.
EDITORIALS. importance, still leave room for the Institute certificate. Although the need for something like the proposed Institute is a need that exists in every business high and low, yet there is no business where it is needed so badly as in analytical chemistry, for there is no other business where the employer is more absolutely incapable of judging for himself whether or not his employee is capable and deserving, and as matters now stand, the analytical chemist must rely for advancement and appreciation rather upon his engaging personal qualities, if he has any, than upon his professional capabilities. I n fact, i t is hardly too much to say that in the iron trade at least, the latter are a bar rather than a help to advancement, if the employer has no outside sources of information about his chemist, for if the chemist has professional capability, and professional pride, in his work, he will rarely succeed in satisfying his employers in the matters of speed and output of work. A far closer approach to the steel man's ideal in these respects would be made by a laboratory boy ignorant of chemistry, and innocent of conscientiousness, and it is by no means a reckless or random statement that in the iron trade, the better the chemist, the lower his employer's opinion of him, if the employer has nothing to guide him but his own impressions. With conditions as they are to day, with employers almost unanimous in the conviction that chemical analysis is quick and easy work, a n d with the gr at majority of chemists seeking t o humor and adapt themselves to this foolish misconception, rather than to combat and correct it, the lot of the conscientious analyst would be hard indeed without the testimony and the support of college degrees and other honors that he may succeed in gaining. Let us have more on the same principle as the College degree ! The College degree is the first thing; it is most important but it is not enough. It certifies to college study. But study does not end with the closing exercises of college. At that point it may be said to begin. What have we now to certify to this real serious life study that begins only as college ends? We have the Ph.D. degree and it is a glorious thing. But that it leaves nothing more to be desired, and that its testimony represents the acme of human effort, and human achievement, we have to deny. Admirable as is the Ph.D. degree, and of more dignity and impsrtance than anything else in the same line, still there is room for more in the same line. The question confront-
65
ing the chemical profession in America is t'his: Since the College degree is a good thing, shall we develop the underlying principle of it further, or shall we stop there and be content? GEORGEAUCHY.
ORIGINAL ARTICLES. [CONTRIBUTION
FROM T H E LABORATORY O F T H E FUEL COMPANY, CHICAGO, ILLINOIS.]
ENGINEERING
THE AMOUNT OF INERT VOLATILE MATTER IN THE MINERAL CONSTITUENTS OF COAL. BY W. BRINSMAID. Received November 5 , 1908.
Chemists working on the analysis of coal have long known that the non-volatile mineral matter that they weighed and called ash did not truly represent the weight of the inorganic matter, when in its original form in the unburned coal. They have also known some, if not all, of the Sources of error but have not been able to calculate the amount. The combined water in fire clay and gypsum, and the presence of carbonates that give off carbon dioxide on heating may be mentioned as probably the principal sources of error in weighing an ash correctly. Pyrite also loses weight when burned to ferric oxide and is thus a source of loss. In case the amount of pyrite present in the coal were known, then the addition of five-eighths the weight of its sulphur content to the ash would correct for this loss. Unlike the other cases, however, this loss is accompanied by combustion and develops some heat. We might, therefore, call the iron in the pyrite inert matter and the sulphur a combustible and deduct the weight of the oxygen that unites with the iron, when the pyrite is burned to iron oxide. However, the determination of the amount of pyrite in coal is attended with some difficulty. The usual method has been to calculate the pyrite from either the iron present or from the total sulphur. As coal may have iron present in other forms than pyrite, and generally has organic sulphur and sometimes gypsum present, it can readily be seen that any determination of pyrite in coal that is based on total iron or total sulphur may be the reverse of accurate. I n speaking of the determination of oxygen by difference in ultimate analysis of coal, Prof. Lord
66
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y ,
says:' "The result so obtained is always inaccurate, the error increasing with the percentages of the ash and sulphur. The weight of the ash does not represent that of the mineral matter in the coal, the pyrite in the coal being burned t o Fe20,, and the sulphur passing off as SO,. Thus 4 of sulphur in 2 FeS, (pyrite) is replaced by 3 of oxygen in Fe,O,, and the loss of weight is equal t o five-eighths of the sulphur. For this reason many chemists use five-eighths sulphur instead of sulphur in the determination of oxygen by difference. As coals contain sulphur in other forms than FeS, and also frequently other compounds that lose weight on burning, such as FeCO, and CaCO,, it is doubtful whether the results obtained in this way are any better than those given by the simple formula first given." In fact a calculated correction made in this way may be a greater source of error than the absence o€ any such correction. These various losses in weight, of the mineral constitutents of coal, are an important factor in some calculations and vary considerably in different coals. I will endeavor t o show how we can arrive at the sum total of these losses, although we cannot tell of what they consist or a t present make any special corrections. Sometime ago it became necessary for the laboratory of the Fuel Engineering Company t o get out a set of tables for various coals, by which one coal could be compared with another, and their relative value shown in heat units. Under the advice of Mr. E. H. Taylor the following method was worked out in the laboratory. A sample of coal was taken and the whole sample (which was usually about thirty pounds) was turned on to a clean table. There were then picked out by hand some of the very best pieces in the whole sample. These were laid aside and another sample was picked by hand in such a way that it would have about a 2 0 per cent. ash. Care was used to see that this sample had all its ash constituents present in the proportion natural to the sample. This could be done by proper crushing and mixing and was necessary for the reason that there were often present pieces of both the roof and bottom which varied widely in their character. We had then two samples of coal, one of which was the very best coal that could possibly be gotten from that mine in a commercial way, and the other rep1-"Notes on Metallurgical Analysis,'' N. W. Lord, page 170.
