Chloride. by JOHNEDWIN MACKENZIE, D.Sc., PI1.D

View Article Online / Journal Homepage / Table of Contents for this issue

I204 MACKENZIE: THE ACTlON OF SODIUM MEETHOXIDE ON Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42.

CXXVIIL-The Action of Sodium Nethoxide and its Homologues on Benxophenone Chlo&le and Bmxal Chloride. By JOHNEDWIN MACKENZIE, D.Sc., PI1.D. INa previous communication to the Society (Trans., 1896, 69, 985), a description was givsn of the preparation of dimebhoxydi phenylmethane by the action of sodium methoxide on benzophenone chloride according to the equation (C,H,),CCI, + 2CH,*ONa= (C,H,),C(O*CH,), + 2NaC1, and of di-ethoxy- and dibeazoxy-diphenylmethane,using sodium ethoxide View Article Online

BENZOPHENONE CHLORIDE AND BENZAL CHLORIDE. 1205

and benzoxide respectively instead of methoxide. In continuing this investigation, dipropyloxy-, diisobutyloxy-, and dihydroxy-tetraphenyl- methane have been prepared, and for the sake of comparison a number of experiments with benzal chloride have been carried out. The action of sodium methoxide and its homologues on benzal chloride was first investigated by Wicke (Anncden, 1857, 102, 356), who prepared dimethoxy-, diethoxy-, and diamyloxy-benzylidenes, the reaction pro- ceeding thus : C,H,*CHCI, + 2CH,*ONa = C,H,-CH(O*CH,), + 2NaCI. Some years later, Limpricht, in a research on the chlorine substitution products of toluene (Anncderh, 1886, 139, 319), repeated Wicke's work, but was unable to obtain products free from chlorine. This being so, and no record of the yield of the products obtained by Wicke being given, it was thought advisable to try these experiments agzin under various conditions. Following Wicke's instructions, it was found that the reaction took place under the ordinary pressure as represented, but that the benzal chloride only disappeared on repeated treatment with sodium methoxide." On the other hand, when benzal chloride mas heated with sodium ethoxide under pressure, a different reaction took place, ethyl chloride, benzaldehyde, and sodium chloride being the products. The change may be represented thus : C,H5*CHC12+ C,H,*ONa = C2H,C1 + C6H5*CH0+ NaCl. By using the sodium compounds of different alcohols, the corresponding chlorides mere produced. That this reaction does not apply to phenols would appear from the work of Fosse (Compt. rend., 1900, 130, 725 and 1194), who obtained diphenoxyethylidene by heating ethylidene chloride and sodium phenate in sealed tubes at 120°, according to the equation,? Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. CH,*CHCJ, + 2C,H,*ONa = CH,,*CH(O*C,H,), + 2NaCl. As in benzophenone chloride and benzal chloride both halogen atoms are attached to the same carbon atom, it was thought that itl would be interesting to see what happened when the halogen atoms

* In the case of sodium benzoxide, the only product separated was stilbene, and that only in minute quantity, The reaction would thus appear to be expressed in part by the equation : 2C,H5*CHCI, + 4C,H;CH;ONa = 4NaC1 f C,H,*CH:CH-C',H, + 2CGH,*CH,*OH -t- C,H,*CHO. .t. That this should be so is rather curions in view of the fact that when benzal chloride and phenol are heated together, they form clihydroxytriphenylmethane according to the equation : CGH,*CHCI, + BC,H,*OH = CGH,*CH(CGH,'OH), + 2HCl9 and that benzophenone chloride gives n similar componiid with either phenol or s'oclinm phenoxidc. View Article Online

1206 MACKENZIE: THE ACTION OF SODIUM METHOXIDE ON

are attached to different carbon atoms. In the case of ethylene dibromide, the reaction would be expected to take place thus : CH2Br*CH2Br + 2C,H5*ONn = CH,(O*C,H,)*CH,(O*C,H,) + ZNaBr, but no record of experiment showing this to be so is known to the author. Diethoxyethylidene has, however, been prepared by Henry (Compt. rend., 1885, 100, 1007) from an intermediate product in this way: CH,I*CH,(O*C,H,) + C,H;ONa = CH,(O~C,H,)*CH,(O*C,H,) t NaI. On heating ethylene dibromide with sodium ethoxide under pressure, however, it was found that acetylene was produced, reaction taking place thus : CH,Br*CH,Br + 2C,H,*ONa = CH:CH + 2C,H,*OH + 2NaBr. The production of acetylene in a similar fashion, using sodium isoamyloxide instead of ethoxide, is recorded by Sawitch (Juhresber., 1861, 646).

I. Action of Xodium AlkyZoxides on Benxoyhenons CI'LZoride. Diln.o~~yZoxydipT~enzylmetl~url.le,(CGBs),C!O.C)H,*CHz'C)H,),. In preparing this substance, 3.7 grams of sodium were dissolved in 40 grams of propyl alcohol, and to the cold product 19 grams of benzophenone chloride were added. The mixture was gradually heated to a temperature of 95O, when a violent action took place, and then kept at 120° for 7 hours, at the end of which time the supernatant liquid still gave an alkaline reaction. The liquid was filtered from the salt which had separated, the latter being washed three times with small Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. quantities of propyl alcohol. The filtrate mas then distilled under reduced pressure to remove the excess of propyl alcohol. The residue, which was a viscous, brown oil, was diluted by the addition of a little ethyl alcohol, filtered from some salt, and distilled, 7.9 grnma of a colourless oil boiling at 200-210° under 40 mm. pressure being collected. The yellowish, semi-solid, strongly alkaline substance remaining in the distilling flask was distilled with steam, From the milky distillate, crystals separated which melted at 32-34O, and boiled at 204O under 40 mm,pressure. They amounted to nearly 1 gram. By cooling the fraction boiling at 200-210° under 40 mm. pressure in a freezing mixture, a white, powdery mass wasobtained, which on recrystallisation from light petroleum formed colourless, prismatic crystals melting at 33-34.5'. On analysis : 0.2026 gave 0.5982 GO, and 0.1576 H,O. C= $0.53 ; H= 8.64. C19H2402requires C = 80.28 ; H = 8.46 per cent. View Article Online

