FENTON : DIHYDROXYTAKTARIC ACID. PART I. 11 I: V. -Pq-operties and Relationships of Dihydroxy- tartaric Acid. Pwrt I. By HENRY J. HORSTMAN FENTON, M.A. THIS acid mas first observed by Griiber, who obtained it, in solu- tion, by the action of nitrous acid on protocatechuic acid (Ber., 1879, 12, 514). It was afterwards prepared from pyrocatechol by Barth (Sitx. Acad. Feim, 82, ii, 1024); from guaiacol by Werzig, and from '' nitrotartaric acid " by Kekuld (Annalen, 1883, 221, 230) in a similar manner ; these authors isolated it in the form of the spar- ingly soluble sodium salt, and from this the free acid was subse- quently obtained by Miller (Ber., 1889, 22, 2015). The salt was de- composed by hydrogen chloride under ether, excess of hydrogen chloride being avoided and moisture carefully excluded ; but since its isolation by Miller, the free acid appears to have been scarcely examined.I n a former communication (Trans., 1895, 6 7 , 48), it mas shown that a solution of dihydroxytartaric acid may very easily be obtained by the oxidation of the new acid C,H,G, (dihydroxymaleic acid) with bromine in presence of water. The change takes place nearly quanti- tatively according to the equation C,H,O, + 2H,O + Br2 = C,H,O, + 2HBr.72 FENTON : PROPERTIES AND RELATIONSHIPS OF The solution was precipitated with sodium carbonate, and the free acid prepared from the sodium salt by Miller's method. The yield of free acid obtained by this method appears to be small, probably owing to its very sparing solubility in anhydrous ether. It was mentioned, however, that the free acid might be directly obtained from the solution after oxidation with bromine, without first preparing the sodium salt ; this was effected by concentrating the solution in a vacuum desiccator over solid potash, but the purity of the pro- duct was uncertain.Further experiments have now shown that., by modifying, the details previously given, this process affords a very simple and productive method for the preparation of the acid in a state of purity. Having regard to the very interesting constitution of this acid and to the close relation which has been shown to exist between it and dihydroxymaleic acid, it was considered desirable to take advantage of this new method of preparation and to make a study of the proper- ties of the acid.Prepurution of the Acid. Crystallised dihydroxymaleic acid, C4H406,2H20,* is well triturated with from 4 to 5 times its weight of glacial acetic acid; and rather more than the calculated quantity of bromine, dissolved in a little glacial acetic acid, is added to the mixture in small portions a t a time. The first portions are almost instantly bleached, but the action after- wards becomes more sluggish and apparently ceases-a few drops of water are then added, whereupon the colour of the bromine is again im- mediately discharged. The addition of bromine is continued until the colour is quite permanent on standing, even when a drop or two of water is added. It has been previously shown (Zoc. cit.) that this final stage is reached when the bromine has been added in about the calculated proportion (1 mol.acid to 1 mol. bromine) ; fumes of hydrogen bro- mide are freely evolved during the operation. The dihydroxymaleic acid is nearly insoluble i n cold glacial acetic acid, but when the oxidation is finished complete solution takes place. The solution is allowed to stand for an hour or two, and then vigorously stirred, when the dihydroxytartaric acid quickly, sometimes suddenly, sepa- rates as a heavy, white, crystalline powder. The product is now collected and drained with the aid of the pump, washed once or twice with small quantities of anhydrous ether, and kept in a vacuum desiccator over solid potash and sulphuric acid to remove the last traces of hydrobromic acid, acetic acid, bromine and ether. The yield of purified product thus obtained is '70 per cent.or more of the theoretical. Thus, using 18.4 grams of dihydroxy- * For the preparation of this acid, see Trans,, 1894, 65, 901.DIHYDROXYTARTARIC ACID. PART I. 73 maleic acid, 80 C.C. of acetic acid, and 5.5 C.C. of bromine, the yield of pure product was 13.3 grams. Again, with 28 grams of dihydroxymaleic acid, 120 C.C. of acetic acid, and 9 C.C. of bromine, the yield was 21 grams. The remainder may of course be recovered as sodium salt by neutralising with sodium carbonate. The product thus obtained, when heated in a capillary tube, melts sharply, and decomposes, at 114-1 15". The acid previously prepared by Miller's method melted and decomposed a t 98". I. 0.1492 gave 0.1421 CO, and 0.0443 H,O. C = 25.97 ; H= 3.29.Theory = 18.2 grams. Theory = 27.7 grams. 