INTERACTION OF GLYCEROL AND OXALIC ACID. 151XX. -Interaction of Glycerol and Oxalic Acid.By FREDERICK DANIEL CHATTAWAY.ALTHOUGH the interaction of glycerol and oxalic acid is no longeremployed to prepare formic acid on a large scale, it is still thesimplest process for obtaining a small quantity in the laboratory,and is a practical exercise habitually performed by students. Themore complicated decomposition which takes place when the initialproduct is heated to a higher temperature is by far the mostconvenient source of ally1 alcohol, and is invariably used for itspreparation.The explanations, which are given in most textbooks, of thesefamiliar reactions are fundamentally incorrect.Oxalic acid reacts with glycerol as it does with other alcohols,both an acid and a normal oxalate being produced.The former,* like all such compounds, is unstable at a slightly* This explanation of the production of formic acid has been suggested as anobviocs alternative in Wade’s and Richter’s text-boolrs, but no evidence in supportof it is brought forward152 CHATTAWAY : INTERACTION OFelevated temperature, and decomposes when this is reached intocarbon dioxide and monoformin.The oxalic acid subsequentlyadded displaces the formic acid from the monoformin, and the cycleof operations is repeated.That this is the correct explanation of the reactions leading t othe production of formic acid is shown by the observations thatglycerol and oxalic acid interact readily at temperatures below thata t which carbon dioxide begins to be evolved, and that, althoughthe acid oxalate which must be formed has not yet been isolated,the products of its interaction with aniline and with ammonia,oxsnilic acid, and oxamic acid respectively are readily obtainable.That the whole course of the reaction is as above stated isrendered practically certain by the fact that a precisely similarcycle of operations can be carried out with ethyl alcohol and oxalicacid when the products can easily be isolated a t every stage.As is well known, ethyl hydrogen oxalate, which is formed whenethyl alcohol and oxalic acid are heated together, and which can bedistilled under diminished pressure, decomposes into carbon dioxideand ethyl formate when heated under ordinary atmospheric pres-sure, this being the source of the ethyl formate always obtained insuch quantity when the product of heating together oxalic acid andethyl alcohol is distilled.Oxalic acid, when heated with ethyl formate, displaces the formicacid, producing ethyl hydrogen oxalate.It is possible, although it seems unlikely, that the peculiar decom-position of monoformin, invariably stated in textbooks to be thesource of allyl alcohol, can occur to a very limited extent, but thechief, if not the sole, source of the allyl alcohol is the normal oxalicH,--$!H *CH,*OH, This on heating decomposes o-co*co*o ester, dioxt-lin,into carbon dioxide and allyl alcohol.The presence of this compound in the reaction mixture, after thefirst evolution of carbon dioxide has ceased, is shown by the produc-tion of oxamide or oxanilide when ammonia or aniline is added.These can ocly be produced from a normal ester of oxalic acid, andthe more complicated esters in which two glyceryl residues areunited by two oxalyl residues, although they may exist, are unlikelyto be produced in large amount.As the quantity of oxamide obtainable always corresponds, withinthe limits of experimental error, with the amount of allyl alcoholobtainable, the correctness of this theory of the process is estab-lished.The allyl formate, a little of which is always obtained as aby-product, results from a similar decomposition of monoformoGLYCEROL AND OXALIC ACID.15.3(?H2--?H* CH2*o*CHo, which is produced from mono-0 * co*co * 0 dioxalin,formin by the action of oxalic acid o r from dioxalin by theformation and decomposition of an acid osalate.The small amount of acrolein and the large quantity of carbonmonoxide which also are formed as by-products in the reactionresult from the decomposition of glycerol and monoformin respec-tively by heat, the latter yielding carbon monoxide and glycerol,as formic acid yields carbon monoxide and water, when heated.The main reactions, resulting in the production of formic acid,allyl alcohol, allyl formate, and carbon monoxide respectively,should therefore be formulated thus * : y H ,' 0 * (10- I IC;H=OH +(TO,yH2*OH FH,*O*CO*COPH------*+ cIi,.oIl(?H*OH C O 2 5 (?H*OH .CH,*O+ 70CI1,*OEi c H ,*OH '-9 I 7 H-0-CoC 1 I . p H$!H,*OIl7 H,*O-CO*H CH;OHC11;OH >G+ 7 H,-O*CO*CO,H,--+ YEI-OH + CO-+ YH*OHc02rf C: t i 00 H + 1I.C L2HI Ct4 ,*onp 2CH +3CO,--? I s]tI,*O*$!O 1- CH,.C)H-+ y€-c)--COCI I 2 * 0 € I >p $!H,*O.p p 2 c:O,a YH-O-CO -+ C;H + 3C0,CH,*O*CO~CO,ll CH,-O*CO*HAction of Anhydrous Oxulic Acid on Ethyl Formate.One hundred grams of ethyl formate (2 mols.) were heated toboiling for five hours with 60 grams (1 mol.) of anhydrous oxalicacid.Eleven grams of oxalic acid, which crystallised out on cooling,were filtered off, and as much as possible of the unchanged ethylformate distilled off under the ordinary pressure on a water-bath.