504 WRIQHT AND LUFF’S RESEARCHES LX-Second Report to the Chemical Society on ‘( Research.es on Some Points in Chemical Dpamics. ” By C. R. ALDER WRIGHT D.Sc. (Lond.) Lecturer on Chemistry, and A. P. LUFF Demonstrator of Chemistry in St. Mary’s Hos-pitd Medial School. IN the first report it was shown that when oxide of copper and ferric oxide are reduced by the three agents carbon oxide hydrogen and amorphous carbon the temperatures at which reduction first becomes perceptible after a time varying from a few minutes to half an hour or so depend firstly on the physical condition of the metallic oxide employed and in the case of carbon on that of the carbon; and secondly on the amount and character of the heat disturbance ensuing during the reaction the rule being uniformly noticeable in comparable cases that the greater algebraically speaking is the heat evolution (Le.the greater the actual evolution if the heat disturbance be of + sign and the less the heat absorption if the heat disturbance be of -sign) the lower is the temperature of initial action. Moreover it has been shown that the temperatures at which these three agents first, begin to act on cuprous oxide are no further removed from those at which they begin to act on various forms of cnpric oxide than may be reasonably supposed to be due to difference in physical state ; this being a particular case of the above rule since the heats of combus-tion of copper to Cup0 and to CuO are practically the same (calculated per equal amounts of oxygen consumed). In order to see if these rules are also obeyed by other metallic oxides experiments have been made as described below with different kinds of oxides of various other metals ; the determinations were made in the main in the same way as those described in the first report the chief differences being that in determining the temperatures of initial action of carbon oxide the paraffin bath was mostly discarded the air-bath arrangement described in the first report as employed in certain cases for temperatures above 250° being used instead as more convenient ; and that in some cases the criteria adopted as to the tem-peratures at which action commenced were for reasons dependent on the nature of the oxides examined different from those employed in the case of the oxides of copper and iron ; thus in many cases espe-cially with hydrogen the temperature of initial action was determined by finding a t what temperature a perceptible loss of weight first took place after 15 minutes or more exposure to the reducing agent ; in some instances (with super oxides) the loss of oxygen was arrived a ON SO51E POINTS IN CHEXICAL DYNAMICS.505 by boiling the substance with hydrochloric acid receiving the evolved chlorine in potassium iodide solution and titrating the iodine liberated ; by comparing the numbers thus found before exposure and after the amount of oxygen removed was readily deducible (this method is re-ferred to briefly hereafter as the " iodine process "). I n some instances samples of amorphous carbon prepared from carbon oxide by ferric oxide but different from the specimen used in the first report were employed wherefore the corrections for amount of occluded gases are not always the same ; and in a few cases a bath of sulphur vapour was employed instead of a melted solder bath with test pieces of metal when a temperature of about 420" was required.9 1. OXIDES OF IRON LOWER TMN FERRIC OXIDE. (A.) Ferrous Oxide. It has been showu by Lowthian Bell (Chemical Phenomena of Irm Smelting p. 28) that when a mixture of equal volumes of carbon oxide and dioxide is passed over ferric oxide at a red heat reduction goes on until about one-third of the oxygen in the ferric oxide is removed ; whilst conversely spongy metallic iron heated in the same gaseous mixture to the same temperature becomes oxidised to an extent indicating an addition of about two-thirds of the quantity of oxygen requisite to form ferric oxide; in other words that whether metallic iron or ferric oxide be originally employed the resultant sub-sfance is an oxide of composition tolerably close to that indicated by FeO.A quantity of this stable oxide was prepared by passing equal volumes of carbon oxide and dioxide (prepared from oxalic acid and sulphuric acid) over pure ferric oxide at a bright red heat for eight hours. After cooling down in a stream of the mixed gases, 1.3680 gram of the product gave 1.42370 E'e203 whence the composition was nearly that indicated by FeO ; in order to make sure that the action had gone on as far as possible the whole was well intermixed, and heated again to a bright red heat in the same gaseous mixture for eight hours more.After cooling 1.0425 gram gave 1.1360 Fe205, Calculated. Found. - h f \ For Fe203. For FeO. After 1st 8 hours. After 2nd 8 hours. Fe'= 70.00 77.78 76.09 76.28 whence scarcely any change had been produced during the second eight hours. The substance finally obt>ained contained a perceptible amount of occluded gases; on heating in a Sprengel vacuum it be 506 WRIGHT AND LUFF'S RESEARCHES haved in just the same way as carbon Le. on heating to a given temperature gas was given off for a few minutes and the evolution then ceased ; on raising the temperature to another limit more gas came off for a short time and then entirely ceased coming off; whilst on further raising the temperature the same result again occurred.The following amount of gas was extracted per gram of substance :-At 360" (in mercury vapour). . 360-420" (up to melting point 420460" (antimony thorough-1.83 C.C. at 0" and 760 mm. of zinc) 1-50 , 9 , ly melted) 0.67 , ¶ -4.00 Of this gaseous mixture 12.5 per cent. was absorbed by potash the remainder being almost wholly carbon oxide ; whence it results that the composition by weight of the ultimate product was-Iron 76.28 Occluded CO + CO,. . 0.55 Oxygen (by difference). . 23-19 Or omitting the occluded gases the composition approximated to FeO, being much more nearly FeuO17 = 14Fe0,Fe203 = 13Fe0,Fe304. Calculated. Found. 7-For FeO. For FeuOg. Iron . 77.78 76.71 76.68 Oxygen 22.22 23-29 23.32 -~ 0 0 ~ 0 0 100~00 100~00 No iron in the metallic form was present as on digesting with boiled water and iodine no trace of iron was dissolved out.The action of carbon oxide on this substance when examined as described in the first report was found to be imperceptible at all temperatures below 270" ; at 275" the baryta-water just began to grow turbid after 3-4 minutes' passage of gas and at 290' it became very turbid after the passage of only a few bubbles of gas. In order to find out how far determinations by loss of weight would agree with those made with baryta-water as a test of initial action, 0.794 gram was heated in anarrow U-tnbe containing the oxide at the bend with a plug of glass wool in front of the oxide to prevent an ON SOME POINTS IN CH.EMICAL DYNAMICS.507 mechanical loss of particles by the current of carbon oxide which was passed through the tube heated by an air-bath. No appreciable loss of weight was observed below and up to 270" ; at 280" the loss after 15 minutes was 0.0025 = 0.31 per cent. Hence both the baryta-test and the loss of weight experiment indicated about 275" as the tern-perature of initial action ; but as gases containing COz were expelled at this temperature on heating in a Sprengel vacuum it is possible that the turbidity of the baryta-water and the loss of weight mere partly due to this cause although not wholly inasmuch as the gas expelled on heating to 360" in a Sprengel vacuum corresponded to only 0.24 per cent. by weight and hence the amount expelled at 280" would have been less than this and consequently less than the 0.31 per cent.lost on heating in carbon oxide for 15 minutes. To settle this point definitely the experiment was repeated a current of pure nitrogen (from ammonium chloride and potassium nitrite) being substituted for carbon oxide. After passing through baryta-water for 15 minutes the effluent gas produced a trace of turbidity in baryta-water the substance being heated to 290-300"; the amount of occluded gas expelled however was so minute that no perceptible loss of weight was produced in 1.4330 gram of substance the loss being less than 0.25 milligram and hence less than 0.02 per cent. The same result was found on heating to 320" in a current of nitrogen for 15 minutes ; no appreciable loss of weight occurred either in this case or in another experiment in which air was substituted for nitro-gen the temperature being kept at 300" or 15 minutes.It hence results firstly that about 275" is the true temperature of initial action of GO on the ferrous oxide used ; and secondly that the gas occluded by the ferrous oxide is removed far more rapidly by heating in vacu.0 for a short time than in air or nitrogen; i.e. gases do not under all circumstances act as vacua to one another. The temperature of initial action of hydrogen was determined by the loss of weight process ; the weighings were always made with the tube full of hydrogen the exit end being first closed by a tightly fitting india-rubber cap and the entrance end being then disconnected from the hydrogen generator and rapidly closed in the same way ; it waB found that with due care entrance of air could always be avoided, so that no error was introduced by the displacement of hydrogen by the heavier air.In this way no loss of weight was noticeable after 15 minutes' action at various temperatures below 300" or at that tem-perature. At 310" 0.9200 gram lost in 15 minutes 0.0010 gram = 0.11 per cent. At 360" (boiling mercury) 0.7460 lost in 15 minutes 0.0100 gram = 1.34 per cent 508 WRIGHT AND LUFF'S RESEARCHES Range of temperature Carbon dioxide collected Other gases Time of heating in minutes Total gas collected Gas occluded by ferrous oxide Ditto by carbon Total gas due to occlusion . Gas due to reduction of FeO. --At 360" (boiling mercury) 0.7460 lost in 30 minutes 0.0160 gram = 2.14 per cent A visible condensation of moisture was observed in the last two ex-periments ; that the small loss of weight in the first experiment was not due to the removal of occluded gafies is evident from the circum-stance above mentioned that on heating in a current of nitrogen for 15 minutes a t 320" no appreciable loss of weight occurred.Hence 305" may be taken as the temperature of initial action of hydrogen on this ferrous oxide. In order to determine the temperature of initial action of carbon, 0% gram of the oxide and 0.3 of carbon (from CO) were ground together in an agate mortar and treated as described in the first report. The following numbers were obtained the amounts of gas collected being given in C.C.at 0" and 760 mm.:-~ Up to 360" 3 6 0 4 0 " 440-460" 30 15 16 ------0.6 - 1.5 1.2 - 2.8 1.8 2.3 43 ---.--1.1 2.0 2.0 0.3 0.4 0.4 - 1.4 - 2.4 - 3.2 0.4 0.1 1.1 - -Up to 440" the expulsion of gas was fitful going on for a short time on raising the temperature and then stopping till the temperature was again raised; in short the gas collected was simply that due to occlusion. At 450" and upwards however a slow regular evolution of gas was observed whence the temperature of initial action may be taken as being 460". It results from these experiments that the temperatures of initid action of carbon oxide hydrogen and carbon are situated respectively at near 275" 305" and 450" following the same order as that observed with ferric oxide and the oxides of copper.