HENZOPHENONE CHLORIDE AND BENZAL CEILORIDE. 1 20'9

Dipropyloxydiphenylrnethane is insoluble in water, but extremely easily soluble in methyl or ethyl alcohol, ether, benzene, light petroleum, $c. It is rather unstable, the faces of crystals very soon becoming dull. Well formed crystals are very difficult to obtain. The amount of the fraction boiling at 2UO-21O0 under 40 mm. pressure corresponds to 35 per cent. of the theoretical yield.

The preparation of this substance mas effected in the way previously described, the quantities taken being 3.7 grams of sodium, 41 grams of isobutyl alcohol, and 19 grams of benzopheaone chloride, When the sodium compound melted, a brisk action took place, but was not complete after heating to 130--150° for 7 hours, the liquid still showing an alkaline reaction. The salt was removed by filtration, and the liquid then distilled under diminished pressure. After removal of the excess of isobutyl alcohol, the temperature rose rapidly to 199O, and 16.5 grams of a thick oil were collected, having a boiling point between 199' and 210" under 35 mm. pressure. When placed in a freezing mixture, this oil became a thick jelly in which minute needles appeared on scratching, After standing at the ordinary temperature for 18 days, the small bunches of needles, which had grown, were filtered off, dried on a porous plate, and, as they showed no sharp melting point, dissolved in ether., The ethereal solution deposited silky needles after being kept in a desiccator for a week. The dried crystals, which weighed 0% gram, now melted at 62--64O, but on exposure to the air the melting point fell, and, as mill be shown later, the substance decomposed with the formation of benzophenone. A combustion per-

Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. formed a week after the separation of the crystals gave the following figures :

0.2682 gave 0.8313 CO, and 0.1657 H20. C = 84.5 ; H = GbS6. C,,H,,O, requires C = 80.76 ; 13 = 8-97 per cent. C13€Ilo0 ,, C=S5.71 ; €I=5*49 ,, The quantity being so small, no experiments could be carpied out to ascertain the reason for the difference in melting point between these crystals and those described below. The oil from these crystals was again distilled under a pressure of 35 mm., when eleven grams, or 44 per cent. of the theoretical quantity obtainable, came over between 203" and 215", mostly at 210'. By cooling in a freezing mixture, small, diamond shaped plates crystsllised from the oil, which gradually became solid throughout. After drying on a porous plate, the solid melted at 36-38', but the melting point was lowered by exposure to the air. From an ethereal solution, View Article Online

120s MACKENZIE: THE ACTION OF SODIUM METHOXIDE ON crystals melting at 35-37", and from light petroleum, crystals melting at 35-36', were obtained. On analysis of the crystals melting at 36-38', the following numbers were obtained :

0,2041 gave 0.6017 CO, and 0,1644 H,O. C = 80.40 ; H = S.95. 0,2076 ,, 0.6131 CO, ,, 0.1669 H20. C=80.54; H=8*92. C,,H,,O, requires C = 80.76 ; H = 8.97 per cent, The following experiment is a further proof that the substance has the formula attributed to it. 0.2660 gram of the crystals in a platinum boat was placed in a desiccator over sulphuric acid and weighed every second day. The crystals soon deliquesced, then became entirely liquid, meanwhile steadily losing weight for nearly three weeks. At the eud of the fourth week, the weight had become constant, and on introducing a particle of benzophenone, the reddish- brown oil crystallised. The loss in weight amounted to 0.10SO gram, or 40.6 per cent. The loss in weight caused by splitting off diiso- butylic ether from diisobutyloxydipheny1methaneis 41.6 per cent. An analysis of the residual product proved it to be impure benzophenone. 0.1562 gave 0.4858 CO, and 0.0794 H,O. C = S4.82 ; H = 5-65. C,3H,,0 requires C=S5*71 ; H=5.49 per cent. In this respect, therefore, the dibobutyloxy-derivative agrees with the dimethoxy- and diethoxy-compounds. Diisobutyloxydiphenylmethane is exceedingly soluble in the ordinary organic solvents. Although it crystallises more readily than the propyloxy-compound, it is difficult to work with in summer weather on account of its low melting point and ready dissociation. Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. Attempt to prepare Diurrt y Zoxydiphenylmethane. The quantities used in this experiment were 2.8 grams of sodium, 50 grams of amyl alcohol(b. p. 129*5-130.5"),and 13.6 grams of benzophenone chloride. The mixture was heated-after a violent action had taken place-in an oil-bath to 135-145' for 5 hours. Dry ether was then added to the still alkaline product, and the solution filtered from the salt. On Fractional distillation under 40 mm. pressure, 3 grams of oil boiling between 200' and 220", 10.2 grams between 220' and 235*, and 4.3 grams between 235' and 260° were obtained. The two latter portions formed a jelly when placed in a freezing mixture, but no crystals separated. On exposure to winter cold during the Christmas vacation, small groups of stellate needles were obtained, which after drying between filter paper melted at 66-67', Only a very small quantity was obtained, and further efforts to get more were fruitless. An analysis gave figures which View Article Online

