V O L U M E 2 0 , NO. 11, N O V E M B E R 1 9 4 8 Table I. Cr as Na*CrO,,
%
100 99 95 90
70 50 20 10 0
Cr as KazCr20r,
%
0 1 5 10 30 50 80 90
100
Transmittancies
1% Cr (A = 528 mp)
T
Log T
96 92 79.2 66.5 30.5 14.5 5.75 3.80 2.58
1 ,982
1.964 1.899 1.823 1.484 1.161 0.760 0,580 0.412
15% Cr (A = 555 m p ) T Log T 9.5.4 1.980 90 1.954 73 1.863 55 1,740 18.5 1.267 6.7 0.826 2.1 0.322 1.6 0.204 1.2 0.079
length for each concentration of chromium in order to obtain the most satisfactory curve. A shift of approximately 2 mp towards the longer wave lengths occurred for each 1% increase in concentration. A family of curves may be plotted on the same graph for convenience in using the method which is described in detail below, and a notation made as to the most suitable Tyave length to be used a t the given concentration.
1045 Accurately matched or calibrated Corex cells are then filled with the solvent and the solution to be analyzed. Readings are taken until constant within a small fraction of a scale division. The logarithm of the per cent transmittancy is then found and the ratio of chromate to dichromate is read directly from the proper curve. It is necessary to interpolate when the percentage of chromium found falls between two of the curves plotted. From the known concentration of chromium and the ratio of chromate to dichromate obtained spectrophotometrically it is a simple matter to calculate the actual concentrations of the two salts. For most convenient operation, a nomograph may be constructed which will relate the three variables: chromium concentration, log T , and ratio of sodium chromate to sodium dichromate; the unknown variable is then instantly computed by laying a straightedge on the two known values. ACKNOWLEDGMENT
The authors wish to express their appreciation to Marvin Carmack of the Chemistry Department, University of Pennsylvania, for his constructive criticisms during the development of the above procedure and for his checking of the spectrophotometric data. LITERATURE CITED
DESCRIPTION O F METHOD
Barnard and XIcMichael, IND. ENG.CHEM.,AXAL.ED., 2, 363
Beckman Model DU Spectrophotometer. Previously standardized Cores cells were used. The cells n-ere cleaned with distilled water and spectroscopically pure methanol each time the solution was changed. Analytical Procedure. An aliquot portion of the solution to be analyzed is first titrated for its chromium content by any standard method. Only one determination need be made for each batch of material. When the chromium content has been determined, the spectrophotometer is set a t the wave length which has previously been established as the one that gives the greatest range of readings, and therefore the greatest over-all accuracy, and also a curve most nearly approaching a straight line.
(1930).
Hantssch, 2 . physik. Chem., 72,366 (1910). Kitson and Mellon, ISD.ERG.CHEM.,ANAL.ED.,16, 42 (1944) Plotnikow and Karshulin, 2. Physik, 38, 502 (1926). Rossler, Ber., 59, 2606 (1926). Sconzo and Marsengo, I n d u s t r i a Chemica, 9, 1163 (1934). Silverthorn and Curtis, Metals and Alloys, 15, 245 (1942). Tawde and Paranjpe, I n d i a n J . Phys., 4, 533 (1930). von Halban, T r a n s . Faraday Soc., 21, 620 (1926). Westburg, Tek. Tid., Upp2. C.K e m i , 59, 41 (1929). Zscheile, J . P h y s . Chem., 38, 95 (1934). RECEIVED February 28, 1948.
Lamp Method for Quantitative Determination of Total Sulfur WILLIA3ZI H. LANE Monsanto Chemical Company, Texas C i t y , Texas
An investigation of the variables of the lamp method has shown that it cannot be used for the quantitative determination of elemental sulfur, or for total sulfur unless elemental sulfur is known to be absent. Complete recovery of sulfur from the combustion of solutions containing volatile organic sulfur compounds has been effected only under highly specialized conditions.