(B.) Ferroso-ferric Oxide. Magnetic oxide of iron was prepared from the rolling mill scales of boiler plate manufacture by finely powdering sifting and digestin ON SOME POINTS IN CHEXICAL DYNAMICS. 509 Temperature 360" Time in minutes 30 GO2 produced -Other gases --- -with a solution of iodine in potassium iodide to dissolve out particles of metallic iron ; about 0.3 per cent. of metal was thus extracted ; the whole was then quickly filtered and washed successively with water, alcohol and ether and exposed to the air. The oxide thus obtained gave the following numbers on analysis 1.55 per cent. of Si02 being subtracted; the ferrous iron was determined by titmtion with per-manganate :-Calculated for Fe30+ Found. Fe as FeO 24.14 24.1 7 Fe as Fe203 48.28 48.76 0 27-58 (by diff.) 27-07 100~00 100*00 On heating to redness in a Sprengel vacuum no gas was given off.The action of carbon oxide was found to begin at 200" by the Hydrogen caused no loss of weight at temperatures below and up At 300" 1.36410 gram lost in 15 minutes OoOOIO gram = 0.07 per At 360" (mercury vapour) 1.163 gram lost in 15 minutes 0°0050 At 360" (mercury vapour) 1.163 gram lost in 30 minutes 0.0150 Hence about 290° may be taken as the temperature of initial action. With amorphous carbon the following numbers were obtained the experiments being conducted prccipely as described in the first report. baryta-water test GO2 being copiously formed at 213". to 280'. cent. gram = 0.43 per cent. gram = 1-29 per cent.360-420" 42044" 15 15 --- -- -Totalgas Gas due to occlusion Gas due to reduction --~ ~~ ~~ -23 *5 *75 -3 -4 -7 Nil Nil Nil -Hence 450" may be taken as the temperrtture of initial action. On comparing the foregoing results with those obtained with ferric oxide prepared by ignition processes described in the first report, 510 WRlGHT AND LUFF’S RESEARCHES the following numbers are obtained as the temperatures of initial action :-Ferroso-ferric Reducing agent. Ferric oxide. oxide. Ferrous oritlv. -’--From ignition From ignition of precipitated of FeS04. Fe,O,. Carbonoxide . . 202 220 300 275 Hydrogen 260 245 290 305 Carbon (from CO) 430 430 450 450 from which it appears that the difference in temperature of initial action of each reducing agent is hardly greater than may reasonably be ascribed to difference in physical texture due to differences with modes of preparation of the different oxides; that is the oxides of iron like those of copper are first acted on by the three reducing agents carbon oxide hydrogen and carbon at sensibly the same temperatures when in the same physical states whatever the composi-tion of the oxides ; whence it probably results that the heats of corn-bustion of iron to FeO to Fe30p and to Fe203 are sensibly the same (calculated per equal amount of oxygen combined) just as the heats of combustion of copper to Cu20 and to CuO are sensibly the same (Andrews).That this is so is indicated by the valiies obtained by J u l i u s Thomsen as the ‘‘ heat of formation” of ferrous and ferric hydrate from iron oxygen and water (Juhresbricht 1873 p.Sl), viz. 68280 for FeO,H,O and 1911.30 for F20j 3H20 or 68280 and 63710 per 16 grams of oxygen it is not-iceable that the heat of for-mation of the lower oxide is somewhat but not much greater than t h a t of the higher one whence the heat evolution during reduction of the first must be slightly less than that during reduction of the second ; whilst accordingly if there is any noticeable difference between the temperatures of initial action of reducing agents on FeO and on higher oxides it is in the direction indicated by the thermal rules arrived at from the previous experiments the FeO requiring i% slightly higher temperature than the other oxides. In the case of ferroso-ferric oxide the presence of small quantities of silicate titanate &c.and the denser texture produced by fusion, appears to raise considerably the temperatures of initial action of these three agents ; thus the following numbers were obtained with New Zealand iron-sand picked out by a magnet after crushing and well washing to get rid of particles of dust &c. and finally powdering finely in an agate mortar. Neither carbon oxide nor hydrogen had any action whatever on the iron-sand at the temperature of boiling mercury ; both however acted at the temperature of boiling sulphur, weight being lost and COZ being formed on passing CO over the or ON SOME POINTS IN CHEMICAL DYNAMICS. 511 at that temperature and a yet greater loss of weight ensuing when hydrogen was used instead of carbon oxide.1.256 gram heated in CO for 15 minutes in sulphur vapour lost 0.0030 = 0.24 per cent. A copious precipitate was formed in baryta-water by the exit gas. 1.0995 gram heated in H for 15 minutes in sulphur vapour lost 0.0035 = 0.32 per cent. 1.0995 gram heated in H for 30 minutes in sulphur vapour lost 0.0105 = 0.95 per cent. It was found impracticable to obtain accurately the temperatures of initial action in these two cases ; it does not follow that because at about 420" hydrogen acts somewhat more energetically than carbon oxide therefore the initial temperature of action of hydrogen is below that of carbon oxide ; for i t has been noticed in other cases that the loss of weight on heating in CO to a temperature somewhat above the point of initial action takes place more slowly than in hydrogen under analogous conditions a circumstance probably connected with the superior lightness and diffusibility of hydrogen moreover the follow-ing experiment indicates that the initial action of CO is really mani-fested at a lower temperature than that of hydrogen.A gram of finely powdered iron-sand was introduced into each of two similar tubes holding about 4 c.c. one of which was filled with CO and care-fully sealed and the other similarly filled with hydrogen and sealed : both tubes were then heated in mercury vapour for three hours. After .cooling no deposition of moisture was apparent in the hydrogen tube, and no absorption took place on opening the tube under mercury ; whence evidently no action had here taken place.With the other tube however action had evidently taken place ; for on opening the tonbe no absorption ensued but 25 per cent. of the contained gas had become converted into COz being absorbed by potash and preci-pitating baryta-water. It is noteworthy that an analogous result as regards the effect of long continued heating to a temperature somewhat below the point of initial action as deduced from experiments lasting a few minutes only has been observed with other metallic oxides thus cupric oxide pre-pared from ignition of the nitrate is not visibly acted on by pure CO during 5 to 10 minutes at temperatures below 125"; but if a large quantity of this substance he sealed up with CO and heated to 100" for several hours a copious prodtiction of COz ensnes.Thus in two experiments in each of which 6 grams of this cupric oxide were sealed up with about 30 C.C. of CO the following numbers were obtained: After 8 hours at looo gas contained 70 per cent. of GO,. 9 9 72 , 9 ) 3 9 90 , 512 WRIGHT AND LUFF'S RESEARCHES 360420" 15 --. -It is to be remarked however that the conditions were not precisely the same in the experiments for a few minutes at 125" and those for several hours a t looo irrespective of temperature difference for in the first cage a stream of gas afi the ordinary pressure was employed, whilst in the second the pressure was increased to about -2 atmosphere since the tubes were filled and sealed cdd and then heated to 100"; some experiments on the absorption of CO by lead monoxide (0 3A) indicate that increase of pressure is correlative with an alteration in rate of chemical action in such cases.In connection with these points we are endeavouring to obtain the data for the construction of curves representing the correlations between the deoxidation taking place with given oxides and reducing agents the time of action and the temperature and pressure the other conditions being as fa.r as possible the same. On heating the iron-sand with carbon (from CO) it was evident that in this also a higher temperature was requisite before action com-menced than with the other samples of iron oxides the following numbers were obtained :-420-460" 460" to about antimony 500° fully well melted. red-hot in daylight. 15 I 15 I_-- 1.8 -Temperature Time in minutes ~ ~~~ Carbon dioxide formed Other gases > ? -Total gas formed Gas due to occlusion.. Gas due to reduction ---360" 30 No continuous evolution of gas took place until the temperature was considerably higher than the fusion point of antimony so that 4430" at least may be taken as temperature of initial action. The temperatures obtained with this iron-sand then are-Carbon oxide Between 360" and 420'. Hydrogen Carbon (from G O ) . . Ditto but higher than with GO. About 480". 9 2. OXIDES OF MANGANESE. It has just been shown that in the case of iron and copper the tem-perature of init.ia1 action (during some 5-30 minutes) of carbo ON SOiW POINTS IN CHEMICAL DYNAMICS. 513 oxide hydrogen and carbon are practically the same whatever oxide of the metal examined be employed provided the physical state is about the Bame ; whilst coincidently the heat evolved in converting a given quantity of oxygen into any one oxide of either metal severally is practially equal to that evolved in the formation of any other oxide of the game metal.A priorz' it would seem unlikely that the heats of formation of the higher oxides of metals which like manganese form superoxides decomposed by heat alone and not corresponding to stable well-defined series of salts would be as great as the heats of formation of the lower oxides of these metals respectively; whilst as shown by J. T H o M s E N (Jalwesbericht 1873-80) the heat of formation of manganous hydrate Mn0.H20 is far greater in proportion to the oxygen employed than that of hydrated manganese dioxide Mn02.H20, the values being 94,770 and 116,280 respectively or 94,770 and 58,140 per 16 grams of oxygen.It should hence result if the rules hitherto traced out connecting the initial temperature at which an action of the form-AB + C = A + BC takes place be generally applicable that the action of a given reducing agent on manganese dioxide begins to take place at a lower tempera-ture than that at which the reducing action on the lower oxide is first noticeable. To Ree if this is the case the following experiments were made from which it clearly results that the thermic rule applies here also :-(A.) Precipitated Maaganese Dioxide. Two specimens of manganese dioxide were prepared by precipitation processes ; the first by adding to solution of pure manganese chloride excess of caustic Roda adding bromine to supersaturation leaving the solution a t rest for 48 hours and filtering off after repeated washings by decantation ; the second by boiling potassium permanganate so-lution with dilute nitric acid and thoroughly washing on a glass-wool filter.Jet-black and somewhat glossy substances were obtained after drying at 100"; neither product however was pure manganese dioxide ; for in addition to water of hydration a certain amount of an oxide lower than MnOz was present in each. Thus the following analytical numbers were obtained :-I. Specimen prepared by Bromine process.