RENZOPHENONE CHLORIDE AKD EICNZAL CHLORIDE. 120'3

lie between those required for the amyloxy-compound and benzophenone. 0.1268 gave 0.3887 CO, and 0.0745 H,O. C = S3.60 ; H = 6.53. C,,H,,O, requires C = 81-65 ; R = 8.87 per cent. C,,H,,O ,, C=85.71 ; H=5.40 ,,

It is curious that this compound is the product of the action of sodium phenoxide on benzophenone chloride, and that no diphenoxy- compound could be isolated, Two methods of preparation are described. In the first, sodium phenoxide was prepared by adding 4.6 grams of sodium to a solution of 18.9 grams of phenol in dry ether, long con- tinued heating and frequent shaking being necessary to convert the sodium into sodium phenoxide (which is almost insoluble in ether) ; 23.7 grams of benzophenone ch!oride were then mixed with the product. On heating, no action took place until the ether was distilled off, when violent ebullition set in, vapours of phenol being evolved. The residue was heated for 1& hours in an oil-bath at 115", and when cold extracted with ether, The ethereal extract was repeatedly shaken with water until the washings no longer had an alkaline reaction. From the first washings, phenol separated as an oil on the addition of hydrochloric acid, After drying the ethereal extract and distilling off the ether, there remained an uninviting, thick, brown oil, which would neither solidify on cooling in a freezing mixture nor on the introduction of particles of benzophenone or phenol. After trying to crystttllise it from alcohol, ether, chloroform, and other solvents, it was found that by the addition of light petroleum to the ethereal solution, Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. small crystals separated, and after several days a product very like treacle filled the bottom of the beaker. This was spread on a porous plate and then dissolved in a large volume of chloroform and boiled under a reflux condenser with animal charcoal to decolorise it. Repeated boilings reduced the colour of the solution to a light yellow, and on the addition of light petroleum, light yellow, sandy crystals separated. After several recrystallisations, the substance softened at about 260" and melted at 285-2879 Analysis showed that the substance had the peculiar composition of 2 molecules of ether combined with 1 of dihydroxytetraphenyl- methane." 0.1330 gave 0.3862 GO, and 0.0943 H,O. C= 79.19 ; H= 7.S9. C,,H,,0,,2C,H,o0 requires C = 79.20 ; H = 8-00 per cent.

* Bseyer has obtained dichloral pcroxidc hydrate with one niolecule of ether of crystallisation, C4H4C1604,C4H1,,0(Bey., 1900, 34, 2482). View Article Online

1210 MACICENZIE: THE ACTION OF SODIUM MlWIIOXIIIE ON

The ether of crystallisstion was expelled by heating in a steam oven. A portion of the substance, which had already lost in weight by standing in an exhausted desiccator for several days, lost more than 20 per cent. by heating, the theoretical loss being 29.6 per cent. The tesults of the analysis of a portion of the substance so obtained were :

0.1418 gave 0.441 2 C'O, and 0 -01300 I€,O. C = S4.85 ; H = 6.26. C,5H2002requires C = 85.22 ; H = 5.68 per cent.

The following method of preparation WAS found to give much better results: 6.1 grams of sodium dissolved in 130 grams of phenol and 31.6 grams of beneophenone chloride were heated in a water-bath until hydrogen chloride ceased to come off, which was at the end of 22 hours. The excess of phenol was then distilled off under 40 mm. pressure at 130-140°, and the residue, mixed with water, repeatedly extracted with ether, The ethereal solution was washed with caustic soda solution and then with water, dried, and boiled for 6 hours with animal charcoal, After filtration and concentration, the solution on cooling became almost solid. Tile greenish-brown, crude substance so obtained weighed 61 grams, or 91 per cent. of tho theoretical quantity, After further boiling with animal charcoal and recrystallisation, the crystals melted at 284-385". Qualitative experiments show that benzophenone chloride and pheuol react directly with production of. the above product, but no quantita- tive experiment has yet been performed. Dihydroxytetraphenylmethanc is a colourless, crystalline substance which dissolves very easily in ether, easily in alcohol, and moderately in chloroform, but only sparingly in light petroleum. It is soluble in caustic soda without coloration, and is reprecipitated by the addition of

Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. an acid. It does not give the bromine colour reaction mentioned by Russanoff for the corresponding dihydroxytriphenylmethane. By heating the substance with its own weight of fused sodium acetate and four times as much acetic anhydride, an acetyl derivative was obtained, wliicb, after recrystallisation from acetone, melted at 170-171". Analysis showed it to be the diacetyl compound.

0.2194 gave 0.6386 GO, and 0.1202 H,O. C = 79-38 ; H = 6.08. C,,U,,O, requires C = 79.81 ; I€ = 5.60 per cent.

Action of Ethyl Alcohol on Benxophenone. Prom the easy dissociation of the above compounds, it was thought possible that they might be formed directly from benzophenone and the corresponding alcohol, but this proved not to be the case, as the following experiments showed. Five grams of benzophenone were heated with 20 grams of ethyl View Article Online

BENZOPHENONE CHLOItIDE AND I3ENZAL CHLOP,IDE. 1 211

alcohol in a water-bath for 3 hours and allowed to stand overnight. No crystals havicg appeared, most of the alcohol was distilled off. The residual oil did not crystallise on standing, but on the addition of a particle of benzophenone, a rapid growth of crystals took place. After drying they melted at 48-49', and were evidectly unchanged benzophenone. A further crop of crystals from the mother liquoi- brought the amount recovered up to 4% grams.

