Influence of Sulfur Compounds on Octane Number and Lead Susceptibility of Gasolines J
JULIAN G . RYAK Shell Oil Company, Incorporated, Wood River, Ill.
Numerous investigators (2, 10, 14, 17, 19, 21, 26) have called attention to possible economies obtainable through application of various desulfurization processes. Previous efforts ( 3 , 6) to correlate simultaneously lead susceptibility, sulfur content, and hydrocarbon composition have not taken into account differences among sulfur types. Evaluation of the improvement resulting from desulfurization would be expedited if the quantitative effect of various types of sulfur compounds and the influence of hydrocarbon composition were known.
Antiknock characteristics of gasolines are impaired by the presence of sulfur compounds to an extent dependent upon hydrocarbon structure and sulfur type. Susceptibility to tetraethyllead as affected by structure of sulfur compounds has been investigated quantitatively. A relation which is independent of hydrocarbon composition has been established between concentration of various sulfur types and lead susceptibility. This mathematical expression permits accurate evaluation from group-sulfur analyses of the improvement obtainable by partial or complete desulfurization, the effect of gasoline sweetening methods, and the effect of blending stocks of similar hydrocarbon composition but different sulfur contents. The effect of sulfur compounds upon octane number in the absence of tetraethyllead varies with size and structure of the hydrocarbon portion of the molecule as well as the sulfur type.
Method of Study The procedure involved in this study comprised three steps: addition of purified sulfur compounds to a reformed gasoline of very lorn sulfur content; evaluation of the quantitative effect upon octane number (A. S. T. M. Motor Method) and lead susceptibility; and application of the results to gasolines of different hydrocarbon composition. PURIFICATION OF SULFKR COMPOUNDS. The sulfur compounds employed were, u ith the exceptions noted, obtained commercially and purified by fractionation in a column of ten theoretical plates at a reflux ratio of 10 to 1. Three polysulfides, n-butyl trisulfide, tert-butyl trisulfide, and tert-butyl tetrasulfide, were prepared in these laboratories according to the method of Klason ( 2 1 ) and purified by distillation under high vacuum. GROUPSULFURANALYSES. Determinations of the various sulfur groups in gasoline are based on the analytical scheme of Faragher, Morrell, and Monroe (5) and the test methods of the Universal Oil Products Comoanv (85).as modified and imaroved by the Shell Devclopment company '($3). Total Sulfur. A. S. T. h1. Method D90-41T is used only for sulfur contents above 0.05 per cent by weight and is modified by application of a nitric acid correction. Reproducibility in total sulfur content by this method is approximately =t0.004 per cent by weight. Sulfur contents below 0.05 per cent by weight are determined by a special method which is a combination of the A. S. T. M. procedure for burning the sample and absorhing the products of combustion and a turbidimetric method for determining barium sulfate ($2). Reproducibility in total sulfur content is about *0.0004 per cent by weight. Mercaptan, Sulfur. The potentiometric titration method devised by Tamele and Ryland (24) and described by U. Q. P. as Method H163-40 (65) is employed. Reproducibility is about *0.0002 per cent by weight. DisulfLcle SuJfzLr. A method similar in principle to U. 0. P. Method A19-40 (26) is used, the difference being determination (by potentiometric titration) of the mercaptan sulfur obtained on reduction. Special precautions are taken to prevent loss of mercaDtans. Reoroducibilitv is of the same order as in the mercaptan sulfur determination.Thioether Sulfur and Ring Sulfur. U. 0. P. Method A119-40 with minor modifications is used. Reproducibility is dependent on the total sulfur method. Polysulfide Sulfur. No quantitative method is available. A qualitative test for polysulfides formed during doctor treatment is described in the literature (12).
ASOLINES of high antiknock quality involve in their manufacture the use of large quantities of tetraethyllead by the petroleum industry. One of the factors requiring serious consideration is the influence of sulfur compounds. Conclusive evidence has been compiled during the past decade that sulfur compounds decrease antiknock rating and response to tetraethyllead. Hebl and Rendel (8),in studies with the Ricardo variable-compression engine, pointed out in 1932 that susceptibility to tetraethyllead varies with the crude source, volatility, and manufacturing process, and that certain sulfur types may be detrimental. Schulze and Buell (20)and Birch and Stansfield ( 1 ) added sulfur compounds to straight-run gasoline and to blends of isooctane with n-heptane, respectively, and demonstrated that the deleterious effect varies with the class of sulfur compounds and that this effect per unit increment of sulfur may be greater, the loa er the sulfur content. Indications also were obtained that structural differences among individual members of a class may have some influence.
G
824
INDUSTRIAL A N D ENGINEERING CHEMISTRY
July, 1942
825
OF SLOPETO LEADSUSCEPTIBILITY TABLE I. CONVERSION
Slope Susceptibility 0.00 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.60 1.70 1.80 1.90 2.00 2.10 2.20 2.30 2.40 2.50 2.60
0.03 0.05 0.08 0.10 0.12 0.15 0.17 0.20 0.22 0.25 0.27 0.30 0.33 0.36 0.39 0.42 0.45 0.48 0.51 0.54 0.57 0.61 0.64 0.68 0.71 0.74
Slope Susceptibility
Slope Susoeptibility
2.70 2.80 2.90 3.00 3.10 3.20 3.30 3.40 3.50 3.60 3.70 3.80 3.90 4.00 4.10 4.20 4.30 4.40 4.50 4.60 4.70 4.80 4.90 5.00 5.10 5.20 5.30
5.40 5.50 5.60 5.70 5.80 5.90 6.00 6.10 6.20 6.30 6.40 6.50 6.60 6.70 6.80 6.90 7.00 7.10 7.20 7.30 7.40 7.50 7.60 7.70 7.80 7.90 8.00
0.78 0.81 0.85 0.88 0.92 0.96 0.99 1.03 1.07 1.11 1.15 1.19 1.23 1.27 1.31 1.35 1.39 1.43 1.47 1.52 1.56 1.60 1.65 1.69 1.74 1.79 1.83
1.87 1.92 1.97 2.02 2.07 2.11 2.16 2.21 2.26 2.31 2.36 2.41 2.46 2.52 2.57 2.62 2.67 2.72 2.78 2.83 2.89 2.94 3.00 3.05 3.11 3.17 3.23
40
35
REPRODUCIBILITY OF LEAD SUSCEPTIBILITY VALUES. Lead Improved methods designed to eliminate precision errors in translatin engine results into lead susceptibility values, according to the fefinition and chart of Hebl, Rendel, and Garton (9),were devised. These procedures comprise an analytical method for highest precision and a graphical method for routine work. In principle the analytical method consists of the evaluation, from octane number ws. lead concentration data, of coordinates which when plotted on a rectangular system determine the same line that the original data determined on the chart. These coordinates are obtained from a graphical relation between ordinate ( y value) and octane number, and from a table relating abscissa (x value) and concentration of tetraethyllead a t any given ordinate. (A linear ordinate scale ranging from 0 to 100 covers the range of 0 to 100 octane number, and a linear abscissa scale ranging from 0 to 6 covers the range of 0 to 6 cc. tetraethyllead per gallon.) With the coordinates of the points thus precisely determined, the slope of the best straight line is evaluated by the method of least squares. Finally, from the slope the lead susceptibility value can be determined from the relation between these two quantities, which has been calculated upon the basis of the present definition of lead susceptibility (9). In the graphical method the lead susceptibility chart is constructed exactly according to the instructions of Hebl, Rendel, and Garton (9),and a vertical scale of y values ranging from 0 to 100 is accurately drawn along the 0-cc. and 6.0-cc. lines with the result shown in Figure 1. To use this chart, the data relating octane number and concentration of tetraethyllead (cc. per gallon) are plotted; the best straight line is passed through the points and extended to the 0-cc. and 6.0-cc. lines. The slope is then calculated as the difference between the y values (at 6.0 and 0 cc. tetraethyllead per gallon) divided by 6.0 (the difference in x values of the two points). The lead susceptibility value corresponding to this slope is obtained from the relation between these two quantities, given in Table I. For cracked gasoline as an example:
3
Susceptibility Chart.