11. 0.1736 ,, 0.1667 CO, ,, 0.0516 H,O. C=26*1S; H=3*30. C4H,08 requires C = 26.37 ; H = 3.30 per cent. Action of Heat. Preparation of Turtronic Acid. Dry Acid.-It might perhaps be expected from the constitution of dihydroxytartaric acid that it would tend to lose water on heating, giving the diketonic acid, (?'o'cooH especially as Anschutz has CO*COOH' shown (AnnaZen, 1891, 261, 130) that the ethylic salt has a composi- tion corresponding t o this acid; it is found, however, that the free acid when heated a t 90" in a current of dry hydrcgen for about one hour, undergoes no loss in weight. Heated in a vacuum on a water bath a t 90-loo", it loses weight slowly and somewhat irregularly, and gradually turns dark brown without melting ; the loss of weight after 5 hours heating was about 42 per cent., that is, considerably more than is represented by the loss of all the hydrogen as water.The residue, moreover, when dissolved in water and tested with sodium carbonate, no longer gives the reaction of dihydroxytartaric acid. Aqueous SoZzction.-Gruber showed that the sodium salt, when heated with water, is resolved into sodium tartronate and carbon dioxide. An aqueous solution of the free acid is found to undergo a similar change ; when this is gently heated, carbon dioxide is freely evolved, and the solution, after concentration, yields crystals of pure tartronic acid, as will be seen from the following results. A few grams of dihydroxytartaric acid were dissolved i n water, the solution heated on a water bath until carbon dioxide ceased to be evolved, and then concentrated to a small bulk and allowed to stand in a desiccator; in a few days, long, transparent prisms separated, which, after being drained on filter paper, air dried, and then heated a t 100" until the weight was constant, were snalysed.G=29*76 ; H=3.29. I. 0.1603 gave 0.1750 CO, and 0.0476 H,O. Tartronic acid, C,H40, requires c! = 30.00 ; H = 3.33 per cent.74 FENTON : PROPERTIES AND RELATIONSHIPS OF 11. 0.6067 substance, on titration, required 12.4 C.C. of a solution of caustic soda contaiuing 0.0187 Na per C.C. Theory for a dibasic acid, C,H,O, = 12.0 C.C. The crystals melted a t 158-159O ; the melting point of tartronic acid has been very variously given by different authors, but Griiber (Zoc.c i t . ) , and Rlassol (Conzpt. rend., 1892, 114, 422), both obtained anhydrous crystals melting a t 155'. [The crystals appear t o separate from the solution in the anhydrous state, and not with &H,O as sometimes stated, since they undergo scarcely any change in appearance when heated a t 100'. The u i ~ clTiecl crystals gave, on analysis, C = 29.14 ; H = 3.27. C,H,O,,iH,O would require C = 27.9 ; H = 3.8. The water is, therefore, probably only '' mechanical. "1 This reaction, then, affords an extremely easy and productive method for the preparation of pure tartronic acid; the yield is almost theoretical, and the product is pure without recrystallisation. Thus 0,9696 gram of dihydroxytartaric acid was dissolved in water, the solution heated as described above, taking care t o avoid loss by spirting, and then allowed to evaporate in a vacuum desiccator; the crptalline residue dried a t 100" until the weight mas constant, melted at 157-1558", and its weight was 0.6237 gram, theory requiring 0.6392 gram.Titnxtion by Alkalis. Judging from the observation, above referred to, that sodium dihydr- oxytartrate, when heated, decomposes into sodium tartronate and carbon dioxide, Griiber considered that the acid was " carboxytar- tronic " acid, c o o H > ~ ( ~ ~ ) - ~ ~ ~ ~ , COOH and consequently that the sodium compound was an acid salt'. Attempts to prepare the normal salt were, however, unsuccessful. Griiber found that the sodium salt was not acted on by a dilute solution of caustic soda, and Barth showed that dry ammonia was also without action on it ; a strong solution of caustic soda dissolves it, but apparently decomposes it, since the sparingly soluble salt cannot be again obtained from the solution after acidification. Bnrth, however, prepared and analysed the barium salt and found its composition to be Ba3(C4H07)2, 3H,O, corresponding with the normal salt of a tribasic acid C4H407.Since this acid had been obtained from benzenoid