T
HE differentiation between elemental sulfur and sulfur compounds when they are dissolved in hydrocarbons has been the subject of considerable study. There are several references to the fact that the standard lamp method does not give satisfactory recoveries of elemental sulfur (1, 6).but recommended methods for the determination of total sulfur apparently ignore these results (3, 4, 9 ) and several systematic procedures (4, 6, 9) are based upon differences in total sulfur as determined by the lamp method before and after the removal of other sulfur compounds by extraction, The tentative A.S.T.M. method ( 3 ) describing the sulfur lamp states simply that it is intended for the determination of sulfur in gasoline, kerosene, petroleum naphtha, and other petroleum oils that can be burned completely. Ball ( 4 ) carried out a comprehensive investigation on the determination of types of sulfur compounds in petroleum dis-
tillates and pointed out that the method for total sulfur most widely used in the United States is the A.S.T.lI. lamp method ( 3 ) . Ball noted that not only is the total sulfur determination important in itself, but it also forms the basis for those determinations of groups of sulfur compounds that depend on difference in sulfur content before and after treatment to remove a specified group. This same author used the butyl mercaptan (butanethiol) method (8, IO) to determine free sulfur, but employed lamp sulfur determinations before and after shaking with metallic mercury as a check on the elementary sulfur content. This paper presents the results of a critical study of the lamp method which demonstrate that it is not applicable to aromatic hydrocarbons containing elemental sulfur, and presents certain improvements in the method
ANALYTICAL CHEMISTRY
1046 Table I. Expt. No.
223 224 213 214 215 216 221 222
Combustion of Dilute Solutions of Elemental Sulfur in Isopropanol-Toluene Mixtures Sulfur Theoretical Concn. in Sulfur Mixture Present"
Sulfur Found Recoverv
%
Sg.
Mu.
%
0.0117 0.0117 0.0117 0.0117 0.0117 0.0117 0.0117 0.0117
0.70 0.62 0.77 0.82 0.79 0.84 0.52 0.53
0.44 0.32 0.20 0.21 0.21 0.16 0.13 0.14
63 52) 19 25 27J
Remarks Wick flush with burner top Wick hack
pulled inch
1/16
Calculated from known concentration of sulfur in isopropanol-toluene mixture and amount of mixture burned.
CRITICAL STUDY OF LAMP METHOD
Method. The details of the method and description of apparatus and reagents have been given by Zahn (11). An important variation in this method as applied in the present investigation was the use of sulfur-free isopropanol as a diluent for the aromatic hydrocarbon sample for the purpose of obtaining a smokeless flame. Another variation in this method RW the use of no special apparatus for air purification. Apparatus. The apparatw employed has becn described in detail ( 3 ) . Reagents. Special reagents to serve as standard sources of volatile organic sulfur compounds used in this work were prepared by dissolving small amounts of carbon disulfide, thiophenol, and thiophene in isopropanol. The volatile organic sulfur compounds mere weighed on the analytical balance in sealed glass ampoules which were later broken beneath the surface of the isopropanol The purities of the three sulfur compounds, as determined by combustion in the oxygen bomb (I) were found to be: thiophenol, 97.6%; thiophene, 97.9Te; and carbon disulfide, 98.4% Blank determinations made on the isopropanol, toluene, and air used in these experiments indicated that they were all sulfur-free. Experimental Work. In his method Zahn (11) made use of a Betz-Hellig turbidimeter for determination of sulfate in the absorbing solution. This instrument and its use have been described in detail by Sheen, Kahler, and Ross ( 7 ) . A variation of Zahn's method in which a Klett-Summerson photoelectric colorimeter is used for turbidity determination of the final barium sulfate sol was used in this work. The sensitivity of this turbidimetric method is *0.01 mg. of sulfur in the range 0.10 to 1.00 mg. of sulfur. The salt acid solution recommended by Sheen, Kahler, and Ross ( 7 ) was used for stabilizing the barium sulfate sol. Ball ( 4 ) has reported data which indicate the lamp method to be only 88.5c7, efficient