resenting the same coal but very high in ash. These two samples were then ground to pass a Ioo-mesh sieve and the ash was determined on each sample. From the ash of these two samples there were calculated four more coals to make a series having per cents. of ash increasing in regular order, and these were then made up from mixtures of the low and high ash samples. For instance, if we found our low ash coal t o be 5 per cent. ash and the high ash coal to be 2 0 per cent. ash, four mixtures of these two coals would be made having respectively 8, 11, 14 and 17 per cent. ash. Thus we got samples of a certain coal having 5, 8, 11, 14, 17 and 20 per cent. ash. There is nothing artificial about this set of coals but on the contrary i t is all the natural coal and everything is in its proper proportion. This set of six coals was then carefully run in duplicate in a Mahler oxygen calorimeter and the ash was also run in duplicate. These results were then plotted on cross-section paper, and if the work had been carefully done the result was a straight line. The line, however, showed a peculiarity. /MRT V O L A T Z E flATT€R %ASH
fN
ASHCONSTITUE~TS OF COAL
As an example I will use an Illinois coal that is quite common in the Chicago market. The line of this coal shows that it has a t 5 per cent. Ash -13978.5 4 per cent. -14149 0 3 per cent. -14319.5 2 per cent. -14490.0 1 per cent. " -14660.5 0 per cent. '' -14831.0 ' I
British
Thermal
Units
' I
' I
"
I'
'' I'
'*
'* or pure:cod.
T H E JOLRNAL OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y . This shows that each addition of I per cent. ash means a loss of 170.5 British Thermal Units. If the line is carried along through increasing per cents. of ash it will be found that at 87 per cent. ash the B. T. U. are used up. We can also say that if each I per cent. of ash represents the loss of 170.5 B. T. U. (which the line shows to be the case), then the loss of IOO per cent. ash or the pure coal would be IOO >: 170.j B.T. U. or 17050 B. T. U. However, we have already seen that our pure coal is 14831 B. T. U. Now as to this difference. We rely much on the accuracy of the Mahler oxygen calorimeter. Much use of the instrument has proved that it is accurate and reliable, and if properly handled will give close and concordant results. We, therefore, conclude from this that our determinations of the B. T. U. are correct and that the discrepancy is caused by the ash, and as our line shows that we have not sufficient B. T. U. to carry the line to IOO per cent. ash, then also we conclude that what we have been weighing and calling I per cent. ash represents matter that in the original coal weighed more than I per cent. This being so, then the only correct figure we have in our table is o per cent. ash or pure coal, as this is the only figure in which all errors, due to loss of volatile non-combustible matter in the ash constituents, are eliminated. Our figure for pure coal then being 14,831 B. T. U.,our factor of loss for each I per cent. ash will be 148.31B. T. U. instead of 170.5. Our line having used up all B. T. U. a t 87 per cent. ash shows us that we have lost 13/100 of our total ash, or in other words, each I per cent. ash that we have weighed as such really represented 1.13 per cent. inert matter in the original coal, of which amount 0.13 per cent. was volatile and non-combustible. While we have good grounds for stating the amount by weight of this volatile non-combustible matter in the ash constituents, we have nothing to show of what it is composed, and so we are as far as ever from this very desirable and interesting point. This loss, of which no account has been taken, throws some light on certain discrepancies which have never been explained. Much work has been done on certain lines in the endeavor to get a basis on which coals might be compared. The pure coal basis has been used, with and without a correction for pyrite. Ash- and moisture-free coal has usually been called pure coal. Now pure coal must necessarily
67
be only one thing for a certain coal and should be the same figure when calculated from any per cent. of ash. The fact is, however, that it varied with each difference in the ash, and the same coal from the same mine would give as many different figures for pure coal as there were different per cents. of ash in the different samples. As stated in the first part of this article, the correction for pyrite in the coal based on the total iron or total sulphur may in some cases be misleading. A glance a t our original line with uncorrected ash will show the cause of some errors. As the error in the ash increases regularly for each I per cent. of ash as weighed, then the error in a I O per cent. ash would be twice as great as in a 5 per cent. ash, and the error not being recognized as present, they would never figure back to a common basis. This method has never proved of any practical value and has been generally discarded. It has also been stated that in calculating this pure coal, a correction should be made for water of constitution in the mineral constituents. While this is a fact commonly recognized, still we are as yet unable to make such a determination. Until we are able to determine this point with a reasonable degree of accuracy, there is not much that we can say in regard to it. I n the ultimate analysis of coal there are three of the principal determinations that we know may be in error. Carbonates if present will give up carbon dioxide and this will be calculated as carbon, so this determination will be too high. Combined water will be calculated to hydrogen and cause this determination to be also too high. Oxygen being determined by difference is a very uncertain figure as all the errors may effect it. Having no ultimate analysis on this particular coal that I a m using as an example, I will take the average oxygen and ash of thirty Illinois coals on which the ultimate analysis has been run, to show what difference a corrected ash would make in the oxygen figure alone. The average ash is 14 per cent. and the average oxygen figure is 8.75 per cent. I n the coal taken as an example when I per cent. of ash as weighed equals 1.13 per cent. ash in the coal, the corrected ash would be 15.82, thus lowering the oxygen figure from 8.75 per cent. to 6.93 per cent. Such corrections as this show why a t times heat values calculated by Dulong and Pettit’s formula vary so widely from the heat values as determined by the oxygen calorimeter.