y value at 6.0 cc. tetraeth llead/ al. y value a t 0 cc. tetraethylTead/g$.
Difference Slope = 16.2/6.0 = 2.70 Lead susceptibility value (Table I)
=
59,. 1
42.9 16.2 0.78
Humidity Control. Throughout this study the moisture content of the intake air was maintained a t 80 k 5 grains of water per pound of dry air (72 * 4 per cent absolute humidity a t 70" F,). With controlled humidity the average deviation in lead susceptibility was 0.04 unit. Engine Test Method. The results of this investigation apply only to the A. 8. T. M. Motor Method. For highest precision all possible octane numbers have been determined by direct comparison at each concentration of tetraethyllead. Properties of Base Stock. Because of its low sulfur content a Rodessa reformed gasoline containing 0.0033 per cent sulfur by weight was selected as a base stock. The properties are given in Table 11.
2 2 I I
I l i l i l l
o T'
.5
I
1.0
1
1
1
2.0 I I
TETRAETHYL
I
,
30 I
I
4.0 I
I 5-0
P L-
LEAD, CC./GAL.
FIGURE 1. LEADSUSCEPTIBILITY CHART
TABLE 11. PROPERTIES OF RODESSA REFORMED GASOLINE Sulfur content, wt.% Total Mercaptan Disulfide Thioether Ring Bromine No. (18) Sp. dispersion at 25' C. Aromatic content, wt.% (7) Induction period, min. Sp. g r . at 25'/4O C.
s
0.0033 0.0000 0,0003 0.0009 0.0021 21
128y6 33 320 0.784
Rqid vapor pressure, lb./sq.
4.0 185 225 270 333 392 71.0 79.1
Effect of Sulfur Compounds on Octane Number The influence of various types of sulfur compounds upon octane number in the absence of tetraethyllead was investigated by adding purified sulfur compounds t o the Rodessa reformed gasoline and obtaining octane number comparisons (A. S. T. M . Motor Method) with the base stock. Thioethers and ring-sulfur compounds at a concentration of 0.10 weight per cent of sulfur or lower were found to have no measurable effect upon octane number. On the other hand, mercaptans, disulfides, and polysulfides exhibited a definite and increasingly severe adverse effect, in the order named upon antiknock properties. MERCAPTANS AND DISULFIDES. The influence of various mercaptans and disulfides upon octane number is shown in Figures 2 and 3. Decrease in octane number (from an original of 71.0) resulting from the addition of mercaptans and disulfides at a concentration of 0.05 per cent sulfur by weight is expressed as a function of molecular size of the sulfur compound (i. e., number of carbon atoms per molecule).
INDUSTRIAL AND ENGINEERING CHEMISTRY
826
Comparison of Figures 2 and 3 reveals that disulfides have a slightly more detrimental effect upon octane number than the corresponding mercaptans. Thus i t is evident that in the doctor treatment of gasoline a slight loss in octane number cannot be avoided, even under ideal conditions in which mercaptans are converted quantitatively to disulfides.
CARBON ATOMS/
MOLECULE
Vol. 34, No. 7
0.002 weight per cent of sulfur as tert-butyl tetrasulfide. Approximately six times this amount of sulfur as tert-butyl trisulfide is required t o produce an equal effect. The importance of avoiding polysulfide formation, particularly tetrasulfide formation, during doctor treatment is clearly evident.
CARBON ATOMS/
MOLECULE
4. EFFECTOF POLYSULFIGURE2. EFFECTOF MERCAPTANS OF FIGURE 3. EFFECT OF DISULFIDES OF VARYINGFIQURE SIZE ON OCTANENUMBEROF FIDES ON OCTANE NUMBER OF LowVARYINGMOLECULAR SIZE ON OCTANE MOLECULAR RODESSA REFORMED GASOLINE SULFURREFORMED GASOLINE NUMBER OF RODESSA REFORMED GASOLINE
Among the normal or straight-chain mercaptans and disulfides shown in Figures 2 and 3, the detrimental effect upon octane number increases uniformly with increase in molecular weight or boiling point. Branched-chain isomers, however, exhibit a somewhat greater effect than do the corresponding straight-chain compounds. Thus, i t appears that in order to predict accurately the improvement in octane number to be gained by removal of mercaptans as compared with their conversion to disulfides, a knowledge of not only the concentration of mercaptans present but also of their molecular structure and distribution in the gasoline would be necessary. From the foregoing discussion, the relation between molecular structure of mercaptans and disulfides and the tendency to cause knock may be expressed by the statement that in a homologous series the tendency to knock increases with increase in the length of the carbon chain or boiling point, and in an isomeric series the tendency to knock increases with increase in the number of side chains. It is interesting to note that this behavior is in qualitative agreement with the relative thermal stabilities of the individual mercaptans and disulfides. (Mercaptans and disulfides are reported to decrease in thermal stability with increase in chain length, and branched-chain mercaptans are indicated t o be less stable than are their straight-chain isomers, 4, 13.) This suggests the possibility that the effect of various mercaptans and disulfides upon octane number is determined by the relative rates of decomposition of these sulfur compounds in the engine. POLYSULFIDES. The effect of the polysulfides, tert-butyl trisulfide and tert-butyl tetrasulfide, upon the octane number (originally 71.0) of the Rodessa reformed gasoline is shown in Figure 4. The octane numbers are expressed as differences from the result for the base stock. A decrease in octane number of 0.5 unit is obtained with the addition of only
As in the case of mercaptans and disulfides, the relative effect of the two polysulfides upon octane number is in the same order as their relative thermal stabilities. Because the polysulfides readily liberate sulfur upon being heated, the effect of 0.005 weight per cent elementary sulfur upon octane number was determined. A decrease in octane number of only 0.2 unit was obtained. However, the possibility that the polysulfides are not active as such, but rather through a product of their decomposition, is not necessarily disproved. Substances formed in situ may be in a much higher state of activity and, hence, much more effective in promoting detonation. Their behavior may be analogous, although opposite in effect, to that of the suppression of detonation by tetraethyllead. The mechanism for the action of tetraethyllead is known to involve decomposition to elementary lead, and yet the high degree of activity of tetraethyllead as an antiknock agent is in direct contrast with the inactivity of colloidal lead sols added to gasoline (16,16').