Action of 1 per cent. Solutioiz of Hydrogen Cldoi-itle iiz HetlyZ Alcohol on Benxophenone. As E. Fischer (Bey., 1897, 30, 3053 ; lS9S, 31, 545, 1989) mentions that the use of 1 per cent. alcoholic solutions of hydrogen cbloride is particularly adapted to the production of sldals, and only infers that it is not applicable for the production OF ketals, the following experiments were performed. These shorn that the process is inapplicable in the case of benzophenone. Five grams of bonzophenoue were dissolved in 100 C.C. of 1 per cent. solution of hydrogen cliloride in methyl alcohol, and the mixture allowed to stand for 5 days. Most of the alcohol was then removed by distillation, and the residue crystsllised on cooling. The crystals were separated from the mother liquor, which afforded a second crop on standing, the amount recovered being 4.9 grams. The crystals melted at 48-49' (benzophenone m. p. as0),hence the benzophenone was unacted on. A similar experiment in which 5 ptms of bmzophenone and 40 c.c, of 1.5 per cent, solution of hydrogen chloride in methyl alcohol were heated in a sealed tube at a temperature of 75-90' for 12 hours also

Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. resulted in the recovery of the whole of the benzophenone unchanged.

Dimet?~ox?/d~~~?~en~l~~et~~c~~e. Action of Nitk Acid on Diinel?~ox~c~i~~~e7~~~~~~t?~une.-The following experiment was performed in the hope of obtaining the nitro-derivative of dimethoxydiphenylmethane, but dissociation took place, and dinitrobenzophenone resulted, Two grams of the powdered crystals were slowly added to 25 C.C. of fuming nitric acid cooled by ice, the crystals dissolving apparently without change. After standing for a day, the solution mas poured on to crushed ice, by which means a white, curdy precipitate was formed. The weight of dried precipitate was 2.1 grams. It dissolved easily in hot ethyl alcohol, and from the solution there separated simultaneously pale yellow leaflets and needle-like crystals, neither of which showed a constant melting point. As this might be due to a View Article Online

1212 MACRENZIE: THE ACTION OF SODIUM METHOXIDE ON

mixture of isomerides, analyses of different portions of the substance, which had been kept in desiccators until of constant weight, were made.

0.3082 gave 27.9 C.C. moist nitrogen at 13Oand 757 rnm. N = 10.66.

0.2287 ,, 20.4 C.C. 99 13" ,, 759 mni. N- 10.52 C,,H,0,N2 requires N = 10.30 per cent. As dinitrobenzophenone exists in several modifications, further efforts were made to separate tho possible isomeric forms. From a fairly concentrated nitric acid solution, glistening crystals resembling sand mere deposited. They melted at 189-190', and crystallised from hot glacial acetic acid in fan-shaped groups of needles, thus agreeing with the properties of 4 : 4'-dinitrobenzophenone (Stadel, Ber., 1894, 27, 2110). On heating the mother liquor from the above experiment and diluting it with hot water, a second crop of crystals consisting of rosettes and needles was obtained. Recrystallised from glacial acetic acid, they softened at 137Oand melted at 149", thus conforming with the description of the properties of 3 : 3'-dinitrobenzophenone. Neither of these substances is appreciably volatile with steam, and, both are very sparingly soluble in ether. Di~~~enylmetlLyZeneanilid~,(C,H,),C:N*C,H,.-This substance was formed when dimethoxydiphenylmathane was heated with three times its weight of aniline. At about 60', the solid had all dissolved, and at 175" crystals began to separate and a slight ebullition occurred. Atter being kept at lS0" for 5 minutes, the product mas allowed to cool, filtered with the aid of a pump, and the residue mashed with dry ether. From a hot methyl alcoholic solution of the residue,

Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. glistening, yellow plates separated, showing a melting point OC 113-113*5°; Rohde gives 112-113' (Ber., 1892, 25, 2056).

11. Action of Sodium Alkyloxides on Benaccl Chloride undev the ordinwy Pressuye. Action of #odium &fethoxiJe." According to Wicke's instructions (Arcnalert, 1857, 102, 366), 23 grams of sodium were dissolved in 250 C.C. of methyl alcohol, 80.5 grams of benzal chloride added to the solution, and the mixture heated in a water-bath for 15 hours. As it was thought possible that methyl chloride might be formed, this being the case when the action takes place under pressure, an apparatus for collecting gas was attached to * As the products obtained in these experiments were not all of the same char- acter, they will be described nndcr the hending of the sodium derivative of the alcohol which acted 011 beuzal chloride. View Article Online

BENZOPHENONE CHLORIDE AND 1;lCNXA L CHLORIDE. 121 3

the end of the reflux condenser, but no gas was found to be evolved. After removing the alcohol by distillation, the residue, which was alkaline, was mixed with water, and the oil which separated extracted by ether. The ethereal extract was dried over calcium chloride, the ether removed, and the residue subjected to fractional distillation under diminished pressure. In this way, a colourless oil which distilled between 190" and 2 10' under the ordinary pressure was obtained. On testing the various fractions between 190" and 210" for chlorine, it was found to be present in all. Only after heating this oil with fresh quantities of sodium methoxide in alcoholic solution several times, could it be obtained free from chlorine. In another experiment,. potassium was substituted for sodium, and zinc dust and alcoholic potash were used to remove the last traces of chlorine, as described under sodium ethoxide. In this way, 12 grams of an oil boiling between 180' and 210' were obtained free from chlorine. Supposing it to be pure dimethoxybenzylidene, this would be a yield of 15 per cent. of the theoretical. Analysis of the portion boiling at 194-196' gave the following figures :

