CHEMICAL NEWS, Feb. 3, 1911

THE CHEMICAL NEWS.

VOL. CIII., No. 2671.

THE WORK OF THE CHEMICAL LABORATORIES OF THE BUREAU OF MINES.* By J. K. CLEMENT.

THE chemical work is divided among a number of separate laboratories, each carrying on its own lines of work under the direction of its own chief, the whole forming a group of more or less independent units. The relation of the work of the several chemical laboratories to that of the other departments of the Bureau varies with individual cases. In general, however, the problems of the chemists are closely connected with those of the mining and mechanical engineers.

The Fuel Testing Laboratory is occupied mainly with the analysis and calorimetric testing of fuels, including coal, coke, lignite, and peat. In addition to analysing samples of all fuels used in the boiler and gas producer tests of the Bureau, ultimate analyses and calorific value determinations are made on mine samples of coal collected by the U.S. Geological Survey, as well as by certain State Geological Surveys. The data on these latter tests are of value in establishing the composition and heating value of

the coals in connection with the classification of the coalfields of the United States.

In addition to the laboratory at Pittsburg, there is located in Washington, D.C., a laboratory in which are tested samples of coal delivered to the various buildings, arsenals, navy-yards, and military posts within the district of Columbia, and in various parts of the country, and of the coal purchased by the Panama Railroad,

Fusibility and Clinkering of Coal Ash.

In the use of coal under steam-boilers, the property next in importance to its calorific value is perhaps the fusibility of its ash. Indeed, some coals, which have a high heating value, are worthless for making steam on account of their tendency to clinker and adhere to the grate bars. The relation between the fusibility and clinkering properties of coal-ash and its chemical and mineralogical composition is now being investigated.

It is interesting to note that TiO2 was found in all the clinkers examined in amounts varying from 1 to 3 per cent. Determinations of the fusion-point of various ashes give values ranging from 1150° to 1400° C.

Chemistry of Petroleum Technology.

The Bureau is making a study of the commercial bodies contained in the crude petroleums of the United States, of the methods for their separation and purification, and of their economic uses. The California fields, because of their showing at this time the greater promise of a large and continued production, their proximity to naval stations, and the peculiar adaptability of their product as a maritime and a locomotive fuel, have been selected for first study. Besides the determination of the properties and uses of the various products of the petroleums of the country, an investigation is being made of the processes of distillation and of the methods of refining.

Combustion Investigations.

The processes of combustion in the boiler-furnaces are being investigated in a furnace specially designed for the purpose. By taking simultaneous samples of the combustion gases, the progress of the reactions may be followed, and the time or space necessary for the complete com

Abstract of a Paper read before the Minneapolis meeting of the American Chemical Society.

49

bustion of various coals, and under varying conditions of

operation, may be determined.

The process of producer gas formation is being studied from a physical-chemical standpoint, and an attempt will be made to apply on a commercial scale the results of laboratory experiments an the rate of formation of carbon monoxide and water-gas.

The Composition of Coal.-Our scientific knowledge of the chemical character of coal is limited almost entirely to its chemical analysis and its adaptation to certain industrial operations. The object of one of the investigations of the Bureau is the isolation and identification of some of the constituents of coal. By the use of inert solvents, it has been found possible to extract as much as 35 per cent of been isolated, and the analysis and molecular weights of the original coal. A number of different substances have

some of these substances have been determined. In a few cases it is believed that the materials are practically pure substances.

The Volatile Matter of Coal.-The quantity and composition of the gases evolved from various coals when heated to temperatures of from 400° to 1000° C., have been determined. In the experiments which are now in progress, particular attention will be given to the influence of the rate of heating on the character of the gases produced, liberation, and to the thermal decomposition of these gases to the initial composition of the gases at the instant of during passage over heated surfaces.

Weathering and Deterioration of Coal.—In co-operation with the Navy Department, the Panama Railroad Company, and the University of Michigan, the Bureau is conducting

an extensive series of tests on the deterioration of various in fresh water and sea-water. coals in storage, both in the open air and when submerged

The Accumulation of Gas from Coal.-The quantity and rate of formation of inflammable gas from freshly mined coal, at ordinary temperatures, and the rate of absorption

of oxygen by the coal have been determined.