68
T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y .
We find t h a t coals vary, as we would' naturally expect, in the amount of this volatile non-combustible that is included in the mineral constituents. On the few coals that have been calculated in this manner we find the largest amount t o be as in the table used as a n example. I n this table each I per cent. of ash as weighed equals 1.13 per cent. of ash in the coal unburned. The smallest amount found, I per cent. ash as weighed, equals 1.067 per cent. ash in the unburned coal. It is very probable that further work along this line will show a wider variation than is given above.
filtered barium hydroxide solution. Another wash bottle containing a solution of sodium hydroxide was placed between the barium hydroxide tube and the aspirator, a suction pump. A blank experiment with this apparatus showed no precipitation of barium carbonate after drawing air through for one hour. A repetition of Schwalbe's experiment showed a copious precipitation of barium carbonate. The possibility that this evolution of carbon dioxide might be due to the action of the oxygen of the air upon the heated rosin, aided by the presence of slight traces of spirits of turpentine in the rosin, led to a repetition of the experiment using spirits of turpentine alone instead of rosin. With a specimen of old spirits of turpentine an THE STABILITY OF ROSIN AT SLIGHTLY even heavier precipitation of barium carbonate ELEVATED TEMPERATURES. occurred than with rosin. No question of the B Y CnAS. H. HSRTYAND W. S. DICKSON. splitting off of a carboxyl group could arise here. Received November 16, 1908. A specimen of freshly distilled turpentine showed On heating American rosin to 12o0--14oo C. in a also a precipitation of barium carbonate, but not current of air freed from carbon dioxide, Schwalbel so marked as with the old specimen. obtained a copious precipitate of barium carbonate Having proved that the spirits of turpentine by conducting the gases from the flask in which alone was capable of giving the precipitation the rosin was heated into a solution of barium a current of steam was observed by Schwalbe, hydroxide. He interpreted this as evidence of passed through molten rosin for eight hours in the decomposition of the abietic acid in the rosin order to completely remove all spirits of turpentine. with consequent formation of the hydrocarbon Repeating Schwalbe's experiment with this rosin abietene, and pointed' out the effect such a decomthe precipitation was still observed. Evidently position must have upon the melting point and the presence of slight traces of spirits of turpentine saponification number of rosin. was not alone responsible for the precipitation From evidence obtained during the course of observed. another investigation we were inclined to doubt It remained therefore to determine the possible the accuracy of Schwalbe's interpretation. Acinfluence of oxygen and of moisture on the formacordingly the following investigation was under- tion of carbon dioxide from the molten rosin. taken, the results of which show that rosin which Accordingly, the current of air drawn through the has not been long exposed to the oxygen of the flask was freed first from carbon dioxide by sodium atmosphere can be heated indefinitely a t 140O hydroxide, then dried by passing through sulphuric without showing any evidence of the formation acid. A marked precipitation of barium carbonate of carbon dioxide, provided oxygen and moisture was again observed. Then moist nitrogen was are excluded from the flask in which the rosin is substituted for air. The nitrogen was prepared heated. by drawing air through three wash bottles filled EXPERIMEXTAL. with an alkaline solution of pyrogallic acid. Again At the outset Schwalbe's experiment was re- a precipitation of barium carbonate occurred. peated. For heating the rosin a 2 0 0 cc. Erlenmeyer Finally a current of dry nitrogen was drawn through flask was placed in a beaker containing cotton- the flask and after all air had been expelled the seed oil. The air entering the flask was freed rosin was heated t o 140' and kept at this temperafrom carbon dioxide by being drawn through three ture for seven hours without the slightest prewash bottles filled with a strong solution of sodium cipitation in the tube containing barium hydroxide. The above experiments were carried out onza hydroxide. After leaving the flask the air was specimen of freshly distilled rosin from the oleopassed through a test tube half filled with freshly &sin of Pinus heterophyEZa (Cuban Pine). This leeit. angew. Chem. 18, 1852.