compounds and was regarded as tribasic, with one carbon atom directly associated t o three others, arguments were advanced from its supposed constitution which were a t variance with Kekulb's well known benzene symbol. Kekule then made an exhaustive study of the sodium salt obtained from various sources, H e showed also t h a t it could be prepared fromDIHYDROXYTXRTARIC ACID PART I.7 5 ( ( nitrotartaric ” acid, and that by reduction with zinc and acid, modifi- cations of tartaric acid were produced. His results indicated that the acid is in reality dibasic, having the formula C,H,O, (dihydroxytartaric acid, or tetrahydroxysuccinic acid), and that the sodium compound is a. nornzal salt ; he suggests that the ba.rium salt obtained by Earth, if it is a homogeneous substsnce, may be a basic salt, or that replace- ment may have taken place in the liydroxj-1 groups. The acid being now available in quantity, and in a pure state, i t was considered that the results of titration by various alkalis might he of interest, as affording further evidence with regard to the basicity of the acid.Experiments were accordingly made, using sodium, potassium, and barium hydroxides, as well as ammonia and sodium carbonate, with the following results. I. 0,3778 gram of acid dissolved in 5 C.C. of water required 6.9 C.C. of a solution of caustic soda, prepared from metallic sodium, containing 0.01877 Na per C.C. ; phenolphthalein was used as indicator, and the most minute precautions were taken in order t’o ensure the exclusion of carbon dioxide, not only during the preparation of the solution, but also from the water employed, and during the operation of titration. One mol. of acid required, therefore, 2.7 mols. NaOH for neutralisa- tion. IT. 0,6035 gram of acid in 5 C.C. of water required 8.7 C.C. of normal KOH solution. (Phenolphthalein as indicator.) One mol. of acid neutralised 2.6 mols.of KOH. 111. 0,6035 gram of acid in 5 C.C. of water required 6.4 C.C. of normal sodium carbonate. (Methyl-orange as indicator.) One mol. of acid neutralised 0.96 mol. of Na,CO,. IV. 0.2910 gram of acid required 15.5 C.C. of barium hydroxide solution containing 0.0225 gram of Ba(OH), per C.C. (Phenolphthalein as indicator.) V. 0.4218 gram of acid in about 10 C.C. of water required 11.3 C.C. of a solution of pure ammonia containing 7.5555 gram NH, per litre. (Litmus as indicator.) One mol. of acid neutralised 2.16 mols. of NH,. The colour indications with phenolphthalein, although sharp a t first, quickly faded, and another drop o r so of alkali was required t o restore the colour; in the case of litmus, the blue colour quickly changed t o wine-red in a similar way.The numbers given represent the amount of alkali required to give a colour change which was permanent for a few minutes; the differences between the initial and final colour- change, however, were small, amounting only t o about 0.2 to 0.3 C.C. Those results would appear t o indicate that dihydroxytartaric acid behavei normally as a dibasic acid towards sodium carbonate and One mol. of acid neutralised 1.27 mols. of Ba(OH),.76 FENTON : PROPERTIES AND RELATIONSHIPS OF ammonia, but that with the hydroxides of sodium, potassium, and barium, its behaviour is intermediate between t h a t of a clibasic and a tribasic acid. A t first it seemed probable that the high results obtained when " strong " bases are employed might be explained by supposing that one or more of the alcoholic hydroxyl groups in the acid exerted " acid " functions and that the compounds formed are more or less hydrolysed; or that the acid might in reality be tribasic carboxy- tartronic acid as was formerly believed, the replacement of the third atom of hydrogen giving rise to an unstable salt as in the case of orthophosphoric acid.Dilution of the solution, however, has but little effect, as seen by the following experiments. VI. 0.3158 gram of acid dissolved in 50 C.C. of water required 5.45 C.C. of NaOH solution containing 0.0189 gram Na per C.C. One mol. of acid neutralised 2.6 mols. of NaOH. VII. 0,3391 gram of acid dissolved in 150 C.C. of water required 5.8 C.C. of the same NaOH solution. One mol. of acid neutralised 2.5 mols.of NaOH. The high results might, on the other hand, be due t o the partial decomposition, at the ordinary temperature, of the acid into dibasic tartronic acid and carbon dioxide; such a decomposition would not influence the result when methyl-orange was used as indicator, but would give a high result with phenolphthalein. Various experiments were therefore made in order t o throw light upon this question. VII1.