for the recovery of elemental sulfur, and preliminary experiments by the author indicated that whenever solutions of elemental sulfur in aromatie hydrocarbons were burned by the lamp method, the amount of sulfur found as barium sulfate in the absorbing solution was always on the low side. In the first two experiments of Table I, the tip of the wick was flush with the top of the burner, while in the remaining six experiments the wick tip was pulled back into the burner 0.156 em. ('/,6 inch). These data show that in the combustion of elements1 sulfur solutions, the degree of sulfur recovery as barium sulfate is very low and is markedly dependent on the position of the wick tip. I t is conceivable that low sulfur recoveries can be attributed to the nonvolatile nature of sulfur as compared to the hydrocarbon solvent, possibly only part of the sulfur actually passes into the flame region and is burned It was concluded that the lamp method is not quantitative for elemental sulfur, and hence not quantitative for total sulfur unless elemental sulfur is known to be absent. Variables. The principal variables of the lamp method that were investigated during the course of this study were: number of absorbers, t.ype of absorbing solution, effect of rinsing the
burner chimney, glass wool versus cotton wicking, wick height, character of flame, and elemental versus volatile organic sulfur compounds. The first variable investigated in an effort to find the reason for low sulfur recovery was the effect of more than one absorber in series. Repeated tests in which both two and four absorbers in series were used showed that in no case was the amount of sulfur absorbed increased by the additional absorbers-Le., no appreciable amount of sulfur was ever found beyond the first absorber. Sodium hydroxide varying in strength from 0.1 to 1070, both with and without added bromine water, was investigated as the absorbing liquid. In no case was any variation found in sulfur recovery as barium sulfate. An absorbing liquid consisting of 25 ml. of 1% sodium hydroxide plus 10 ml. of saturated bromine water was used in the study of other variables. The A.S.T.M. method (3) specifically mentions that the chimney of the sulfur lamp must be rinsed after each determination and the rinsings combined with the absorbing solution. In four separate tests the amount of sulfur obtained from the rinsings naq 2% of the amount found in the absorbing liquid. These results clearly testify to the necessity of this step for accurate results. One of the organic sulfur compounds tested by the lamp method to determine degree of sulfur recovery was thiophenol. When burning dilute solutions of thiophenol in sulfur-free isopropanol, the degree of sulfur recovery averaged only about 90% of theoretical until Pyrex woo1 wicks were substituted for the cotton wich that had been in use up to thiq point. Immediately, the sulfur recovery rose to practically lOOyeof theory. This indicated that the cotton wicking had been responsible for the selective removal of the thiophenol from its isopropanol solution. and the data of Table I1 illustrate this point This finding would seem to be in disagreement with the results secured by Zahn (111, who obtained almost perfect sulfur recoverv when using cotton wicking
Table 11. Comparison of Cotton and Pyrex Wool Wicks Expt. NO.
99 100 101 102 103 104 105 107 108 109 110 111 112 113 114 115 116 117 233 234
Wick Material Cotton
Sulfur
Thiophenol Theoretical Concn. in Sulfur 1-opropanol Present
%
Found .WQ. 0.28 0.39 0.29 0.36 0.70 0.48 0.75 0.44 0.70 0.48 0.72 0.38 0.52 0.37 0.47 0.57 0.50 0.73 0.71 0.69
MQ.
0.31 0.45 0.32 0.42 0.76 0.54 0.81 0.51 0.83 0.56 0.79 0.42 0.57 0.43 0.54 0.64 0.59 0.77 0.74 0.74
Recovery
IV.
119 121 122 123 124 129 130 131 132 133 134 135 136 235 236
Standard deviation of sulfur recovery = Thiophene Concn. in Mixture % 0.78 Pyrex wool 0.0389 0.68 0.0389 0.68 0.0389 0.69 0 0389 0.75 0.0389 0.72 0.0389 0.77 0 0389 0.78 0.0389 0.94 0.0389 0.78 0.0389 0.80 0.0389 0.86 0.0389 0.93 0.0389 0.65 0.0285 0.73 0.0285 Standard deviation of sulfur recovery
-
% 90 87 91 86 92 89 93 86 85 86 91 91 91 86 87 89 85 95 96 93 89
3.3 0.79 0.71 0.68 0.69 0.74 0.72 0.72 0.79 0.93 0.80 0.78 0.83 0.93 0.64 0.73 Av.
