OF THE CHEMICAL SOCIETY, PAPERS READ BEFORE THE CHEMICAL SOCIETY. I.-First Report to the C‘iLemicul Society on “Researches on some Points in Chemical Dynamics.” By C. R. ALDER WRIGHT, DSc. (Lond.), Lecturer on Chemistry, and A. P. LUFF, Demonstrator of Chemistry, in St. Mary’s Hospital Medical School. 9 1. INTRODUCTORY. From certain theoretical speculations, it appeared probable to us that in cases of actions of single decomposition typified by the equa- tion- A + BC = AB + C, the temperature at which the action commences is a function (I) of the physical condition of the bodies experimented with ; (11) of the heat-disturbance (i.e., evolution or absorption-warmetonung) occur- ring during the reaction ; and (111) of other chemical habitudes of the substances concerned, possibly expressible as numerical values con- stant for each substance.(I.) That the temperature of initial action of a given body A, on a given compound BC, varies to an appreciable though not an enormons extent with the physical condition of the compound BC, is not only highly probable, h priori, but has been demonstrated to be the case in the instance of carbon oxide acting on ferric oxide, by I. Lo w t h i an Bell,” with whom one of us co-operated in the inquiry. According as the ferric oxide had been prepared by a wet process, or by ignition of ferrous sulphate, and according as it was pure or ictermixed with * “Chemical Phenomena of Iron Smelting,” by I. Lowthian B e l l , F.R.S., M.P. Routledge and Sons. Page 13, et sep. VOL. XXXIII. B2 WRIGHT AND LUFF OK RESEARCHES OX inert earthy matters (as in the case of various iron ores), it was found that the temperature at which pure carbon oxide first began to take away oxygen from the ferric oxide forming carbon dioxide varied between 141” and 211”.It is a matter of everyday experience that, hard compact substances are little if at all affected by various chemical agents, under circumstances where action is readily brought about if these substances be in different physical states as to aggregation of particles, &c., although determinations of the temperatures at which action commences in these different instances have for the most part not yet been made. (11.) As regards this speculation, little if any trustworthy information of an experimental kind is extant.Lowthian Bell has shown (7oc. cit., p. 91) tliat carbon oxide does not act on zinc oxide at 420°, whilst it reduces lead oxide completely at that temperature. Now the heat disturbance during the reaction- ZnO + CO = Zn + C02, must have a lower value than that during the reaction- PbO + CO = Pb + C02, because the “ heat of combustion ” of zinc is greater than that of lead, whence less heat must be evolved in the reduction of zinc oxide by a given reducing agent than in that of lead oxide. In order to see if it is a general rule that, ccete~is paribus, the temperature at which a reducing action commences is lower the greater the algebraic value of the heat disturbance taking place during its occurrence (i.e., the less the heat absorption, or the greater the heat evolution taking place), the experiments detailed in this report were made.(111.) That the temperature at which action commences is influenced by other chemical habhdes of the bodies involved besides their “ heats of combustion” (and their physical state), is rendered probable by the observation of Lowthian Bell (Zoc. cit., pp. 13 and 91), that stannic oxide is unaffected by carbon oxide at 420°, whilst ferric oxide is acted o u at much lower temperature (141” to 211°, according to the physical state) ; for the “ heat of combustion ” of tin is nearly the same as that of iron (reckoned per equal quantities of oxygen consumed): viz., according to Aiidrews :- Tin. Iron, 4230 4153 Heat evolved in uniting with 1 gram of oxygen . . from which it might be expected that the action of carbon oxide on tin oxide and on iron oxide would commence at temperatures not far apart. It would result from this that if numerical values, “constat~ts of chemical activity,” be really assignable to each substance, the constaut, for sitannic oxide must differ considerably from that for ferric oxide.SOMX POINTS IN CHEMICMJ DYNAMICS.3 In order to trace out the relationships (if any) which exist between the heat disturbances during reactions of the form- A + BC = AB + C, and the temperatures at which these actions are first noticeable, a series of observations have been made as to the temperatures at which the reducing actions commence of the three agents, carbon oxide, hydro- gen, and finely divided carbon, on the oxides of copper and iron in dif- ferent physical conditions, with the result of substantiating the general correctness of the rule, that the lower the algebraic value of the heat dis- turbance, the higher the temyeratwe of initial actiovt, so far as these bodies are concerned.Thus during the reactions- (1) ...... CUO + GO = cu + co, (3) ...... 2 c u o + c = 2cu + coz, (2) ...... CuO + H, = Cu + H20 the heat disturbances are (per 1 6 grams of oxygen transferred) + 30.05, + 19.52, and + 9.48 kilogram heat-units respectively ; with a given specimen of copper oxide (i.e., with a constant physical state), the reducing action of carbon oxide is uniformly manifested at a lower temperature than that of hydrogen, and that of hydrogen is noticeable at a lower temperature than that of carbon. With cuprous and ferric oxides, and these three reducing agents, the same order of sequence is observed; whilst on comparing the temperatures of initial action of these reagents on oxide of copper and oxide of iron in analogous physical states, it is found that each reducing agent begins t o act on copper oxide at a lower temperature than on iron oxide, the heat disturbances during the reduction of ferric oxide by these three agents being respectively + 1.!,0, -8.63, and -18.61, or considerably less in each instance than with copper oxide and the same reducing agent respectively, $ 2. CALCULATION OF HEAT DISTURBANCES DURING REDUCTION OF OXIDES OF COPPER AND IRON BY CARBON OXIDE, HYDROGEN, AND CARBON.The numerical values just cited as the calculated heat disturbances are deduced from the following data ; they represent the heat evolu- tions or absorptions that would ensue if the metallic oxidcs were reduced in such a way that all the materials and products were examined at the same temperature, via., about 15" C.; as shown further on, little difference in the relative values is caused by recalculating the numbers, on the supposition that the action takes place at some more elevated temperature, say 200" or 300". Let H, be the heat evolved by copper in uniting with 16 grams of oxygen to form cupric oxide. B 24 WRIGHT AND LUFF ON RESEARCNES ON H, that evolved by CO in uniting with 16 grams of oxygen to form H, that evolved by hydrogen in uniting with 16 grams of oxygen to And H4 that evolved by carbon in uniting with 16 grams of oxygen Then the heat disturbances taking place during the reactions- co,.form HaO. to form CO,. (1) ...... CUO + co = Cn + co, (3) ...... 2cuo + c = 2cu + COa, (2) ...... CUO + Hz = CU + HZ0 are respectively :-(1) Hz-HI ; (2) H, -HI ; and (3) H4--H1. For HZ the following values have been found:- Now H, was found by Andrews to be 38-30 kilogram heat-units. 