* In the first action, either the a- or the P-hjdroxyl grouii may irlteiact, the firstonly is represented as acting.Di- and possibly tri-forruin may also Lq Yimilarlyproduced in snrall quantity and react similarly154 CHATTAWAY: INTERACTION OFThe mobile, strongly acid, pungent-smelling residue, which weighed71 grams and still contained some ethyl formate, was fractionatedunder diminished pressure, when 12 grams of formic acid, 20 gramsof ethyl hydrogen oxalate, and 12 grams of ethyl oxalate wereobtained.Actiod of Anhydrous Oxalic Acid on Glycerol.Nine grams of finely-powdered anhydrous oxalic acid (1 mol.)were thoroughly mixed with 184 grams (20 mols.) of glycerol(D 1*2638), and the acid dissolved by warming to about 50° for ashort time. The liquid was then allowed to remain for three monthsa t the laboratory temperature, small weighed quantities beingremoved from time t o time, and the free acid titrated withN/lO-potassium hydroxide.The titre fell rapidly a t first, and moreslowly afterwards, until it became practically constant at about54 per cent. of its original value. No recognisable amount of carbondioxide was at any time given off.An excess of concentrated aqueous ammonia was added to aquantity of the h a 1 product, when a copious precipitate consistingof oxamide and ammonium oxamate was formed. The additionof concentrated aqueous ammonia t o a similar quantity which hadbeen neutralised as rapidly as possible by N / 10-potassium hydroxidegave no oxamide, showing that the normal eater is very rapidlypartly hydrolysed t o a salt of the acid ester.When concentrated aqueous ammonia is added to precipitate theoxamide this partial hydrolysis takes place to a considerable exf,ent,so that a much larger yield of oxamide with a correspondinglysmaller yield of ammonium oxamate is obtained when a saturatedalcoholic solution of ammonia is used.When the product obtained as above by the interaction of oxalicacid and glycerol is warmed for some hours to about 90° with excessof aniline, a mixture of oxanilic acid and oxanilide is produced,which can easily be separated on account of the sparing solubilityof the latter.The fall of the titre and the production of these compounds showthat an a.cid ester and a normal ester are formed without evolutionof carbon dioxide when glycerol and oxalic acid interact.If such a mixture is heated until the first evolution of carbondioxide ceases and is then allowed to react with alcoholic ammoniaor a.niline, although oxamide or oxanilide is obtained as before,neither oxamic acid nor oxanilic acid is produced, showing thatduring the first evolution of carbon dioxide the acid oxalic ester isdecomposed.I n the normal ester the oxalyl residue must be attached in thGLYCEROL AND OXALIC ACID.155C]H,-YH*CH,*OH 9 o*co~co*o or in some manner shown in the formulasimilar way. This normal ester decomposes during the secondevolution of carbon dioxide, for if the product left when thedisengagement of gas has ceased is treated with ammonia nooxamide is obtained.I n order to prove experimentally that the allyl alcohol is formedby the decomposition of this normal ester, it is necessary to showthat the amount of the latter formed corresponds with the amountof allyl alcohol obtainable by further heating the product.Sixty-three grams of anhydrous oxalic acid and 252 grams ofglycerol, the relative proportions found most advantageous byTollens and Henninger, were mixed and heated for a few hoursto 80-90° on a water-bath, and then under diminished pressureuntil the temperature of the liquid reached 1 8 0 O .The productthus obtained weighed 280 grams. A tenth of this (28 grams)was cooled, mixed with a cold saturated solution of ammoniain absolute alcohol, and a rapid stream of dry ammonia passedthrough the liquid for some minutes. The heavy, white precipitateof oxamide which was deposited was collected, well washed withhot water and alcohol, and dried.It was found to weigh 1.8 grams.The remainder (252 grams) was then heated under the ordinaryatmospheric pressure until the temperature of the liquid reached270O. The distillate, which weighed 24.5 grams, was allowed t oremain over dry potassium carbonate, then separated, the potassiumcarbonate washed with a little ether, and the mixed liquids frac-tionated, using a distilling column. Nine grams of allyl alcohol ofcorrect boiling point were obtained.The amount of oxamide precipitated from the one-tenth showsthat in the remainder which was distilled there must have been26.8 grams of the normal ester, which shovld have yielded 10.6grams of allyl alcohol.The result is within the limits of experimental error consideringthe conditions of thO experiment, and proves that the formationof allyl alcohol is due to the decomposition of the normal ester.The quantity of allyl formate produced was too small to be isolatedsatisfactorily ; if its amount could have been determined the approxi-mation would have been somewhat clmr.On treating with alcoholic ammonia, the residue left after heatingto 270°, from which no more allyl alcohol could be obtained, nooxamide was precipitated,The carbon monoxide so freely liberated during the later stagesof the heating is formed by the decomposition of the mono- andpossibly di-formin produced during the first evolution of carbo156 CAIN AND SIMONSEN: NITRO-ACIDS DERIVED FROMdioxide, whilst the acrolein also produced is due to a decompositionof the glycerol itself at the high temperature of the reaction.Thewhole of the monoformin ie not destroyed a t the temperature atwhich the normal ester decomposes, and if the residue, even afterheating t o 250--260°, is distilled in a, current of steam a smallquantity of formic acid can be obtained.I n conclusion, the author desires to express his thanks to Mr.Dalziel, General Manager of Price’s Candle Company, who kindlysupplied him with the specially distilled glycerol (D 1.2638) withwhich the work was carried out.UNIVERBITY CHEMICAL LABORATORY,OXFOKD