purified crystals were tested for boiling point, this turned out to be 167-169° (5 mm.). This was somewhat of a surprise, since the greater portion of the crude product had come over below 160°. The yield of the two fractions, 150-160° and 16c-180°, a total of 151.9 3., is 27.6% of theory, though one is not Justified in classing all of 18 this total as the yield of the pure p- compound. The first direct evidence, aside from the question of boiling point, that the two products were not identical was an unmistakable difference in solubility and in crystal form,- points which will be more fully treated in a later ' section. Then it was discovered that the mother liquor, or washings, obtained at the beginning of the second crystal- lization of the A1013 condensation product produced crys- tals very similar in form to those of the Claisen product; and a test of these crystals showed that they consisted chiefly of a substance whose boiling point was 150—152° un- der 5 mm. pressure, thus affording further evidence of the identity suggested by the crystal form. These facts, together with the additional fact that the entire fraction of the A1013 product from 150-1800 formed crystals which were comparatively difficult to pur~ 1fy to constant melting point, are evidence that this prod- uct was a.mixture: and, While more proof will be given.lat- er, we may say at this Juncture that the mixture consisted of a large proportion of the p- product and a small propor- tion of the 0- product. Not having foreseen this outcome sufficiently early, it was too late to determine definitely the preportions of the two products by means of fractional distillation, so that a clearer idea of this point must come from a repetition of the experiment. l9 Comparative Tests and Further Data Reference has already been made to some evidence that the two compounds which were the chief subject of our in- vestigation are not identical. For the sake of brevity the A1013 condensation product will hereafter be designated by A.and the Claisen's method product by.§. We have seen that A and‘g‘have the same melting point, but that their boiling points differ by 17°, and that there is an apparent difference of crystal form. Tests of solubility in ligroin were made. 50 cc. of solutions of each compound, saturated at 20°, were pipet- ted into tared beakers , evaporated to dryness, and the weight of the residues ascertained. These are the results: .A 1 Trial I. .5715 g. Trial II. .5743 g. ‘g Trial 1. 2.7210 5. Trial II. 2.7392 g. Thus we see that'g is approximately five times as soluble in ligroin as‘A. A.more concrete idea of comparative sol- ubility may be obtained by looking at Plate 1., page 20, which is a photograph of the two beakers containing the residues of Trial II. This solubility test was also applied to the small quantity of impure crystals which came from the first washings of‘g, which had the same boiling point and crys- tal form asig. The available quantity was not sufficient to provide 50 cc. of solution, so 25 cc. was used. The residue weighed 1.3796 g. on first trial and 1.3889 g. on Plate I. Views of beaker: containing the dried res- idues from ligroin solubility tests. flate‘II. Above and to the right are shown natural size views of the crystal forms as they collect on the sides of the beakers during crystaliz- ation. The photographs were taken with the camera lens pointing obliquely downward toward the bottom of the beakers, A and the tangled masses of A and the bundles and parallel arrangement of B are both clearly seen. 21 second trial. Since these crystals still contained some oily impurity, which was quite soluble in ligroin, the re- sults check closely enough with those of the other solubil- ity tests to amount to further evidence that the crystals were identical with those of.§. A difference between‘g and g which may not seem to be of much significance at present, but which may at some time prove to have a bearing on the relation of color to Chemi- cal constitution, was also noted. While both compounds were colorless when freshly prepared, upon standing for some time, especially if exposed to the light,‘A began to develop a trace of pink color, and‘g of yellow. The pres- ence of the oily impurity normally found in the crude prod- ucts greatly hastened the appearance of these colors. The crystal forms characteristic of‘g and g are hard to describe adequately. .5 gives White, flattened.needles, or elongated platelets, with a hard, glassy glitter and a tendency to form in.much¥tangled masses. .g gives long creamy-white needles, or fibers, with a soft, silky sheen and a tendency to lie parallel or form in bundles. This parallel arrangement can best be seen on the side of the beaker Where the fibers cling during the process of crys- tallization. Plates II., III., IV., and V., found on pages 20, 22, 23, and 24, are natural size photographs s; crystal masses formed under comparable conditions, and they help to give a clearer idea of comparative crystal forms. 