Effect of Sulfur Compounds on Lead Susceptibility The effect of various types of sulfur compounds upon lead susceptibility was determined by adding purified sulfur compounds and obtaining octane number comparisons with the base stock a t each concentration of tetraethyllead. Mercaptans, disulfides, polysulfides, thioethers, ring-sulfur compounds, and hydrogen sulfide were investigated. A complete tabulation of results is given in Table 111. Lead susceptibility (as defined by Hebl, Rendel, and Garton) is plotted against concentration of added sulfur compounds (expressed in per cent by weight of sulfur) in Figure 5. Figure 5 and Table I11 show that the detrimental effect of sulfur compounds upon lead susceptibility increases in the order: ring-sulfur compounds, thioethers, mercaptans and
July, 1942
INDUSTRIAL AND ENQINEERING CHEMISTRY
827
disulfides, and polysulfides. No significant difference is observed between the effect of the TABLE 111. EFFECT OF SULFUR COMPOUNDS ON OCTANE NUMBER AND LEADSUSCEPTIBILITY OF LOW-SULFUR REFORMED GASOLINE various mercaptans and disulfides; likewise the polysulfides tested exhibited the same effect Total S Lead Added S Content A. S. T. M. Octane No. Susceptiupon lead susceptibility. The class to which Compound % b y W i . 0 c c . a 0 . 5 ~ 0 . 1.0~0. 2 . 0 ~ 0 . 3 . 0 0 0 . bility(8) a sulfur compound belongs determines its inNone 0.0033 71.0 75.8 78.9 81.9 83.5 1.29 Thiophene 0.023 $0.1 -0.1 -0.4 -0.6 -0.8 1.20 fluence upon lead susceptibility. I n all cases Thiophene 0.063 0.0 -0.4 -0.7 '1.0 -1.2 1.13 the effect per unit increment of sulfur is greater, Thiophene 0.103 0.0 -0.7 -0.9 -1.0 -1.6 1.08 Thiophene 0.292 -0.5 -1.9 -3.5 -3.8 -3.8 0.88 the lower the sulfur content. I n contrast with n-Propyl sulfide the effect of these same sulfur compounds upon 0.023 +O.l -0.3 -1.3 -1.6 -1.1 1.11 %-Propylsulfide 0.053 '0.2 -0.5 -2.1 -2.2 -1.5 1.02 octane number in the absence of tetraethyllead, n-Prop 1 sulfide 0.103 -0.1 -0.6 -2.4 -2.2 -2.1 0.97 n-Butyf sulfide 0.310 -1.9 -4.1 -5.5 -5.8 -6.3 0.68 i t is evident that structural differences within a given series have little or no influence upon n-Butyl mercaptan 0.053 -0.2 -1.3 -2.6 -3.5 -3.3 0.82 n-Prop 1 mercaptan 0.053 -0.2 -2.5 -1.1 -3.2 -3.7 0.82 lead susceptibility. n-AmyYmercsptan 0.053 -0.4 -8.3' -3.9 -2.3 -0.9 0.80 Isobutyl mercaptan 0.053 -0.3 --1:4 -3.4 -4.0 -4.0 0.76 Attention is directed t o the fact that the tmt-Butyl mercaptan 0.053 -0.4 -1.0 -3.4 -2.8 -3.3 0.83 Hydrogen sulfide effect of hydrogen sulfide upon lead suscepti0.053 -0.2 -1.2 -2.8 -3.6 -3.5 0.82 bility is identical with that of the mercaptans n-Propyl disulfide 0.0075 O'.O 0.0 -0.5 -0.6 -0.6 1.18 0.0105 +o. 1 .. -0.5 -1.3 -0.8 and disulfides. Since hydrogen sulfide and 1.14 0.023 -0.3 -2.0 -2.1 -0.8 -2.4 1.03 gaseous olefins are the primary thermal de0.063 -0.4 -2.9 -3.7 -0.9 -3.7 0.84 -4.1 0.053 -2.5 -3.8 -1.0 -3.9 0.85 composition products of the mercaptans and -2.4 0.063 -1.0 -0.5 -3.0 -3.4 0.84 0.053 -1.4 -3.1 -3.9 -0.7 -3.1 disulfides, it is possible that formation of hydro0.86 -4.1 -4.8 0.103 -1.9 -1.5 -4.7 0.74 gen sulfide in the engine is responsible for this 0.278 -2.6 -6.4 -4.9 -8.0 -8.6 0.49 similarity in effect upon lead susceptibility. n-Butyl trisulfide 0.0667 0.0 -0.1 -0.4 -1.1 -0.5 1.15 I n so far as the other classes of sulfur com%-Butyltrisulfide 0.0166 -0.4 -0.7 -1.8 -2.4 -2.1 1.00 tert-Butyl trisulfide 0.013 -0.4 -0.8 -2.0 -2.7 -1.8 1.02 pounds are concerned, it may be significant tert-Butyl tetrasulfide 0.005 -0.8 -0.4 -1.1 -0.9 -0.9 1.21 tert-Butyl tetrasulfide 0.008 -1.0 -1.5 -2.1 -1.7 -1.9 1.11 that the ease of pyrolysis of the various sulfur tertButy1 tetrasulfide 0.013 -1.8 -2.3 -2.7 -2.8 -3.4 0.99 types is in exactly the same order as their tert-Butyl tetrasulfide 0.033 -2.4 .. -4.6 -5.8 -5.5 0.78 Elemental sulfur 0.008 -0.2 .. . . -0.5 -0.3 1.27 relative effect upon lead susceptibility, the a Tetraethyllead per ' gallon. thermal stability decreasing in the order: ringsulfur compounds, thioethers, mercaptans, and I disulfides. There is obviously a difference in mechanism from that involved in the effect on obtained on a sulfur-free basis; the constants in the logarithoctane number since the behavior of different polysulfides, mic equations would then indicate disulfides t o be approxidisulfides, and mercaptans with respect t o octane number mately 2.3 times as effective as thioethers and 5.5 times as varies within the groups, with possible overlapping between one group and the next; however, the effect upon lead effective as ring sulfur. To test this possibility, a careful extrapolation of lead susceptibility t o zero sulfur content is susceptibility falls distinctly into groups, with very little if any variation within these groups. required. EXTRAPOLATION OF LEADSUSCEPTIBILITY TO ZEROSULFUR This is done b y means of successive approximations; the CONTENT.When concentration of added sulfur compounds value 1.29 shown in Table 111is taken as the lead susceptiis plotted directly against per cent decrease in lead susoeptibility of the sulfur-free gasoline as a first approximation. bility of the Rodessa reformed gasoline, a family of concaveThis value is the average of thirty-three determinations made downward curves passing through the origin is obtained. in the course of this investigation and is therefore well Since the same data plot almost linearly on a logarithmic established. The 0.0033 weight per cent sulfur in the Rodessa scale, it appears that an exactly linear relation would be gasoline consisted of 0.0003 per cent disulfide, 0.0009 thioether, and 0.0021 ring sulfur, and corresponds to a total reduction of 8.5 per cent in lead susceptibility according to the approximate curves. From this result a sulfur-free lead susceptibility value of 1.29
(a) 100-8.5 or 1.41 is obtained.
