Response of Aircraft Fuels to Tetraethvllead

stituents to tetraethyllead is therefore an integral part of the characteristics which determine the limits of performance. After considerable effort ...
1 downloads 0 Views 500KB Size
Response of Aircraft Fuels to Tetraethvllead

A. G. CATTANEO AND A. L. STANLY

J

Shell Development Company, Emeryville, Calif.

The Ethyl blending chart for evaluating lead response of gasolines in the A. S. T. M.C. F. R. engine has been extended into the region above 100 octane number. The extension is consistent with the original chart in that the line representing the response of a given fuel is continuous and of constant slope throughout its length. The extended chart permits a cpantitative measure of lead susceptibility throughout the range from 0 octane number to isooctane plus 3 cc. of tetraethyllead per gallon.

E

XCEPT for a small quantity of special-purpose fuel, all aviation gasolines used today contain some tetraethyllead. The susceptibility of the various gasoline constituents to tetraethyllead is therefore an integral part of the characteristics which determine the limits of performance. To this fact justice is only rarely done. After considerable effort on the part of all industries concerned, it has become possible t o measure the antiknock value of gasolines within a few tenths of an octane number, but the susceptibility of a gasoline to tetraethyllead is still expressed by such vague terms as poor, fair, or good. Thus the interpretation is left to the individual, and comparison of results is impossible. The reason for this situation lies in complexity of the response to lead. The rise in octane number on any one fuel is different for each cubic centimeter of tetraethyllead added. Also, the shape of the curve is different for fuels of the same octane number and different lead response as well as for fuels of different octane number but equal lead response. Figure 1, in which conventional lead response curves are plotted for a number of common fuels, shows these effects clearly. By far the most confusing element, however, is the nonuniformity of the octane scale. An 85-octane f u e l , showing an increase of 15 oct a n e numbers from the addition of 3 cc. of tetraethyllead, is co n s i d e r e d to have a good lead s usc e p t i b ili t y, but a 50-octane fuel, showing an equal increase in TETRAETHYLLEAD -CC. PER GALLON octane number. FIGURE 1. LEAD SU~~CEPTIBILITY CURVE is not. This is (A. s. T. M.-C. F. R. METHOD) because an in-

crease from 85 t o 100 octane number is more than twice as effective in terms of engine performance as an increase from 50 to 65 octane number. In 1933 Hebl, Rendel, and Garton (2) first showed an empirical way to deal with this situation. On the Ethyl blending chart which they proposed, the horizontal scale of cc. tetraethyllead per gallon and the vertical scale of octane number were so chosen that a lead response curve becomes a straight line. This was later revised (3) for A. S. T. M.C. F. R. motor method octane numbers, and also to take better account (by the use of slanting ordinates) of the effect of the antiknock value of the base stock upon the relation between lead concentration and increase in compression ratio. Figure 2 shows the data of Figure 1entered upon the revised chart. This chart has been in use for several years, and we have found i t reliable for lead concentrations up to 3 cc. per gallon. For higher concentrations of tetraethyllead the chart is useful but is not so well established, partly owing to lack of sufficient data and partly to the large influence which small variations in the sulfur content may have, particularly in the region above 60 octane number. Below 3 cc. of tetraethyllead the chart can be considered more reliable than any one individual octane number determination. The slope of the lines representing two gasolines can easily be compared on this chart, either visually or by use of the scale of lead susceptibility marked along the right-hand side. A given slope-i. e., lead susceptibility-represents a greater

W

0

on

-y.I r

t

-m

-c LL

W

0 ln

ln 3 0 4 WI -

0 LCenter o f

Lsod suIc)pt i b i l i t y Soole

05

IO

3.0

2.0

TETRAETHYLLEAD

-

4.0

5.0

CC. P E R G A L L O N

FIGURE 2. ETHYL BLENDIKC CHART

1370

6.0

INDUSTRIAL A N D ENGINEERING CHEMISTRY

November, 1941

increase in octane number for a given addition of lead when the unleaded fuel is of low octane number. Although the chart has been developed empirically, and its scales do not suggest any obvious relations, reference to its derivation will show that the relation between lead concentration and octane number was obtained by elimination from expressions relating both to compression ratio. While the relative size of octane numbers OCTANE NUMBER (A.S.T.M.-C.F,R. METHOD1 used in the Ethyl blendFIGURE 3. OCTANENUMBER ing chart is thus linked us. ALLOWABLE BOOSTRATIO to a specific enginefacSupercharged C. F. R. engine, P / 6 tor, further coniirmation inoh-diameter oylinder, 1800 r. p. m., of its significance should 7 t o 1 compression ratio, 330° F. jacket be looked for.

