889 LXX V.- Combustion by 3Ienn.s of Chromic Aizhydyide. By C. F. CROSS and E. J. BEVAN. SONE time since we published a preliminary note of results obtained on the combustion of carbohydrates with chromic anhydride and sulphuric acid as a method of ultimate analysis (Chem. News Octo-ber 23 1885). Having observed that carbonic oxide was formed and evolved with the carbonic anhydride we found it necessary to abandon gravimetric methods and to measure the volume of the evolved gas the proportion of carbon per unit of volunie being inde-pendent of the ratio CO CO,. We have employed mercury as the confining liquid and an apparatus consisting of a U -tube graduated in the one limb which is in connection with the combustion flask the other limb being open to the air and connected with a reservoir of mercury by means of which the levels are adjusted for reading (Rep.A n d Chern. 1887 37). In some cases we have employed a Lunge's nitrometer but the particular form of apparatus to be employed for the purpose is rather a matter of detail. Before proceeding to deal with the results obtained we must briefly notice a well-known paper of Ladeiiburg's published in 1865 ( A r m a h , 135 l) on the subject of ultimate analysis by the few methods other than incandescent combustion which are available. In this paper we find chromic acid mentioned but dismissed as unavailable by reason of its decomposition on heating with sulphuric acid with evolution of oxygen. This decomposition was specially studied by one of us (Cross and Higgin Trans.1882 113) with the result of showing that it took place only at high temperatures the evolution of oxygen being imperceptible up to a temperature some degrees above loo" and becoming brisk only as the boiling point of the acid was reached. At the same time it was noticed that) the course of the de-composition differed somewhat when potassium dichromate was snbstituted for the chromic anhydride taking place more readily and yielding a sulphate of different character and composition. Laden-burg having dismissed chromic acid apparently as a result of a general impression of its instability finally selected iodic acid added in the form of silver iodate to the sulphuric acid used as the liquid auxiliary. His combustions were performed with this mixture in sealed tubes, and his estimations consisted in determining the loss of weight after complctely removing the gaseous products and the total oxygen consumed.The results he obtained leave nothing to be desired in point of accuracy and we can only conclude that it is the trouble an 890 CROSS AKD BEVAN COMBUSTION risk attending operations wit,h sealed tubes which have deterred chemists from adopting this method. On the other hand with an open combustion such as that we have employed the operation is a very simple one completed in a few minutes and if proved to be accurate would be a useful addition to our laboratory methods even if its applications were limit,ed. But apart from any question of application the results of such combns-tions partial for certain groups of compounds complete in others, cannot fail to throw some light on the problems of the molecular coiistitution of carbon compounds.We shall therefore in the mean-time content ourselves with recording the results obtained and leave it to the progress of research to determine their value. The proportion of carbonic oxide formed i n the combustion of the carbohydrates depends very much upon the conditions of the decom-position. We have found under those which we have perhaps arbitrarily selected that it is reduced to a minimum ; this being so, and a strict uniformity in the conditions being observed the “ error of the apparatm ” we find satisfactorily constant. The conditions we may note here are :-(1) A quantity of substance to be used yielding from 90 to 100 C.C.of gas; (2) the same volume of sulphuric acid, namely 9 c.c. must be employed ; ( 3 ) an excess of chromic acid of about 30 per cent. beyond that required for complete combustion. The substance cellulose for example being dissolved in the acid, and the chromic anhydride introduced in a small tube into the neck of the flask the latter is connected in the horizontal position with the U -tube and after a suitable interval the level of the mercury is brought to zero The thermometer having been read off the tube containing the chromic anhydride is allowed to fall to the bottom of the flask, and is well shaken with the acid solution. The combustion proceeds rapidly the temperature rising to 60-70” ; after about two-thirds of the gas has been collected it is necessary t o apply heat the remainder of the gas being rapidly expelled a t 100”.The conclusion of the combustion is well marked by the cessation of the frothing which ac-companies the evolution of the gas in the viscous medium. The flask is allowed to cool and the gas volume read off with the usual pre-cautions. If again heated and a second reading taken after cooling, it will be found identical with the first. We may take this together with the numerical results about to be described