2.3
101 104 100 100 99 100 94 101 99 103 97 97
100 99
100 108
1047
V O L U M E 20, N O . 11, N O V E M B E R 1 9 4 8 in the combustion of dilute solutions of dimethyl disulfide in a hydrocarbon. However, as the Pyrex wool wicks are as easy to prepare and use as the ordinary cotton wicks, and their use makes it impossible for the liquid being burned to come in contact with any surface other than Pyrex. it was decided to use them throughout the remainder of this work. All the data shown, other than in Table 11, \yere obtained using Pyrex wool wicks. The Pyrex ~ o owicks l are fashioned by hand rolling and twisting a section of glass wool about 17.5em. (7 inches) long cut from a conventional batting of the material. The width of the section is only‘approximated and skill in choosing the proper size comes from practice and experience. The quantity of wool must be such that the wick will fit tightly enough to fill the burner when in position and yet allow for forcing it all the way through. The wick is forced through the burner by a combination of twisting and pushing until it protrudes slightly beyond the burner tip. Thc excess is cut off and the wick pulled back until its end is exactly flush with the burner tip. The effects of wick height and character of the flame are closely related and must be discussed together. Wick height has been found to be without influence on the degree of recoverv from the combustion of volatile organic sulfur compounds; however, it is an important variable in the combustion of solutions of elemental sulfur. Organic liquids, such as isopropanol, burn with an almost nonluminous flame. Aromatic hydrocarbons, such as toluene, burn with the smoky flame indicative of incomplete combustion. However, a luminous but smokeless flame may be secured by mixing the aromatic hydrocarbon with an equal volume of an alcohol such as isopropanol. Even under these conditions, only 89 to 94% of the sulfur present could be recovered, as shown in Table TIT. This low recovery is thought to be caused by incomplete combustion in such a type of flame. When the volatile organic sulfur compound is present in dilute, solution in isopropanol, combustion leads to nearly theoretical recovery of sulfur (Table IVf. The amounts of sulfur theoretically present are based on the purities shown earlier. The data obtained for carbon disulfide are rather erratic, and the recovery of sulfur is low as compared to the results for thiophene and thiophenol. This may be due to high volatility of carbon disulfide. CONCLUSIONS
The lamp method gives very low and erratic recovery of sulfur from the combustion of organic solutions of elemental sulfur The necessity of rinsing the chimney of the apparatus used in the lamp method after each determination and combining these rinsings with the absorption liquid as specified in the A S.T M. method ( 3 ) has been verified The lamp method described by Zahn (11) can be improved by the use of Pyrex wool wicks in place of the usual cotton ones.
Table IV. Combustion of Dilute Solutions of Thiophenol, Thiophene, and Carbon Disulfide in Isopropanol Expt. NO.
157 158 159 160 161 162 163 164 165 166 167 168 169 170 172 173 174 175 176 235 236
177 178 179 180 240
217 218 219 220 229 230 23 1 232
Theoretical Sulfur Present
Sulfur Found
Sulfur Recovery
%
.Mg.
Mg.
%
0,0379 0.0379 0.0379 0.0379 0.0379 0.0379 0,0379 0.0379 0,0379 0.0379 0,0379 0,0379 0.0335 0.0335 0.0335 0,0338 0.0335 0.0335 0,0335 0.0285 0,0285
1.12 1.37 1.48 1.63 0.67 0.66 0.67 0.70 0.51 0.48 0.46 0.48 0.40 0.36 0.39 0.60 0.56 0.54 0.59 0.64 0.73
1.06 1.30 1.46 1.74 0.69 0.67 0.67 0.72 0.51 0.50 0.47 0.48 0.42 0.37 0.39
95 99 99 107 103 101 100 103 100 104 102 100 105 103 100 98 98 96 97 99 100
Thiophenol Concn.