This value is used as the basis for a second approximation, The effect of the residual sulfur calculated from the new set of curves is found t o be 9 per cent, which is in substantial agreement with the result based on the first approximation. The value 1.41for the lead susceptibility of the sulfur-free stock is therefore used for subsequent calculations in this work.
RELATION BETWEEN SULFUR CONTENTAND DECREASE IN LEADSUSCEPTIBILITY. From the lead susceptibility values given in Table I11 for the various sulfur compounds added to
FIGURE 5. EFFECT OF VARIOUS SULFUR COMPOUNDS ON LEAD SUSCEPTIBILITY OF LOW-SULFUR REFORMED GASOLINE
the reformed gasoline, the per cent decrease in lead susceptibility from the sulfur-free value (1.41)is computed, with the results shown in Table IV. The relation between per cent decrease in lead susceptibility and concentration of the five sulfur types (polysulfides, mercaptans and disulfides, thioethers, and ring-sulfur compounds) is plotted on a rectangular scale in Figure 6 and on a logarithmic scale in Figure 7. A linear logarithmic relation is obtained. Since on rectangular coordinates the curves intersect only a t the origin, identical
Vol. 34, No. 7
INDUSTRIAL AND ENGINEERING CHEMISTRY z70
/
zm
/O /
X 0
9
AND DISULFIDE
THIOETHERS RING- SULFUR COMPOUNDS
0 SULFUR,
WT. %
FIGURE 6. RELATIOK BETWEEN COKCENTRATIOX OF VARIOUSTYPESOF SULFURCOMPOUNDS A N D PER CENTDECREASE IK LEAD SUSCEPTIBILITY
OF RODESSAREFORMED GASOLINE
slopes are to be expected on a logarithmic scale, and this is seen to be the case. Figure 6 also indicates the agreement between the experimental data and curves calculated according to the logarithmic relation. The general equation of the curves shown in Figures 6 and 7 is of the type:
+
log y = log kl ICs log c (1) where y = yo decrease in lead susceptibility C = by weight of a given sulfur type, multiplied by 1000 for convenience kl, kz = line constants
FIGURE 7 . RELATION BETWEEN LOGARITHM OF CONCEKTRATIOS OF VARIOUSTYPES OF SULFUR Conwouxns AUD LOGARITHM OF PERCENT DECREASE IS LEAD SUSCEPTIBILITY OF RODESSA REFORMED GASOLINE
The coordinates of the point$ given in Table IV and shown in Figure 7 are employed to calculate the slope and intercept (or line constants k 2 and k l ) in Equation 1 by the method of least squares. The ICl values-i. e., the y values at C = 1 (0.001 per cent sulfur by weight)-are found to be 12.0, 8.9, 6.6, and 4.8 for the polysulfides, mercaptans and disulfides, thioethers, and ring-sulfur compounds, respectively. An average slope (kz value) of 0.36 is obtained. The equations thus become:
+
(polysulfides) log yl = log 12.0 0.36 log C1 (mercaptans and disulfides)log YZ = log 8.9 0.36 log Cz (thioethers) log y~ = log 6.6 0.36 log CS (ring sulfur) log y4 = log 4.8 0.36 log Cq
+ +
TABLE IV. RELATION BETWEEX CONCESTRATIOU OF VARIOUS TYPESO F SULFUR COZlPOUNDs I N REFORMED GASOLINEAKD DECREASE I N LEADSUSCEPTIBILITY FROM THE VALUE OF THE SULFUR-FREE STOCK Added Sulfur Compound None Thiophene Thophene Thiophene Thiophene
Concn.. Wt.70 of S 0 000 0.026 0.056 0.108 0.296
n-Propyl d f i d e n-Propyl sulfide n-Propyl sulfide n-Butyl sulfide
0.022 0.052 0.102 0.309
Mercaptansa and HIS
0.051
n-Propyl disulfide n-Amyl disulfide n-Propvl disulfide Disulfidesb n-Propyl disulfide n-Butyl disulfide
0.0053 0.0083 0.021
n-Butyl trisulfide n-Butyl trisulfide tert-Butyl trisulfide tert-Butyl tetrasulfide tert-Butyl tetrasulfide tert-Butyl tetrasulfide tert-Butyl tetrasulfide
7Lead Value 1.41 1.20 1.13 1.08 0.88 1.11 1.02 0.97
Susceptibility-Decrease 70 Decrease 0 0 0.21 15 0.28 20 0.33 23.5 0.52 38
0.68
0.30 0.39 0.44 0.73
21 28 32 52
0.81
0.60
42
0.23 0.27 0.38 0 56 0.67 0.92
27
0.101 0.276
1.18 1.14 1.03 0 85 0.74 0.49
0.003 0.013 0.010
1.02
1.15 1.00
0.26 0.41 0.39 0.20 0.30 0.42 0.63
18 29 28 14 21 30 45
0 051
0,002
0.005 0.010 0.030
1.21
1.11
0.99 0.78
16
19
(3) (4) (5)
The ratio of the concentrations a t which exactly the same per cent decrease in lead susceptibility is obtained with each of the sulfur types can be calculated from the above equations. For convenience the ratios will be based on mercaptan and disulfide sulfur (which will be referred to simply as disulfide sulfur). If log y1 = log y2 = log y3 = log y4, and C2 = I, the 1
CI= 2.29 Cs
2.30;
CI
= 5.56
t7
7
40 47.5 65
Average results for n-butyl, n-propyl, n-amyl, isobutyl, and tert-butyl mercaptans. b Average results for n-propyl, tert-butyl, n-amyl, and isoamyl disulfides. Q
(2)
+
'
.050
~ ~ G ~ L E N ~ . % S U L F ~ DSULFUR E
.DO
.XI0
CONTENT,^^: To
FIGURE8. RELATIOKBETWEEN EQUIVALENT DISULFIDE SULFURCONTENTASD PER CEXT DECREASEIN LEAD SUSCEPTIBILITY FROM VALUE OF SULFUR-FREE STOCK
July, 1942
INDUSTRIAL AND ENGINEERING CHEMISTRY