The Spontaneous Combustion of Coal is being investigated by the Bureau. Statistical information will be combined with the results obtained in the laboratory.

The Burning of Coal in Mines under a Diminished Supply of Oxygen.-The factors governing the propagation or extinguishing of fires in mines are being investigated. Chief among these are variations in temperature and in the

oxygen content of the surrounding atmosphere. Examination of Mine Gases. -Examination is made of samples from normal mine air, from the after-damp following explosions, from stagnant areas, and from burning areas during mine fires. Particular attention has been given to the detection of small amounts of carbon monoxide. By analysing samples during the progress of mine fires, the chemist has assisted in combating fires. The effect of variations of barometric pressure on the exudation of methane and the influence of carbon dioxide on the explosibility of mine gases are being investigated.

The Chemistry of Explosives.

Chemical analyses are made of all explosives submitted to the Bureau for test, of the products of combustion of explosives, and of electric detonators, blasting-caps, and fuses. All explosives, blasting-caps, electric detonators, and fuses purchased by the Isthmian Canal Commission are inspected by representatives of the Bureau, and all shipments of such explosives are sampled and analysed. Coal-dust Explosions.-The two greatest sources of danger encountered in mining operations are the explosive gases given off by the coal, and the finely divided coal-dust which exists throughout most coal-mines. The first danger can be overcome by increasing the ventilation in the mines. Unfortunately, this increases the danger from the coaldust by the removal of its moisture.

A laboratory method has been devised to test the inflammable character of samples of coal-dust, and to classify them according to their inflammability. This method is based on determining the amount of combustion which

takes place when clouds of dust of the same density are ignited under the same conditions; the amount of combustion being determined by the pressure developed within the explosion vessel. In this way it is possible to obtain results on any one sample of coal-dust which agree to 3 to 5 per cent of the total pressure developed.

One af the proposed means of lessening the inflammable character of coal-dust is to add a non-inflammable dust. The laboratory method used to investigate the inflammability of coal-dust has been extended to various percentages of coal-dust and finely-ground shale in order to determine to what extent the combustion is limited by the presence of the inert dust. The experiments indicate that a marked diminution of pressure is not obtained until about 25 per cent of inert dust is added, the pressure then falling off rapidly with a further increase in the amount of shale-dust added.

SEPARATION OF OXYGEN BY COLD.* By JAMES SWINBURNE, F.R.S. (Concluded from p. 42).

So far I have discussed the subject on what are sometimes incorrectly called theoretical lines, that is to say ideal lines, in which all sorts of assumptions are made so that the discussion should be easier. Thus I have discussed perfect heat insulation, and completely reversible changes and perfect thermodynamfc engines. The principles are the same on a large scale; but there is leakage of heat, and all sorts of irreversible changes of heat, and a compression pump does not come up to the thermodynamic ideal in efficiency. In the industry as now carried on the question of efficiency is of very little importance. Oxygen is sold in small quantities in bottles, and the costs of compressing, handling, selling, delivering, and so on, and the collection of the empties, are enormously large in proportion to the mere cost of making. Thus we have discussed oxygen taking 34,000 joules for 230 grms., or about 2 k.-w. hours per cubic foot.

The great future for the separation of the gas is in the direction of making nitrogen for the manufacture of calcium cyanamide, and oxygen for use in blast-furnaces, and other work where a high temperature may be worth the extra cost. It is therefore important to see how far the gases can be produced on a large scale at prices within a reasonable margin.

We may therefore sum up the rectifying part of the plant as an apparatus which gives up heat at a low temperature, say, 82° A, and absorbs heat at, say, 90° A, the heat taken in at 90° A being equal to that given out at 82° A. A refrigerator must therefore be a thermodynamic engine capable of giving heat at 90° A, absorbing heat at 82 A, and giving out heat at the temperature of the works, say, 290° A.

It is generally supposed that Linde depended on an irreversible expansion of a gas. This is not the case, and I would put the matter in another way.