-O3615 gram heated with hydrochloric acid and the vapours received in potassium iodide solution liberated iodine equal to 65.5 C.C.of decinormal iodine solution (titrated by thiosulphate) corresponding to 0.0524 oxygen = 14.49 per cent. 0.3885 gram corresponded to 70.9 C.C. = 0.05672 oxygen = 14.59 per cent! 514 II-RIGHT AND LUFF'S RESEARCHES 0.7030 gram ioited over a gas blowpipe left 0.6110 gram of residual oxide ; this however contained somewhat more oxygen than Mn;04, since on titration by the iodine process it corresponded to 73.9 C.C. of decinormal iodine = 0.0591 gram oxygen = 9-67 per cent. whilst had it been Mn304 it would only have corresponded to 53-4 C.C. of de-cinormal iodine = 0.0427 oxygen = 6.99 per cent. ; the amount> of MnO present therefore. was 0.6110 - 0.0591 gram = 0.5519 gram = 78.51 per cent.0.3420 gram was heated in a current of air for several hours to 200" till perfectly constant in weight; the loss of weight was 0.0240 gram = 7.02 per cent. That no material portion of this loss was due to the evolution of oxygen was shown by titrating the residue by the iodine process when it represented 60.7 C.C. of decinormal iodine = 0.0486 gram oxygen = 1421 per cent. reckoned on the undried substance or practically the same as that found before heating to 200'. These percentages are close to those reqriired for the formula-hfn11020,4Hz0 = 7MnOZ,2Mn2O3,4H2O, Calculated. Found. l l M n O . . 781 78.33 78-51 0 9 144 1443 14-49 14-59 4H,O 72 7.22 7.02 7MnO2,2Mn20,4H20 . . 997 100*00 11. fipecimew prepared from Perrnaaganate.-0-7510 gram ignited over a large Bunsen lamp left of residue 0.6620 gram ; on repeating the ignition the residue slightly gained in weight now weighing 0.6650 gram ; whilst after heating almost to whiteness for 15 minut,es over the gas blowpipe the crucible being surrounded by a clay jacketting furnace it weighed 0.6545 gram.Even this final pro-duct however was not Sln304 inasmuch as 0.6485 gram of it treated by the iodine process corresponded to 69.1 C.C. of decinormal iodiiic = 0.05528 oxygen or 8.52 per cent. ; whilst Mn304 requires 6.99 p e ~ cent. Since however the product of the last heating contained 100 - 8.52 per cent. of &In0 = 91.48 per cent. the original dioxide employed contained 0'6545 - x 91.48 or 79.73 per cent. of Mn0.S U 7510 1.6725 gram in Fresenius and Will's apparatus gave bJ- the * In connection nith thew numbers and those obtained previously with thc firlt specimen of precipitated MnOz indicating that the composition of the oxiclu obtained by igniting Mn02 is variable and contains more oxygen than it is noticeable that thc proclncts obtained by strongly calcining in the air manpnebe carbonate and manqmie~e monoxide MnO exhibit similar peculiarities.T-ide infru 9 2 C D and E ON S031E POINTS IN CHEMICAL DYNSMICS. 515 oxalic acid process 1.4070 C02 representing 0.2558 oxygen = 15.29 pel. cent. 01965 gram corresponded to 37.6 C.C. of decinormal iodine when titrated by the iodine process = 0.03008 oxygen = 15.31 per cent. 0.5435 gram heated to 200" in a current of dry air lost 0.0230 gram, a drying tube attached gaining 0.0250 (hence as with the first speci-men no oxygen was expelled at 200" for otherwise the loss of weight would have ezceeded the water collected; it is shown below that no oxygen was evolved on heating this specimen in a Sprengel vacuum to any temperature short of 260O).0.5165 gram lost at 200" 0.0230 gram. These percentages are close to those required for the formula-Mn2,0,,5H20 = 15Mn02,3Mn203,5H20. Calculated. Found. 21Mn0 1491 79.78 79.73 0 1 8 . 288 15.41 15.29 15.31 5H20 . 90 4.81 4.23 4-60 4.45 l5MnO2,3Mn2O3,5H20 . 1869 100.00 Comparing this formula with that arrived at for Specimen I, 7MnO2.2Mn2O3.4H20 it is at once evident that Specimen No. I1 contains a somewhat smaller quantity of oxide lower than MnO than is present in No.I. For this reason Specimen I1 (prepared from per-manganate) was chosen for the subsequent experiments detailed below; the point of initial action of hydrogen being determined by loss of weight the substance was rendered anhydrous by heating to 200" in a current of dry air for some time no further loss of weight ensuing on continuing the heating for an hour. Carbon oxide wa.s found to act vigorously on the anhydrous oxide at temperatures below 60" whilst even at the ordinary temperature (30 was slowly formed the issuing gases forming a perceptible preci-pitate in baiytn-water; a tube containing about 20 C.C. of pure CO, and about 1.5 gram of the manganese dioxide was sealed up and left to itself for 10 days at Christmas in a room without a fire (temperature probably from - 10" to + 10").When opened the gases contained iJ2.1 per cent. of COz. It is noteworthy that the non-dehydrated oxide (dried only at 100") acted in just the same way and not apparently any more vigorously ; the other specimen of dioxide (prepared by the bromine process) bc-haved in precisely the same way a perceptible precipitate being formed in baryta-water by passing pure CO over it for two or three minutes at, n temperature of about 18" 516 WRIGHT AND LUFF'S RESEARCHES To determine the temperature at which hydrogen first began to act, weighed quantities of the dioxide dried at 200" were heated in a slow current of hydrogen precisely as above described. No loss of weight was produced by heating for 15 minutes to any temperature below and up to 140".0.2920 gram heated to 150" for 15 minutes lost 0.0015 gram = 0.51 per cent. On further heating to 170" for 15 minutes this lost in addition 0.0040 gram = 1.37 per cent. And on further heating to 190" for 15 minutes this lost in addition 0.0085 gram = 2.91 per cent. Hence the temperature of initial action lies between 140 and 150", and therefore is near 145". In another similar experiment no loss occurred at 140" 0.48 per cent. at 150" €or 15 minutes and a further loss of 2-91 per cent. on again heating to 190" for 15 minutes. As a further check on these numbers 0.3060 gmm were heated in hydrogen to 160" for 30 minutes whereby the loss in weight waq 0.0070 gram = 2-29 per cent. On titrating the '' available oxygen " in the residue by the iodine process it was found to represent 51.9 C.C.of decinormal iodine = Q.04152 gram oxygen = 13-57 per cent. reckoned on the original sub-stance. As the substance originally yielded by the same process 16.04 per cent. the loss of oxygen during reduction was 2-47 per cent., which agrees very fairly with the loss of weight determination. The determination of the temperature of initial action of carbon was complicated by the circumstance that although no oxygen was evolved on heafing to 200" in air yet on heating in a Sprengel vacuum to temperatures below 360" oxygen was continuously evolved. Experi-ments were therefore made simultaneously with the manganese dioxide alone (0.6 gram) and the same quantity of dioxide mixed with half its weight of carbon (from CO) in the same way as those previously described with metallic oxides mixed with carbon.The following values were obtained each pair of simultaneous experiments being exactly alike in all respects save the presence of carbon :-(1.) Heated in mercury-vapour (360") for 30 minutes gas evolved reduced to 0" and 760 mm.-Dioxide alone. Dioxide and carbon. CO, Nil 7.9 C.C. Oxygen . . . . . . 3.5 C.C. Other gases trace Total gas 3.5 8. ON SOME POIXTS IN CHEMICAL DYNAMICS. 517 (2.) Heated in air-bath to 270" for 30 minutes-Dioxide alone. Dioxide and carbon. C02 Nil 2.3 0.7 Other gases . . Nil Oxygen } 0.1 - -Total gas 0 7 2.4 (3.) Heated in air-bath to 250" for 30 minutes-Dioxide alone. Dioxide and carbon. Total gas evolved Kil 0.15 just that due to occlusion by the carbon.It hence results that at a temperature between 250" and 270" and therefore very close to 260" this form of manganese dioxide first begins to evolve oxygen in vuczco and that the action of carbon thereon is also first manifested in vacuo at the same temperature That the carbon does actually act is shown by the circumstances firstly, that in presence of carbon no free oxygen was evolved but only carbon dioxide ; and secondly that the amount of oxygen present in the C02 thus formed is considerably greater in each case than that evolved in the free state in the companion experiments respectively. Thus reckoned per gram of manganese dioxide employed the oxygen eliminated was-Tcmpera-ture. Evolved as COB. Evolved as free 0.Exp. (I) . . 360" 0.0187 gram = 1.87 p. c. 0.0083 gram = 0.83 p. c. , (11) . . 270" 3.0054 , = 0.54 , 0.0016 , =0.16 ,, The ratios of the oxygen evolved as C02 and as free 0 in these two experiments are (since carbon dioxide "occupies the same space as the osygen contained therein)-(1.) = 2.26 to 1 at 360" (24 0.7 2'3 = 3.28 to 1 at 270" whence it would seem that the reducing action of carbon increases less rapidly from 270 to 360 than does the tendency to evolve oxygen i n the free state the increase in the former case being from 2.3 to 7.9 or from 1 to 3.43 and that in the latter from 0.7 t o 3.5 or from 1 to 5.0. On the whole then the temperatures of initial action of carbon oxide hydrogen and carbon on this manganese dioxide are-7.9 CO Below 15" Carbon.. , 260" Hydrogen Near 145" \rOL. XXXIIT. 2 518 WRIGHT AND LUFF'S RESEARCHES or in the same relative order as that hitherto found with iron and copper oxides. (B.) Pyrolusite. A sample of finely crystallised compact pyrolusite was found on qualitative analysis to contain practically nothing but manganese, oxrgen water and SiO,; especially no more than traces of iron or other reducible metaliic oxides were present. At 200" this sub-stance did not become anhydrous although it lost a little moisture ; on analysis the substance dried at 200" till per€ectly constant in weight .gave the following numbers :-1.4330 gram gave 0.0070 Si02 = 0.49 per cent. 1.1860 gram ignited in a current of dry air communicated to 1.2530 gram by Fresenius and Will's process gave 1.078 COz 0.3230 gram by the iodine process corresponded to 63.7 C.C.of calcium chloride tube 0.0420 gram H20. = 0.1960 oxygen. decinormal iodine = 0.05096 0. Mean. SiO 0.49 0-49 H2O 3.54 3.54 0 15.64 15.78 15.71 MnO (by difference) 80.26 100~00 On ca,lculating these numbers info percentages reckoned on the material excluding the 0.49 per cent. of Si02 they are found to agree with the formula Mn,O6, 5H20 = 22Mn02,4Mn20s,5H20, Calculated. Found. 30MnO 2130 80.81 80.66 (by difference) 0 416 15.78 15.78 5H20 90 3.41 3.56 22Mn02,4Mn203,5H20. . 2636 100.00 100.00 so that the natural pyrolusite had a composition closely resembling the precipitated manganese dioxide (prepared from permanganate) so far as ratio between manganese and oxygen is concerned the anhydrous substances being indicated respectively by-Pyrolnsite 22Mn02,4MnzOs = 11Mn02,2Mn20s, Precipitated dioxide 15MnOz 3Mn203 = 10MnO2,2Mn2OS, each being practically manganese dioxide but actually containing somewhat less oxygen than pure Mn02.The pyrolusite however, parted with only minute quantities of combined water at 200° th ON SOME POINTS IN CEI'EMICAL DYNAMICS. 