-A standard solution of the acid was prepared and a portion titrated immediately with soda; the remainder of the solution was allowed to stand for about 2 hours and a n equal portion again titrated with the same soda. 1X.-Air, carefully purified from carbon dioxide by passing it through strong caustic soda solution and then through baryta-water, was allowed to bubble through a freshly prepared solution of 0.3208 gram of acid in about 10 C.C.of water contained in a small flask, The issuing gas was carefully tested for carbon dioxide by passing it through a series of bulbs containing baryta water; no trace of turbidity could be detected for the first 10 or 12 minutes, after which, however, a faint cloudiness was perceptible in the first bulb. From the results of these last two experiments it appears to be very improbable that the high results on titration can be due to the decomposition of the acid itself. But it may be that the snlt produced is less stable than the acid, and breaks up in the manner indicated, X.-0*3223 gram of acid was dissolved in about 10 C.C. of water and the experiment conducted exactly as in IX, but with the alteration that, as soon as the acid had dissolved, standard caustic soda was run in The two results were practically identical.DIHYDROXYTARTARIC ACID.PART I. 7'7 from a burette i n quantity insufficient for neutralisation (about 0.054 gram Na). Carbon dioxide was, in this case, given off almost im- mediately ; after 3 minutes there was a dense turbidity in the first baryta bulb and after 5 minutes all three bulbs were turbid. The soda solution had been prepared from metallic sodium with great precautions to exclude carbon dioxide, and the apparatus from which it was supplied was constructed so as to avoid the possibility of contamination. Still it was considered advisable to make a blank test with the whole apparatus, and this was done, dilute sulphuric acid being partially neutralised with the same soda solution and the experiment conducted exactly as before.No trace of carbon dioxide could be detected after passing the air for 10 minutes. It is tolerably certain therefore that the high results on titration are due to the splitting up of the salts produced into tartronate and carbon dioxide; if this is so, one would expect that lower, if not normal, results should be obtained on cooling the solutions, the titrations mentioned all having been performed a t the ordinary temperature of the laboratory. XI.-0*4498 gram of acid in about 10 C.C. of water was cooled by ice and standard caustic soda containing 0.02183 gram of Na per C.C. was slowly run in, phenolphthalein being used as indicator.5.4 C.C. were required for neutralisation. One moL of acid neutralised 2.07 mols. of NaOH. XII.-0,3201 gram of acid in about 10 C.C. of water was cooled, as above, and titrated with pure ammonia solution containing 7.5555 grams of NH, per litre; 7.9 C.C. were required. Litmus was used as indicator and the final blue colour was permanent for 5 minutes or more. One mol. of acid neutralised 1.99 mols. of NH,. XIII.-0*3076 gram of acid in about 10 C.C. of water, cooled as before, required 14.7 C.C. of baryta water containing 0,02244 gram of Ba(OH), per C.C. (phenolphthalein as indicator). One mol. of acid neutralised 1.14 mols. of Ba(OH),. This indeed, is found to be the case. Reduction of Dih?/cli.oxytcLrtcLric Acid to the Acid C,H,06. It has already been pointed out (Trans., 1895, G7, 48) that the new acid, C,H,O, (dihydroxymaleic acid and its isomeride), may be regarded as intermediate between tartaric and dihydroxytartaric acids ; that is, the anhydrous acid contains 2 atoms of hydrogen less than tartaric acid, and the hydrated acid contrains two atoms of hydrogen moye than dihydroxy tartaric acid 11.C,H,0G,2H,0. n r . c,H,o,. Or, if it be assumed, as is not improbable (TOG. cit., 67 7777, that a78 FENTON : PROPERTIES AND RELATIONSHIPS O F solution of the acid C,H,06 contains trihy droxysuccinic acid, the relation is perhaps better illustrated thus, I. C,H,(OH),(COOH),. 11. C213(OH),(COOH),. 