0.60
0.56 0.54 0.59 0.64 0.73 AV. Standard deviation of sulfur recovery = 3.1 Thiophene Concn., % 0.95 0.0357 0.95 0.79 0,0357 0.77 0.97 0.0357 0.96 1.23 0.0357 1.20 0.69 0.0229 0.70 .4v . Standard deviation of sulfur recovery = 1.7 Carbon Disulfide ’ Concn., % 0.44 0.0100 0.41 0.44 0.0100 0.47 0.36 0.0100 0.40 0.36 0.0100 0.41 0.53 0.0100 0.58 0.52 0.0100 0.57 0.54 0.0100 0.58 0.46 0.48 0.0100 AV. Standard deviation of sulfur‘recovery = 5.5
lo0
100 103 101 103 99 101
107 94 90 88 91 91 93 96 94
Complete recovery of sulfur from the combustion of solutions of volatile organic sulfur compounds has been effected only under highly specialized conditions-i.e., only when they are dissolved in isopropanol so as t o give a nearly nonluminous flame. Otherwise, solutions high in aromatic content burn with a luminous flame and sulfur recovery is about 90% of the theoretical. ACKNOWLEDGMENT
Grateful acknowledgment is made to H. E. Morris for his advice and guidance during the course of this investigation. LITERATURE CITED
Table 111. Combustion of Dilute Solutions of Thiophenol and Thiophene in Mixtures of Toluene and Isopropanol XO.
Thiophenol Concn. in Mixture
Isopropanol in Mixture
Theoretical Sulfur Present
%
%
185 187 188 197 198 200
0.0251 0.0218 0,0240 0.0300 0.0236 0.0251
67 58 64 80 63 67
Mg. 0.74 0.90 0.99 0.69 0.47 0.57
Expt.
Found
Sulfur Recovery
7%
Me.
0.64 0.83 0.88 0.64 0.41 0.48 A \, Standard deviation of sulfur recovery = 4.7
.
86 92 89 93 88 85 89
Thiophene Concn. in Mixture
% 189 190 191 192 237 238
0,0229 0.0196 0.0143 0.0278 0.0121 0.0112
64
0.69
40 78 53 49
0.48 1.02 0.62 0.57
0.64 0.57 0.44 0.99 0.58 0.53 AY. Standard deviation of sulfur recovery = 1.7 55
0.60
93 95 92 97 94 93 94
(1) Altieri, V. J., “Gas Chemist’s Book of Standards for Light Oils and Light Oil Products,” 1st ed., pp. 119-24, New York, American Gas Association, 1943. (2) h i . SOC.Testing Materials,” Standard Method of Test for Sulfur in Petroleum Oils by Bomb Method,” D129-44. (3) Am. Soc. Testing Materials, ”Tentative Method of Test for Sulfur in Petroleum Oils by Lamp Method,” D90-41T. (4) Ball, J. S., U.S.Bur. Mines, Rept. Invest. 3591 (1941). (5) Carnegie Steel Co., “Methods of Chemists of United State8 Steel Corporation for Sampling and Analysis of Coal, Coke, and By-Products,” 3rd ed., pp. 222-36, Pittsburgh, 1929. (6) Faragher, W.F., Morrell, J. C., and Monroe, G. S., Ind. Eng. Chern., 19, 1281-4 (1927). (7) Sheen, R. T . , Kahler, H. L., and Ross, E . M., IND.ENG.CHEM., A N A L . ED., 7, 262-5 (1935). (8) Universal Oil Products Co., “Laboratory Test Method for Elementary Sulfur in Gasoline,” H-19-40. (9) Universal Oil Products Co.. “Laboratory Test Method for Sulfur and Sulfur Derivatives of Hydrocarbons in Petroleum Distillates,” A-1 19-40. (10) Wirth, C., 111,and Strong, J. R., IND. ENG.CHEW,ANAL.ED., 8, 344-6 (1936). (11) Zahn, V., Ibid., 9, 543-7 (1937). RECEIVEDAugust 16, 1947. Presented a t t h e Texaa Regional Meeting, AMERICAS CHEMICAL SOCIETY, Dallas, Tex., December 12 and 13, 1946.