22 Plate III. Natural size views of the crystal forms of A and B as they appear when looking down into a large beaker. The crystals shown in this plate were formed from saturated ligroin solutions in a room where the temperature was about 25°, and the air circulation good. Being formed rapidly, they are comparatively small and fine, those of B felting to- gether somewhat. 23 These views were taken in the same way as those of Plate III, but the crys- tals were formed at a temperature of about 20° instead of 25°. Since evaporation was slower, there was more time for the crystals to build up into larger forms. In B there is no further sign of felting, but the crystal bundles begin to appear. This plate gives perhaps the clearest idea of the difference between the crystal forms of A and B. The photographs pre taken in the same Way as those of the two previous plates. The crystals were formed slowly. Start- ing with solutions hardly saturated at 10°, placing the beakers in a cool place,- 15° or under,- and adding a little more saturated solution from time to time after the crystals began to form, al- lowed the crystals to build up to such a size as to enable one to discern their peculiar characteristics very plainly. 25 Further attempts were made to distinguish‘g from‘fi by means of bromine and benzoyl derivatives. Here consider- able difficulty was encountered. The bromine derivative of .g was readily formed by treating it in Gael} solution with an equivalent quantity of bromine. When recrystallised from ligroin it formed characteristic rosette-shaped crys- tale with a.melting point of 63-64°. The reaction in can3 solution seemed to take place with‘g as readily as with‘g, but no crystals would form. Finally, in order to get some comparative data, larger quantities of the bromine deriva— tives were made and tested for boiling points. That of‘A boiled at 180-182°, (5 mm.), but that of.§ proved to have the uneXpectedly high boiling point of 2fl-2§?° under the same pressure. This great difference in boiling points led to a question as to Whether the two compounds were really analogous bromine derivatives; and to settle this question a Parr bomb bromine determination was made, following the general method suggested by Lemp and Broderson. (J. Am. Ch. soc. 39, 2069) The following are the results: Sample AgBr % Bra ‘5 I. .2582 g. .1763 g. 28.68 II. .2071 g. .1426 g. 28.87 .p I. .2259 g. .1550 g. 28.82 II. .2867 g. .1961 g. 28.65 Theory,— for mono- brom derivatives,- is 28.48% of Bra. This determination, therefore settled the question as to 26 the relation between the two compounds, indicating clearly that they were of the same empirical composition, and, taken in connection with their greatly differing boiling points, affording more evidence that the phenols from Which they were derived, while doubtless isomeric, were surely not identical. The benzoyl derivative of.g was prepared without any difficulty. The phenol dissolved readily in an equivalent amount of 5% KOH. When this solution was treated with the calculated quantity of benzoyl chloride, a sticky mass of pale pink oil was formed. This gradually hardened When put in the ice-box. When washed with water till free from K01, dissolved in ligroin, and purified by recrystallization from the same solvent, it formed small, colorless, transparent, rhombic crystals, with a melting point of 54-55". The Quan— tity available was too small to permit of making a boiling point test. On the other hand, all attempts to prepare a crystal- line benzoyl derivative of‘g resulted only in the formation of benzoic acid crystals and a small quantity of a nearly colorless 011. .p would not dissolve in 5% KOH, even when five times the calculated quantity was used. It would dis- solve in an equivalent amount of 50% KOH with a little warm— ing, but separated again when the solution was diluted to 25% or less. The solution in 50% solidified on cooling, but appeared to react readily with benzoyl chloride if warmed again. 