111. C,(OH),(COOH),. The conversion of I into I1 is brought about by the oxidation of tartaric acid in presence of ferrous iron, of I1 into I by hydrogen iodide, and I1 is easily oxidised to I11 by bromine and water.The missing .transformation was that of I11 into 11, that is, the re- duction of dihydroxytartaric acid to the acid C,H,06. This trans- formation has now been effected in several ways, and can be very easily recognised owing to the striking difference in properties between the two acids ; for example, dihydroxyta,rtaric acid is very easily soluble in cold water, deliquesces slowly on exposure to the air, and its aqueous solution gives no colour reaction with ferric chloride. On the other hand, both the a and forms of the acid C,H406 are very sparingly soluble in cold water, and the hydrated crystals, C,H,06,2H,0, are quite permanent in the air ; moreover, their aqueous solutions give a transient, emerald-green coloration with ferric chloride in presence of mineral acids, and a beautiful blue-violet with ferric chloride followed by caustic alkali in excess.Action of Zinc and Dilute Acid-KekulB (Zoc. cit., 239) showed t h a t sodium dihydroxytartrate, when treated with excess of zinc and hydro- chloric acid, gave rise to racemic and inactive tartaric acids (together with a small quantity of n substance assumed to be tartronic acid), This change is usually explained by supposing the dihydroxytartaric CO*COOH acid to behave as a diketonic acid, I CO C 0 0 H' It is now found that, by using a limited quantity of zinc, the acid C,H,O, may be isolated as an intermediate stage. To dihydroxy- tarta.ric acid (1 mol.) dissolved in water, and mixed with granulated zinc (1 atom), dilute sulphuric acid was gradually added, the mixture being kept cold by ice; when all the zinc had dissolved, a portion of the liquid, on being tested with ferric chloride, was found to give strongly marked colour reactions characteristic of the acid C4H406.The remainder of the liquid was carefully mixed with about one-tenth of its volume of strong sulphuric acid, twice extracted with ether, and the ethereal solution evaporated ; the white residue thus obtained was very sparingly soluble in cold water, but dissolved readily in warm water, and the aqueous solution showed all the reactions of the acid C,H,O,. On cooling this solution, crystals separated which, when ex- amined under the microscope, mere found t o consist entirely of the @modification of the acid." it Commercial sodiiini dihydroxytartrate gave an exactly similar result.DIITYDROXYTARTARIC ACID.PART I. 79 The constitution of this P-modification has not yet been ascertained, but it was shown (loc. cit., 69,561) that, not improbably, it is dihydroxy- fumaric acid. I f so, its productmion in the manner just described would seem t o be somewhat analogous t o the formation of pinacones, the group -Y=O -E-OH OH-C- becoming o=c- But this /3-acid may, instead, have the alternative formula suggested, namely, yH(OH) *COOH CO*COOH ' in which case the change would be readily understood. Hydrogen sulphicle or stunnous chloride, when used in limited quan- tity, effects a similar reduction. Action of Hydyoyen BronuzicEe.-Attempts were made by various methods to bring about the dehydration of dihydroxytartaric acid, and, with this object in view, the action of hydrogen bromide was studied. It has been previously shown that, in the case of dihydroxymaleic acid, the action of hydrogen bromide in glacial acetic acid solution brings about, as a final effect, the loss of 1 mol.H,O with production of the substance C,H,O, (which appears to be the lactonic acid corresponding with the p-form of the acid). Dihydroxytartaric acid, however, when submitted to this treatment, gave an altogether unexpected result. About 25 grams of the acid were mixed with 250 C.C. of glacial acetic acid, and the mixture satu- rated with dry hydrogen bromide a t about 15", when the acid dissolved completely after standing and shaking.The solution, after about 2 days, was heated in a pressure-bottle a t 60-7'0' for 2 hours, and the product, which was bright orange-red, was then distilled down to a small bulk under reduced pressure. It was observed that practically all the colour passed over with the first half, or so, of the distillate, leaving a colourless residue in the distilling flask. This residile, which set to a crystalline mass on cooling, was found to consist of two sub- stances, one of which was very easily soluble in ether, glacial acetic acid, or cold water, whilst the other was very sparingly soluble in these solvents; the appearance of the product was exactly like that which had been obtained in the case of dihydroxymaleic