Thus, the disulfides are 2.30 times as detrimental as the thioethers and 5.56 times as detrimental as the ring-sulfur compounds, but the polysulfides are 2.29 times as detrimental as the disulfides. That is, to obtain a given per cent decrease in lead susceptibility with disulfides, thioethers, ring-sulfur compounds, and polysulfides 5.56 times as much ring sulfur, 2.30 times as much thioether sulfur, and 1/2.29 times as much polysulfide sulfur is required as disulfide sulfur. These factors permit the conversion of the sulfur compounds to a common type, which simplifies the calculation of the per cent decrease in lead susceptibility in a mixture of the four types. For example, to convert all four types to disulfides :
TABLE V. EFFECT OF A MIXTUREOF SULFUR COMPOUNDS UPON LEAD SUSCEPTIBILITY OF REFORMED GASOLINECONTAINING ADDEDHYDROCARBONS Hydrocarbon Added None 10% diiao-
butylene
10% benzene 10% Ca-
polymer 0
equivalent wt.% of S as RSSR = (wt.% of S as polysulfides X 2.29) (wt.% of S as RSSR or RSH) wt.% of S as RSR wt.% of S as rings (6) 5.56 2.30
+
+
+
Expression 6 includes all sulfur types ordinarily encountered in gasoline. The weighting constants are approximately equal t o 2.3 raised to the 1, 0, -1, and -2 power, respectively. The simplified equation for the calculation of per cent decrease in lead susceptibility thus becomes: log
?J
+
0.949 0.36 log (polysulfides X 2.29
=
ring S + RSSR + RSR 2.30 + m) (7)
where the respective concentrations of polysulfide, disulfide, thioether, and ring sulfur are expressed as weight per cent X 1000. For example, if a gasoline of 0.130 per cent total sulfur content by weight should contain no mercaptans or polysulfides, 30 per cent of disulfide sulfur (0.039 per cent by weight), 25 per cent of thioether sulfur (0.032 per cent by weight), and 45 per cent of ring sulfur (0.059 per cent by weight), the per cent decrease in lead susceptibility would be: log y = 0.949
+ 0.36 log
39
GRAPHICAL METHODOF CALCULATION. The per cent decrease in lead susceptibility for a given type can also be obtained graphically, although with less precision than is obtainable by calculation. I n this method the sulfur types are converted to disulfides (or to any one of the other types) in the usual manner (Equation 6), and the per cent decrease in lead susceptibility is read from the disulfide curve (or other appropriate curve). For convenience, a curve from which the per cent decrease in lead susceptibility for a given “equivalent disulfide” sulfur content can be read directly is shown in Figure 8. Saom _ _ OF . THE METHOD.This method is limited to sulfur contents giving less than about an 80 per cent decrease in lead susceptibility from the value of the sulfur-free gasoline. This entails no limitation on the usefulness of the method in general, however, since few gasolines contain sufficient sulfur t o cause a decrease of this magnitude (80 per cent decrease corresponds to 0.45 per cent by weight of equivalent disulfide sulfur or 2.5 per cent by weight of ring sulfur). ~
COMPARISON OF EXPERIMENTAL WITH CALCULATED VAL~JES. To test the relation developed between “equivalent disulfide” sulfur content and per cent decrease in lead susceptibility, the effect of adding the mixture of sulfur types described in the above-mentioned example to low-sulfur reformed gasoline
Total S Content, Wt.% 0.003 0.133 0.003 0.133 0.003 0.133 0.003 0.133
A. S. T . M. Octane No. 0.5 1.0 2.0 cc. cc. cc. 71.0 75.8 78.9 81.9 -0.3 -2.3 -3.6 -3.7 74.8 80.0 81.8 84.1 -0.6 -2.5 -2.7 -3.6 72.8 77.2 80.3 82.7 -0.8 -2.2 -3.3 -4.0 73.0 77.9 80.0 83.3 -0.5 -2.1 -3.6 -3.2 0
C C . ~
3.0
cc.
Lead Susceptibility (9)
83.5 -4.3 85.8 -3.5 84.9 -3.5 84.3 -2.8
1.29 0.85 1.17 0.78 1.26 0 81 1.24 0.80
Tetraethyllead per gallon.
was determined (composition in weight per cent: 0.039 disulfide, 0.032 thioether, and 0.059 ring sulfur). Similar experiments were carried out in which the hydrocarbon composition was varied somewhat by the addition of 10 per cent by volume of an aromatic or unsaturated hydrocarbon mixture. The resulting octane number and lead susceptibility values are shown in Table V. The lead susceptibility values for the gasolines containing 0.003 per cent sulfur by weight (Table V) are extrapolated to zero sulfur content on the same basis as before (i. e., by correcting for a 9 per cent decrease in lead susceptibility due to the presence of 0.0011 per cent by weight of “equivalent disulfide” sulfur). The extrapolated lead susceptibility values and the calculated per cent decrease in lead susceptibility obtained are shown in Table VI.
TABLEVI. EFFECT OF SULFUR COMPOUNDS UPON PER CENT DECREASE IN LEADSUSCEPTIBILITY Lead Susceptibility Hydrocarbon Added None l o 7 diiaobutylene 1 0 8 benzene 10% Cs-polymer
0.000% S by wt. 1.41 1.26 1.38 1.36
0.133% S by wt. 0.85 0.78 0.81 0.80
~
Decrease in Lead Susceptibility Value Per cent 0.56 0.48 0.57 0.56
40 38 41 41
Average 40
32 59 + 2.30 4- m)
l o g y = 1.598 y = 39.6 per cent decrease
829
The close agreement between the experimentally determined results (average 40.0) in Table VI and the calculated value in the above example (39.6 per cent) for per cent decrease in lead susceptibility indicates that the relation is applicable to mixtures of sulfur compounds and is unaffected by small changes in hydrocarbon composition.