519 water of hydration being only expelled a t ft higher temperature, whilst the water of hydration in the precipitated dioxide (not removed at 100") was practically all lost at 200"). The action of carbon oxide was not noticeable at any temperatme below and up to 8.5". At 87" a slight turbidity was produced in baryta-water on passing the gas for some few minutes and at 95" the reduc-ing action was very manifest.Hydrogen did not begin to act at any temperature below and up to 180" no loss of weight being occasioned after 15 minutes' exposure. At 200' 1.1705 gram lost in 15 minutes 0.0005 gram = 0.04 per cent. At 210" 1.1545 gram lost in 15 minutes 0.0075 gram = 0.17 per cent. At 210" 1.1895 gram lost in 30 minutes 0.0075 gram = 0.63 per cent. At 220" 1.1700 gram lost in 15 minutes 0.0100 gram = 0.85 per cent. Hence the temperature of initial action lies between 180" and 200", and therefore is near 190". In order to prove that the trifling loss of weight at 200-210" was not due to expulsion of combined water a similar experiment was made with air instead of hydrogen ; no loss of weight at all occurred. Moreover whilst as just stated on heating 1.1895 gram in hydrogen to 210" for 30 minutes the loss of weight was 0.0075 gram = 0.63 per cent.on further examining by the iodine process 0.3530 gram of the reduced substance representing 0.3535 of the unreducod substance it corresponded to 66-0 C.C. of deciuormal iodine = 0.0528 oxygen = 14.85 per cent. As the original substance contained oxygen = 15.71 per cent. the loss on heating in hydrogen is 0.86 per cent. which fairly agrees with the loss of weight determination. The temperature of initial action of carbon was determined in the manner above described simultaneous experiments being made with 0.6 gram of the dioxide and the same weight of dioxide mixed with half its weight of carbon. The following numbers were obtained:-Temperature.At 360" far 30 minutes --At 420" for 15 minutes Pyrolusite alone. No gas evolved. GO,. . Nil Oxygen 0-6 Other gases trwe 0.6 ---Pyrolusite and carbon. 0.2 C.C. (due to occlu-sion) . COZ 6.5 } 0.2 Oxygen Other gases -6.7 2 P 520 IVRIQHT AND LUFF'S RESEARCHES Hence between 360" and 420° say at 390° is the temperature at which reduction by carbon and evolution of free oxygen on heating alone in n Sprengel vacuum both commence the former action being at 420" yery much more powerful than the latter. The temperatures of initial action then are-Carbon oxide 87" Hydrogen Near 190 Carbon , 390 again following the usual order. (C.) Mangaiaoso-maizganic Oxides. With the object of preparing the oxide Mn304 the usual process employed in the quantitative estimation of manganese was tried per-fectly pure manganous chloride being precipitated by sodium car-bonate and the resulting manganous carbonate thoroughly washed, dried and calcined for two hours at a pretty bright red heat with fre-quent stirring.The brown-red powder obtained however did not possess the desired composition but corresponded much more closely to Mn,O = Mn304 Mn&. 0.6005 gram heated by the iodine process corresponded to 62.4 C.C. of decinormal iodine = 0,04992 oxygen. 0.4385 gram corresponded to 46.0 C.C. = 0.0368 oxygen. Calculated. Found. / n For M11304. For Mn,Op 3Mn0 213 9301 5Mn0 355 91-73 0 16 6.99 0 32 8.27 8.31 8-39 \ Mn304 229 100*00 Mn507 387 10000 Another specimen of manganous carbonate furnished analogous results; the product of several hours' calcination at a full red heat over a large Bunsen burner finally contained of '' available oxygen " 8.52 per cent.By long-continued strong ignition over a powerful gas blowpipe, especially in a capsule surrounded with a clay-jacketting furnace the substances thus prepared lost in weight forming ultimately substances closely approximating to Mn30a; thus the product of the heating of the second sample above described for three successive periods of 10 minut,es each became almost constant in weight and then gave these numbers :-0.483 gram titrated by the iodine process corresponded to $2.8 C.C. of decinormal iodine = 0.03424 oxygen = 7.09 per cent. 5.533 gram titrated by the iodine process corresponded to 47.1 C.C ON SOME POINTS IN CHEMICAL DYNAMICS.521 Temperature 360" 80A-00° zinc not melted but softened Time in minutes,. 30 15 = 0.03768 oxygen = 7.07 per cent.; calculated for Mn304 6.99 per cent. Another specimen similarly treated finally contained 7.13 per cent. The action of carbon oxide hydrogen and. carbon on each of the above two substances was tried with the following results :-(I.) Manganoso-manganic oxide containing 8.31-8.39 per cent. of available oxygen corresponding nearly to Mn507 = Mn304,Mn203. The action of carbon oxide was first noticeable at 97" by the baryta test the action being well defined at 110". Hydrogen caused no loss of weight in 15 minutes a t any tempera-ture up to and including 230". At 250° 0.5575 gram lost 0.0025 gram in 15 minutes = 044 per cent.At 300" 0.4345 gram lost 0.0035 gram in 15 minutes = 0.80 per cent. At 320° 0.3165 gram lost 0.0095 gram in 15 minutes = 1.42 per .cent., whence the action commenced between 230" and 250" or at about 240'. With carbon from CO the following numbers were obtained 0.6 gram being heated with 0-3 of carbon as usual. 400-420" zinc just melted 15 C02 formed -Other gases - Total gas evolved 0.25 Gas due to occlusion 0.3 Gas due to reduction. . Nil -------- 3.7 - I 0.1 -- -3-8 0.4 i -0.4 0.4 Nil 1 3.4 Hence about 410" may be taken as the temperature of initial action of carbon thak of hydrogen being 240' and that of carbon oxide being 97" Le. the order is the usual one.(11.) Manganoso-manganic oxide containing 7.07 to 7.09 per cent. of available oxygen corresponding nearly to Mn304. By the baryta test the action was first noticeable at 240" being well marked at 260"; on attempting to corroborate these results by loss of weight however no noticeable difference in weight was perceptible afher 15 minutes' beating at any temperature lower than that of boil-ing mercury at which temperature 2.14 per cent. was lost in 1 522 WRIGHT AND LUFF'S RESEARCHES minutes and 3.90 in 30 minutes. That this imperceptible variation in weight (observed at 200" 300" and 320") was due to absorption of part of the CO formed so as about to compensate for the loss of oxygen is rendered highly probable by the following experiment :-0.5 gram of the manganese oxide was sealed up in a tube containing 8-0 C.C.of C O ; the tube was then heated to 250" for 3 hours and after cooling opened under mercury the mercury rushed in filling the tube all but 2.0 c.c. of which 0.5 C.C. was absorbed by caustic potash. Hence the C02 formed by action at 250" had been to a large extent, absorbed with formation of manganese carbonate ; on adding dilute nitric acid to the particles of manganese oxide after the experiment effervescence was distinctly noticeable. Hence 240" is the temperature of' incipient action of CO as nearly as could be determined but it is not at all improbable that this tern-perature is a little too high the non-evolution of CO at temperatures below 240' being due to its retention by the manganese oxide; the action at these lower temperatures however could only have been.very faint no gah in weight being observable. Hydrogen brought about no loss of weight in 15 minutes at any femperature up to and including 250". At 260" 0.2480 gram lost in 15 minutes 0*0005 gram = 0.20 per cent. At 280" 0.2480 gram further lost in 15 minutes 0.0020 gram = 0.80 per cent. At 300" 0.1600 gram lost in 15 minutes 0.0055 gram = 3-44 per cent. At 300" 0.1600 gram lost in 30 minutes 0.0085 gram = 5.31 per cent. At 360" 0.3500 gram lost in 15 minutes 0*0200 gram = 5.71 per cent. On sealing up in a tube about 8 C.C. of hydrogen with 0.5 gram of the manganese oxide and heating to 250" for 3 hours a slight absorp-tion of hydrogen was noticeable on opening the tube about 10 per cent.having disappeared. In a precisely similar companion-experiment in mercury Fapour however the action was much more decided for on opening the tube under mereury the mercury rushed up and filled the whole tube. Hence 255" is the temperature of incipient action as nearly as it can be fixed. On heating 0.6 gram of the MnaOa with 0.3 of carbon from CO the following values were obtained : ON SOME POINTS IN CHEMICAL DYNAMICS. 523 Temperature 360" 360420" Time inminutes 30 15 4 2 0 4 0 " Zinc melted. Antimony not melted. 15 ~~ ~~ ~~~ ~ ~~~~~ - C02 formed Other gases I - 1 Total gas evolved Gas due to occlusion Gas due to reduction. . ---0.25 0.3 Nil -Hence about 430" is the temperature of initial action of carbon. (D.) Ma-nqgntzozcs Oxide.On heating precipitated manganese dioxide to redness for several hours in a platinum boat in a gentle stream of hydrogen it lost weight rapidly at first becoming of a greenish-grey colour and subse-quently more slowly gradually becoming more and more green in shade ; finally a substance of an emerald green tint was obtained re-sembling oxide of chromium but of a somewhat lighter and brighter shade. On heating this product for several hours to a full red heat in hydrogen no further loss of weight ensued and precisely the same result was obtained on substituting carbon oxide for hydrogen ; whence it is evident that the points of initial action of carbon oxide and hydrogen on the green substance lay considerably above 300-600". That this green substance was pure MnO was shown by determining the increase of weight on roasting for some time in the air a t a red heat and also determining the " available oxygen" in the snperoxide thus formed by the iodine process senaibly equal values were obtained, which could not have been the case had the green substance contained manganese and oxygen in any different proportion than that indicated by MnO.Gram. Per cent,. 0.2655 gram roasted for three-quarters of an hour at a full red heat in a small platinum capsule gained in weight On igniting over the blowpipe for five minutes it further gained On further calcining at full red heat for three-quarters of an hour it further gained. . 0.0240 = 9.04 0.0010 = 0.37 0,0035 = 1.32 0*0285 10.73 - 524 WRIGHT AND LUFF'S RESEARCHES Precipi-tated.MqOl3 -02915 gram of the roasted product corresponding to 0.2633 of original green oxide when titrated by the iodine process corresponded to 344 C.C. of decinormal iodine = 0.02752 gram of oxygen = 10.45 per cent. calculated on the original green oxide or sensibly the same percentage as that found by direct weighing. It is worthy of note in passing that the gain in weight had M304 been formed by the roasting would only hat-e been rx-.l = 7.51 per 100 of MnO ; the actual gain 10.45-10.73 corresponds to the com-position Mn15022 = Mn304,6Mn203 which requires 10.52 per 100. On mixing 0.6 gram of the green monoxide with half its weight of carbon from CO and heating in a Sprengel vacuum to various tem-peratures from 360" up to a full red heat continued for half an hour (the tube being almost collapsed) only i.0 C.C.of gas was obtained, that due to occlusion by the carbon employed being as nearly as could be estimated exactly the same 1.0 C.C. Moreover the gas came off fitfully on raising the temperature from stage to stage and not con-tinuously. Hence the point of initial action of carbon also lies upwards of 500-600". On comparing the temperatures of initial action of carbon oxide, hydrogen and carbon on these different kinds of manganese oxides, it is at once noticeable that whilst in every case the first reducing agent begins to act at a lower temperature than the second and the second than the third the several temperatures of initial action are notably higher the less the oxygen in the manganese oxide.16 Py~o-lusite. MnuOa8 Composition by analy-sis of substance re-garded as anhydrous Physical texture . . . . Carbon oxide . . . . . . . . Hydrogen . . . . . . . . . . Carbon. . . . . . . . . . . . . . Fine powder, amor-phous Below 15' Near 146' , 260° Very compmt, crystal-line 87' Near 190" , 390' Manganoso-manganic oxide. Mn&0; Fine powder, amor-phous 97' Near 2 4 0 , 410' Fine powder, amor-phous 240' 255" 4!30" Man-ganous oxide. -MnO Anlor-phous No action at 600" Ditto Ditt OX SOME POINTS IN CHEMICAL DYNAMICS. 