acid, and on testing it with ferric chloride and aikali, both the easily and the sparingly soluble portions were found to give the reactions of that acid.The sparingly soluble substance was then dissolved in the smallest possible quantity of hot water and cooled as quickly as possible, to avoid loss by decomposition, which began on heating, when crystals began to separate almost immediately ; these, on examination under the microscope, were seen t o consist of the characteristic diamond-80 FENTON : PROPERTIES AND RELATIONSHIPS OF shaped plates of dihydroxymaleic acid. A few crystals of the P-acid could also be distinguished, but this, of course, might be expected since the a-form is slowly transformed into the p-form by hydrogen bromide. The crystals mere collected, washed with a small quantity of cold water, air dried on filter paper for 24 hours, and analysed.0.1538 gave 0.1459 CO, and 0.0577 H,O. C,H,O,,BH,O requires C = 26.0s ; H = 4.02 per cent. That part of the residue which dissolved easily in ether, &c., cor- responded exactly with the product obtained when dihydroxymaleic acid was acted on by hydrogen bromide in acetic acid solution; the ethereal solution, on evaporation, gave radiating, feathery crystals which were extremely deliquescent. I t dissolved in alkalis giving a bright lemon-yellow solution, and on adding ferric chloride, the cha- racteristic blue-violet coloration wasmproduced. The aqueous solution, on standing for some hours, gave short prisms of the P-acid. It follows, therefore, from these results, that when dihydyox~turturic acid is acted on by excess of hydrogen bromide in presence of acetic acid, and the mixture afterwards distilled, a crystalline product is left which is identical with that obtained when dihydyoxyrnaZeic acid is treated in the same manner.The liquid distillate, however, is quite different. I n the case of dihydroxymaleic acid, it is practically colour- less and consists only of acetic and hydrobromic acids; but with dihydroxytartaric acid the first portions of the distillate are bright orange-red. When this distillate is diluted with water and shaken with ether or with carbon bisulphide, the orange-red substance may be extracted, and is found, by all the usual tests, to be bromine. Dihydroxytartaric acid therefore acts as an oxidising agent towards hydrogen bromide, liberating bromine, and becoming itself reduced to the acid C4H40A.This is the converse of the change by which dihydroxytartaric acid was prepared in the manner described above. C = 25-87 ; H = 4.16. The reaction C4H,06 + 2H,O + Br2 C4H,0, + 2HBr is therefore a reversibIe one, the direction depending on the masses of the reacting substances, and probably to some extent on the tempera- ture, The forward * change is brought about at the ordinary tem- perature, when bromine is added in excess, some water introduced, and the operation conducted in an open dish so as to allow of the free escape of hydrogen bromide. The reverse change takes place when a large excess of hydrogen bromide is used, and the mixture is heated in a closed vessel and afterwards distilled; it can take place, t o some * The terms “ forward ” and (‘ reverse ” are used, for shortness ‘ in the sense indicated by the arrows.STEREOCHEMISTRY OF UNSATURATED CARBON COMPOUNDS. 81 extent at any rate, at the ordinary temperature, for if dihydroxy- tartaric acid is mixed with excess of a saturated solution of hydrogen bromide in glacial acetic acid, allowed t.0 stand a few hours, and then kept in a vacuum desiccator over solid potash and sulphuric acid, the residue gives strongly marked reactions of dihydroxymaleic acid as well as of those of dihydroxytartaric acid. If the orange-yellow solu- tion, referred to above, be treated with water, it is instantly bleached, whereas the orange-yellow distillate obtained from it is not affected by water. It is possible that, both in the forward and reverse changes, the unstable compound, C,(OH)2Br2(COOH),' (dibromotartaric acid), is first produced as an intermediate stage, and that this, by the action of water, gives dihydroxytartaric acid, or by loss of bromine gives dihydroxymaleic acid. The reversibility of this action explains the fact, above mentioned, that, on adding bromine to the crystallised acid, the action appears to cease considerably before the calculated quantity of bromine has been introduced, but again proceeds rapidly on the addition of a small quantity of water.