If gas is compressed, cooled to, say, 290° A, and sent into an interchanger, out of which it comes by a return pipe at the pressure of the air, there must be cooling going on inside, as the air delivered has higher internal energy than the air going in. You have therefore an apparatus for producing cold—that is to say, for absorbing heat at a temperature below the room. The question is, What is the lowest temperature at which this apparatus will take in heat? The lowest temperature is that of the boiling-point of the liquefied gas at the pressure of the atmosphere. If the apparatus is working in such a way that it can take in heat H at temperature 02 the increase of entropy is H/02. But if instead of working in this way the heat is taken in at 01, which is a higher temperature, the increase of entropy

* A Paper read before the Faraday Society, December 13, 1910.

is H/01, which is less, so the apparatus makes up the deficit of entropy by irreversible expansion. On starting up a Linde machine it begins by abstracting heat from its own air and metal-work at the temperature of the air, sothe entropy is made up by irreversible expansion. As the temperature at which the heat is absorbed falls the irreversible expansion gets less and less, until finally the whole of the gas is turned into liquid. Then there is some production of heat by the liquid passing from a high to a low pressure through an obstruction, and also some slight evolution of gas to cool the released liquid from the boiling-point under the high to the boiling-point under the low pressure. The irreversible loss due to liquid is pv/10 in joules, where v is the volume of the liquid. If the compression is small this is small, but if it is high pv may be considerable. It would be quite easy to get this energy out by means of a little engine, as there is no expansion and no difficulty in lubrication, and it would be worth while on a large scale. The liquid may also be cooled to a lower boiling-point while still under pressure, so that there is no irreversible evolution of gas on its being exhausted from the engine.

The refrigerator has to do the work of two thermodynamic engines, for it has to abstract a lot of heat at one temperature, give a large proportion back at a higher temperature, and give out heat at the temperature of the works. The ideal apparatus consists of a compressor with water-cooling delivering gas at the temperature of the works and a high pressure to an interchanger. The pressure is chosen so that the gas liquefies under it at 90° A, giving up its heat to the rectifier. The liquid at 90° A is then cooled in an interchanger to 82°, and it is then passed through a hydraulic engine, coming out at atmospheric pressure and temperature 82°. It is now vaporised, taking in heat at 82°, and goes out by the delivery side of the interchanger.

This process described is not quite reversible, because the specific heat under constant pressure of the gas coming down the interchanger is greater than that of the gas coming up under atmospheric pressure; there has therefore to be an abstraction of heat all along the interchanger to make up for this difference.

The idea of this apparatus is something like that of an ordinary carbonic acid machine worked under different conditions. If a carbonic acid machine were worked in the tropics and the cooling water were above the critical-point of CO2, the compressor would supply compressed gas only. Some sort of refrigerator would be necessary to liquefy the CO2, and the liquid CO2 would then be evaporated at a We thus have three temperatures to low temperature. consider the temperature of the works at which heat due to compression is given out, the temperature at which the CO2 is liquefied under pressure, and the temperature at which the CO2 is finally evaporated. In a refrigerating plant three temperatures would be a nuisance, so a fluid is chosen with a critical-point above the temperature of the works. But for rectifying three temperatures are necessary-the temperature of the works, the temperature of the boiling-point of the least volatile liquid, and the temperature of the liquefaction of the mixture. We can therefore use a fluid which is compressed at a temperature which may be much above its critical-point.

Fig. 4 shows an ideal Linde refrigerator. The compressor A delivers a fluid, say, air under pressure, to the cooler B. The air goes on down through the interchanger c until it is Heat comes out there at the temperature of the works. cooled to 90° A. Then it is cooled in D, which is really a coil of pipe immersed in the oxygen to be evaporated at the bottom of the rectifier. From D the liquid, still under pressure, goes down the interchanger until it is cooled to 82° A, and it then works a little engine E, and comes out F is really a coil of pipe in the liquefier of the The air is evaporated in F, and passes up the interchanger, and escapes at G. There are some points to be noticed. As already pointed out the interchanger must take in heat by leakage, which

into F. rectifier.