28 x 2.431 (Andrews.) 2,s x 2.40'3 (Favre and Silbermann.) 28 x 2.490 (Dnlong.) Mean 28 x 2.441 = 68.35. Again, for H3 the value 2 X 34.275 = 68.55 is found as t,he mean ~7alue deduced from the experiments of Hess, Dulong, Grassi, Favre, and Silbermann, J u l i u s T homsen, Andrews, and Joule." This number expresses the " heat of combustion of hydrogen," assuming that water is formed; if, however, aqueous vapour be pro- duced, as is the case in the experiments that follow, this value is too high by 18 x 0.596 = 10.73, where 0.596 is the latent heat (in kilo- gram heat-units) of aqueous vapour at 15" (Regnault).Hence H3 = 68.55 -10.73 = 57.82. Similarly for H4 the following values have been found (for amor- phous carbon) :- 6 x 8.080 (Fnvre and Silbermann.) 6 x 7.912 (Despretz.) 6 x 7.900 (Andrews.) Mean 6 x '7.964 = 47.78 Q Vide Alder W r i g h t , Phil. Zag., Dec., 1874. This number is practically the same as t,hose found recently by von T h an, and by S c h u 11 e r and v. Wart, h a (Dew5 Chem. Bes. Bey., 187'7, 947 and 1298) ; the first of whom obtained the values 2 x 33.982 and 2 x 34.041, whilst the latter observers found 2 x 34126, when 2 grams of hydrogen and 15.96 grams of oxygen (S tas) at 0" are burnt to water a t 0".SOME POINTS IN CHEMICAL DY NAISIKCS.5 Hence the heat disturbances during the above three reactions are- (1) H, - HI = 68.35 - 38.30 = + 30.05 (2) H, - HI = 57.82 - 38.30 = + 19.52 ( 3 ) H4 - Hi = 47.78 - 38.30 = + 9.48 I n just the same way the heat disturbances during the reduction of iron oxide are calculated; the "heat of combustion" of iron, per 16 grams of oxygen added on, was found by Andrews to be 16 X 4.153 = 66-45 kilogram heat-units ; whence the heat disturbances during reduction by carbon oxide, hydrogen, and carbon are respec- tively- 68.35 - 66.45 = + 1-90 57.82 - 66.45 = - 8.63 47.78 - 66.45 = - 18.67 I n order to recalculate from these numbers the heat disturbances taking place at any other temperature, say To, the following formula must be employed :- rr, = HI, + 72.1 + h.2 - (h, + h*), where Hk is the recalculated heat disturbance at T".H15 is that found as above at 15". hl is the heat required to raise the temperature of the metallic oxide h2 is the heat required to raise the temperature of the reduciiig ha is the heat required to raise the temperature of the reduced metal And h4 is the heat required to raise the temperature of the comple- h,, 1x2, h g , and h* are obtained by means of the formula- from 15" to T. agent from 15" to T, from 15" to T. mentary product from l 5 O to T. h = W x s x (T-15) Where W is the weight of substance heated (in kilos.), and S its mean specific heat between 15" and To. Since W and S are tolerably small fractions, the value of h,, h,, h3, ha are necessarily not very k r p , so that the differential correction, hl + h2 -(h3 + h4) is but small Thus if '1' be taken as 215" (so that T - 15 = 200", this number being taken to simplify calculation) the values of H, for the three cases of reduction of copper oxide by carbon oxide, hydrogen, and carbon become respectively-6 WRIGHT AND LUFF ON RESEARCHES ON For Carbon Oxide- Value of W.Value of s. + 0.028 x 0,245 x ZOO}={ - + 1*37}=~,~+0.52 1.21 - 0.0635 x 0.095 x 200 - 0 . 0 4 x 0.216 x 200 - 1.90 H215 = HI5 + 0.0795 x 0.142 x 200 €115 + 2.26 For Hydrogen- H21j = H15 + 0.0795 x 0.142 x 200 H,, + 2.26 + 0.002 x 3.409 x ZOO)=( - + 1*36}=-~+0.68 1.21 - 0.0635 x 0.095 x 200 - 0.018 x 0.4805 x 2,OO - 1.73 For CaThon- Hz15 =HI, + 0.0795 x 0.142 x 200 H,, + 2.26 + 0.006 x 0.2415 x ZOO>=( - + O*?y)_ir..+o.3g 0.21 - 0,0635 x 0.095 x 200 - 0.022 x 0.216 x 200 - 0.95 It is manifest that the three values H,, + 0.52, HI, + 0.68, and HI5 + 0.39 (or 30.57, 20.20, and 9-87) are in the same order of mag- nitude as the values of HI5 itself, viz., 30.05, 19.52, and 9-48.The 0.52 + 0.68 + 0.39 = 0.53, is awerage increment of the three cases, 3 so near to each one increment severally, that the difference is much less than may readily be attributed t o inexact values being taken for S, the values assumed above not representing exactly the mean spec/@ 7wafs between 15" and 215", but only approximations thereto, being mainly the mean specific heats found for somewhat different ranges of temperature. 9 3.PREPARATION OF METALLIC OXIDES I N DIFFERENT PHYSICAL STATES. (I.) Cupric Oxide. Three varieties of cupric oxide in as pure a form as possible were employed throughout the whole series of experiments ; these were obtained as follows :- Specimen A .-Electrolytic copper was dissolved in nitric acid and precipitated by a solution of caustic potash, well boiled, filtered, and thoroughly washed and dried at 130". In order to prevent the taking up of organic matter by the copper oxide, it was found absolutely indispensable to employ washing water freed from organic matter by careful re-distillation, and potash that had been fused at a red heat before use ; glass-wool, or asbestos filters only were employed. In several preliminary experiments much time was lost through the useSOXE POIXTS IN CHElIIChL DYNAMICS.7 of oxides of copper and iron containing traces of organic matters de- rived from one source or other ; these oxides were found t o evolve small quantities of carbon dioxide on heating to rednws in VCLCZLO, which necessarily vitiated the determinations of the temperature a t which carbon begins to reduce these metallic oxides (8 6). Specinze~ B.-Copper nitrate (from electrolytic copper), carefully ignited till completely converted into oxide and retaining no oxide of nitrogen. Spechen C.-Electrolytic copper, in thin slips, roasted for many days a t a bright red heat in a porcelain tube, through which air was aspirated. The lumps of oxide thus produced were crushed, metallic particles picked out, and the whole again roasted in a current of air, and finally in oxygen, a t a bright red heat, for several hours.Even after several repetitions of the roasting, it was not found practicable to obtain a material containing exactly the amount of oxygen to form cupric oxide ; the substance ultimately obt,ained was finely powdered, thoroughly intermixed, and analysed with the result of finding that the oxygen present was exactly 98.5 per cent. of that required to form CuO, i.e., that the composition was Cu2m0191, o r 194Cu0,3Cu20: the deficiency in oxygen was determined by dissolving known weights in a mixture of hydrochloric acid and ferric chloride, diluting with boiled water, and titrating by permanganate. As stated below (§ 5 ) , this method was found t o give very constant results, and was employed, in many cases, to decide whether deoxidation of copper oxide had commenced, and to what amount it had gone on.Besides these, several other specimens of copper oxide were used for comparison in some of the experiments (vide $ 4). (11.) Caprous Oxide. According t o Andrews, the " heat of combustion " (per constant quantity of oxygen) of copper to cuprous oxide is almost identical with that to cupric oxide, viz., per 16 grams of oxygen :- Copper to cupric oxide . . . . . . . . 38-30 kilogram heat-units 7 9 cuprous ,, .. .. .... 