Influence of Hydrocarbon Composition and Octane Number Level
EFFECT OF ADDEDSULFUR COMPOUNDS ON LEADRESPONSE LOW-SULFUR GASOLINES.The effect upon octane number and lead response of various sulfur compounds added to a OF
low-sulfur (0.005 per cent by weight) straight-run gasoline of 51 octane number and to a blend of 65 per cent isooctane in n-heptane has been reported by Schulze and Buell (20) and Birch and Stansfield (1). The octane numbers and the measured and calculated lead susceptibility values are shown in Table V I I . (The 0.005 per cent by weight of residual sulfur is assumed to consist of half thioether and half ring sulfur, which corresponds to a reduction in lead susceptibility of 10 per cent.) Close agreement between the lead susceptibi1it.y values determined experimentally and those calculated from the sulfur content is observed for the straight-run gasoline in Table VII. Only in two instances is the difference greater than 0.10, both cases having a high sulfur content outside the
830
Vol. 34, No. 7
INDUSTRIAL AND ENGINEERING CHEMISTRY
that sulfuric acid treatment, in addition to decreasing the sulfur content and increasing lead susceptibility, increases the calculated 65 P E R CENT ISOOCTANE I N HEPTANE (1, 20) lead susceptibilitv a t zero sulfur content. This Concn. Octane No. 8 Compound Lead Susoeptibility can be attribuied to removal of diolefins Wt.% of's 0 cc.0 1 cc. 2 CC. 3 eo. 6 cc. Measured GaIcd. -.. Added, Wt.% and perhaps some of the more reactive olefins. Straight-Run Gasoline This interesting application of the method 0.005 51.0 63.5 70.7 74.4 79.5 1.93 None illustrates its usefulness in removing the 0.000 51.0 63.5 70.7 74.4 79.5 .. 2:i4 obscuring effect of sulfur in evaluating a refining 0.024 50.6 60.2 66.2 70.3 76.3 1.53 1.54 E t h y l disulfide 1.39 procedure. 0.046 50.3 59.6 64.9 68.5 74.5 1.35 0.113 50.3 57.8 61.6 64.2 70.7 1.00 1.09 Table X gives results taken from the same 0.412 48.4 52.4 55.6 58.2 64.5 0.48 0.80 source (6) and indicates the effect of the 0.066 50.6 58.7 1.28 1.20 r,-Butyl disulfide addition to the two more severely treated 50.3 59.6 0.030 1.48 1.49 Isobutyl disulfide 50.6 62.2 0.024 1.71 1.69 Methyl sulfide samples of a mixture of sulfur compounds of 0.238 1.24 51.0 58.6 1.13 Ivf ethyl sulfide 0.030 51.0 61.9 1.64 1.66 E t h y l sulfide the types and in the amounts present in the 0.031 50.6 60.4 1.55 Isoamyl sulfide 1.64 10.5 pound per barrel treatment. Sulfur com51.0 60.5 0.095 1.54 1.61 Thiophene pounds used included n-amyl disulfide, n65% Isooctane in n-Heptane propyl sulfide, and dimethyl thiophene. ... 64.0 65.0 76.0 * . 84.0 .. 2.3 None 0.10 . . 76.0 .. 1.2 E t h y l mercaptan 1:i The per cent decrease in lead susceptibility 63.0 70:O 0.10 1.5 .. 76.5 .. Isoamyl mercaptan 1.2 63.5 68.5 .. 74.5 .. 0.10 1.2 from the value of the sulfur-free stocks, as 1.0 Ethylene mercaptan 65.0 69.5 .. 75.0 0.10 .. 1.0 E t h y l disulfide 1.2 calculated from the results given in Tables I X 64.5 70.5 0.10 .. 76.0 .. 1.1 Isoamyl disulfide 1.2 65.0 72.0 .. 78.0 .. 0.10 1.3 E t h y l sulfide 1.5 and X, was found to be 46, 47, and 47 per .0.10 65.0 72.0 .. 78.0 .. Isoamyl sulfide 1.3 1.5 cent for the 10.5, 21.0, and 31.5 pound per Pentamethylene 0.10 65.5 71.5 .. 78.0 .. 1.2 sulfide 1.7 1.5 barrel treatments, respectively. The close 0.10 65.0 72.3 .. 81.0 .. 1.7 Thiophene agreement among these results indicates again Tetraethyllead per gallon; results for isooctane and n-heptane are 00. per Imperial that the effect -of sulfur compounds is-ingallon, which is equivalent t o 0.84 cc. per U. S. gallon. dependent of the changes in hydrocarbon composition involved. These results also offer confirmation of the theory that the effect of range ordinarily encountered. Inasmuch as only meager the various sulfur types upon lead susceptibility is indedata for the blend of 65 per cent isooctane in n-heptane are pendent of the structure of the sulfur types. available (two concentrations of tetraethyllead, octane numbers being reported to the nearest 0.5), a less precise correlation is observed in this case. However, the results serve to TABLE IX. EFFECTOF SULFURIC ACID TREATMENT AT 20" F. show that the relation between sulfur content and lead susUPON EQUIVALENT DISULFIDESULFURCONTENTAND CALCUceptibility may be entirely independent of hydrocarbon comLATED LEADSUSCEPTIBILITY AT ZERO SULFUR CONTENT position.
TABLEVII. EFFECTOF SULFURCOMPOUNDS UPON OCTANE NUMBER AND LEADRESPONSE OF STRAIGHT-RUN GASOLINE AND
(I
EFFECTOF SULFURIC ACID TREATMENT ON LEADSUSCEPTIBILITY OF CRACKED GASOLINES.Data reported in the literature by Graves (6) also indicate that the relation between lead susceptibility and concentration of various types of sulfur compounds is independent of hydrocarbon composition, if the change in lead susceptibility is expressed as the per cent decrease from the value of the sulfur-free stock. Table VI11 shows the effect, upon concentration of various sulfur types and upon lead susceptibility, of treating Dubbs cracked gasoline from mixed California residuum with 98 per cent sulfuric acid a t a low temperature (20' F.).