CHEMICAL NEWS,

Feb. 3. 1911

is an irreversible process; and the heat taken in at F will be greater than that given out at D, as it takes more heat to vaporise a liquid at a low pressure and temperature. The processes in this refrigator are, however, substantially reversible.

If schemes for dealing with liquefied gases are examined, it will often be found that a liquid is being condensed or evaporated by a gas in the interchanger. That is always a piece of bad design, involving irreversible change. It may be inevitable, of course; but it is an evil which must be carefully watched.

(NOTE.-There is another fallacy very common with regard to gas issuing from a nozzle. Suppose it comes out into the atmosphere, it is thought that as it has to expand against the pressure of the atmosphere, doing work in joules equal to a tenth of its increase of volume, it must be cooled to a corresponding degree. This is a mistake nearly everyone makes until he thinks of the matter a second time. As such an error would give rise to many miscalculations in connection with the liquefaction of gas, it is not impertinent to call attention to it).

Linde points out that the mechanical work necessary may be reduced by working between two high pressures. But there is a limit to this, as if both pressures are very high the lower temperature limit of the plant is reduced; so the limit is reached by the lower pressure, which must be low enough for the fluid to evaporate at 82° A, or whatever temperature the heat has to be extracted from the rectifier. If we choose a fluid-say, nitrogen-this gives the lowest pressure that can be used. Then the other pressure must be high enough for the liquid to condense at 90° A. Some latitude can be got by control of the pressures in the rectifying plant, but it is simpler for the present to take the rectifier as working at atmospheric pressure for the sake of ease of calculation. The refrigerator, then, has to supply for each kilo of air split into pure gases 135,000 joules at 90° A, absorbing 135,000 at 82° A. The liquefying of this fluid is in contact with a fluid that is being evaporated, so that that part of the process is ideally reversible. The re-evaporation is in contact with the air which is being liquefied, so that part of the process is ideally reversible. The passage of the liquid from the high to the low pressure can very easily be made to give out work. It is not the same problem as dealing with expanding air, as the volume is small, the pressure constant, and the liquid acts as a lubricant. The process, so far, is essentially reversible. The specific heat of compressed gas is higher than that of the same gas at a lower pressure, so a little allowance must be made for that. It is thus possible to use a process which has no external work done by expansion of cooled gas in such a way as to get little loss by irreversible changes.

To work out an example, nitrogen can be taken as the

51

fluid, as it is easier to discuss than a mixture. To get 135,000 joules, taking heat of vaporisation of nitrogen 200 joules per grm., we need 675 grms. of nitrogen, and we can liquefy this at 90° A under about four atmospheres, evaporating it again at 82° A at two atmospheres. Taking the density of liquid nitrogen as o-8 the hydraulic engine gives out 170 joules, a mere trifle in this case, and the compression of the nitrogen from two to four atmospheres needs 37,000 joules, so we need under 37,000 joules. As the separation needs 40,000 this is clearly an erroneous figure, but the data are very uncertain. We must need more than 40,000.

To work on a large scale with atmospheric pressure in the separator and four and two in the refrigerator would be quite out of the question. The volumes of air would be enormous, and the interchanges would be very inefficient. Those who have had to work with gases that have to be heated or cooled will realise the extent of this trouble at once. To make the plant practical it will be necessary to work the rectifier under pressure. This raises the temperature of the rectifier; but that does not mean that we can separate the gases more economically; the cost of separation depends, as already stated, on the difference of entropy of the gases entering and leaving the plant, not on what goes on in the plant, provided there are no irreversible changes. Using pressure allows the refrigator to be worked at higher pressures, so that the interchanger may become practical.

Assuming the ideal energy needed to separate a kilo of air into its gases to be 40,000, the question is, What will it cost in practice? On a large scale it may be safe to say 100,000 joules. We would thus get 230 grms. for 100,000 joules; or a ton would need 120,000 k.w. hours. Taking the cost of a killowatt hour as low as o'rd., this means £50 a ton.