325 (E.) Composition of the Oaide of Hawganese foriiaed by Ignition in Air @' Higher and Lower .Manga.lzese Oaides.In the preceding sub-sections A C D numbers have been quoted showing that the substances formed by the calcination at a full red heat (over a good Bunsen burner in a thin platinum capsule) of man-ganous oxide manganous carbonate or manganese dioxide contain much mom oxygen than that due to the formula Mn304. In the c u e of the product from manganous carbonate half -an-hour's in tense ignition over a gas blowpipe caused the expulsion of oxygen and the reduction to an oxide nearly approaching to Mn304 ; but in the case of manganese dioxide conversion into Mn304 was not so nearly eff'ected a larger surplus of oxygen always remaining. In other experiments not quoted above precisely similar results were obtained thus prolonged ignition did not convert the pyrolusite employed above into Mn304 whilst another specimen of MnO roasted in the air over ft Bunsen burner became a brown oxide containing 8.17 per cent.of oxygen. All the numbers obtained indicated a composition lying between Mn@4 and Mn2O3? many of them being close to those required for the formula Mn507 or Mn30d,Mn203. Dittmar has already shown (this Journal 1864 294) that when manganese dioxide is ignited in air at the ordinary pressure or in a mixture of oxygen and nitrogen in which the oxygen tension is near to that in the air (0.21 atmosphere) the oxide ultimately formed is extremely variable in composition between these limits ; when how-ever the oxygen tension exceeds 0.26 atmosphere the product appGoxi-mates more or less closely to Mn@ and when it falls below about 0.19 atmosphere the product is nearer to Mn3O4.Thus on ignition in air at the ordinary pressure (tension of oxygen 0.21 atmosphere) D i t t m a r obtained in two experiments oxides MnO, in which x = 1.352 ftnd 1.384 respectively ; the first corresponding to 4&OC,M2o3 = M n J b = Mno1.357; the second to 3&!h30,,2Mn20 = Mnl3OI8 = MnO1.385; and on ignition in mixtures of oxygen and nitrogen in which the oxygen-tension was between 6.87 and 7.38 inches (0.23 to 0.25 atmosphere), mmetimes Mn& sometimes MnsO3 predominated in the product. The following experiment well illustrates the influence of difference of temperature on the oxide of manganese formed under otherwise constant conditions as to oxygen-tension when that tension is 0.21 atmospliere.About 14 gram of mauganous carbonate was calcined for an hour over a large Bunsen burner in a thin platinum capsule ; the product was then roasted for some hours over the same burner, being weighed at intervals; finally the whole was strongly ignited .over a blast gas furnace and again weighed at intervals. The fol-lowing numbers were obtained : 526 WRIGHT -WD LUFF’S RESEARCHES Gram. Per cent. 1.4095 gram of the product of the calcination for one hour gained on further calcination for two hours 0.0055 = 0.39 0.0040 = 0.29 , 0.0035 = 0.24 Total gain . . 0.0130 0.92 Per 100 of After two hours more the further gain was 9 9 7 7 - -Gram. original substance. The product was then ignited for ten minutes over the blast gas lamp and lost 0.01‘25 = 0.89 Ignited ten minutes more lost further 0,0065 = 0.46 , 0.0040 = 0.28 Total loss 0.0230 1.63 > 9 >, The product of the last heating was found on titration to contain ‘7.08 per cent.of available oxygen (Mn30c requires 6-99 per cent,.), whence the substance ultimately produced by the first series of calcina-tions over the Bunsen burner contained in 100 + 1.63 parts 7.08 + 1.63 of “available oxygen” = 8-52 per cent. corresponding to 2Mn304,3Mn203 which requires 8.58 per cent. From these experiments it results that in the quantitative estimatJiolr of manganese the usual directions of the text books to precipitate as carbonate and ignite until there is no further gain in weight may lead toserious errors if as is usual an ordinary large Bunsen lamp be em-ployed for the ignition it being assumed that the end product is Mn304 ;.whilst if the manganese be precipitated as hydrated peroxide and ignited the same result may be brought about. Either a prolonged exposure of the substance to the highest temperature of a good gas blowpipe is essential to convert it into something approaching to Mn304 o r preferably the “ available oxygen” in the weighed calcined substance must be determined and the true amount of MnO present thence deduced by subtraction from the total weight. The MnO pre-sent may also be obtained by prolonged heating in hydrogen at a bright red heat until weight is no longer lost and weighing the residue. Tbe error produced by reckoning the roasted product just described containing 8.52 of oxygen as Mn304 would amount to an over estima-tion of manganese t o the extent of about 16 parts in the thousand, = 1.6 per cent.; for every 100 - 8.52 = 91.48 of MnO really present would be estimated as 100 - 6.99 or 93.01 parts. A curious phenomenon was observed with a specimen of nearly pure &In0 prepared as just described by rednction in a crucible at a brigh ON SOME POINTS IN CHEMICAL DYNAMICS. 527 red heat by hydrogen. After cooling down in an atmosphere of hydro-gen (the crucible being plunged into at jar of GOz to extinguish the flame) the product was of a tolerably fine green underneath with a brownish film on the surface; on standing exposed to the air for a couple of days under a shade the film became black and much thicker, whilst an increase in weight of 3.93 per cent.took place indicating a spontaneous absorption of oxygen. Nothing of the kind was observed with pure MnO prepared by long-continued heating of the dioxide in a boat in a current of hydrogen until no further loss of weight ensued even after several hours’ further heating ; no gain in weight or alteration in appearance ensued even after several days’ exposure to air. Another specimen of not quite perfectly reduced MnO just resembling the first in appearance was kept over sulphuric acid for a fortnight without altering in appearance or weight; it was then freely exposed to the air under a shade for a few days without notice-able alteration ; at the end of a week however the brownish film had become sensibly darker and thicker and an increase of weight to the extent of 1.06 per cent.had taken place. tj 3. OXIDES OF LEAD. In extension of the preceding experiments with the oxides of man-ganese the following observations were made with various lead oxides :-(A,) Monorcide. Pure lead nitrate was recrystallised and ignited at first gently and afterwards somewhat more strongly but without €usion until no more fumes of oxides of nitrogen were evolved. On heating the product with pure sulphuric acid no fumes were evolved capable of acting on potassium iodide and starch ; and on boiling with hydrochloric acid, and passing the vapours into potassium iodide and starch only a trace of blue coloration was formed showing that the lead oxide was prac-tically free from any higher oxide. Moreover on dissolving 0.9515 gram in glacial acetic acid adding pure sulphuric acid evaporating the poured off fluid to dryness and weighing the total lead sulphate formed 1.2920 gram of PbSOa were obtained.Calculated. Found. P b 207 92-83 92.77 0 16 7-17 (By difference) 7.23 PbO 233 100.00 100~00 On heating this oxide in carbon oxide it was found that a perma-nent change of tint from reddish-yellow to slatey-grey ensued when even only a very minute amount of reduction had taken place thi 528 WRIGHT AXD LUFF'S RESEARCHES change of colour being accompanied by the production of a percep-tible turbidity when tb e partially reduced substance was dissolved in hot acetic acid (the original substances dissolved to an almost perfectly clear fluid under the same conditions).This change of colour was faintly noticeable at temperatures somewhat below tJhose at which any appreciable formation of CO or loss of weight was first discernible and therefore was presumably a more delicate test of in-cipient reduction than these criteria especially as it was not produced at all at any temperature when air was substituted for carbon oxide. Thus on leading pure carbon oxide over the lead oxide for 15 minutes at a time no change of colour took place at 155" or below whilst a slight change was noticeable at 160" anda more decided alteration of tint at 170° and much greater ones at 180' and 190" respectively. The loss of weight in 15 minutes at 160° however and also at 180" was less than 0.25 milligram on about 1.5 gram of substance or less than 0.02 pel.cent. ; whilst no appreciable turbidity was produced in baryta-water by the effluent gases. At 190" the loss of weight first became perceptible in 15 minutes 1.228 gram losing 0*0010 gram = 0.08 per cent. the baryta test first giving a perceptible turbidity at this temperature a very decided action being noticeable at 200". It hence results that the temperature of incipient action during 15 minutes, as determined by the colour test is about 160° whilst that from the loss of weight and the baryta test lies between 180" and 190° say at 185". This circumstance suggests the possibility that the nascent CO formed by slight action at 16U0 was absorbed by the PbO form-ing PbCO, so as to prevent any action on baryta-water ; the fact that there was 120 appccinble gain in weight in 15 minutes however proves that such action if occurring at all was extremely slight'.That the heated PbO could take up CO when not in a nascent condition, although not very readily was proved by passing CO over the PbO for 15 minutes at 200" when no alteration in weight whatever ensued ; after one hour's passage of gas no alteration in weight had taken place but after four hours' passage 1.750 gram of PbO in-creased in weight 0*0020 = 0.12 per cent.; the resulting product effervesced perceptibly with dilute nitric acid whilst the original oxide did not do so. It is especially noteworthy that the slow absorption of CO by PbO at 200" thus manifest is greatly quickened by increase of pressure ; thus about 2.5 grams of the same PbO was sealed up in a tube containing 10 C.C.of CO at the ordinary temperature and was then heated t o 200" for four hours. On opening the tubes under mer-cury the mercury rushed up and comgdetely $Ued the tabe indicating perfect absorption whence it would seem that the increased pressure (about p atmosphere) obtaining at first caused the action to commence sooner and that it went on after being once started so thoroughly a ON SOME POINTS IN CHEAIICAL DYNAMICS. 