36.61 ,, 7 7 T t should therefore result, if the hypothesis be correct that the temperature of initial action is ( c u h r i s ycirlbus) lower the grc,ater the heat evolution (algebraically), that cuprous and cupric oxides should be acted on by each reducing agent at about tlie same temperature, their physical states being identical.As shown below, this appears to be the case. Caprous oxide mas prepared perfectly free from organic matter by dissolving cane-sugar in water, filtering, and inverting by boiling8 WRIGHT AND LLWF ON RESEARCHES ON with sulphuric acid ; electrolytic copper converted into nitrate was boiled with potash (recently fused a t a red heat) and this inverted sugar, and the red ouprous oxide precipitated, thoroughly washed with water free from organic matter. Several other samples of cuprous oxide were prepared in other ways, such a s reduction by a filtered solution of commercial glucose, $c. ; but these invariably re- tained more or less organic matter, and evolved CO, on heating to redness in vacuo; some absorbed oxygen whilst drying, and niucli labour was expended before a good sample was obtained.On analysis, the pure cuprous oxide was found to contain 99.7 per cent. of CE~O, the process adopted being titration with permanganate as above de- scribed ; it gave off no trace of CO, on heating to redness in a Sprengel vacuum. (111.) Ferric Oxide. Three varieties of ferric oxide were used in the experiments with this substance. Specimen A.-Pure ferrous sulphate was recrystallised several times, dehydrated, and strongly calcined for some hours till wholly con- verted into ferric oxide ; the compact granular substance produced was ground in a mortar to a fine powder. Specimsrz B.-Ferric oxide prepared by calcining the sulphate was dissolved in boiling hydrochloric acid, and the solution poured into excess of ammonia water : the precipitat,e was boiled and thoroughly washed till no trace of chlorine was present in the wash waters ; after drying in the water-bath, it was rendered anhydrous, or almost abso- lutely so, by long-continued heating a t 180-200".Specimen C.-Ferric oxide 13 was strongly ignited in a platinum crucible for a few minutes; it shrank considerably in bulk, and became darker and less readily soluble in hydrochloric acid. Several other specimens were previously prepared by precipitation of ferric salts, such as the filtered solutions of commercial chloride, &c. ; but these were found to contain traces of organic matter derived either from the materials or the wash water, as they gave off more or less carbon dioxide on heating to redness in a Sprengel vacuum.The above three specimens, however, were found to be free from organic matter by this test ; and after heating in a test-tube to 300-400", they were not i n any way reduced, no blue colour being given on solution in hydrochloric acid and addition of ferricyanide. 5 4. DETERXIEATION OF TEMPERATURE OF INITIAL ACTTON OF CARBON OXIDE ON COPPER AND IRON OXIDES, The metallic oxides were placed in a test-tube arranged vertically in a paraffin bath, about 0-5 gram being taken for each observation :SOME POINTS IS CHEMICAL DTNANICS. 9 the mouth of the test-tube was closed by a cork with three perfo- rations ; one for a thermometer, the bulb of which reached down to the bottom ; one for an entrance-tube for leading in pure carbon oxide, also reaching to the bottom; and the third for an exit-tube just passing though the cork, and bent externally to a right angle and coupled by an india-rubber joint to a Will and Varrentrapp’s ammo- nia-bulb apparatus filled with clear baryta-water.In order to make sure that the entering carbon oxide had the same temperatiire as the bath, the entrance-tube was bent over downwards and upwards so as to form a U reaching down to about the bottom of the test-tube, and lying close to the outer surface ; the upper bend of this tube between the cork and the U was wrapped with lamp-wick to avoid aGrial cooling. By keeping the baths a t varions fixed temperatures and passing carbon oxide through for a few minutes, the temperature of initial action was approximately found by noticing at what temper- ature the baryta-water first became turbid ; and then by repeating the observations several times with fresh portions of metallic oxide, still nearer temperature-readings were obtained : the differences observed between several careful determinations rarely exceeded a degree above or below the mean.The carbon oxide employed was generated from ferrocyanide and sulphuric acid being washed through caustic soda, silver nitrate, and mlphuric acid, and stored in a gasholder over water : the gas thus kept in stock invariably contained a little air, probably derived from that absorbed by the water. As some of the experiments showed (§ 4 infra) that a partially reduced metallic oxide sometimes causes a kind of catalytic oxidation of carbon oxide to dioxide when a mix- ture of carbon oxide and a little air is led over it, it is possible %hat the original unreduced metallic oxide might to some extent possess the same power ; wherefore, before use, the free oxygen in the carbon oxide was remove4 by passing the gas from the gasholder through a tube filled with fragments of well-burnt charcoal, and heated red-hotl by a small gas-furnace, and then through a series of Liebig’s bulbs containing caust,ic potash and potassium pyrogallate.In this way carbon oxide, absolutely free from all trace of admixed oxygen, was obtained, but containing a small percentage of free nitrogen. It is noteworthy that a red-hot tube filled with pumice-stone did n o t cause the removal of all free oxygen (by conversion into carbon dioxide) from the air containing carbon oxide, whilst the same tube filled with charcoal fragments compjetely removed all free oxygen unless the stream of gas were considerably rapid.The issuing gas, however, contained carbon dioxide in notable quantity, the reducing action of the charcoal on the carbon dioxide not being suSciently great to re-convert all of that gas into carbon oxide; hence, potash-bulbs toLO KRIGHT AND LUFF ON RESEARCIES ON purify the issuing gas from carbon dioxide were indispensable, although the pyrogallate tubes were only requisite as an additional precaution. I n this way, the following numbers .were obtained as the mean results from several trials in each instance :- C1L)IYZ'C O.?~id& Specimen A.Specimen B. Specimen C. Temperature at which turbidity 1 } 60' 125" 146" was first produced in baryta- water after from three to five minutes. . . . . . . . . . . . . . . . . . . . J Temperature a t which action was bidity produced by a few bub- bles of issuing gas). . . . . - . . . . Very similar numbers were obtained with other samples of cupric oxide prepared in similar ways, but not so pure ; although some of these contained traces of organic matter, yet none of them evolved any carbon dioxide on heating in a current of air t o a temperature higher than that a t which carbon dioxide was freely formed on heat- ing in carbon oxide. Thus, three other specimens gave these num- bers : D, prepared by precipitation from recry stallised commercial copper sulphate, washing, and drying a t 130" ; E, by ignition of commercial copper nitrate; and F, roasted copper wire, as sold for combustion analyses, powdered in a mortar.Temperature of first turbidity . . . . 