It does not look as if oxygen at this order of figure could be used for blast-furnace work. Taking a blast-furnace as using four and a-half tons of air per ton of iron, to increase the oxygen content by 1 per cent of the oxygen content comes out at Ios. per ton of pig. This does not look as if it could be of use.

The principle of separating gases by cold can be applied to other gases. Thus hydrogen may be purified, so that water-gas can be used as source of balloon gas. Methane can be separated out by solution in water under under pressure. Chlorine can be separated from the impure gas from the anode chambers in electrolytic works. Such chambers are generally worked under a slight exhaust to avoid troubles due to leaks. Air thus leaks in, and the chlorine coming off is moist and impure. Chlorine can be condensed into water, and it might appear that condensing such gases as chlorine and methane into water would allow of their being separated with less expenditure of energy than the entropy consideration demands. This is not so. It takes greater pressure to get the gas into the liquid out of a mixture than the solution exerts in an atmosphere of the pure gas.

Chlorine, by the way, is easier to handle than is generally supposed. It can be pumped in iron pumps flooded with sulphuric acid. It can also be dried and then compressed in iron pumps lubricated with sperm-oil saturated with chlorine. It is not necessary to dehydrate it absolutely for compression up to, say, 50 lbs.

There may be other gases that can be separated or purified by liquefaction and rectification; it is difficult to think of any case where the principle cannot be applied.

Crystallised Chlorophyll.-M. Tswett.-The green crystals which Borodin separated by means of alcohol from many species of plants, and to which the name "crystallised chlorophyll" has recently been given, can be shown by adsorption analysis to be an isomorphous mixture of two chlorophylline derivatives, a- and 3-metachlorophylline. These agree spectroscopically with their mother-pigments, and obviously contain their unaltered chromophores.-Berichte, xliii., No. 16.

CHEMICAL NEWS,

Feb. 3, 1911

down and the pycnometer filled to overflowing. Any powder or air-bubbles floating on the top were swept off with a glass rod. Air-bubbles did not usually adhere to the sides, but were always looked for and removed with a wire if present. The pycnometer was placed in a tank of water and kept at 20° C. for twenty minutes. The stopper was then inserted, the top quickly wiped off, and the whole dried and weighed at once. The contents were transferred to a platinum dish, evaporated to dryness, and ignited finally over a blast for five minutes. In transferring the contents of the pycnometer to the dish it was found that all loss could be avoided if the pycnometer were inverted over the dish and shaken slightly till the powder left the bottom. By then bringing the mouth against the side of the dish the mixture will run out quietly. Two or three rinsings with water are necessary. The pycnometer usually contained a trace of oxide sticking to the sides, which might weigh as much as 2 mgrms. It was therefore dried at 120° C. and weighed to determine the small amount of oxide. This method gave results which were entirely satisfactory for our purpose. The average difference between duplicates in forty determina tions of density amounted to o 22 per cent. This result includes several determinations in which the amount of material used was under 1 grm., which increased the percentage error considerably. When 2 grms. or more of material were used, duplicates seldom disagreed by more than 0.15 per cent. To illustrate the densities found for a mixture containing 90 per cent tantalic oxide were 8.103 and 8.078 when 15 grm. were used. Upon repeating the experiment with 3'3 grms. of material, the densities found were 8.090 and 8.092.

We next determined that the mixed oxides became constant in density after heating for an hour over the blast-lamp. For this purpose a sample of the mixed oxides from a Branchville columbite was ignited over the blastlamp for an hour in a platinum crucible. The density was determined, and the material then ignited for five-minute periods, a density determination being made after each ignition. The results were: 4'908, 4'908, 4'919, 4'921, 4'923, 4'912.

The material was then heated an hour longer, and the density re-determined, the results being 4'924 and 4'923. In these and the following determinations the ignition was accomplished in 30 grm. platinum crucibles over a fairly powerful blast-lamp. The question whether different preparations of the oxides in the same proportions had the same specific gravity could only be answered by making a series of such determinations. Complete duplicate series of mixtures were not made on account of lack of material, but the results given in the accompanying table show what agreement was obtained in each case. The average difference in density between different preparations of oxides having the same composition was II per cent. This appears to be the greatest source of error in the determinations, and shows that the method of preparing the oxides must be fixed as definitely as possible.