as sufficient evidence of the stability of the chromic anhydride under the conditions of the combustion. But we shall revert subsequently and more specifically to this point. I n order to “standardise” the apparatus and to determine the error due to the absorption of carbouic acid we made several series of combustions of oxalic acid.We may cite the results of one of these BY MEAXS OF CHROMIC ASHTDR1L)E. 891 consisting of 16 experiments. The weight of substance taken varied from 0.05 gram to 0-2s gram. The mean percentlage of cmbon caleu-lated from the gas volume obtained was 26.0 (C2H20 = 26*6) with a probable error for a single observation of The absorption of GO by H2SOa is 0.75 vol. at ordinary tempera-tures. After reading off and allowing an interval of some minutes before taking a second reading we found that the latter showed no perceptible difference. But after half an hour a difference of level was matnifest; in two hours the diminution of volume amounted to from 3-5 C.C. This is no doubt attributable to a gradual absorption of the gas by the acid.The absorption therefore under the con-ditions of observation adopted is only a fraction of the theoretical maximum and as we find about 217. Supposing this absorption constant and the quantity of substance weighed for combustion to 0.2 per cent. --Cellulose. Swedish filter-paper . Oxalic A c i d . Citric Acid. . Qly cerim Phthalic Acid Salicylic Acid. . Phthalic Anhydride . . Benzoic Acid Weight taken. { 0'1341 0 *1143 U -1147 0 '1280 (0 '1344 0 '1305 0 -1750 0 * 2080 0 a2020 0.1625 0 '1337 0 * 0962 0 -0966 0 -0945 0 *0881 0 -0922 0.0895 0 -0860 1 i { { Gas Polume cor-rected to 0" arid 760 mm. 94'6 C.C. 94.1 ,, 94.5 ), 93.5 ), 94-0 )) 94.9 ) ) 104.3 ), 66.1 )) 64-0 ,, 86.2 ), 102.2 )) 98 *O .>, 108-9 ,, 92.6 ), 101'6 ,, 108% ,, 104.8 ,, 105'0 ,, 109 -0 )) 113-1 ), 107 '5 )) Percentage carbon.Found." 44 -0 43 -6 43.4 43.9 44 .O 44 -3 43 -6 26.3 26 *2 26 -3 26 - 2 26 -0 35-9 37.1 56 *6 60 '2 59 -4 63 -8 63 -8 67 -8 67.8 * The correction t o be applied to these calculated numbers will be dealt with subsequently 892 CROSS AND BEVAN COMBUSTION vary the error in defect should vary in inverse proportion. We found the evidences of such a variation in our numbers but the pro-portion was not sufficiently strict to be made the basis of a correction. We therefore found it necessary to eliminate the variation of the gas volume. Keeping this within the limits of 80-110 c.c.and com-paring substances of varying carbon percentage we find that the error is with sufficient approximation directly proportioned to the carbon percentage. We give a selection of results (p. SSl) obtained in series showing the degree of approximation attainable in combustion according to t.his method. A very large number of our earlier determinations were made with quantities of substance yielding gas volumes varying from 50-120 C.C. We do not think it necessary to reproduce these since they merely serve to indicate the conditions necessary to secure a constant error. Under these conditions already defined the error may be empirically expressed by the following fmction : x 0.4 Carbon percentage of substance 25 _ l _ l _ _ ~ . that is the percentage calculated from the gas volume must be in-creased by this amount to give the true percentage.I n order to in-vestigate the process from the point of view of the oxidising substance the following combustions of cellulose purified bleached cotton were carried out under the above conditions but with a weighed quantity of pot'assium dichromate in excess. The residual chromic acid was determined by titration with ferrous sulphate in the usual way. The chromic acid used is expressed as the equivalenf, of oxygen ; the quantity necessary for complete combustion to COz is given side by side for comparison:-Weight of Oxygen Oxygen to eel lulo ye. consuined. burn to C02. CO coa. 0.1805 0*2010 0.21U7 0.01.69 0.2630 0.1803 0.2060 0.2106 0.0080 0.2770 0.1~40 0.2126 0 2297 0.0294 0.2865 (a.) 0.1164 0.1350 0.1387 0.0065 0.18C5 (6.) 0.1120 0.1300 0.1325 0.0044 0.1755 I n the two latter the combustion-was carried out in connection with the gas collecting apparatus ; the following were the volumes estimated corrected to 0" and 760 mm.(a) 94.5 = 43.5 C per cent. (b.) 91.4 = 43.6 , BY MEANS OF CHROMIC ANHYDRIDE. 893 Taking the mean of these we may apply the correction as above described :-43-55 + 7 43*.55 x 0.4 = 44.25 per cent. carbon. 