T H E JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY
Julv. 1921
633
Determination of the Volatile Matter in Graphite I By Owen L. Shinn TOWNE SCIENTIFIC SCHOOL, UNIVERSITY OF PENNSYLVANIA. PHILADELPHIA. PA.
I n the commercial analysis of graphite it frequently happens that dispute arises between two analysts as to the peroentage of volatile matter in a given sample. In the present investigation this question was studied with the purpose of determining the cause of the different results obtained and of finding some remedy. The literature contains very little upon this subject. The only recommendation to be found is to make the determination in the same manner as that used in determining the volatile matter in coal. This method consists in heating a sample for a definite time, in a closed platinum crucible, a t a definite distance above a burner of definite size. The size of the burner, time of heating, placing of the crucible, etc., are the result of arbitrary selection, and the results are not accurate and are comparable only when obtained under exactly the same conditions. When this method of heating was applied to graphite, the results varied greatly. The variation of time by a few seconds introduced a large error. It was definitely noted that the variation was not proportional to the size of the sample, but was proportional to the time of heating, the burner used, or the temperature. These effects are plainly shown by the following resuIts, obtained by heating 1-g. samples in a platinum crucible provided with a well-fitting lid. The crucible was the same in each case, but a good Bunsen burner and a 30-mm. MBker burner were used. The losses are given in per cent: -
.............
Time of heating, min.. Over Bunsen burner.. ............. Over MCker burner..
..............
4 1.52 2.30
8 1.75 3.56
30
....
9.45
A method for the determination of the volatile matter in coke and anthracite coal, which has been in more or less general use for many years, is to heat the sample in an atmosphere of nitrogen. When 1-g. samples of the same lot of graphite as was used above were heated in nitrogen the losses were 0.88, 0.93, 0.94, and 0.94 per cent. The time of heating was 15, 30, 45, and 60 min., respectively. This shows that much of what was called volatile matter in the first tests was due to oxidation of part of the graphitic carbon. TEMPERATURE OF HEATING I The first question to be answered was: At what temperature will the volatile matter be driven off when heated in nitrogen? Two methods of heating were tried: (a) a platinum boat in a silica tube, and ( b ) a Rose crucible. The first was found to be preferable. Below 400" C. the loss in weight was negligible, even though the heating was continued for over 2 hrs. A temperature of 600" C. was found to be sufficient, as no further loss was sustained by heating to a higher temperature. Heating samples to varying temperatures in a platinum crucible provided with a well-fitting lid gave the following results: TEMPERATURE
c.
150-170 250-270 300-325 400-450 450-500
TIME Hours 1 1 1 1 1
Loss Per cent 0.05 0.11 0.11 0.31 3.34
This shows that below 450" C. the oxidation is very slight, but above that temperature carbon is burned off rapidly. It further shows that the volatile matter cannot be driven off a t a temperature below that a t which the carbon begins to oxidize. I n order to reduce the percentage of oxidation the crucible was filled almost to the top with graphite, and heated for 8 min. The loss was 1.32 per cent. Attempts were made to drive the air out of the crucible by a layer of 1 Received
March 10, 1921.
pure ammonium chloride, oxalic acid, or some organic substance which was non-coke-forming, but the values obtained were all about double that obtained by heating in nitrogen. TIMEOF HEATING: As will be noted from the figures given above, t h e loss on heating in a closed crucible for 4 min. over a MBker burner is about three times that on heating in nitrogen. The time was reduced, until it was found that, by heating for exactly 30 see. to a temperature of about 700" C., the loss in weight was just about equal to that obtained by heating in nitrogen. Five samples of a purified American graphite were heated in nitrogen to about 600" C. for 30 min., and the losses were 0.89, 0.85, 0.93,0.91,and 0.90 per cent, or a mean of 0.896 per cent. Five other samples of the same substance were heated for exactly 30 sec. in a closed platinum crucible to about 700" C., and the losses were 0.95, 0.96, 0.99, 0.97, and 0.96 per cent, or a mean of 0.966 per * cent. Similar samples heated under the same conditions for exactly 60 sec. gave the following results: 1.03, 1.02, 1.01, 1.04, and 0.99 per cent, or a mean of 1.018 per cent. The "coke" obtained from each of the samples heated for 30 sec. was further heated in nitrogen, and an additional loss of from 0.07 to 0.10 per cent was noted, showing that even in this short time of heating slight oxidation had taken place. The results obtained in this way were considered close enough for ordinary commercial determinations. The same procedure was repeated with samples of commercial graphite from different localities, and the results obtained are noted in the following table: LOCALITIES Alabama.. Connecticut.. Canada Madagascar., Norway.. Ceylon (lump). Ceylon (chip).