68' 125" 137" Temperature of well marked action 85" 133" 142" well marked (considerable tur- } 68" 133" 150" Specimen D. Specimen E. Specimen F. Cq~rnus Oxide. Besides the pure cuprous oxide described above, two other speci- mens were also examined so far as the temperature of initial action of carbon oxide is concerned ; one prepared by precipitation of copper nitrate by glucose and potash, and one purchased from Messrs. Hop- k i n and Williams : each of these two contained small quantities of organic matter, but gave off no carbon dioxide on heating in a current of air to temperatures above those where the reducing action of carbon oxide is well marked.Pure cuprous I m p r e cuprous oxide. oxide. -1 -7 From H o 1) li i 11 and W i 11 i a 111 s. r 33s glucose. Temperature of first turbidity.. . . 110" 128" 140" Temperature of well marked action 120" 136" 150"SOME POINTS IN CHEMICAL DYNANICS. 11 Specimen A. Specimen B. Spec*iinni C . Temperature of first turbidity . . . . 202" 90" 220" Temperature of well marked action 210" 100" 2:30° Two other specimens of ferric oxide, prepared by precipitation, but retaining traces of organic matter, were first acted on a t 85" and a t 97", the action being well marked a t 100" and 105" respectively. The above numbers show how considerable an influence on the tern- perature of initial action is exerted by difference in physical state ; this is especially well instanced by specimens B and C of ferric oxide, where simple ignition of B for a few minutes raises the temperature cf initial action by 130'.I n one experiment some of the copper oxide B had been heated in a porcelain crucible over a Bunsen burner, and the surface had become reduced by the diflusion into the crucible of gases from the flame ; on allowing to cool with the lid off, the partially reduced oxide again became black and apparently reoxidised ; after cooling, however, the portion constituting the upper film was found to be first acted on by carbon oxide a t 97", the action being well marked at 100" ; so that a lowering of some 28" in temperature of initial action was brought about by tlie alteration in physical structure due to partial reduction and reoxidation a t a temperature not very elevated.The numbers found by Lowthian B e l l (Zoc. cit.) as the tempera- ture of initial action of carbon oxide on ferric oxide, prepared by cal- cining the sulphate, and by precipitation and gentZe ignition, were respectively 208" and 141" ; the above results agree well with the first of these numbers, whilst tlie numbers found for precipitated oxide dried a t 180", and the same strongly ignited, lie on opposite sides of the value found by B e l l in the latter instance. It deserves notice that the action of carbon oxide on oxide of copper prepared by precipitation, well marked a t temperatures conside~nbly below loo", becomes extremely energetic a t 100" : tubes of 50 to 100 C.C.capacity, containing about a gram of the copper oxide and filled with carbon oxide, were sealed and heated to 100" for a few hours ; a t the end of the time the carbon oxide was wholly converted into carbon dioxide ; in one case this result was observed after four hours' heat- ing, in another after two hours a t 100". I n order to gain some idea as to the rapidity with which the action goes 011 a t loo", a slow stream of purified carbon oxide was lcd over 0.5 gram of cnpric oxide (speci- men A) in a tube heated in boiling water, and the carbon dioxide pro- duced collected and weighed ; in this way the iollowing numbers were obtained :-12 WRIGHT AND LUFF ON RESEARCHES ON Oxygen removed, Time.that originally present being 100". 15 minutes 22.6 60 7 7 49.3 90 > 7 71.8 120 7 7 79.8 The loss of oxygen deduced from the loss of weight of the cupric oxide a t the end of the time was 82.4, the slight difference being due in all probability to the carrying away of a milligram or two of moisture from the potash bulbs for absorption of CO, by the issuing excess of CO. The reduced copper oxide was black and distinctly pyrophoric whilst warm. In the course of some other experiments of the same kind, it was noticed that when the carbon oxide was not purified from small quan- tities of admixed air by the method above described, the amount of carboil dioxide collected greatly ezceedstb that corresponding to the loss of weight ; thus in two experiments the following numbers were obtained :- Oxygen removed, calculated Oxygen removed, Time of exposure.from COO found, that calculated from originally present bzing looo. loss of weight. (I) 90 minutes 155.0 71.9 (2) 220 ,) 113.0 72.0 whence it is evident that either the partially reduced oxide acted catalytically in causing combination between the free oxygen and the carbon oxide in the gaseous mixture, or the nearly reduced oxide took up the free oxygen, and was then again reduced in virtue of the reactions- cu,o, + 0 = CaxOy+l cu,o, + 1 + co = cu,o, + co,, thus forming more carbon dioxide than that corresponding to the difference between the original oxygen in the cupric oxide employed, and that left associated with the copper after the action was com- pleted.§ 5. DETERMINATION OF TEMPERATURE OF INITIAL ACTION OF HYDROGEN ON COPPER AND IRON OXIDES, The same arrangement was adopted as that employed as above described in the case of carbon oxide, the baryta-water bulb being omitted ; in the case of some few of the observations where thc tem- perature was above about 260" (so that there might be some danger of the cork being charred and vapours given off which might perhapsSOME POIBTS IN CHENICXL DYNAMICS. 13 reduce the metallic oxide), a simpler arrangement was employed con- sisting of a narrow glass tube passing horizontally through a hot-air bath, the metallic oxide being placed in the central part of the tube, and the bulb of a thermometer just level with i t : to avoid currents of air i n the bath, the tube and thermometer-bulb mere enclosed in a small inner chamber with perforations just big enough to allow the ends of the tube and the thermometer-bulb to pass through.The hydrogen was prepared from zinc a'nd sulphuric acid, being passed successively through (J -tubes containing pumice-stone soaked in silver nitrate, caustic potash, and sulphuric acid. In order to de- termine as nearly as possible the temperature of initial action, qnanti- ties of about 0.3 gram of metallic oxide were used for each obseri ation ; the air-bath having been heated to some given temperature, the hydro- gen was turned on and allowed to pass through for 15 minutes ; the tube containing the oxide was then removed from the bath and allowed to cool, a current of hydrogen still passing through.The cooled con- tents were then boiled with hydrochloric acid (in the case of the copper oxides with hydrochloric acid and ferric chloride), the soliltion diluted with boiled water and tested by ferricyanide or by permanganate ; generally weighed quantities of metallic oxide and a standard per- manganate solution were employed. Careful tests showed that neither the cupric oxide prepared by precipitation (A), nor that by ignition of the nitrate (B), nor any one of the three kinds of ferric oxide, caused the faintest trace of blue with ferricyanide when tested in this way before exposure to hydrogen or after heating per se. As to the copper oxide from roasting of metal (C) and the cnprous oxide, weighed quantities were taken for each trial, and the amount of permanganate corresponding subtracted from that ultimately consumed ; it was found that absolutely concordant numbers were obtained in several consecu- tive blank experiments made to determine the amount of permanganate solution required for the copper oxides before exposure to hydrogen.In this way the following numbers were obtained, 15 minutes' exD0- sure to hydiogen a t the case :- Temperat me. All temperatures up I ternperat'ure indicated being allowed in each Coyper Oxide (A). Percentage of oxygen removed (that in original oxide being 100'). to 83" No trace. At ............ 