0.1671 gave 0.4345 CO, and 0.1160 H,O. C= 70.91 ; H = '7.71. C,H,,O, requires C = 71.06 ; H = 7.89 per cent. Dimethoxybenzylidene is a colourless, limpid oil, with an odour of geranium. It boils at 194-196', Fischer (Bev., 1898, 31, 549) giving 19So, and Wicke 208O, as its boiling point. The last is probably in- correct, as Wicke gives no analysis, and his product probably contained unchanged benzal chloride.

Action of #odium Ethoxide. Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. In this experiment, 23 gramsof sodium were dissolved in 230 grams of ethyl alcohol, and to the cold solution 80.5 grams of benzal chloride were added. The mixture was heated in a water-bath for 24 hours, no violent action taking place, but salt gradually separating. At the end of this time, the liquid mas still alkaline. Having dist,illed off most of the alcohol, the residue was mixed with water and extracted three times with ether. The ethereal solution mas washed with water until it showed a neutral reaction, dried over calcium chloride, filtered, and the ether evaporated. The residue mas then distilled under a pressure of 45 mm., nearly the whole coming over between 90° and 175". On redistilling this under the ordinary pressure, 3.6 grams were collected below 206", 59.3 grams between 206" and 220', and 6.3 grams between 220° and 300'. As all these fractions contained chlorine, those boiling up to 220' were added to a solution of 9 grams of sodium in 100 grams of alcohol, and the mixture heated on the water-bath for 17 hours. A View Article Online

1214 MACKENZIE: THE ACTION OF SODIUM METHOXIDE ON

Separation of salt again took place, and the liquid darkened in colour. The product was treated as before, and 51-1 grams of a colourless liquid boiling at 135-140' under 75 mm. pressure mere obtained, which, under tbe ordinary pressure, distilled between 300' and 220'. As the various fractions still contained chlorine, they were heated a third time with sodium ethoxide (4.5 grams sodium in 45 grams alcohol) for 16 hours, and subsequently treated as before. The fractions collected were: 6.2 grams at 212-215'; 13.3 grams at 215--217'; 12.6 grams at 217-219O; 4.3 grams at 219-222'. As they all still showed the presence of traces of chlorine, the action of zinc dust and alcoholic potash was tried on the two latter .fractions. Sixteen grams of the oil mixed with 5 grams of potassium hydroxide dissolved in 20 grams of alcohol, diluted with water to make x solution, ancl 5 grams of zinc dust mere allowed to stand at the ordinary temperature for five days, and then heated under a reflux condenser for an hour in a mater-bath. The alcohol having been removed by distillation, the residue mas mixed with water, extracted with ether, dried, and the ether expelled. The resulting product now distilled at 216' to 220°, chiefly at 217', 13 grams being so obtained. In order to make certain of the absence of chlorine, n Carius estimn- tion was carried out. 0,5890 gave 0.0004 AgCl and ash. Ash = 0.0001. It is very curious that Wicke makes no mention of the difficulty of removing the chlorine compounds in this reaction. On the other hand, Limpricht, who repeated Wicke's work, but heated the mixture of benzal chloride, sodium methoxide, and alcohol at 140°, was unable to obtain products free from chlorine, the amount of the latter varying between 7.5 and 18.7 per cent. Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. The absence of chlorine having been established, the liquid mas once more fractionally distilled, and 7.5 grams of oil boiling at 216-217" (uncorr.) * were collected, Wicke, with whom Fischer agrees, mentions 222' as the corrected boiling point of diethoxybenzylidene. A combustion T of this portion gave the following figures : * The temperatnres throughout this investigation are uiicorrected. t The combustion of this substance was a matter of some difficulty, because, on heating it apparently splits off ethyl ether, which, if allowed to pass over rapidly, Causes the percentage of carbon to come out too low. The following analyses of portions (I and I1 boiling at 217-219", 111 at 216-220", and IV at 216-217") which were free, or as nearly as possible free, from chlorine illustrate this fact. I. 0'3745 gave 0.9706 COz and 0'2933 H,O. C=70'68 ; H=8.70. 11. 0.1690 ,, 0.4397 CO, ,, 0'1360 HzO. C=70*90; H=8'94. 111. 0'6398 ,, 1'6868 CO, ,, 0.5056 H,O. C=71.93 ; H=S*77. IT. 0.5438 ,, 1'4243 CO, ,, 0'4255 H,O. C=71'43; H=8*67- C,,H,,O, requires C = 73-33 ; H = 8 $8 per ccnt. In the above four analyses, the combustion was carried out at an ordinary rate, an View Article Online

UENZOPIIENONE CHLOItlDE AND EENZAL CHLORIDE. 1215

0.24 gave 0.645 GO, and 0.2 H,O. C = 73-29 ; H = 9.25.

