ANALYTICAL CHEMISTRY
1050
Results obtained when a number of steroids were heated LITERATURE CITED separately with 10% tetraethylammonium hydroxide are shown (1) Callow, N. H., Callow, R. K., and Emmens, C. W., Biochem. J., 32, 1312 (1938). in Table I. Related to These indicate that ~ ~ ~3-ketosteroids ~ ~when ~ (2) Fiefler, t ~L. F.,and ~ Fieser, ~ M.3t “Natural ~ Products d Phenanthrene,” 3rd ed., p. 101, New York, Reinhold Publishheated with alkali show a characteristic absorption maximum in ing Corp., 1949. the area 373 to 375 mw. (3) Langstroth, G. O., and Talbot, S . B., J . Bid. C h a , 128, 759 (1939). (4) Tansey, R. P., and Cross, J. M., J . Am. P h a r m . Assoc., 39, 660 ACKNOWLEDGMENT (1950). The authors wish to thank Merck & Co., Inc., for generosity in supplying the cortisone acetate used in this work. RECEIYED for review October 15, 1951. Accepted February 7, 1952.
Determination of Vanadium in Fuel-Oil Ash G. L. HOPPS AND A. A. BERK U . S . Bureau of Mines, College Park, M d .
TILIZATION of residual fuel oils in power boilers has, in some instances, resulted in a special type of corrosion. The trouble appears most likely to occur where combustion conditions and design cause oil ash from high-vanadium oils to pass over metal surfaces a t comparatively high temperatures (6). The ash content of fuel oil is generally much less than 0.1% ( S ) , and only a small proportion of the ash would be deposited on a probe placed in the hot-gas stream; a laboratory study of this problem therefore would require a method for determining vanadium in very small samples. Previously described analytical methods include one by Karchmer ( 2 ) for determining vanadium in petroleum ash by fusing the ash with carbonate to separate the vanadium from the interfering metals and subsequently determining vanadium colorimetrically by the color formed with hydrogen peroxide. Wrightson (8) employed a bisulfate fusion and determined traces of vanadium in petroleum ash by the color formed with diphenylbenzidine reagent. Sandell ( 4 ) has determined vanadium in silicate rocks colorimetrically by using phosphotungstic acid, after fusion and separation of the vanadium with 8-quinolinol. An extenuive spectrophotometric study was made of this phosphotungstvte method by Wright and Mellon ( 7 ) , and it was adapted to the determination of vanadium in alloy steels. .ifter attempts to adapt several methods for a colorimetric procedure in this laboratory, the phosphotungstate method was chosen as offering the best possibilities, because it was sensitive enough to be used on very small samples and was subject to a minimum of interference from other constituents of fuel-oil ash. The procedure described herein is relatively rapid, a single determination requiring approximately 20 minutes.
fore attempted by the usual fusion procedure in which the ash is fused with a mixture of sodium carbonate and potassium nitrate. Vanadium recovery was usually low by this method, and results were inconsistent, possibly because of volatilization of the vanadium salts a t the high fusion temperatures used (800’ to 900” ‘2.). When sodium peroxide was substituted aa the fusion medium, results were much better, as fusion temperature could then be dropped to 500’ C.; however, further study showed that iron caused no interference and a tedious fusion separation waa therefore unnecessary. The acid-extraction procedure described herein has been shown epectrographically to bring the vanadium content of the ash into solution. REAGENTS AND APPARATUS
Sulfuric acid, concentrated, C.P. Nitric acid, concentrated, C.P. Phosphoric acid, 85% C.P. Sodium Tungstate Solution, 0.5 M . Dissolve 16.5 grams of Folin’s sodium tunestat.e dihydrate in distilled water t o make 100 mI. of solution. Standard Vanadium Solution. Dissolve 0.535 gram of purified vanadium pentoxide in 50 ml. of 5% sodium hydroxide. Acidify
-
SAMPLING
iish deposits used in this investigation were obtained as s c r a p ings from the gas side of boiler tubes, superheaters, air preheaters, and stacks in installations firing No. 6 fuel oil. All deposits were ground to -50 mesh. Spectroscopic examination of the ash samples showed that major constituents were iron, sodium, nickel, and vanadium, with lesser amounts of magnesium, cop er, chromium, silica, calcium, and other elements normally founx in trace concentrations in heavy fuel oil. The most abundant molecular combination was the sulfate, especially in the lowertemperature deposits. The cold-water solubility of the vanadiumcontaining components of these ash samples ranged from zero in screen-tube deposits t o 100% in air-preheater scrapings. The most recent card files contained no lines that would aid in identifying such x-ray diffraction patterns as were obtained for these vanadium compounds.