2a These numbers indicate a complete cornbustion to gaseous products, in which t8he proportion of carbonic oxide is very small. The forma-tion of carbonic oxide we had verified at the time of publishing our first communication by finding a combustible gas burning with the characteristic blue flame left after absorbing the carbonic acid by potash; the quantity of cellulose treated for the purpose being sufficient to yield 500 C.C.of gas. We have since made two observa-t,ions on the gases evolved (u) in the earlier stages ( b ) in the latter stages of the combustion under the conditions obtaining in our quantitative experiments :-(a) 40.5 C.C. treated with Cu2Cl in HCl gave a contraction of (b.) 30.8 C.C. treated with KOH gave a contraction of 30.4 C.C. The corresponding ratio by weight are ( a ) 1 24 The proportion of CO therefore varies with the conditions of the combustion. We have now briefly to notice those cases in which our investiga-tions have shown that the combustion is partial these are more particularly acids of the fatty series and compounds containing nitrogen.Combustions of palmitic and stcaric acids with chromic anhydride yielded from 60-70 per cent. carbon in t'he form of gaseous products, and the results are variable. The partial nature of the combustion is referable t o molecular structure rather than to the relatively large pro-portion of carbon far we have obtained good results with rromatic compounds containing as much as 90 per cent. of carbon. The second group of compounds which only partially burns is that of the organic bases. We propose to investigate this subject further : i t is probable that useful information will be pined by a study of the residual products of combustion. It is worthy of note that under the conditions we have described urea is not attacked.In the course of this research we have made a number of observa-tions on the behaviour of chromic anhydride and potassium dichromate when heated with sulphuric acid. 2.5 C.C. co. co,. co. co,. ( b ) 1 117. We may cite the following :-(a) 0.010 gram G O 3 heated at 100" with 9 C.C. sulphuric acid for three hours. No appreciable reduction 89.2 CROSS ASD REVAX COMBUSTION (b.) 0.300 Cr03 heated with 9 C.C. H,SO1 a t 100" one hour flask connected to gas apparatus. Increase of vol. a t 16" (corr.) 0.75 C.C. ( c . ) 0.300 K,CrzO7 heated with 9 C.C. H?SO two hours a t 105"; residual K,Cr,O,-estimated with FeS04-0.286 ; 0.014 decom-posed in two hours. (d.) 0,590 K2Cr2O7 heated with 9 C.C. H,SO1 two hours a t 105". Residual KZCr,O7 determined 0.5700.( e . ) 1.574 K,Cr,O7 heated with 9 C.C. H2S04 two hours a t 105". Cr,O formed 0.101. (f.) 1-904 CrO? heated with 9 C.C. H,SO two hours at 105O. Cr,Os formed 0.085. Eqt. of 0 evolved 0.0023. 0.020 decomposed. These results are sufficient to show that the instability* of the chromic anhydride begins to be manifest a t 105" ; the decomposition however even at this temperature is extremely slow. It is more rapid with the dichromate than the anhydride and in both cases, increases with the mass. The explanation of this we take to be that the portion remaining undissolved is more liable to decomposition than that dissolved in the acid. It i H t o be observed however that the conditions in the blank experiments above cited differ from those in the combustion process by the absence of the sesquioxide and the presence of a large proportion of undissolved trioxide.The sesquioxide is peroxidised by the ttrioxide as is well known and its presence therefore is an additional element of stability. Finding no mention of investigations of the oxidation of the sesquioxide in presence of sulphizric acid we made observations on this po!'ntl. We find that it is rapidly oxidised to the trioxide by permangannte under this condition ; also though lcss rapidly by manganese dioxide. The comparative stability of the trioxide under the conditions of the combustion is probably therefore due in part to the presence of the sesquioxide as also to the relatively small mass of residual trioxide, and its being dissolved in the acid solution.We made observations on the stability of t'he acid mixture which remained a t the conclusion of a combustion of 0.14 gram of cellulose. (9.) The combustion having been finished by raising the tempera-ture to go" and agitating until no rncre frothing appeared the acid solution (9 c.c.) was sealed in a tube of 20 C.C. capacity. This as heated at 104" for 1 hour. The tube after cooling, was opened in gas-tight connection with a U-tube of mercu1.y. A difference of level of 6 mm. was observed. The total rolume * The anhydride heated by itself begins to decompose at 250" BY NEANS OF CHROMIC ANHYDRIDE. 895 between the sulphuric acid in the tube and the mercury in the U-tube being 30 c.c. the increase of gas in the tube is 0.3 C.C. at 15". (h.) The tube was again sealed and heated for two hours a t 100" The difference of level in the U-t'ube on opening it after cool-ing was 40 mm.; the corresponding increase of volume in the tube 1.6 C.C. (i.) Again sealed and heated for two hours a t 190" the gas evolved in the tube amounted to 2.3 C.C. G.) A similar mixture of sulphuric acid chromic sulphate and chromic anhydride was heated in a flask connected with a gas apparatus as in a combustion for 14 hours a t 105". The in-crease of volume was 3 C.C. These experiments are sufficient to show that the error due to de-composition of the chromic anhydride during the compnrat'irely short period of heating necessary to complete the combustion is so small that it may be neglected. We wish to express oui- thanks to our friends Mr. R. Merrirnan for his assistance in the research and to Mr. A. Green for kindly supplying pure preparations for analysis