C,,HI,O, requires C: = 73.33 ; H = S.58 per cent. 8 Diethoxybenzylidene is a colourless liquid with a fragrant odour. It is not soluble in water, but is miscible with the ordinary organic solvents. It is stable in the presence of alkalis, but is readily decomposed by acids. With a saturated solution of sodium bisulphite it soon yields the beazaldehyde compound. Assuming that the fractions 215-222", weighing 30.1 grams in all, consisted of pure diethoxybenzylidene, the yield would be 33 per cent. of the theoretical. This is, therefore, not so convenient a method as E. Fischer's for the preparation of this substance.

Action of Sottiurn Bemoxiclc. The quantities used in this experiment were 4.6 grams of sodium, 50 grams of benzyl alcohol, and 16.1 grams of benzal chloride. The mixture was heated in an oil-bath at a temperature of 206-210' for 5 hours, during which a quantity of solid matter separated. When cold, a large voliime of ether was added and the solution filtered, the salt being washed several times with ether. After removal of the ether, the yellow oil which remained was distilled under a pressure of 55 mm., fractions being collected between 100" and 270". The residue in the distilling flask was a viscous mass. The last fraction of distillate, namely, that boiling at 200-2'i0°, smelt of benzaldehydo, and contained some chloride. On distillation under the ordinary pressure, the temperature steadily rose to 300'. A further effort to obtain a constant boiling liquid under a reduced pressure of 60-65 mm. was

Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. also ~n~uc~es~ful,the distillate heing collected in the following fractions : up to 115O, 9.3 grams; 115-120", 12 3 grams ; 120-195', 11.9 grams; 195-260", 5.3 grams, and 360-300', 6.8 grams. The first two fractions were found to be free from chlorine, and as they distilled at 195-202' under 749 mm. pressure, were probably benzyl alcohol (b. p. 204'). The other portions all contained chlorine. By cooling the last fraction in a freezing mixture, a small quantity of crystals separated. These were filtered off, dried on a porous plate, and then melted at 112-115'. As the presence o€ stilbene and benzoic acid, which both melt at 120°, was suspected, the following tests mere performed. The crystals did not dissolve in solutions of either sodium hydroxide or carbonate, nor did they react with litmus,

hour to an hour and three-quarters being taken between tho introduction of tho substance and the removal of the absorption apparatus. In the combustion mentioned in the text, the time was extended to two hours and a half. This ob- servation was also made by Dr. A. J. Walker when burning dirnethoxybenzylidene. View Article Online

1216 MACICENZIE : THE ACTION Ok' SODIUM METHOXIDE ON

but they decolorised a solution of bromine in carbon disulphide, thus indicating the absence of benzoic acid, and the presence of stilbene. They were then purified by recrystallisation from hot ethyl alcohol. The glistening leaflets, of which only a minute quantity was obtained, melted sharply at 120', and on analysis gave the following figures : 0.0316 gave 0.1078 CO, and 0.0246 H,O. C = 93.02 ; H = 8.65. C1,H,, requires c! = 93.33 ; H = 6.67 per cent. With the idea of removing any benzoic acid, and possibly chlorine compounds, from the fractions 195-260' and 260-300°, an ethereal solution of them was shaken with aqueous solutions of sodium car- honatle and hydroxide, dried, and, after the removal of the ether, distilled under 85 mm. pressure. In this way, 5.5 grams were obtained, distilling between 200' and 240'. On mixing with light petroleum and cooling in a freezing mixture, there separated some crystals, which when dry weighed 0.11 gram. As they resembled the crystals of stilbene above mentioned, the bromide was prepared. An ice-cold ethereal solution of bromine was immediately decolorised by an ethereal solution of these crystals. After standing 4 hours in the cold, the silky needles which had been deposited from the soln- tion, were filtered off and dried on a porous plate. The weight of crystals obtained mas 0.124 gram, and they melted at 236-237'. As stilbene dibromide melts at 237', this experiment confirms the results of the analysis. From.the mother liquor of the stilbene, no liquid of constant boiling point could be isolated, and it still contained a considerable quantity of chlorine. The action of sodium benzoxide on benzal chloride would thus appear to be a rather complicated one, the formation of stilbene only taking place in very small quantity. Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42.

Action of Phenol on Benxal Chlode. As it had been found that benzophenone chloride acted directly on phenol, the same result was expected in the case of the above substances. A mixture of 55 grams of phenol and 16.9 grams of benzal chloride was heated gently at first, then to 120' for 10 hours, when hydrogen chloride ceased to come off. The loss in weight of hydrogen chloride was 7 grams, or 95 per cent. of the theoretical. From the dark brown residue, phenol was distilled off under 50 mm. pressure at 100-107', the temperature being allowed to rise to 210'. On removing the capillary tube from the ruby-red, transparent jelly remaining in the flask, a sharp click was heard, and crystals shot out radially from the centre. The weight of crystalline solid was 24 grams, or 87 per cent. of the theoretical. After boiling with animal charcoal and re- crgstallisi ng from chloi oform, the colourless crystals of dihydroxy- View Article Online

BENZOPHENONE CHLOlZlDE AND BENZAL CHLORIDE. 1217

triphenylmethane melted at 160-1 61". As mentioned by Russanoff (Ber., 1889, 22, 1944), when gently warmed with bromine water, washed, and dissolved in alcohol, this substance gives a beautiful, blue colour on the addition of caustic alkalis. When heated with acetic anhydride and fused sodium acetate, it affords a diacetyl derivative, which crystallises from acetone in colourless crystals melting at 10s-110'. In this respect, therefore, it also agrees with the product obtained by Russanoff by the condensation of benzaldehyde and phenol by means of sulphuric acid.