87-90' 1 - 7 ,, .............. 100" 7.0 ,, .............. 150" Almost complete reduction ; much red spongy metal formed not readily dissolved by hydrochloric acid and ferric chloride.14 WRIGHT AN) LUFF ON RESEARCHES ON Cupuic Oside (B).Temperature. Percentage of oxygen removed. All temperatures up to 170" At .......... 178-180" 4.3 ), .............. 200" 8.8 No trace. Cupuic Oxide (C). All temperatures up to 170" Nil. At .............. 175" 16.7 ,, .............. 240" Much metal reduced, undissolved by a few minutes' boiling with hydrochloric acid and ferric chloride. Cupous Oxide. All temperatures up to 150" At .............. 160" 2 7-5 ,, .............. 180" Much metal left undissolved by hydrochloric acid and ferric chlnride. No trace. Ferric Oxide (A). All temperatures up to 255" At .......... 2(j.5-2;oo ,, .............. 290" ,, .............. 310" ,, 360" (in mercury va- pour). Femic Oxide (B). All temperatures up to 190" At ..............200" ................ 210" ................ 2,50" All temperatures up t o 240" At .............. 250" ,, 360" ( i r mercury va- pour). No trace. 1.8 5.6 8.0 16.5 No trace. 0.7 1.4 i s 7 No trace. 1-2 14-5 From these observations the following temperatures may be taken as close npprciximations to the temperatures of initial action of hydro- gen 011 these metallic oxides :-SOME POlNTS IN CHEMICAL DYNASIICS. 15 Oaides of Copper. Cupiic oxide (A). Cupric oxide (B). Cupric oxide (C). Cuprous oxide. 85" 175" 172" 155" Fewic Oxides. Specimen (A). Specimen (B) . Specimen (C). 260" 195" 24.5" It is noticeable that by plotting out some of the above results, taking the temperatures as ordinates and the percentages of oxygen removed as abscisse, approximations are obtained to curves which indicate the relative degrees of energy of reducing action of hydrogen at different temperatures.It is proposed to multiply and extend observations of this nature, carefully conducted under precisely the same conditions, save as regards temperature, so as to obtain the data for accurately- determined curves of this kind, not only with hydrogen and copper and iron oxides, but also with other substances. Three experiments mere made by Lowthian B e l l (Zoc. cit., p. 118) by exposing calcined Clevehd ironstone to the action of hydrogen a t 104-127" for 30 minutes, a t 199-227' for 30 minutes, and at a bright red-heat for four hours respectively. In the first case no reduction whatever was perceptible ; in the second a trace of reduction had ap- parently taken place ; whilst in the third much of the ferric oxide was reduced, though not all.The evidence of the trace of reduction in the second case was that the material, after exposure, slightly reduced permanganate when dissolved in acid : it is manifestly possible that this action was due to the presence, in the substance treated, of a minute grain of imperfectly peroxidised ferrous carbonate ; still, ad- mitting that the reducing action of hydrogen in this form of impure ferric oxide is just perceptible between 199" and 227", this result is fairly in accordance with the above observations. Stromeyer found (Pogg. AnnuZerL, vi, 471 [1826]) that ferric oxide (how prepared is not stated) is readily completely deoxidised by hydrogen at a red-heat, reduction commencing a t considerably lower temperatures, but not proceeding very rapidly.31 a g n u s found (il, id., vi, 509) that whilst ferric oxide is not appreciably acted on by pure dry hydrogen at the temperature of boiling water, and whilst no visible production of water ensues even on heating in boiling rape oil, conzpZete deoxidation ensues in two hours a t a temperature close to that of boiling mercury, and tit any rate below the temperature of meltiug zinc.16 WRIGHT AND LUFF ON RESEARCHES ON 6 6 . DETERMINBTION OF TEMPERATURE OF INITIAL ACTION OF CARBON ON COPPER AND IRON OXIDES. The first preliminary experiments in this direction were made by heating intimate mixhres of the carbon and metallic oxides (ground together in an agate mortar) to various temperatures, and then en- deavouring t o measure the reduction by solution in hydrochloric acid (or hydrochloric acid and ferric chloride) and titration by perman- ganate ; but so many difficulties occurred in the way of obtaining any- thing like trustworthy results, that this plan had t c be abandoned.In the first place, all kinds of carbon tried produced more or less reduc- tion of ferric chloride when boiled t'herewith without any metallic oxide a t all ; and the amount, of reduction produced was different after heating to, say, 300" from what it was originally, and was not con- stant after the heating, so that it was not possible to applya correction by using weighed quantities of carbon ; this was ultimat,ely traced to the power which carbon possesses of occluding gases, and in particu- lar carbon oxide, which in this occluded form seems to act readilyas a, reducing agent, much as hydrogen, occluded by palladium, possesses high reducing powers.For this same reason, indications of reduc- tion by carbon were obtained a t temperatures far lower than those subsequently determined by more accurate methods, the heat causing a certain amount of the occluded carbon oxide to be evolved, and so to act on the metallic oxide. (It is not unlikely, moreover, that occluded air on heating would act on the carbon, thus producing addi- tional carbon oxide.) Direct experiments proved, as described below, that the gases occluded by carbon are not entirely given off in a Sprengel vacuum a t the ordinary temperature, notable quantities of carbon oxide and dioxide being extracted by this means on heating after complete exhaustion a t the ordinary temperature.The mode of operation finally adopted was the following :-Batches of carbon were prepared in quantity sufficient to last for a whole series of experiments : the amounts of gas given off by weighed quantities of each kind on heating through certain ranges of temperature after pre- vious complete exhaustion in a Sprengel vacuum at the ordinary tern- perature were determined once for all (on samples drawn from the bulk after thorough intermixture). Weighed quantities of carbon and metallic oxide were then thoroughly incorporated by con h u e d grinding together in an agate mortar ; the mixture was transferred to a suitable tube connected with a Sprengel pump, and the whole ex- hausted till no more gas was extracted by 5 or 10 minutes' action of the pump.The tube was then heated to a given temperature mid maintained thereat for 15 to 30 minutes, and the gas evolved pumpedSOME POINTS IN CHEMICAL DYNAMICS. 