Loss on Heating in Nitrogen Per cent 1.17 0.72 0.49 1.39 2.06 1.25 2.83
..................... .................. ........................ .................. ...................... ................. ..................
Loss on Heating for 30 Sec. in Closed Pt Crucible Per cent 1.19-1.27 0.77-0.80 0.56-0.60 1.42-1.48 2.07-2.08 1.22-1.29 2.82-2.83
The problem under consideration seems to be quite analogous to that of determining the volatile matter in coke and anthracite coal. Mead and Attix' found that the factors noted above influenced their results. They claimed that the size of the crucible affected the amount of loss. A few isolated instances seemed to indicate this same influence in our experiments, but repeated attempts to prove this failed. Four crucibles holding 16, 17, 24, and 30 cc. were used. One-gram samples of three different graphites were heated in exactly the same way for 4 rnin., the losses being as follows:
4
1.52 1.53 1.43 1.43
1.16 1.22 1.18 1.20
2.88 2.79 2.75 2.85
The fit of the lid or the porosity of the platinum may have. caused the variation, but evidently not the size of the crucibles, as they were all of the same relative shape and the surface exposed increased from 1 to 4, while the variation is not regular in that direction. I n the analysis of coke and anthracite, Mead and Attix recommend heating for a certain time, weighing, heating for exactly the same time, and weighing again. They then consider the second loss oxidation, deduct a like amount from the first loss, and call the difference true volatile matter. This method was tried repeatedly with graphite, but the results were farther from the true value than when the sample was heated in a closed crucible for exactly 30 see. 1
J . Am. Chcm. Soc., 2 1 (1899), 1137.
634
T H E JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY
THENITROGEN METHOD From what has been said it is evident that any method of crucible heating by which air is in contact with the graphite is not accurate. The true value of the volatile matter in graphite can be determined only by heating in an atmosphere of pure nitrogen, hydrogen, or some other inert or nonreducible gas. This is best carried out by placing about a 1-g. sample in a platinum boat, and placing the boat in a silica tube through which a steady stream of pure nitrogen is passed. When the air has been completely expelled, the tube under the boat is heated for 30 min. with a 30-mm. MBker burner. It has been shown that 15 min. is long enough, but a half hour is recommended as a precaution. Various forms of heating and temperature control were tried, such as electric and gas heated combustion furnaces, electric tube furnace, multiple burners, etc., but the single MBker burner was found to be sufficient. It is important that pure nitrogen be used. Attempts to use commercial nitrogen containing 3.5 per cent of oxygen gave inaccurate results, the amount of loss depending upon the time of heating. SUBSTITUTE METHOD The nitrogen method will consume over an hour and will
Vol. 13 No. 7
require the preparation and purification of nitrogen, so that as a working method the following is suggested: Place a 1-g. sample of the graphite in a platinum crucible provided with a well-fitting lid. Support this in such a way that it can be heated to 700" to 750" C. (about 10 mm. above a good MBker burner), and heat for exactly 30 sec. As a method for the determination of volatile matter in coke and anthracite, the short heating process is entirely unsatisfactory. SUhlnIARY 1-It is pointed out that the method of determining volatile matter in graphite by heating in a crucible is rendered inaccurate by the oxidation of the graphite. 2-The error is dependent on the temperature, and time of heating, as well as on other factors. 3-The method of heating in a boat in an atmosphere of nitrogen gives accurate results, but is time-consuming, and requires pure nitrogen. 4-Results very close to the true value may be obtained by heating in a crucible in the air under certain stated conditions.