529 __---Temperature Time in minutes COz collected. . Other gases Total gas Gas due to occlusion Gas due to reduction - to render the tube vacuous the amount thus absorbed representing about 0.0195 gram of CO = 0.78 per cent. of the weight of the PbO employed. (Vide 6 6.) On heating the PbO in hydrogen no alteration in tint was notice-able after 15 minutes' action at any temperature below and up to 185", at 190" a faint change was noticeable after 15 minutes and at 200" the change was much more decided in 15 minutes.No loss of weight was noticeable at 190" in 15 minutes but at 200°,-1.900 gram lost in 15 minutes 0.0010 = 0.05 per cent. 9 9 30 , 0*0020 = 0.10 ,, Hence manifestly 190-195" is the temperature of initial action, both colour test and loss of weight test nearly agreeing whence it is evident that the difference between the values for the point of initial action of CO found above by these two tests (160' and 185" respec-tively) is really due to the retention by the PbO of part OF the first traces of CO formed at the lower temperature thus preventing any perceptible loss of weight being perceptible and diminishing the turbidity produced in baryta-water by the issuing gas below the point at which it is readily recognisable.Carbon first begzm to act at a considerably higher temperature the following numbers being obtained with carbon from CO and the sugar-charcoal described in Dhe first report :--_ -360" 360-420' 420-450" 360' 360-4.20" 30 I 15 - __ 30 - - 23.4 Ti 27.8 0 -2 23.6 0'3 28-0 0 - 5 0 . 5 0 . 5 5 0 . 5 0 '53 0 -6 - -----I_____-OV2 - I - ---- --Nil Nil 23.0 zz 1 27.5 ~~ ~ In the first case no action occurred until zinc was well melted and mitimony was softening so that about 435" is the temperature of initial action. I n the second case no gas was evolved until the temperature was very close to the melting point of zinc when action began and went on rapidly so that here about 415" is the temperature.On the whole then the temperatures of initial action are : 330 WRIGHT AND LUFF'S RESEARCHES By baryt,a and loss of weight tests. By colour test. Carbon oxide . . . . . . . . . . 160 185 Carbon from carbon oxide Hydrogen 190 195 Carbon fmm sugar . . . . 435 41 5 (B.) Plunzboso-plwmbic Oxide. A finely-coloured specimen of commercial red lead was selected which yielded on qualitative analysis nothing but lead and oxygen with a minute trace of organic dust which led t o the evolution of a small quantity of CO on heating in a Sprengel vacuum. On analysis this yielded numbers indicating a composition much nearer to Pb,Oc = Pb,Ol + 2Pb0 than t o Pb304.1.000 gram gave 1.230 gram of PbCI by moistening with HCI and evaporating down till no more gain in weight was noticed and finally heating to fusion. 0.664 gram left after ignition 0.6530 fused PbO. 0.500 gram titrated by the iodine process represented 0.500 gram titrated by the iodine process represented normal iodine = 0*00708 oxygen. decinormal iodine = 0.00716 oxygen. Calculated . Found. 5Pb0 1115 98.59 98.67 98.35 0 16 1.41 1-42 1.43 Pb,Os 1131 100.00 8-85 C.C. of 8.95 C.C. of Mean. 98.51 1-43 99.94 Carbon oxide did not act on this substance below 200"; at this temperahre a slight action was noticeable (baryta test) whilst at 215' the action was very evident. 1.880 gram heated to 205" in CO for 15 minutes lost 0.0015 gram = 0.08 per cent.On heating in hydrogen no loss of weight was observed at 200" nor at 220". At 240" 2.112 gram lost 0.0045 in 15 minutes = 0.21 per cent. , 250" 2.155 ,) , 0.0050 , ) = 0.23 ?, At 270" the residue from the preceding lost further 0.0080 in I5 minutes = 0.37 per cent. Compa'rative experiments made with red lead alone (0.6 gram' and the same weight of red lead with half its weight of carbon (from CO), in the way described in $ 2 gave the following results : OX SONE POINTS IS CHEMIC-IL DYNAMICS. 531 No gas evolved -_---Temperature. 0.2 C.C. (due to occlusion At 200" for 15 minutes . CO 0.8 Other gases . . . -0.8 -_-250" . . . . . . . 300" , , . - . --GO 2.9 Other gases 0.1 3.0 -360" (mercury vapour) for 15 minutes The evolution for a short time of CO from the red lead alone was doubtless due as above stated to the presence of a trace of organic dust in the red lead ; the non-evolution of free oxygen alone is note-worthy a different result being obtained with lead dioxide (infra) as well as with the superoxides of manganese nickel and cobalt.The temperature of initial action of the carbon may be taken as 330" (mean of 300" and 360" so t'hat the order of reduction is the usnal one :-Carbon oxide 200 Hydrogen 230 Carbon 330 (C.) Lead Dioxide. The red lead used in the preceding experiments was digested with dilute nitric acid for two days and the residual puce oxide thoroughly washed and dried at loo" and finally at 130".On analysis it was found to be nearly indicated by the formula Pbl,021 which may be written 19PbO2,Pb3O4 or 10Pb0,PbO. 0.2000 gram titrated by the iodine process represented 15.2 C.C. decinormal iodine = 0.01216 oxygen. 0.2085 gram titrated by the iodine process represented 16.0 C.C. decinormal iodine = 0.0128 oxygen. llPbO 2453 93%3 Ole 160 6-12 6-08 6-14 6-11 Pb,,Oz 2613 100*00 On ignition 0.6980 gram lost 0.0475 gram = 6.80 per cent ; whence it would appear that a few tenths per cent. of water were still pre-Calculated. Found. Mean 532 WRIGHT AND LUFF'S RESEARCHES sent after drying at 130" ; no loss of weight however took place on heating to 240". On heating to redness in a Sprengel vacuum no CO was evolved. Carbon oxide first began to act at 80" (baryta test) the action being well marked at YO".09655 gram at 100" for 15 minutes lost 0.0035 gram = 0.36 per cent. Hydrogen did not begin to act at any temperature below and up to 135". At 150" 1.365 gram lost in 15 minutes 0.0010 gram = 0.07 per cent. At 170" the residue further lost in 15 minutes 0.0050 gram = 0.36 At 200° 0.967 gram lost in 15 minutes 0.0120 gram = 1.24 per cent. At 240" the residue further lost in 15 minutes 0.0120 gram = 1-24 per cent. Hence the action began at about 140° and was so far energetic at 200-240" that in an hour almost the whole of the " available oxygen " was removed 5.27 per cent. being lost out of 6-11 per cent. As a check on this initial temperature determination 1.2195 gram was heated to 140" in hydrogen for 30 minutes when a loss of 0.0025 gram was observed = 0.20 per cent.0-6090 of the residue representing 0.6105 of original substance on titration by the iodine process corresponded to 45.5 C.C. decinormal iodine = 0.0364 oxygen = 5.96 per cent. 0.608 of the residue representing 0.609 of original substance corre-sponded to 45.5 C.C. of iodine = 0.0364 oxygen = 5.97 per cent. Hence the mean " available oxygen " before heating in hydrogen being 6.11 per cent. and that after 5.96 per cent. the loss was 0.15 per cent. closely agreeing with the loss of weight determination. The temperature of initial action of carbon from CO was found as abore described comparative experiments being made with the dioxide alone (0.6 gram) and the same weight of dioxide with half its weight OP carbon.per cent. 7 3 3 Y 30 $1 0.0270 9 = 2.79 9 ON SOME Dioxide alone. No gas evolved --__.-Ditto Ditto Gas continuously erolved .__---I-co Nil Oxygen 1-5 Other gases trace Total 1 . 5 ----Temperature. Heated up to 150" for 15 minutes --At 200" for 15 minutes . . --250" , > Y * ' 270" , > * . ----Dioxide and carbon. About 0 -1 C.C. (due to occlu-sion) Abozt 0 -15 C.C. (ditto) About 0 -2 C.C. (ditto) Gas continuously evolved -3 -0 C.C. }om2 19 -3.2 >9 -- ~~~~ 360" (mercury vapour) -Composition by analysis Carbon oxide Hydrogen . Carbon (from CO) . Carbon from sugar . POINTS IN CHEMICAL DYNAMICS. 533 Monoxide. Plumboso-plumbic DioGde. oxide. PbO Pb606 PbiiOzi 160-185 200 80 190-195 230 140 415 330 260 435 - --- --_ Gas continuously evolved CO Nil Oxygen 4.9 Other gases 0 -1 5 -0 -Gas continuously evolved 8-3 C.C.}0.2 $7 -8 -5 ~~~~~~ ~ ~~~~ ~ Hence about 260" (mean of 250' and 270") may be taken as the temperature at which carbon first begins to act on lead dioxide and at which this substance begins to decompose in a vacuum evolving oxygen. As with manganese dioxide 9 2 the formation of CO causes a much more rapid withdrawal of oxygen than takes place on simply splitting up; in each case the action increases with the temperature being mughly about three times as great at 360" as at 270" in each instance. On comparing together the temperatures of initial action of carbon oxide hydrogen and carbon on the above different oxides of lead it is evident that the dioxide is much more readily reduced than the mon-oxide or the plumboso-plumbic oxide which being formed at a low Ted heat is presumablymore akin to the monoxide as regards sta-bility.5 4. OXIDES OF COBALT. Commercially pure cobalt nitrate was dissolved in water slightly acidulated and treated with sulphuretted hydrogen ; the filtrate was VOL. XXXIIZ. 2 534 WRIGHT AND LUFF'S RESEARCHES boiled with a little nitric acid and treated with excess of ammonia ; the clear filtered fluid deposited on standing a mass of red crystals of a cobalt-ammonium nitrate ; these were collected dried and ignited at first gently a=d finally very strongly. A black light powder was left, consisting essentially of cobalt monoxide but containing a small excess of oxygen ; no trace of nickel was present therein nor of any metal save cobalt on reduction in hydrogen this black powder furnished in two experiments 77.62 and 77.70 per cent.of cobalt or 77.66 per cent. 3s the mean CO,~O~ requiring 77.63 per cent. (Co = 59). I n order to obtain pure COO from this Russell's process (this Journal 1863 51) was adopted the oxide being strongly heated over a blast lamp in a current of CO,. It was found difficult to prevent reducing gases from the flame passing inwards and effecting further reduction of the COO when an ordinary crucible lid was employed; but by employing a sheet of platinum foil fitting pretty tightly ixdde the crucible and perforated with a central hole for the CO tube and keeping up a rapid current of C02 this difficulty was easily overcome.0.7000 gram of the oxide finally obtained gave 0.5510 metallic cobalt on rodnction in hydrogen = 78.71 per cent. Average found by Russell = 78.59 per cent. Calculated for COO (Co = 59) 78.67 per cent. From the black oxide containing 77-66 per cent. of metal the sesqui-oxide was prepared by dissolving in dilute nitric acid adding pure potash in excess and a considerable excess of bromine and digesting in the cold for some days ; the jet-black precipitate was then washed many times by decantation collected on a glass-wool filter thoroughly washed and dried at 100" ; as thus obtained the oxide formed a light black powder which lost a small percentage of water on heating to 200" in a current of dry air but did not part with any trace of oxygen a t that temperature.After thus finally drying a t 200" it gave numbers agreeing with the composition G(Coz03,H,0),Co0. 0.2185 gram titrated by the iodine process was equivalent to 22-15 C.C. of decinormal iodine = 0.01772 gram oxygen. 0.4225 gram = 43.15 C.C. = 0.03452 gram oxygen. 0.7580 gram gave 0.4935 metallic Co = 0.6273 COO (Co = 59). Calculated. Found. Mean. 13CoO 975 82.70 82.75 82.75 0,. . 96 8.14 8.11 8-17 8.14 6H,O 108 9.16 By diff. 9.11 6( cozO3,H,O),CoO 1179 100.00 100~00 As a check on these numbers the water produced during reduction in hydrogen was determined-0.8930 gram gave 0.3495 gram H,O. 0.7580 , , 0.2960 , ON SOXE POINTS IN Temperature Time in minutes --______-CO produced Other gases ---_- -Total gas Gas due to occlusion .. Gas due to action .-Water present as H,O. . , formedfrom COO 7 7 7 CoaOs 360" 30 -.----0.3 0.3 Nil -CHEMICAL DTNARIICS. 