17 out, measured, and analysed (when necessary). As long as the bulk of gas thus extracted did not amount to appreciably more thari that furnished by the same weight of carbon alone under the same con- ditions, it was inferred that there was no action between the carbon and the metallic oxide, and the temperature of initial action was taken to be that temperature a t which more gas began to be furiaished thaw, that due to the carbon alone, the evolution of gas going om continuously foi.some time. I n thus manipulating it was found absolutely necessary to confine the mixture of carbon and metallic oxide in the closed end of the tube by a plug of recently ignited asbestos firmly rammed down, otherwise the evolution of gas from the carbon on exhausting caused a dancing about of particles of carbon in the tube, thus effectinq a partial mechanical separation of carbon from metallic oxide ; and ivhat was of still more serious moment!, caused particles of carbon to be drawn over into the tubes of the Sprengel pump, thus utterly vitiating all measurements by reabsorbing, to a greater or lesser extent, the gas evolved. It was found that after strong ignition the asbestos used for the plug retained so little occluded air, &c., that when the tube 'was rendered vacuous while cold it practically remained so on heating if nothing but asbestos were present.Many of the earlier observa- tions were spoiled, however, by our omitting this precaution and using asbestos which parted with sensible quantities of gases on heating after rendering the tube vacuous when cold. Several kinds of carbon were experiniented with a t first, but many of these were rejected because they contained considerable percentages of hydrogen (even after strong ignition), or on account of inorganic impurities. T.wo varieties were ultimately selected, which, although not absolutely pure amorphous carbon, were, we believe, as pure as that substance can be practically obtained. One of these was a hard dense coke-like charcoal, prepared by heating recrystallised sugar in closed crucibles, powdering the light charcoal thus obtained, and strongly igniting it in small portions at a time in well-closcd pla- tinum crucibles, then igniting it in it current of chlorine for two hours, and finally again igniting it in closed platinum crucibles over a blowpipe for six hours until no trace of hydrochloric acid escaped.The other was an extremely light porous carbon, obtained by passing carbon oxide over ferric oxide at about 400450" for many hours, whereby (as Lowthian B e l l has shown, loc. cit., p. 44) rednction of the carbon oxide to carbon is brought about, to a large extent, by the agency of the lower oxides of iron first formed, thus :- Fe,O, + CO = C + FexOy+,, VOL. XXXIIL. C18 WRIGHT ASD LUFF ON RESEARCHES ON the higher oxide thus formed being again reduced by another portion of carbon oxide. Fe,O,+ + GO = CO, + Fe,O,.The black mass thus obtained was boiled with dilute hydrochloric acid, thoroughly washed, again boiled with concentrated hydro- chloric acid, washed on the pump-filter till the washings were neutral, and finally dried at loo", and gently ignited in a closed crucible to expel moisture. On analysis these two samples gave the following numbers :- From sugar; 0.2575 gram gave 0.99130 GO, and 0.0105 HzO, and left 0*0040 ash; 0,2375 gram gave 0.8285 CO, and 0.0160 H,O, and left 0.0040 ash. From carbon oxide ; C.2250 gram gave 0.7'360 CO, and 0.0145 H,O, and left 0.0020 ash. From sugar. From carbon oxide. Carbon .............. 96.17 95.1g 96-49 Hydrogen............ 0.84 0.75 0.72 Oxygen (by difference). 1.43 2.44 1.90 Ash ............... 1.56 1-68 0.89 -- 100*00 100*00 100*00 The hydrogen was probably present, a t least to some extent, in tb form of moisture taken up during cooling and weighing ; charcoa, especially of the light pnlverulent character of that from carbon oxide, is far more hygroscopic than oxide of copper, which, as is well known, is very difficult to get perfectly free from hygroscopic moisture. The oxygen was necessarily present to s'ome extent as moisture, the rest being probably in the form of occluded carbon oxide and dioxide ; any occluded air would also be reckoned as oxygen in the above analyses. I n order t,o obtain the corrections for the amounts of gas evolved from these two samples of carbon on heating after pumping out all gases expelled in a Sprengel vacuum a t the ordinary temperature, weighed quantities were placed in tubes with asbedos plugs, and treated exactly as the mixtures of carbon and metallic oxides sub- sequently examined. In about half an hour after first setting the pump a t work, all gas capable of coming off a t the ordinary tempera- ture bas expelled, no appreciable amount being collected during five minutes' action of the punip, the mercury clicking thoroughly.On then heating to 100" in boiling water a little gas was evolved; in a few minutes the vacuum became again perfect, no more gas being evolved in five minutes more. I n the same way a further quantitySOME POISTS IN CHEMICAL DYNAMICS. 19 was obtained on heating to 360" in mercury vapour, the vacuum again becoming perfect in a few minutes.Similarly a little more gas ma$ obtained on successively heating to about 420" and again to near 450", these temperatures being deduced from the effect of heating in a bath of melted solder in which mere inserted glass tubes of the same size and thickness as those employed for the carbon, containing fragments of pure zinc (melting point, close upon 420") and pure antimony (melting point close upon 450"). This mode of approximating to the tem- perature to which the tube was heated was found to be fairly reliable, aiid not subject to greater errors than any other method of determin- ing the temperature that could be conveniently adopted (S i em e n's electrical pyrometer was not tried, however).In this way the following quantities of gas mere obtained from each sample of carbons reckoned in c.cs. a t 0" and i60' per gram of carbon :- Gas collected. Range of temperature. Sugar carbon. Carbon from CO. .............. 1.00 0.20 100" to 360" 0.98 360" to 420" (lead just melted) 0.10 420" to about 460" (somewhat 0.60 3-00 moV) 2.38 4-20 The following analyses were made of the total gas thus collected :- .............. Oe70 1 15" to 100" above melting point of anti- - - * Another sample of carbon was prepared from carbon oxide by passing over pure ferric oxide (Specimen A), boiling the resulting mass with hydrochloric acitl, washing, drying, and heating in hydrogen to redness (any iron compound not washed out), boiling again with hydrochloric acid, thoroughly washing, drying, and gently igniting. This was found to give off a rather smaller quantity of gas, posibly because the additional heating in hydrogen had rendered it somewhat less porous.Gas collected per gram of carbon. 15" to 360". ....................... 0 *60 360" to 420'. ....................... 0.20 420' to about 480' .................. 