IO
EXPERIMENTAL
It w a first thought that the heavy iron concentrations in the ash samples would cause interference with the vanadium determination, and separation of the vanadium from the iron was there-
MQ. VANADIUM per 25 ml. Solution
Figure 1.
Standardization Curves
1051
V O L U M E 2 4 , N O , 6, J U N E 1 9 5 2 with 10 ml. of 1 t o 1 sulfuric acid and dilute t o 1000 ml. with distilled water, Each milliliter of this solution will contain 0.030 m of vanadium and the standardization should be checked with fur dioxide and standard permanganate solution in the usual way (1). Colorimeter. A potentiometer-type photoelectric colorimeter was used with a n American Instrument Co. KO.436 filter and a high-pressure mercury lamp as light source. For high vanadium concentrations, the less sensitive No. 460 filter used wit,h an ordinary tungsten lamp was found to be satisfactory. The sample cells were standardized 50-ml. test tubes with an inside diameter of 2 cm. Distilled water was used in the reference cell.
BUT:
PREPARATIOh OF CALIBRATION CURVE
By means of a Mohr pipet transfer aliquots of standald vanadium solution containing from 0.030 to 0.450 mg. of vanadium into a series of 50-ml. beakers. To each, add 3 ml. of nitric acid, 1 ml. of 85% phosphoric acid, and 1.5 ml. of 0.5 Jf sodium tungstate solution. Dilute to 20 t o 25 ml. with distilled n-ater, and boil 2 minutes. Cool t o room temperature, transfer quantitatively to a 25-ml. volumetric flask, and dilute to volume. Measure transmittance of these solutions with the photoelectric colorimeter. and plot vanadium concentration against per cent transmittance. TEST PROCEDURE
Weigh accurately a 10- to 20-mg. sample of the oil ash into B 50ml. beaker. Add 2 ml. of concentrated sulfuric acid and 0.5 nil. of concentrated nitric acid (conveniently added by means of calibrated eyedroppers). Heat carefully over a small Bunsen hurner flame until all the nitric acid is driven off and fumes of sulfur trioxide are evolved. Continue fuming until a considerable quantity of salts has settled out-5 minutes are usually sufficient. Cool, and add 10 to 15 ml. of distilled water and 3 ml. of concentrated nitric acid. Heat t o dissolve all salts. Add 1 nil. of 85% phosphoric acid and 1.5 ml of 0.5 -11sodium tungstate solution. Boil for a few seconds; then cool to room temperature, transfer the solution t o a 25-1111. volumetric flask, and dilute to volume with distilled water. Measure the transmittance of this solution with the colorimeter, and determine the amount of vanadium present from the calibration curve,
Table 1. Accuracy of Vanadium Determination on Samples of Known Vanadium Content Sample V Present, % V Found, % Error, %
a
Averaee. % Max. -deviation f r o m Standard deviation, %
Characteristic transmittance curves are shown in Figure 1. Of the colorimeter filters available in this laboratory, the 436-mp filter was found to be m o ~ tsensitive to the yellow phosphovanadium-tungstate complex, and reagent absorbency with the filter was only slight when a high-pressure mercury lamp was used as light source. Where a wider concentration range was desired, the less-sensitive 460 mp filter was satisfactory. The calibration curve obtained with the 436-1np filter shows that as little as 0.005 mg. of vanadium may be detected in 25 ml. of solution.