The detailed method of preparing the mixed oxides in condition for specific gravity determination was as follows: -About 3 grms. of the oxides were fused in a platinum dish with six times their weight of acid potassium fluoride till the mass was just liquid. The fusion was dissolved in 200 cc. of hot water in a platinum dish, adding a little hydrofluoric acid to obtain a clear solution. The solution was made alkaline with ammonia, and after allowing the precipitate to settle, filtered on a rubber funnel and washed well with water containing ammonia. The precipitate runs through the paper if pure water is used. The precipitate was re-dissolved in the platinum dish, using as little dilute hydrofluoric acid and water as possible, and evaporated to dryness. Ten cc. of concentrated sulphuric acid were added in such a way that the residue was completely moistened. The liquid was evaporated over asbestos till all hydrofluoric acid was expelled, stirring if there seemed to be danger of spattering. The residue was cooled, 200 cc. of water added, and the solution made

53

alkaline with ammonia, filtered, and the precipitate washed with dilute ammonia. The precipitate was transferred while moist to a 30-grm. platinum crucible and ignited over a blast-lamp for an hour. The residue was ground to a fine powder with water and was then ready for the density determination previously described. This method of preparing the oxides may appear longer than necessary, but we adopted it only after preliminary work had made it appear essential. Acid potassium fluoride was used in decomposing the oxides instead of potassium bisulphate, because the latter does not render the oxides completely soluble in hydrofluoric acid if they have previously been ignited. This fusion with acid fluoride does not involve any loss of either columbium or tantalum by volatilisation. To prove this, 13576 grms. of mixed columbic and tantalic oxides were subjected to the process outlined above, and 13603 grms. were obtained. If the mixed oxides have not been ignited and are completely soluble in hydrofluoric acid, this may be used to dissolve them instead of acid potassium fluoride. The first precipitate with ammonia from hydrofluoric acid solution cannot be ignited and used for density determination, as it contains fluorine after being ignited. Washing the precipitate with dilute ammonia will not remove all the fluorides.

Results.

In preparing the mixtures for density determinations, we were fortunate in obtaining very pure tantalum and columbium oxides. Dr. W. E. Ford supplied a sample of columbic oxide which had been given him by Prof. E. F. Smith. It had been prepared from the chloride, and was exceedingly pure. Dr. Ford also supplied some tantalum oxide from Dr. W. R. Whitney, of Schenectady, who stated that the material was considerably over 99 per cent pure. Dr. T. B. Osborne furnished us with considerable amounts of potassium tantalum and potassium columbium double fluorides, which we re-crystallised and converted into the oxides. The densities were as follows:Columbic oxide (Smith)

..

(Osborne)..

(Whitney) (Osborne)..

4'552

In making the mixtures, it was assumed that Smith's and Whitney's oxides were pure and that Osborne's columbic oxide contained o 61 per cent tantalic oxide and his tantalic oxide o 32 per cent columbic oxide. These values were calculated from the densities. The original intention was to make mixtures of the oxides at intervals of ro per cent. Irregularity in the curve showing change in density with change in composition when there was over 60 per cent of tantalic oxide caused us to repeat some of the determinations in this region, as we suspected an error in the work. We were led to the conclusion that in the mixtures containing a moderate excess (65-85 per cent) of tantalic oxide, the densities are not as constant as when other proportions are present. The table and curve given below will make this clearer. All the results which were obtained upon known mixtures are tabulated below with one exception. This result was obviously in error, as the density was not approximately related to the composition. In most cases, the density of one mixture of any given composition was determined. Where the curve as plotted from the table showed irregularities, duplicate mixtures were made and the results enclosed in brackets in the table. All densities are referred to water at 20°.

The average of the results in the table are plotted in the figure.

The curve has some interest from the standpoint of solid solutions. The fact that the points as plotted do not lie on a straight line indicates that the oxides have formed a solid solution and not a mechanical mixture. The irregularities in the curve, which can hardly be due entirely to experimental error, suggest that the oxides are not soluble in each other in all proportions, but that tan