2 -00 2 *SO Range of temperature. - This gas consisted of- Carbon dioxide ............ '71 -02 Carbon oxide.. ............ 21 -01 Other gases.. .............. '7 *97 100 *oo This sample was not employed in any of the experiments described below.20 WRIGHT ASD LUFF ON RESEARCHES ON Sugar carbon. Carbon from CO. Carbon dioxide .............. 48.2 28.6 Carbon oxide (absorbed by cu- prous oxide) ..............43.4 33.3 Other gases ................ 8.4 38.1 - - 100.0 100.0 3 decigrams of carbon and 6 of metallic oxide were employed in every observation, the temperature determinations being made precisely as above described. Just as with the carbon alone, a practically perfect vacuum was obtained after each further heating of the tube after a few minutes' working of the pump, until a temperature was attained when gas began to come off continuously ; this temperature mas then taken as the temperature of initial action ; on repea,ting the experi- ments the results were never further from those previously obtained than might readily be due to the comparative roughness of the means of estimating the temperature adopted, whilst in most cases almost absolute agreement was noticed.The following are the numerical results obtained (it being under- stood that in every instance the tube was pumped vacuous at the ordi- nary temperature before any heat was applied, and that the tempera- ture WRS never raised above the particular limit assigned till long after a vacuum had been again established, i.e., until the mercury clicked perfectly and no appreciable amount of gas was collected in five minutes ; the temperature was judged to be 420" when the zinc in the companion tube was partially, but not wholly melted, and similarly at 450°, when the antimony was partly, but not entirely, melted; it was estimated a t about 430" when the zinc was just thoroughly melted, and 440" when the temperature was rising and the antimony showed signs of incipient fusion ; a t 460" and upwards, when the anti- mony was thoroughly melted, the temperature still rising I n a few instances temperatures a little below 360" were obtained by a hot air- bath furnished with a thermometer as described in § 5.The volumes of the gases are all reduced to 0" and 760'.SOME POIKTS IN CHEMICAL DYNAMICS. 360-420" 21 420-450* COPPER OXIDES. Carhon. from Sugar. 30 .. .. 0 *50 0 *50 -- Nil. Specimen (A), Cupric Oxide. 15 15 .. lo -90 .. 0 '10 0'55 11.00 0.53 0 -61 ___------ -- Nil. 10.39 __ Specimen (B). .. .. 0 *4 0 50 Nil. 15-360" 30 2 -05 4 *30 0.05 0 '10 2 .lo 4 *40 0 *53 0 -61 ------ 1 -57 3 *'is 360-420" 360450" 15 1 30 15-360" 1360-430" 30 15 ---- .. .. .. .. ---__- 15-360" 430460' 15 7 -90 0 .lo Range of temperature .. Time of heating in mi- nutes .............. 350-360' (hot-air bath) 360" (in mercury vapour) 30 .. .. -- 0 -40 0 .80 0 '10 ----- 30 2 -20 0 .lo 2 *30 0 '30 I__________ 2 '30 Carbon dioxide colIected Other gases Y Y - * ----- Total gas collected. ..... Gas due to occlusion . . , . Gas produced from ac- tion of carbon on metallic oxide .... --.__-- I Total gas collected.. .... Gas due to occlusion.. .. Gas produced by action of carboii on metallic oxide ............ - ~ - - (Specimen (C) . 15-360- 860-420" 1 Range of temperature . . Time of heating in mi- nutes .............. I 30 I 30 - 12 *30 0.50 12-50 0-50 1 0.61 --I__- - 1 0'20 ---- Carbon dioxide collected Otlicr gases 7 7 * . Total gas collected ....... Gas due to occlusion.. .. Gas produced from ac- tion of carbon on metallic oxide...... Nil. I 7'39 Nil. Nil. 11 *89 Carbon f r o m CO. Specimen (B) . Specimen (A). 15-360" 30 360420" 15 15-300" (hot-air bath) 30 Range of temperature . . Tinip of heating in mi- nutes .............. Carbon dioxide collected - Other gases 7 . .. .. .. .. .. 5 '40 0 *lo 5 *so 0 -40 -- 0 -25 0 'YO 0 *20 0 *so Nil. 5 -10 Nil.22 WRIGHT AXD LUFF ON RESEARCHES ON Cnrbosz from CO. I 15-340O (hot-air bath) 30 Specimen (C). I 340-350' (hot-air bath) 30 Cuprous Oxide. 15 ---- .. .. -- 0.40 0 -36 ~ _ _ _ _ _ - I 15 5 -5 0'1 5 -6 about 1 -0 . 36O0 (in mercury Tapour) 30 ----__I__ Carbon dioxide collected. Otlier gases > 7 * * Total gas collected., .... Gas duc to occlusion. ... --- Range of temperature . .I 15-360' I 360-420" 1 420--440' .... ~- 0 *30 0 *30 .. .. -__I 0 *25 0 '30 0 50 0 '10 0 '60 0 *30 Range of temperature. ........... Time of heating in minutes ...... Carbon dioxide collected ........ Other gases ........ 7 ) ~~ Total gas collected .............. Gas due t.0 occlusion ............ 15-360" 15 -- .. .. 0.40 0 *50 About, 360' (in Hg vapour) 30 --~- .. .. __- 0 -50 0 -50 360-420" 15 .. .. 0 -55 0 *53 Gas produced from action of carbon on metallic oxide.. ............ Kil. Nil. Nil. Range of tempernture. ........... Time of heating in minutes Carbon dioxide collected ........ Other gases ........ ...... ____- 7 9 15-360' 15 -8. .. About 360' (in Hg vapour) 30 -- 360-420' 15 Total gas collected .............. Gas due to occlusion ............ C -45 0 5 0 Gas produced from action of carbon on metallic oxide.. ............Nil. Nil. Kil. 1 *70 0 '10 1 *SO 0 *30 -~ 1-50 FERRIC OXIDES. Ci-crbon f r o m Szcgar. Ferric Oxide (A). 420480" 15 2.90 0 -10 3 -00 0 -71 2 -29 I Ferric Oxide (B). 420-480' 15 2 *60 0 -10 .. .. * * I .. 0 -50 0 *55 0'50 1 0.53 2.70 0 9.1 1 '99SOME POINTS IN CHERlICiiL DPXAXICS. 23 Cnrbm f r o m sup^. ,. .. 0 -40 0 '50 Ferric Oxide (C). .. .. -- 0.50 0 *50 15-360' 15 About 360' (in Hg vapour) 30 .. .. 0 *30 0 *30 -- Nil. About 360' (in Hg vapour) 30 360-420' 420-440" 15 15 -- 2.50 0-45 1 2.70 0 -443 1 -00 .. Nil. I 1.70 360-420' 15 Time of heating in minutes ...... Carbon dioxide collected ........ -- Other gases ........ 7 9 _ _ _ _ ~ ~ _ _ _ Total gas collected .............. Gas due to occlusion ............Range of temperature .......... Time of heating in minutes ...... 15 .. .. 0 *25 0 *30 420-480" 15 2 *40 0 *10 -- Gas produced from action of carbon on metallic oxide.. ............ Carbon dioxide collected ........ Other gases ........ 7) Nil. .. .. -- 0 *60 0 -53 -- Nil. -~ Total gas collcctrd .............. Gas due to occlusion ............ Gas produced from action of carbon on metallic oxide.. ............ - e__ --I-- Nil. I Nil. I I Carbon fionx CO. Ferric Oxide (A). Ferric Oxide (B). (Not determined, the specimen being entirely used up.) Ferric Oxide (C). About 360' (in Hg vapour) 30 Range of temperature, ........... 420-440" 15 2.60 0 -20 2 *80 1-00 -- --* -- 1 -80 360420" 15 -- .. .. G -4.0 1 0.40 15-360' 15 Time of heating in minutes Carbon dioxide collected ........Other gases ........ ...... -~ > 7 .. .. -- 0 -20 0 *30 -- Nil. .. .. n 0 -30 0 -30 Total gas collected.. ............. Gas due to occlusion ............ Gas produced from action of carbon on metallic oxide.. ............ Nil. I Nil.24 WRIGHT AKD LUFF OK RESEARCHES ON From these observations the following values are deduced as the temperatures of initial actions :- Copper Oxides. Specimen (A). Specimen (B) . Specimen (C). Cuprous oxide. Sugar charcoal.. . . 390" 430" 440" 390" Carbon from CO . . 350 390 430 345 Iron Oxides. Sugar charcoal . . . . . . 450" 450" 450" Carbon from CO . . . . . . 