1.40 0 79 1.64
Table 11. Reproducibility of Results for Yanadinnl in Ash Deposits
av., %
RESULTS AND DISCUSSION
Ash deposit 1 2 3 Bureau of Standards eample 50aa Bynthetic sample 1 F~SO?.~H~O.NH,‘VO~ Synthetio sample 2 FeSOp.zH*O.Vd
The principal variant in the method is concentration of sulfuric acid in the final solution, due t o the inexact fuming procedure. However, it was found that considerable leeway in acid concentration was permissible. Varying the concentration of sulfuric acid over the entire range, corresponding to from no fuming loss to almost complete fuming loss, had little or no effect on the vanadiuni found. When the maximum possible amount of sulfuric acid is present, ( 2 ml. of sulfuric acid in 25 ml. of solution), the total acidity of the final solution is high enough 80 that the sodium tungstate is partly insoluble, and precipitation occurs when the reagent, is added. However, the precipitate readily redissolves when the solution is brought to a boil in the next step of the procedure and gives no further trouble. Chromium is often present in petroleum ash and niay cause interference due to the formation of difficultly soluble chromium compounds when the sample is fumed with sulfuric acid. -XIthough such insoluble matter may be filtered off, it is simpler to avoid its formation in the first place by avoiding the funling of the ash sample for longer than is necessary. Chromium coricentrations as high as 5% (1 mg. in a 20-nig. ash sample, added as chromium metal) caused no interference with the method when due care was exercised, and this concentration is considerably higher than that found in any petroleum-mh sample in this laboratory.
1.38 0 79 1.62
-1.2
-1.4
0.97
0.96
-1.0
2.43
2.40
-1.2
2.80
2.77
-1.1
n
Hydrofluoric acid added to diwolve tungsten in samplr.
The recovery of vanadium from synthetic samples and from ash samples of known vanadium content is shown in Table I. The “known” vanadium of the ash samples was determined by titration with ferrous sulfate in accordance with a procedure described by Willard and Young (6)-a method considered to be accurate when a large enough sample is available (1 or 2 grams). The reproducibility of results by one operator for ash deposits of both low- and high-vanadium content is shown in Table 11.
Ash 4 2.42 2.45 2.51 2.45 2 31 2.49 2.47 0.05
0 03
Per Cdnt vanadium Ash 5 Ash 6 Ash 7 .ish 8 0.22 0.15 1.73 3.63 0.25 1.75 3.63 0.14 0.25 1.71 3.37 0.13 1.73 3.61 0.24 0.14 1 70 3.55 0.20 0.14 0.0 1.70 :3 60 0.14 0.14 1.72 3.61 0.23 0 01 0.003
0.03 0 02
0.06
0 03
0.03 0.02
.ish 9 0 71 0.69 0.70 0.68 0.71 0.72 0.70 0.20 0.1
The effective range in concentration of vanadium meamred by this method is 0.005 to 0.450 mg. of vanadium in 25 nd. of solution when the No. 436 filter is used in the colorimeter. The amount of vanadium present in the h a 1 25-nd. solution of the ashes tested was usually within the range of the method when a 10- to 20-mg saniple was used. Higher vanadium concentrations would, of course, require either less than a 10-mg. sample or the use of an aliquot of the sample solution. ACKNOWLEDGMENT
This work is part of a program of study developed and correlated by John F. Barkley, chief, Fuels Utilization Branch. Publication review and suggestions by F. H. Gibson of the Coal Constitution and Miscellaneous .4nalysis Section are gratefully acknowledged. LITERATURE CITED
(1) Hillebrand, W. F., and Lundell, G. E. F., “rlpplied Inorganic .An-
alysia.” p. 359, New York, John U’iley & Sons, 1929. Karchmer, J. H., Proc. Am. PetroEeum Inst., 3, 2 9 M , 72-8 (1949). Kottcamp, C. F., and Crockett, L. O., “Some Aspects of the Application of Residual Oils as Fuel for the Gas Turhinc,” ASME Paper 50-A-131 (December 1950). (4) Sandell, E. B., IND. ENG.CHEY.,Ax.4~.ED.,8, 336 (1936). (5) Tibbetts, E. F., et al., “Problems Encountered in Burning Heavy Fuel Oil as Related to Attack of Metals at High and Low Temperatures and the Fouling of Tube Banks,” ASME Paper 50--4136 (December 1950). (6) Willard, H. H., and Young, P., IXD.ENG.CHEM., ANAL.ED..4 , 1 S i
(2) (3)
(1932).
Wright, E. R., and RZellon, A I . G., Ibid., 9, 251 (1937). (8) Wrightson, F. M., ANAL.CHEM.,21, 1543-5 (1949). (7)
RECEIVED for review December 1, 1951.
Accepted Jan?iary 23. IY.52.