111. Action of Sodium Alkploxides on Benxal Chloride uiacler Pressure in Sealed Tubes. Action of 8odium Methoxide. In this experiment, 41.4 grams of benzal chloride were added to a solution of 6 grams of sodium in 60 grams of methyl alcohol, and the mixture heated at 100-105" for 6 hours. On the tube being opened, only a slight pressure of gas was noticed, but a considerable quantity of a white salt had separated. The tube was again sealed and heated at 150' for 3 hours; on reopening, a strong pressure of gas was observed, and the liquid contents were faintly acid to litmus. After standing for a few minutes, the liquid began apparently to boil, and methyl chloride was briskly evolved for more tlian an hour. It may be noted that a corresponding result was obtained with the other alcohols used under similar conditions. The residue was filtered with the aid of the pump, and the liquid thus obtained was distilled under a pressure of 75 mm., the fractions collected being 21.7 grams below 35', and 24 grams between 35' and Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. 105', chiefly 100-105'. On redistillation under the ordinary pressure, nearly 20 grams of crude methyl nlcohol, boiling between 66' and 70", and 23 grams of crude benzaldehyde, boiling between 175' and 195', were obtained. The crude benzaldehyde so obtained corresponds to a yield of 87 per cent. of the theoretical. From the residue, a small quantity of a crystalline acid melting at 131" was obtained (benzoic acid melts at 121"), probably formed by the oxidation of the aldehyde.

Action of Sodium Ethoxide. This experiment was carried out in a similar manner to the previous one, the quantities taken being 4.6 grams of sodium, 47 grams of ethyl alcohol, and 32 grams of benzal chloride. The mixture was heated at 150-160' for 7 hours. No pressure of gas was noticeable on opening the tube, but on heating to 55" n large volume of gas was VOL. LXXIX. 40 View Article Online

1218 MACKENZIE: THE ACTION OF SODIUM METHOXIDE ON

liberated. The gas, which measured 2-25 litres, mas collected in a gas holder. After passing it through two wash-bottles containing sulphuric acid, some of the dry gas was liquefied in U-tubes immersed in a freezing mixture. About four C.C. of the colourless, very mobile liquid which was obtained in this way evaporated very rapidly when removed from the freezing mixture, The vapour burned with a green edged flame, and was evidently ethyl chloride. The volume of gas measured was about half that which should 'theoretically have been obtained. The contents of the tube were next filtered with the aid of the pump, and after distilling off the unchanged alcohol from the filtrate, the temperature was raised to 210°, the oil which came over amounting to nearly 23' grams. The residue was dark brown in colour and resembled pitch. It did not give evidence of the presence of chlorine as shown by the sodium test. On heating alone, it gave off a very acrid odour. A solution obtained by boiling it with water and filtering hot, deposited glistening leaflets, which melted at 121' (benzoic acid melts at 121'). On redistilling the oil, the temperature rose rapidly to 175' and between 175' and 185' more than 12 grams, and between 185' and 210' 5.5 grams, came over. On shaking with a cold saturated solution of sodium bisulphite, both these fractions were shown to consist mainly of benzaldehyde, and the presence of only a small quantity of benzal chloride in the latter fraction was proved by the tests for chlorine.

Act ion of Sodium Propyloxide. In this experiment, 4.6 grams of sodium, 50 grams of propyl alcohol,

Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. and 32.2 grams of benzal chloride were heated at 150'for 7 hours. No gas escaped on opening the tube. The tube was then connected with a condenser, to which two U-tubes, immersed in a freezing mixture, were attached. It was heated in the outer tube of a V. Meyer vapour density apparatus filled with water. When the temperature OF the water rose to 55', bubbles of gas began to come off, and after heating for an hour and a quarter, during which time the temperature gradually rose to 97", all the gas seemed to have been evolved, The liquid, con- densed in the U-tubes, weighing 11.4 grams, was then fractionally distilled, a Linnemann column being used. The temperature, which rose to 43" before any liquid distilled over, remained constant at 46" for some time, and then rose rapidly to 90'. The fractions collected were 43-48' (chiefly 46O), 4.4 grams; 48-97' (chiefly 90-94'), 6.4 grams. The former was a colourless, mobile, strongly refractive and pleasant smelling liquid resembling propyl chloride in all its properties (prop91 chloride boils at 46.5'). A yield of 4.4 grams of propyl chloride View Article Online

UENZOPHENONE CHLORIDE AND BENZAL CHLOXIDE. 1219

corresponds to 28 per cent, of the theoretical. The latter portion was chiefly propyl alcohol. The residue in the tube was mixed with about 250 C.C. of water and extracted with ether eight times. The ethereal extract, which was neutral to litmus, was dried over calcium chloride, and, after expelling the ether, was distilled, 26 grams of a mobile liquid coming over below 105O and 18.5 grams of an oil between 175" and 530'. The first portion consisted mainly of propyl alcohol, whilst the last had a stjrong odour of benzaldehyde, but was found to contain a considerable quantity of chlorine, In order to get a rough estimate of the quantity of benzaldehyde present, the last portion was shaken with a saturated solution of sodium bisulphite, and the crystalline compound thus formed filtered with the aid of the pump, washed twice with ethyl alcohol, and dried on a porous plate. The weight of the dry bisulphite compound was 9.6 grams, or 23 per cent. of the theoretical.