430 - 430 Specimen (A). Specimen (B). Specimen (C). The above numbers indicate further how energetic is the action of carbon oxide on copper and iron oxides ; the gas collected was always practically wholly carbon dioxide, indicating that all carbon oxide evolved from occlusion as well as any due to the action of carbon on carbon dioxide, or on the metallic oxide, was wholly oxidizcd by the metallic oxide.An experiment was made by Lowthian B e l l (Zoc. cit., p. 53), which appears a t first sight to indicate that an extremely intimate mixture of partially reduced ferric oxide and carbon, prepared by leading carbon oxide over heated ferric oxide, is capable of evolving gas at a temperaiure of 250-265", and therefore that carbon begins to reduce this form of iron oxide a t that temperature. One gram of the mixture was placed in a tube, air displaced by a current of' nitro- gen, and the tube sealed and heated in an air-bath for 44 hours ; a t the end of that time, on opening the tube under rtiercury, it was found to contain about 7.0 c.cs.of carbon oxide and 0.16 of carbon dioxide. As the mixture cont]aiiied carbon 52.14, iron 35-13, and oxygen 12.73 per cent., this amount of gas represents about 13.4 c.cs. of carbon oxide and 0.3 of carbon dioxide per gram of carbon. Admitting that this gas was really formed by t h e action of the carbon on the partially reduced iron oxide, it would not in any way vitiate any coiiclusions drawn from the results abo-re described, noT is it necessarily opposed to them in any way, since it is highly probable that the oxide of iron, partially reduced at about 400" by carbon oxide, would be excessively spongy and in an entirely different physical state from the ferric oxide examined by us, and therefore that action would commence a,t a lower temperature ; but the following experiment ren- ders it somewhat doubtful as to how far tlhis gas really did arise from the action of the carbon on the iron oxide, and how far it WRS simply occluded gas displaced by the heating and not perfectly reabsorbed as the tube cooled. We prepared an analogous mixture containing-SOME POIXTS IN CHEJIICBL DYNARIICS.25 Carbon ....................................... 39.4 Metallic iron (soluble on digesting with iodine and boiled water) ................................ 23.3 Iron existing as oxide (insoluble in iodine water) .... 31% Oxygen (by difference) .......................... 5.7' 100.0 - and then placed a gram of it in a tube, displaced all air by nitrogen, and exhausted cold with the Sprengel pump.After subtracting from the total bulk of gas collected that due to the capacity of the tube, &c., a residue of 1.2 c.cs. appeared, and on analping the expelled gas just 1.2 c.ca. of carbon oxide were found, or 3.0 c.cs. per gram of carbon present, and not more than traces of carbon dioxide. On successively heating to loo", to 250", and to 360", gas was pumped out at each heating, the tube speedily becoming practically vacuous till the temperature was again raised. The following quantities of gas mere thus extracted, calculated per gram of carbon, so as to com- pare the figures with Lowthian Bell's results :- At ordinary temperatures. 15' tco 100". 100" to 250'. 250" to 360". Totd.Carbon dioxide. . Nil. 2.0 2 *o 3.3 7.3 Carbon oxide. ... 3.0 0.5 traces. 0.5 4-0 11.3 -- It is evident from the circumstance that carbon oxide was actually extracted at the ordinary temperature, that this gas, at any rate, must have been occluded ; whilst the fitful expulsion of the remaining gaq on raising the temperature by successive stages, renders it much more probable that the rest of the gas was also occluded than that it was due to the action of the carbon on the iron oxide. The gradually in- creasing amounts of carbon dioxide are doubtless due to the partial oxidation of the carbon oxide by the hot. iron oxide, The much smaller amount of occlusion observed in the carbon experimented with by us as above described, is probably due simply to alteration of texture during the process of freeing from reduced iron, &c.§ 7. DISCUSSION OF THE ABOVE RESULTS. It is abundantly manifest that the temperature of initial action of carbon oxide and of hydrogen on all the metallic oxides exaniined, varies considerably with the physical state of the metallic oxide ; that of carbon varies not only with the physical state of the metallic oxide, but also with that of the carbon, a light pulverulent carbon beginning to act at a much lower temperature than a hard, dense, coke-like26 WRIGHT AND LUFF ON RESEARCHES, ETC. charcoal, even though the latter be reduced to very fine powder. In the case of ferric oxide, although a difference in physical structure of the carbon produces a notable difference in temperature of initial action, yet no appreciable difference in temperature appears to be caused by any initial difference in the physical state of the ferric oxide, the reason for this presumably being that at a temperature of 400' and upwards the " precipitated " oxide alters its texture, and becomes in structure the same, o r nearly the same, as that produced in the first instance by ignition processes. This alteration in physical texture by heat alone does not seem t o take place, or at any rate not to so great an extent, with oxides of copper. Nextly, on comparing the temperatures of initial action on a, given kind of metallic oxide of different reducing agents, it is invariably found that that reducing a p z t b e g i m t o act at the lowest temperature which has the greatest heat of combustion); so that the heat disturbance during its reaction has (algebraically) the greatest value. Thus hydrogen always begins to act a t a lower temperature than carbon, and carbon oxide at a lower temperature than hydrogen :- Carbon oxide. Cupric oxide A . . .. 60" ,, C . . . . 146 Cuprous oxide .... 110 Ferricoxide A .... 202 ,, B . . .. 125 ,, B .... 90 ,, c .... 280 Hydrogen. 85" 175 172 155 260 195 245 Sugar carbon. Carbon from CO. 390" 350 O 430 350 440 430 390 345 450 430 450 - 450 430 Again, when cupyic and ferric oxides, prepared by analogous pro- cesses, and, therefore, presumably in much the same physical state, are compared, it is uniformly found that the temperature of initial action of a given reducing agent is lower o n oxide of copper than o n oxide of iron; i e . , that the action commences a t a lower temperature the greater (algebraically) the value of the heat disturbance :- Jetcnllic Oxides pwpared by precipitation. Carbon osidc. Hydrogen. Sugar carbon. Carbon from CO. Copper .......... 60" 85" 390" 350" Iron.. ............ 90 195 450 430 ?* Jltallic Oxides prepared by ignition of Xalts, PG. Copper .......... 125' 175" 430" 390" 202 260 4.50 430 { 220 245 450 430 Iron. ............. Q Deduced from the results with the other specimens.MUIR ON THE INFLUENCE EXERTED, ETC. 27 That this rule is not invariable for all metallic oxides has been already indicated (§ 1). It is noticeable that cuprous oxide, having much the same “heat of formation’7 as cupric oxide (0 3, ii)7 and, therefore, causing much the same heat disturbance during its rednction by a given reducing agent, is uniformly first reduced at a temperature no further removed from the temperature of initial action on one or other form of cupric oxide than may readily be due to difference in physical state.