Discoloration in Canned Sweet Potatoes' By Edward F. Kohman RESEARCHLABORATORY, NATIONALCANNERS ASSOCIATION,WASHINGTON. D . C.
The producer of canned foods must impart to his goods at least two prime and essential qualities before he can consider himself in any measure successful in his art. The first requisite is wholesomeness, secured by the sanitary packing of sound raw material when it is in that particular state of maturity most suitable for food. The second is palatability and appearance. On first thought it might seem that palatability and appearance are a necessary consequence of wholesomeness. This is not always the case, as is illustrated by various forms of discoloration in canned foods. The discoloration in canned sweet potatoes is of this character. Sweet potatoes frequently become black in the can, unaccompanied by any other noticeable feature. Then discoloration apparently begins where the potatoes are in contact with the can, but may eventually permeate its entire contents. As there is usually a small amount of semiliquid starch paste in the bottom of the can, the discoloration is more pronounced there in the early stages, as the result of closer contact. How soon after packing discoloration sets in is a variable factor. From the experiments below, it would seem that in extreme cases it may occur very early. All foods evolve hydrogen sulfide upon being heated. This sometimes leads to a black discoloration in canned foods, due to thc formation of ferrous sulfide. From general appearance one cannot distinguish the discoloration of canhed sweet potatoes from that due to the formation of ferrous sulfide, but a few qualitative tests sufficiently demonstrate its entirely different character. If black sweet potatoes are treated with dilute hydrochloric acid, the color is discharged, but when the acid is neutralized with ammonia, even after heating to drive out any hydrogen sulfide, the black color is again obtained. Hydrogen peroxide does not bleach the discolored sweet potatoes. EFFECTOF AIR Into a number of cans of sweet potatoes from each of two distinct commercial packs, both of which were normal in 1 Presented before the Division of Agricultural a n d Food Chemistry, a t t h e 61st Meeting of the American Chemical Society, Rochester, N Y , April 26 t o 29. 1921
every way, air was admitted under sterile conditions by a very small puncture. The cans were placed in boiling water for some time to relieve the vacuum and thus to prevent the rapid inrush of air which would augment the chances of contamination. Alcohol was then burned on the top of the can. Just as the last had burned away, a puncture was made, and this was a t once covered with a pad of sterilized cotton which was held in place by setting another can upon it. These cans mere left undisturbed for periods of time ranging from a week to a month. At the end of a week, typical and distinct black discoloration of the sweet potatoes had set in in the bottom of the can, where there was the most moisture and the best contact between the can and its contents. This became more and more abundant as time elapsed. After varying periods of a week or more, some of the cans were resealed.$ These resealed cans retained the black that had formed in them to the end of the experiment. It could not be decided whether there was any increase of the black in the cans after resealing, because only a limited number of cans were used in the experiment. This experiment makes very evident the necessity of air-tight seams in the cans. It does not answer the question as to the manner in which the admitted air brings about the discoloration, whether by direct action upon the sweet potatoes or by aiding the solution of the iron, which in turn is the immediate cause. EFFECTOF IRON SALTS To throw more light upon this point, sweet potatoes were canned with the addition of iron salts. The sweet potatoes were steamed in the usual way until soft, then peeled, and packed into No. 3 cans and into quart glass fruit jars. A varying amount of concentrated iron solution ranging from 100 mg. to 700 mg. per can was then added. It had already been found by analysis that black sweet potatoes contained amounts of iron as high as this. When this solution of iron was poured over the potatoes they immediately became black. Where a thick peel had been removed or where the ends were cut away, there was no appreciable darkening.