TABLE VIII. EFFECT OF SULFURIC ACIDTREATMENT AT 20' F. SULFUR CONTENT AND LEADSUSCEPTIBILITY OF CRACKED GASOLINE (6)
EPON
Acid Consumption, Lb./Bbl. Sulfur content, wt.% Total Disulfide Thioether Residual (rings)
10.5 0.433 0.013 0.025 0.395
21.0 0.20 0.004 0.006 0.190
31.5 0.084 0.001 0.004 0.079
Octane No. (A. 8. T.M.-C. F. R.) 0 00. T. E . L./gal. 0.5 00. T. E.L./gal. 1.0 ea. T. E . L./gal. 2.0 cc. T . E..&./gal. Lead susceptibility (calcd. from above data)
69.7 71.5 73.5 75.5
69.0 72.0 74.5 76.5
68.0 72.5 75.5 77.5
~
0.60
0.77
0.96
~~~
Table IX gives the equivalent concentrations of disulfide sulfur, the per cent decrease in lead susceptibility, and the calculated lead susceptibility values of the sulfur-free stocks computed from the data in Table VIII. These tables show
TABLE X. EFFECT OF ADDINGSULFUR COMPOUNDS TO SULFURICACID-TREATED GASOLINES TO DUPLICATE THE SULFUR CONTENT OF THE 10.5 POUND PER BARREL TREATMENT (6) Sulfur content, wt.% Total Disu 1fide Thioether Residual (rings) Octane No., A. 8. T. M.-C. F. R. 0 cc. T. E. L./gal. 0.5 cc. T. E. L./gal. 1.0 00. T. E. L./gal. 2.0 cc. T.E. L./gal. Lead susceptibility Measured Calcd. from 8-free values (Table IX)and S content
Acid Consumption, Lb./Bbl. 10.5 21.0 31.5 0.433 0.44 0.42 0.013 0.013 0.013 0.025 0.025 0.025 0.395 0.395 0.395
69.7 71.5 73.5 75.5
69.0 71.0
73.0 75.0
68.0 70.0 72.0 74.5
0.60
0.62
0.67
0.60
0.63
0.63
EFFECTOF OCTANENUMBERLEVELUPON RESULTS OF LEAD SUSCEPTIBILITY CALCULATIONS. The tests of the method for computing lead susceptibility values from sulfur analyses indicated satisfactory results in the octane number range 50-75. Further tests of the method in a higher octane number region were made by adding sulfur compounds to blends of Rodessa reformed gasoline with isooctane and with diisobutylene. Table XI shows the effect of rt-butyl disulfide upon the lead susceptibility of 50-50 blends of Rodessa reformed gasoline with isooctane and with diisobutylene. From group sulfur analyses and the relation between the efl'ects of various sulfur types upon lead susceptibility are
INDUSTRIAL AND ENGINEERING CHEMISTRY
July, 1942
obtained the equivalent disulfide sulfur content and the corresponding per cent decrease in lead susceptibility from the value of the sulfur-free stock. These data, as well as a comparison of the calculated and measured lead susceptibility values, are given in Table XII. The calculated and measured values are in good agreement for blends of Rodessa reformed gasoline with isooctane and diisobutylene (with and without added n-butyl disulfide). Therefore i t appears that the method of computing lead susceptibility from sulfur analyses will give satisfactory results a t octane numbers as high as 84.
831
a W
W
0
20
40
80
80
100
TABLEXI. EFFECTOF n-BuTYL DISULFIDE UPON LEADSUSCEPTIBILITY OF BLENDS S
Compound Added None n-Butyl disulfide
Total S Content, Wt.%
0
cc.0
Octane Number 0.5 1.0 2.0 00. cc. OC.
Lead Susoept5bility
3.0 co.
50% Rodessa Reformed f 50% Isoootane 0.0021 83.5 87.3 90.2 92.1 93.3
1.33
0.062
0.81
-0.9
-2.2
-4.0
-3.9
-3.9
50% Rodessa Reformed f 50% Diisobutylene None 0.0023 84.2 86.4 82.0 88.0 88.5 rt-Butyl disulfide 0.054 -0.6 -1.9 -1.8 -2.1 -2.0 a Tetraethyllead per gallon.
0.46 0.30 % HIQH-S
TABLEXII. COMPARISON OF CALCULATED AND MEASURED
LEADSUSCEPTIBILITY VALUES OF BLENDS CONTAINING W B ~ L DISULFIDE E uivalent %SSR Gasoline Wt.% oi S 60% reformed f 0.0006 0.060 50% isooctane 50% reformed f 0.0007 50% diisobu0.051 tylene
Lead Susceptibility Calod Calcd. Measured value value 1.43 1.33 1.43 O:k7 0.81 0.60 0.46 0.50 O:i2 0.30
% deorease &free) 7 39 8 37
Applications of the Method for Computing Improvement in Lead Susceptibility from Sulfur Analyses IMPROVEMENT OBTAINED BY SOLUTIZER TREATMENT. The following data illustrate the method of calculating the improvement to be gained in lead susceptibility by this process, taken from a study of solutizer sweetening ($6). The average octane numbers at various concentrations of tetraethyllead are shown in Table XI11 for seventeen cracked gasolines before (sour) and after solutizer sweetening. TABLE XIII. EFFECTO F SOLUTIZER SWEETENING UPON AVERAGE OCTANE NUMBER AND LEADSUSCEPTIBILITY Condition of Gasoline8
sour Solutizer-sweetened
Oatane Number 0 (10." 1.0 CC. 2.0 co. 3.0 co. Seventeen Craoked Gasolines 71.9 67.4 74.3 75.9 +0.35 +1.21 f1.46 +1.65
Eight Straight-Run Gasolines 65.9 70.3 Sour 57.9 73.2 Solutizer-sweetened fO.66 $1.71 f2.10 $2.28 0 Tetraethyllead per gallon.
Lead Susoeptibility 0.73 0.95 1.28 1.54
The average mercaptan sulfur content of these gasolines was 0.065 and the average total sulfur was 0.261 per cent by weight. The proportions of thioethers and ring-sulfur compounds was not determined; for illustration, the ratio of ring sulfur to thioether sulfur will be assumed to be 2 to 1. This gives a sulfur content of 0.065 per q n t mercaptan by
CRACKEO IN LOW-8
REFORMED,
FIGURE 9. EFFECTOF BLENDINQ STOCKSOF HIGHAND Low EQUIVALENT DISULFIDE SULFUR CONTENT ON CALCULATED P E R CENT DECREASE IN LEADSUSCEPTIBILITY (above) AND ON LEAD SUSCEPTIBILITY (below)
weight, 0.065 thioether, and 0.131 ring sulfur. The calculated sulfur-free lead susceptibility is 1.44. A calculated lead susceptibility value of 0.91 is obtained for the solutizertreated samples, as compared with the measured value of 0.95. Variation of the assumed ratio of ring sulfur to thioether sulfur does not sensibly alter this result. Thus, the assumption of a 1 t o 1 ratio gives a calculated lead susceptibility for solutizer treatment of 0.90 and a sulfur-free lead susceptibility of 1.48. A similar calculation can be made for the average results of eight straight-run gasolines examined ($6). The octane numbers and lead susceptibility values are given in Table XIII. The average mercaptan sulfur of the straight-run gasolines was 0.046 and the average total sulfur was 0.136 per cent by weight. A 2 to 1 ratio of ring sulfur to thioether sulfur will again be assumed. On this basis the calculated results indicate a 41 per cent decrease due to the presence of sulfur compounds, giving a sulfur-free lead susceptibility value of 2.17. Solutizer sweetening reduces the decrease to 28 per cent, giving a calculated lead susceptibility value of 1.56 as compared with the measured value of 1.54. INFLUENCE OF SULFURCONTENT OF COMPONENTS UPON LEAD SUSCEPTIBIL~TY OF BLENDS. The effect upon lead susceptibility of blending gasolines of high and low sulfur contents was investigated by adding doctor-treated highsulfur cracked gasoline to Rodessa low-sulfur reformed gasoline. The octane number and lead susceptibility values of the two components and of a 50-50 blend are given in Table XIV. The lead susceptibility value for the blend is 0.69, well below the arithmetic mean of the values for the two components-namely, 0.90. Table XV and Figure 9 show the calculated lead susceptibility values and other data for blends of the two stocks. I n the calculations, based on sulfur contents, i t was assumed
Vol. 34, No. 7
INDUSTRIAL AND ENGINEERING CHEMISTRY
832
Acknowledgment
that the lead susceptibility values of the two stocks, whirh were of similar hydrocarbon composition, combine linearly. The calculated value for the 50-50 blend (0.71) is in close agreement with the measured value.