535 Calculated. F O U d . - - 9.11 9.16 19.86 - -- -Total 38.13 39.14 39.05 Hence the substance dried at 200" was essentially monohydrated cobalt sesquioxide. (A.) Cobalt Monoxide. The action of CO on the pure monoxide was found by the baryta-test to be first manifested at 155" after 3 or 4 minutes' passage of the gas; even at 170" and upwards however the action was not rapid a copious production of turbidity in baryta-water by a few bubbles of issuing gas being only caused at about 180". 0.8155 gram heated in CO to 155" for 15 minutes lost 0.0010 gram = 0.12 per cent.no loss of weight being experienced at loo" 120" and 150" in 15 minutes in each case. The action of hydrogen on tht other hand was imperceptible at 155" and also at 160" after 15 minutes but was quite noticeable at 170" ; so that 165" may be taken as the temperature of initial action ; distinctly above the temperature of initial action of CO but only slighttly above when tested in the same way by loss of weight. 0.8190 gram at 160" lost in 15 minutes 0.9400 , 180" , 15 , 0.0030 = 0.32 ,, Nil = 0 ? ? ? 170" , 15 , 0.0910 = 0.12 per cent. ? 7 3 7 7 30 , 0.0080 = 0.85 ,, ? ? , , 45 , 0.0105 = 1.12 ,, 7 7 , 7 , 30 , 0.0105 = 1.12 ,, 0,9360 , 210" , 15 , 0.0075 = 0.80 ,, With carbon (from CO) the following numbers were obtained, 0.6 gram of the oxide and 0.3 of carbon being employed as in all the other experiments :-Nil 1 Nil Hence about 450" is the temperature of initial action.2 Q 536 WRIGHT AND LUFF7S RESEARCHES No gas evolved . . . (B.) Cobalt Sesquioxide. By the baryta-water test the action of CO on the cobalt aesquioxide prepared as above described was not only readily perceptible at the temperature of the laboratory but also at 0" and could even be dis-tinctly noticed after two or three minutes' passage of gas in a freezing mixture at -11". Hydrogen however did not cause any loss of weight in the sub-stance (dried till constant in zl current of air at ZOO") below 110". 0.5380 gram at 100" for 15 minutes lost nil = 0. 0.5380 gram at 120" for 15 minutes lost 0.0005 = 0.55 per cent.0.5380 gram further heated to 135" for 15 minutes lost further 0.6905 gram at 170" for 15 minutes lost 0.0785 = 11.36 per cent. To make doubly sure that the loss at 120" was not due t o the ex-pulsion of water of hydration 0.6285 gram were heated in hydrogen to 125" for 30 minutes the loss of weight being 0.0030 gram = 0.48 per cent. On titrating the residue by the iodine process it was found to be equivalent to 59.1 C.C. of decinormal iodine corresponding t o 0,04728 gram oxygen = 7.52 per cent. ; as the original sesquioxide contained 8.14 per cent. of extra oxygen the loss deduced from the titration is 8.14-7-52 = 0.62 per cent. agreeing as nearly as could be expected with the loss of weight determination and clearly indi-cating that oxygen was actually removed.The temperature of initial action of carbon was determined as described in 8 2 ; it was found that just as with manganese and lead dioxide the temperature of initial action is practically identical with that at which oxygen is first evolved the following values were ob-tained 0.6 gram of Co,O being used in each case alone in the first instance mixed with 0.3 gram of carbon (from CO) in the second:-0.0030 = 0.55 per cent. 0.1 C.C. evolved (due to occlusion by At 200" for 15 minutes. Nil . . . . . . . . . . . . ----CO evolved -Oxygen Other gases . . 0.2 ,, Total 5.4 ,) . . . . 5.2 C.C. --.--At 250" for 15 minutes. 0.15 C.C. evolved (due to occlusion) 5.0 C.C. 0.5 7, 0.2 7 7 5-7 ?, -At 270" for 15 minutes.Hence about 260' is the temperature of initial action ON SOME POINTS IN CHEMICAL DYNAMICS. 537 On contrasting the numbers obtained as above described with cobalt monoxide and sesquioxide the same result is distinctly noticeable as was observed with manganese oxides vie. that each reducing agent respectively begins to act at a considerably lower temperature on the higher than on the lower oxide. Thus :-Monoxide. Sesquioxide. CO 155 Below -11" H 165 110 Carbon 450 260 It may be noticed that the temperature of initial action of CO and H were determined on the cobalt oxide prepared by ignition of the cobalt-ammonium nitrate (indicated nearly by C O ~ ~ O ~ ) with the following results -Carbon oxide Hydrogen 195" (loss of weight test).I n this case the difference between the temperatures for CO and H is not so small as with the pure COO whilst the usual rule is obeyed that GO begins to act at a lower temperature than hydrogen. 120' (baryta-water test,). § 5. OXIDES OF NICKEL. Pure nickel oxide was prepared from commercially pure nitrate by acidulating and treating with sulphuretted hydrogen boiling the filtrate with nitric acid and adding ammonia in large excess to the blue clear filtrate clear solution of pure oxalic acid was added in quantity insufficient to neutralise completely ; on standing nickel oxalate separated ; this was collected and washed dried and ignited, and the residue moistened with nitric acid and again ignited. In this way an oxide was formed approximating to the monoxide more closely indicated by Ni20081 (Ni = 59) ; on reduction in hydrogen it furnished 77-80 per cent.of nickel whilst NizoOz requires 77.83 per cent. This oxide contained no trace of cobalt or of any other metal save nickel. By Russell's process (this Journal 1863 51) the pure monoxide was prepared from this substance (by strong ignition over a blast amp in a crucible surrounded with a clay jacket) as with cobalt monoxide it was found necessary to avoid entrance of reducing gases by employing a crucible lid fitting imside the crucible with a current of carbon dioxide led in through a perforation in the lid. 0.6705 gram of the oxide finally obtained gave 0.5270 metallic Average found by Russell 78.59 per cent. Calculated (Ni = 59) 78.67 per cent.Ni = '78.60 per cent 538 WRIGHT AND LUFF'S RESEARCHES From the nickel oxide containing 77-80 of metal we attempted to prepare the sesquioxide in the same way as that employed for cobalt sesquioxide (§ 4) viz. by dissolving in nitric acid adding excess of pure potash and a large excess of bromine and digesting in the cold for some days a jet-black powder was thus obtained which after thorough washing by decantation and on a glass-wool filter became constant in weight at loo" forming a hydrated substance containing considerably less oxygen than Ni,O,; on heating to 150" it readily parted with oxygen but none was evolved at 140" ; the oxygen evolved at 150" conhined a little C02 indicating the presence of organic matter in small quantity (taken up from the wash-water &c.).0.4740 gram of product dried at 100" till perfectly constant in weight was titrated by the iodine process and found equivalent t o 20.55 C.C. of decinormal iodine = 0*01644 oxygen. 0.4810 gram = 20.9 C.C. of iodine = 0.01672 oxygen. 0-9355 gram gave 0.5440 metallic Ni = 0.6915 NiO. These percentages approximate t o 2Ni203 5Ni0,12,H20. Calcu1at;cd. Found. 9Ni0 675 73.13 73-92 o2 32 3.47 3.47 3-48 12Hz0 216 23.40 2Ni20a,5Ni0,12H20 . . 923 100.00 So that this body was hydrated sesquioxide largely intermixed with monoxide. (A.) Pure Nickel Nonoxide. The action of GO as determined by baryta-water was first manifest Hydrogen caused no loss of weight below or at 215". At 225" 0.7310 gram lost in 15 minutes 0*0010 gram = 0.13 per At 250" 0.6515 gram lost in 15 minutes 0.0095 gram = 1.45 per At 270" 0.6420 gram lost in 15 minutes 0.0560 gram = 8-72 per Carbon (from CO) gave the following nnmbers 0.6 gram of oxide at 120" beingvery evident a t 130".cent. cent. cent. and 0.3 of carbon being employed as in the other experiments : ON SOME POINTS IN CHEBIICAL DYNAMICS. 539 Temperature Time in minutes CO evolved Other gases evolved Total Gas due t o occlusion Gas due to action -_I_-------360" 30 --0.25 0.3 --360420' ' 420-440" 15 1 15 - I -I I I- -I 440-460' 15 3.8 0.2 -Hence the temperatures of initial action of CO H and C are respec-tively about 120°,'220" and 450" or not very far from those found for the similarly prepared cobalt monoxide ; i t is noticeable that the nickel oxide containing 77.80 per cent.of metal (Ni200zl) gave as tempera-tures of initial action of CO and H respectively 1.80' and 215" the action of the CO being but slow even at 200" so that only a few tenths per cent. were lost in half an hour at that temperature. (B.) Nickel Sesquioxide. The sesquioxide containing monoxide above described did not form CO on leading CO over it at the ordinary temperatizre but baryta-water began t o be rendered turbid by the issuing gas at 30°, the action being well marked at 45". Hydrogen caused no loss of weight at 60" or any temperature below that point. At 70" 0.4935 gram lost in 15 minutes 0*0010 gram = 0.20 per cent. At 90" 0.4935 gram further lost in 15 minutes 0.0060 gram = 1 - 2 1 per cent.At 110" 0.4935 gram fnrther lost in 15 minutes 0.0110 gram = 2.23 per cent. So that about 65" is the temperature of initial action ,zs.deduced from the loss of weight test; to make sure that this loss of weight was really due to removal of oxygen and not of water of hydration, 0.5720 gram was heated to 80" for 50 minutes in a current of hydrogen, whereby a loss of weight tcok place of 0.018 gram = 3.15 per cent. On titrating the residue by the iodine process it corresponded to 11.2 C.C. of decinormal iodine = 0.00896 oxygen = 1-56 per cent. ; as the original substance yielded 3.48 per cent. of oxygen by titration, the loss of oxygen by reduction was 3.48 -1.56 = 1.92 per cent. Hence oxygen was actually removed the somewhat higher value found by loss of weight being doubhless due t o loss of a little water in addition 540 WRIGHT AND LUFF'S RESEARCHES At 120' for 15 minutes With carbon (from CO) the following numbers were obtained, 0% gram of the oxide being heated in one tube and 0.6 gram of oxide with 0.3 of carbon in another tube simultaneously :-I Nickel oxide alone.I Nickel oxide Tith carbon. No gas evolved Only a trace of gas, ' due to occlusion. -_--At 140" f o r 15 minutes At 150" for 15 minutes --------No gas evolved A little gas evolved a t first and then a much slower steady evolution of oxygen. CO 1% cc. Ditto. Gas evolved conti-nuously throughout the 15 minutes. 4.2 C.C. Oxygen 1.7 0.2 ,, Total 3.3 j 4.4 Other gases - 1 trace - -i As above stated apparently a minute quantity of organic matter was contained in the nickel superoxide giving rise to the evolution of CO on heating alone; manifestly the evolution of oxygen and the temperature of initial action of carbon lie close together a t near 145".On contrasting the temperatures of initial action of CO H and carbon on these two nickel oxides it is a t once noticeable that they follow the same rule as with the oxides of manganese lead and cobalt, viz. that the higher oxide is first affected at a sensibly lower tempera-t ure. Pure monoxide. Superoxide. co 120 30 H 220 65 c 450 145 8 6. ABSORPTION OF CO AND CO RY METALLIC OXIDES. In § 2 it was incidentally noticed that when an oxide of manganese approximating in composition closely to Mn,Oa was heated to 250' for three hours in a sealed tube containing CO a considerable absorption took place manganons carbonate being formed together with a small quantity of free CO,.This action may apparently be due to one or both of two slightly different causes ; either the CO becomes directly united to the manganese superoxide forming carbonate thus-(1.) Mn,Q4 + CO = 2Mn0 + MnCO, ON SOME POINTS IN CHEMICAL DYNAMICS. 