Action of Xoclium Benxoxide. In this experiment, 4.6 grams of sodium dissolved in 50 grams of benzyl alcohol and 32 grams of benzal chloride were heated at 150-160° for 11 hours. On opening the tube, slight pressure was noticeable, a combustible vapour escaping. A small quantity of glistening leaflets, which had crystallised in the upper part of the tube, proved to be insoluble in cold, but readily soluble in hot, water, and also in caustic soda, from which they were reprecipitated by the addition of hydrochloric acid. After drying, the crystals softened at 90' and melted at 11S0, agreeing generally therefore with the properties of benzoic acid, The contents of the tube were then filtered, and the oil thus obtained subjected to fractional distillation under Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. 60 mm. pressure, the temperature being allowed to rise to 140°, at which point crystals appeared in the condenser, The fractions were then distilled under the ordinary pressure, and 7-2 grams mere collected between 162' and 176O, and 20 grams between 176" and 178'. As the last fraction was expected to be benzyl chloride, chlorine esti- mations were made with the following results : 0.2673 gave 0.2330 AgCl. C1= 21.5. 0.2207 ,, 0*1918 ,, C1=21*5. C,H,Cl requires C1= 28.06 per cent. AS the liquid probably contained benxaldehyde, it was shaken with a saturated solution of sodium bisulph ; the colourless crystalline substance so obtained was filtered off, and proved to be the benzaldehyde bisulphite compound. The oil was then separated by the addition of water and extraction with ether. After drying the ethereal extract and removing the ether, the residue was distilled under a pressure of View Article Online

1220 ACTION OF SODIUM METHOXIDE ON BENZOPHENONE CHLORIDE. 70 mm., when it came over between looo and llOo. Unfortunately some was lost, but of the remaining 8.7 grams, 1.S grams distilled below 176', and 5.7 grams at 176-177O under the ordinary pressure. The latter portion showed all the properties of benzyl chloride (b. p. 179"), and a chlorine determination gave satisfactory figures. 02635 gave 0.2967 AgC1. C1= 27.9. C,H7C1 requires C1= 28.06 per cent. In another experiment, in which the same quantities were used, the components were thoroughIy mixed by stirring with a glass rod before sealing off the tube. After heating at 150-155O for 11 hours, the sodium chloride which had separated mas filtered off, with the aid of a pump from the brown liquid,-washed with a little benzyl alcohol, and dried. It weighed almost 12 grams, the calculated quantity being 11 *7 grams. From the liquid obtained by distillation of the filtrate under 50 mm. pressure, 18.2 grams of sodium bisulphite compound were separated, corresponding to a yield of 43 per cent. of the theoretical. After repeating the distillation and treatment with sodium bisulphite, 11 *S grams of benzyl chloride, or 46 per cent, of the theoretical quantity, distilling at 175-176O under 748 mm. pressure, were obtained. An estimation of chlorine gave the following figures : 09165 gave 0.2403 AgCl. C1= 27.5. C7H,C1 requires C1= 28.06 per cent.

IV. Note on the Analysis of Benxczl and Benxgl Chlorides. As Wicke states that the chlorine in benzal chloride is completely

Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. precipitated by the addition of an alcoholic solution of silver nitrate, it was thought that in the case of this compound and of benzyl chloride the Carius method of estimating halogens might be replaced by the more direct process. The following results, however, show that the precipitation is very incomplete in the case of benzyl chloride (I), and only partially complete in that of benzal chloride (I1 and 111). I. 0.3280 gave 0.3623 AgC1. C1= 22.4. C,H,Cl requires C1= 28.06 per cent. 11. 0.6302 gave 1.0907 AgCl. Cl=42*98. 111. 0.4226 ,, OT314 AgCl. C1= 42.97. ' C7H,CI, requires C1= 44.0'3 per cent. An attempt was also made to estimate the chlorine in benzal chloride by reduction with sodium amalgam and precipitation of the sodium chloride by silver nitrate in the usual way. The figures obtained in two experiments were : View Article Online

DIARYLOXYISOPROPYLPHOSPHOROUS ACIDS. 1221

I, OT012 gave 0.4654 AgC1. C1-16.5. 11. 1.3776 ,? 0,9345 AgC1. C1= 16.8. C7H,C12 requires C1= 44.09 per cent. Both the above methods were therefore inapplicable, and the Carius method had to be resorted to.

Action, of Xodium Ethoxide on Ethylene Dibromide under Presswe. A mixture of 23.5 grams of ethylene dibromide and 5.75 grams of sodium dissolved in 57.5 grams of ethyl alcohol wa8 sealed up in a tube and, after remaining at the ordinary temperature for 2 days, during which a quantity of salt separated, was heated at 120-130' for '7 hours, On the tube being opened, a gas possessing all the properties of acetylene escaped under slight pressure. This gas burned with a brilliant, white, luminous flame, decolorised bromine water, forming,oily drops having a peculiar odour, and produced a deep red and a yellowish-white precipitate in ammoniacal solutions of cuprous chloride and silver nitrate respectively, both of which were explosive when dry. The filtrate from the salt distilled at R temperature of 68-70', and in this way nearly all the alcohol was recovered.

The author desires to express his thanks to Dr. A. Jamieson Walker for carrying out the experiments on the action of sodium methoxide and of isobutyloxide on benzal chloride under the ordinary pressure. The latter, having unfortunately yielded negative results, are not recorded here. BIRKBECKINSTITUTION,

Published on 01 January 1901. Downloaded by University of Connecticut 18/01/2018 18:50:42. LONDON.