The author wishes to express his thanks to George W. Waters and H. H. Zuidema for their generous assistance in the development of the relationships presented in this communication, and to K. R. Edlund and Lowrey Love for their supervision during the progress of the work. TABLEXIV. EFFECTUPON LEADSUSCEPTIBILITY OF BLENDING WITH LOW-SULFUR REFORMED HIGH-SULFUR CRACKED GASOLINE GASOLINE Literature Cited Octane Number
0 C C . ~ 0 . 5 cc. 75.8 Rodessareformed 71.0 High-sulfur cracked 6 6 . 2 6 8 . 4 69.3 80-50 blend 66.6 0 Tetraethyllead per gallon. Gasoline
1. O cc.
78.9
70.1 70.8
2 . 0 cc. 81.9 71.6 73.6
3 . 0 cc. 83.5 73.6 74.9
Lead Susceptibility 1.29 0.52 0.69
Birch, S.F., and Stansfield, R., IND. ENG.CHBM., 28, 668 (1936). Bottomley, H., Refiner Natural Gasoline M ~ T20, . , 526 (1941). Eastman, D., IND. EXG.CHEM.,33, 1555 (1941). Faragher, W. F., Morrell, J. C., and Comay, S., Ibid., 20, 527 (1928).
Faragher, W. F., Morrell, J. C., and Monroe, G. S., Ibid., 19, 1281 (1927).
TABLE XV. EFFECTUPON LEADSUSCEPTIBILITY OF BLENDING WITH RODESSA REFORMED HIGH-SULFUR CRACKED GASOLIKE GASOLINE Volume 9% in Lead Susceptibility Blend High-S Low6 cracked reformed 0 lo 2o 30 40
100 90 80 70
60 50
50 40
70 so 90
20 30 10 0
Equivalent RSSR,
Wt.% of S 0.0011 0,020 0.039 0.058 0.077 0.0955 0.114 0.133 0.152 0.171 0.190
Calcd.. Calcd. Measured % decrease S-frees value value 9 26.5 33.5 39 43 47 40 53 55.5 5s 60
1.41 1.39 1.38 1.37 1.36 1.38 1.34 1.33 1.32 1.31 1.30
1:02 0.92 0.84 0.77 0.71 0.67 0.63 0.59 0.56
1.29
.. 0:69
..
..
..
.,
o:i2 a It is assumed that the sulfur-freelead susceptibility values, 1.41 and 1.30 will combine linearly since these gasolines are of similar hydrocarbon corn:
100
position.
It is clear that the blending of stocks of high and low equivalent disulfide sulfur contents will result in a decided loss in lead susceptibility, since the addition of only a small amount of high-sulfur stock will sharply decrease the lead susceptibility of the low-sulfur stock; conversely, an excessive quantity of a low-sulfur stock is required to improve significantly the lead susceptibility of a high-sulfur stock. Application of this method for calculating lead susceptibility of blends of different sulfur-containing stocks is limited to gasolines having similar hydrocarbon composition. Lead susceptibility values of sulfur-free stocks of widely different hydrocarbon compositon-e. g., cracked and straight-run-do not combine linearly.
Graves, F. G., Ibid., 31, 850 (1939). Grosse, A. V., and Wackher, R. C., IND. EXG.CHEM.,ANAL.ED.. 11, 614 (1939).
Hebl, L. E., and Rendel, T. B., J . Inst. Petroleum Tech., 18, 187 (193%. \ - - - - I
Hcbl, L. E., Rendel, T . B., and Garton, F. L., IND. ENG.CHEM., 31, 862 (1939). Henderson, L. M . , Ross, W. B., and Ridgway, C. IS., I b i d . , 31, 27 (1939).
Klason, P.,' J . prakt. Chem., [21 15, 274 (1877). Lowry, C. D., Dryer, C. G., Wirth, C., and Sutherland, R. E., IKD. ENG.CHEM.,30, 1275 (1938). Malisoff, W.M., and Marks, E. M . , Ibid., 23, 1114 (1931). Mason, C. F., Bent, R. D., and McCulloueh, - J. H., Oil Gas J.. 40, No. 26, 114 (1941). O h , H. L., and Jebens, W. J., IND. ENG.CHEM.,21, 43 (1929). Olin, H. L., Read, C. D., and Goos, A. W., Ibid., 18, 1316 (1926). Ridgway, C. M., Oil Gas J., 36, No. 46, 83 (1938). Rosenmund, 2. angew. Chem., 27, 68 (1924). Schulre, W. A., and Alden, R. C., Oil Gas J . , 38, No. 27, 199 (1939); Refiner Natural Gasoline
Mf?., 18, 474
(1939).
Schulre, W.A . , and Buell, A. E., Oil Gas J., 34, No. 21, 22 (1935): Natl. Petroleum News, 27, No. 41, 25 (1935). Schulze, UT. A., and Buell, A. E., Refiner Xatural Gasoline M f r . , 16, 268 (1937). Sheen, R. T., Kahler, H. D., and Ross, E. M., IND.ENG. CHEM.,Ax.4~.ED., 7, 262 (1935). Shell Development Co., private communication. Tamele, M. W., and Ryland, L. B., IND.ENO.CHEM.,A x ~ L . ED., 8, 16 (1936). Universal Oil Products Co., Lab. Test Methods for Petroleum and Its Products, 2nd ed., pp. 8-37, H-41 (1940) Yabroff, D. L., and Nixon, A. C., Refiner Natural Gasoline Mfr., I
19, 73 (1940). PRESENTED before the Division of Petroleum Chemistry at the 103rd Meeting of the AXERICAN CHEMICAL SOCIETY, Memphis, Tenn.