27 While the attempt to form a crystalline benzoyl deriv- ative of'g failed, yet the visible evidence that a reaction always occurred, and the presence of the oil along with the benzoic acid crystals, led to the idea that fractional dis-I tillation of this oil might yield a definite compound whose boiling point would distinguish it from other possible sub- stances in the mixture, and thus establish it as the sought- for benzoyl derivative. Consequently a new and larger lot of the derivative was carefully prepared and fractionated. It yielded about 3 g. of a very viscous, light yellow oil, boiling sharply at 216-2180 (5 mm.). This did.not corres- pond with the boiling point of the original phenol, or of benzoic acid, or of any other known substance which could be present in the reaction.mixture, hence there was no hesitation in pronouncing the oil the benzoyl derivative of p. A study of the reactions Which resulted in the forma- tion.of.A.and,§ led to the conclusion that they were iso- meric, if not identical. More evidence of their isomerism has been found, notably the bromine content of their bros derivatives. Still further proof was afforded by combus- tion tests made on the two compounds: Theory: Guano: c. 84.8% s. 7.12% Found: Subst. 002 H20 %G 7:11 A (Huston) .1523 g. A746 g. .0971 g. 84.99 7.19 .5 (Swartouw .1443 g. .4473 g. .0915 a. 84.50 7.07 28 A further word should be said about by-products. The ether which was formed along with‘g has already been men- tioned. But during the distillation of the mixture which contained.§, no evidence of the presence of any considerable quantity of any substance with a definite boiling point above 155° was seen. This was not true with the A1013 con- densation product. The large fraction boiling between 215° and 235° was significant, so this fraction was made the basis of further fractional distillation, nearly the Whole of it finally coming over at 225-2270 (5 mm.). It was a thick yellow oil, which slowly darkened to an orange color. The , yield, 77.2 g., was 19.0% of theory. This oil could not be X made to produce crystals. An account of its behaviour to- ,0 ward alkali and beniyl chloride will be found in the conclup sion, because of its bearing on the question of molecular configuration, which is discussed there: but it may be said here that we were led to believe the oil to be 2-methyl #- 6-dibenzyl phenol. A combustion made to test this idea gave the following results: Theory: 021H200: 0, 87.45% H, 7.00% Found: Subst. 002 H20 %0 %H .#294 g. 1.372} g. .2686 g. 87.62 7.01 These results amount to a confirmation of our Opinion. CONCLUSION 0- cresol has been benzylated by two different methods, yielding products which have been designated by A md B. That A and.§ were isomeric is not only theoretically 29 probable, but it has been proved true by analysis of bran! ine derivatives and by combustion tests. That they were not identical was proved by different boiling points, different solubilities in ligroin, different crystal forms, and dif- ferences as to the formation and character of their bromine and benzoyl derivatives. But it was also proved that the A1013 condensation process resulted in the formation of a small proportion of a compound identical with.§ and a larger prOportion of a high-boiling substance, along with the main product, Which was designated byié. That this high-boiling compound was a phenol, and not an ether, was proved by its forming a solution in Claisen‘s methyl alcoholic potash, from which solution ligroin would extract nothing. We were unable, however, to prepare its benzoyl derivative, for though it slightly dissolved and turned blue and gummy in hot 50% KOH, benzoyl chloride did not react with it. It was upon these grounds, in addition to the combustion test, that this compound was named 2- methyl #- 6-dibenzyl phenol. According to the views of Claisen and other investi- gators, A was a p- and.§ an 0— product, and we now have added evidence that these views are correct. The two sub- stances were not identical, and one of then must have been the 0- compound. The comparative difficulty of preparing the benzoyl derivative of‘g was strong evidence that this / 30 was the one, for such a derivative must have the following , configuration: H CH3 The difficulty of overcoming steric hindrance and heaping the methyl, the benzoyl, and.the benzyl groups on adjacent carbon atoms must be considerable. And we should expect the same difficulty with the dibenzyl derivative, whose benzoyl derivative must be constituted thus: H on, H