541 OY the formation of carbonate takes place in two stages the Mn,O, being first reduced to MnO thus-(2.) and the MnO thus formed subsequently combining with the C02 liberated forming carbonate thus-Mn304 + CO = GO2 + 3Mn0, (3.) MnO + CO = MnC03. That certain anhydrous metallic oxides are capable of uniting with dry COO forming carbonates is a well-known fact ; in § 3 it has been shown that in the case of PbO the absorption of COz is facilitated by sealing up in a tube and heating so that a t first the tension of the CO is greater than that of the atmosphere ; for in two experiments in which PbO was heated to 200' for four hours in CO, in the one case in a gentle current of the gas in the other in a sealed tube the amounts of gas absorbed were respectively 0.12 and 0.76 per cent.of the weight of the PbO the latter percentage being manifestly less than the PbO would have absorbed had the supply of CO not been necessarily limited by the conditions of the experiment. In order if possible to throw some light as to the true cause of the formation of manganous carbonate in the aboqe-cited experiment the following trials were made that the reaction represented by equation (1) does not represent the whole chemical change is manifest from the presence of free CO in residual gas.(A.) Pure MnO was sealed u p in a tube with pure CO (about 0.8 gram of MnO and 10 C.C. of CO being employed) and the whole heated to 2:50" for four hours. After cooling no inrush was perceptible on opening the tnbe under mercury and no effervescence of the manganese oxide was perceptible on adding nitric acid; whence apparently re-action (3) did n o t take place ; hence if the second view above referred to be correct the combination of MnO and CO to form MnC03 must be greatly facilitated by both substances being in the nascent state. (B.) Precipitated Mn02 was sealed up in a tube with a quantity of CO insufficient to reduce the MnOz to Mn,Oa by the reaction-SMn0 + 2CO = Mn30 + 2C02, and the whole laid by for a fortnight in a cold room in winter time (temperature about 0-40° throughout).On opening the tube no inrush ensued but on examination almost the whole of the CO was found to have been converted into GO,; the manganese oxide formed did not effervesce with nitric acid. Hence the following reaction-(4.) MnO + GO = MnC03, lmrallel to equation (1) did mot take place 342 WRIGHT AKD LUFF'S RESEARCHES (C.) Precipitated MnOz was heated to 100" for three hours in a current of pure CO ; reduction took place copiously the loss in weight being 6.7 per cent. the residual oxide contained no trace of carbonate, but evolved chlorine on heating with hydrochloric acid.Here again, reaction (4) did not take place. Hence it results that there is no reason for supposing that the for-mation of MnC03 from Mn304 and CO is due t o the direct combination of CO with a superoxide of manganese as indicated by equations (1) and (4) ; the more probable explanation being that equations ( 2 ) and (3) express the changes (3) only taking place with n a s c e d 31110 and GO,. As the nascent state is not requisite for the combination of PI10 and CO to form PbC03 corresponding experiments were made with oxides of lead. (D.) 2.5 grams of PbO were sealed up with 10 C.C. of pure CO at the ordinary temperature and were then heated to 200" for four hours. On opening the tube an absorption of gas t o the extent of 2 C.C.was observed the residual gas being CO with no trace of COa ; the lead oxide effervesced with nitric acid. In this case therefore, every portion of CO formed was absorbed by the lead oxide with for-mation of lead carbonate thus-(5.) PbO + CO = PbC03. (E.) On repenting the last expeyiment with the red lead described above (§ 3) indicated by Pb506 2.5 C.C. of gas were absorbed with formation of carbonate by 2.4 grams of red lead ; 7.5 C.C. of gas left unabsorbed consisted of CO with no trace of GO,; the residual lead oxide effervesced with nitric acid. Here therefore the carbonate may be supposed to be formed either by reduction t o PbO and absorption of the CO formed in accordance with equation ( 5 ) or by direct corn-bination thus-( G .) Pb506 + CO = 4Pb0 + PbC03, this latter view is however rendered improbable by the result of the following experiment :-(I?.) 1.2 gram of the PbO above described (§ 3) was heated in it sealed tube with about 10 C.C. of pure CO to 100" for four hours. On .opening the tube no absorption was evident and no effervescence cf the resulting lead oxide was noticeable on treatment with nitric acid ; 2 C.C. of free COz were however formed. Hence it would appear to result (since the amount of reduction taking place was insufficient to reduce the PbO to Pb,04 and S jo~tior? insufficient to reduce it to YbsOs or PbO-) that CO is not absorbed by superoxides of lead under the conditions of these experiments although it is readily absorbed b OX SOME POINTS IN CHEhlICAL DYNAMICS.543 PbO under these conditions. That this is so is made evident by the following two experiments :-((3.) 1.2 grams of PbO and about 10 C.C. of GO were sealed up together and heated to 100" for four hours. Nn trace of absorption nor of formation of lead carbonate mas noticeable. (H.) Precisely similar negative results as regards absorption and formation of carbonate were obtained on heating 2.5 grams of red lead and 10 C.C. of CO to 200" for four hours. On the whole then these experiment,s point to the conclusion that when a carbonate is formed by heating it manganese or lead super-oxide in CO (which does not take place with all the superoxides of these metals but at most only with those intermediate between mon-oxide and dioxide) the action that takes place is not a direct corn-bination of GO and the superoxide but is accomplished i n two stages, lnonoxide being first formed and this substance subsequeutly com-bining with the CO thus produced the combination taking place in the case of manganese only when the monoxide and CO2 are nascent.Another conclusion that may be drawn also is that the red lead in-dicated by the formula Pb506 was a deJcinite and distinct oxide of lead, a d was n o t a mixture of Pb,O and 2Pb0 as might be supposed ; for in the latter case an absorption of COz must have taken place in experiment (H). § 7. CONCLUSIONS. From the results detailed in Parts I and I1 of these researches the following general conclusions tnay be drawn :-(1.) Whilst differences in physical state are attended with corre-lative differences in the temperatures at which the actions of the reducing agents CO H and C on metallic oxides.are first manifest (after periods not exceeding a few minutes up to 15 minutes or so) the variations in temperature of initial action thus observable are usually not SO great as to interfere seriously with the drawing of certain inferences especially when the substances compared are in approxi-mately the same physical state having been prepared by analogous processes. (2.1 In the case of the oxides of the metals iron and copper which form different oxides each corresponding with definite series of salts and other compounds and stable (Le. not evolving oxygen) on heat-ing the temperature of initial action of a given reducing agent on all the oxides of any one metal is sensibly the same unless the difference in physical structure is very marked ; t h i s fact is clearly correlated with the sensible equality in heat-evolution by the union of a give 544 WRIGHT AND LUFF'S RESEARCHES ETC.quantity of oxygen with either of these metals respectively whether the one or the other oxide be formed. (3.) I n the case of the oxides cf the mettals manganese lead nickel, and cobalt which form superoxides not corresponding with definite series of salts and other compounds and not stable on heating ( L e . , evolving oxygen when heated) the temperature of initial action of a given reducing agent on the superoxide of one of these metals is sensibly lower than that of the same agent on the stable monoxide of the same metal even after making all probable allowance for difference in physical structure in the case of manganese this fact is clearly related to the circumstance that the heat-evolution in uniting a given quantity of oxygen with the metal so as to form the superoxide is sensibly less than the heat-evolution taking place when the monoxide is formed instead ; probably therefore the same relationship also holds in the case of lead nickel and cobalt.(4.) I n no case out of many examined has any exception been found to the general rule that the temperature of initial action (during a few rninntes) of carbon oxide on a given metallic oxide is lower than that of hydrogen on the same sample of metallic oxide; and that the temperature of initial action of hydrogen is similarly lower than that of carbon on the same sample of metallic oxide.I n some few cases however the differences between the temperatures of initial action of carbon oxide and of hydrogen are not very large. (5.) In one or two instances a t present examined a long continued exposure of' a metallic oxide to the action of a given reducing agent a t a temperature some degrees lower than the temperature of initial action deduced from experiments lasting a few minutes only will give rise to a sensible reduction of the metallic oxide especially under an increased pressure in a sealed tube. It is therefore desirable that ex-periments should be made to find out how far the elerneats of time and pressure can be substituted for increase of temperature in such cases ; on these points experiments are in progress.It is noticeable in this connection that in certain instaaces the rapidity of action of' hydrogen appears to augment more rapidly with a slight rise of tern-perature than is the case ifi corresponding experiments with carbon oxide ; a circumstance doubtless connected with the superior lightness and diffusibility of hydrogen. (6.) I n a large number of cases the rule holds that the greater (algebraically) the heat-evolution during the performance of a reaction of the kind indiqated by the symbols-AB + C = A + BC, (when AB is a metallic oxide and C a reducing agent) the lower is the temperature a t which this action is first manifested to a just FRANKLAND AND DOBBIN ON THE COXSTITUTION ETC. 545 measurable extent in the course of a few minutes.All the cases in-cluded in paragraph 4 and some of those in paragraph 3 of these con-clusions are special cases of this rule ; whilst probably all the other cases included in paragraph 3 also come under the rule? although it cannot be said with certainty that they do from want of determina-tions of the actual heat-evolutions or of the data from which they can be calculated. (7.) Whilst experimenting on the above points it has incidentally been found that the products of ignition in air of manganese carbonate and of various oxides of manganese always contain more oxygen than Mn,04 and can be reduced nearly to that composition only by long continued heating over a powerful blast-lamp ; and that in consequence serious errors are apt to be introduced into quantitative determina-tions of manganese unless special care be taken to avoid them. (8.) Evidence has also been obtained tending to show that under certain conditions carbon oxide can apparently combine directly with certain metallic superoxides to form carbonates ; but that the reaction really takes place in two stages viz. first reduction to monoxide with formation of GOz and second combination of monoxide with COz to form carbonate