ONE such source of Confirmation is found in the rating of fuels on a supercharged C. F. R. engine operated a t a fixed compression ratio. The antiknock value of a fuel is found by increasing the intake manifold pressure under constant operating conditions of speed, temperature, and air fuel ratio, until incipient detonation occurs. The rating of the fuel could be expressed in terms of permissible manifold pressure or indicated mean effective pressure, but it is preferable to refer in some way to a reference fuel in order to eliminate engine, atmospheric, and personal variations. A convenient expression, proposed by Boerlage, Peletier, and Tops (1) is "allowable boost ratio", defined as A. B. ratio = manifold ressure (abs.) at inci ient detonation for test gel/manifold pressure gbs.) at incipient

detonation for reference isooctane

I n each case mixture strength was adjusted to give maximum thermal plug temperature. I n Figure 3 the A. B. values of standard reference fuels (a-heptane-isooctane blends) are plotted against their octane numbers. When the octane number scale is adjusted so as to give a straight-line relation, as in Figure 4, the relative size of the single octane numbers becomes nearly the same as that found from their relation to compression ratio on the A. S. T. M . 4 . F. R. unit, although the cylinder sise and operating conditions of the supercharged engine were all different.

1371

From the preceding example it might also be thought that the value of isooctane plus tetraethyllead units would also remain the same relative size under different test conditions. This, however, is not the case as can be seen from Figure 5 which contains lead susceptibility curves derived under four differeat conditions of test. Although the Ethyl blending chart was originally designed for fuels below 100 octane number, it is increasingly evident that the study of aircraft fuels must take account of antiknock quality higher than that figure represents; in fact,

1.60 '

.

'

O

~

1.50

0

I-

i 2

I.3J

3

3

1.20

x

S ~ ' C Y L , P L A I N INTAKE VALVE. I800 R PM,7:I COMP. RATIO 2 S/~"CYL.,SHROUDED INTI~KE VALVE, I800 R P M , 7 I COMP. RATIO

A 2 6/8"CYL.,SHROUDED INTAKE VALVE, 900 R.P.M.,I.I COMP RATIO. 0 3 I/4mCYL,SHROUDED INTAKE VALVE, I800 RP.H>'I:I OOUP. RATIO.

3 TETRAETHYLLEAD

- CC. PER. GALLON

FIGURE 5. FOURTESTMETHODS FOR LEADSusCEPTIBILITY OF ISOOCTANE ON THE SUPERCHARGED C.F . R. ENQINE

1.0 0

E 0.6

s W

0.4b

'

I 70

I

I 80

I

I

1

90

OCTANE NUMBER ( A S T M - C F R

I 100

METHOD)

FIGURE 4. OCTANE-NUMBER SCALE FROM FIQURE 3, ADJUSTED TO GIVESTRAIGHT LINE The same relative size of octane numbers appears to hold under other test conditions as well. The relation is not precisely the same for all methods of test, but no radical deviations have been found in the dozen or so different laboratory methods that have come to our attention.

F I G W R6.~ L ~ A D SUSCEPTIBILITIES OF ISOOCTANE--~-HEPTANE BLENDS

INDUSTRIAL AND ENGINEERING CHEMISTRY

1372 w

e

3.0

Vol. 33, No. 11

A

2.0

0

P

w K

P6

.3

.e

-

n

0 IW

9

a 0

K

0' w

I. W

n

E:

E:

-

4 u).

.o

BO

W

s

u)

a W

m

>

z

2 0

4

W A

0 7

2 W

0

w

70

s a

c

60

I-

50

.O

n 0 W

40

3

30

0

20

2

W 4

io 0 0

0.5

[center o f Lead suscppt i b i l i t y Scale

FIGCRE

7.

EXTEWDED ETHYL BLE\DIhG

CliiRT

the cooperative organs of the interested industries are actively studying various methods of expressing such antiknock values. Meanwhile, expressions such as 120 octane number are often seen in the world's press. Such figures hare no meaning so long as the method of extrapolation is not stated. From Figures 3 and 4 it is evident that a harmonious eutrapolation using octane numbers would be nonlinear, the size of a single unit becoming continually greater; and it has been estimated that an absolute ceiling ~ o u l dbe reached in the neighborhood of 130 octane number. It appears preferable to use a more linear scale and one which has no arbitrary upper limit, yet it should be expressible in terms of a reference fuel in order t o facilitate interpretation of new data and comparison with data already available. Up t o 100 octane number the A. S. T. M. standard reference fuels can be used, and above 100, c. P. isooctane (2,2,4-trimethylpentane) plus specified quantities of tetraethyllead is the most readily available and easily understood reference. To extend the Ethyl blending chart (A. S. T. M . 4 . F. R. method) into this region by the method used in its derivation, it would be necessary to determine the compression ratio a t standard knock intensity for these various blends of isooctane plus lead. However, in this region of high compression ratios, the high sensitivity of the bouncing pin (which results from the high compression pressure and steep combustion pressure rise) is a disturbing factor. Furthermore, the vertical adjustment of the C. F. R. cylinder head for a given dif-

FIGURE 8.

1.0

2.0

TETRAETHYLLEAD

-

3.0

4.0

5.0

6.0

CC. P E R O A L L O N

V E R I F I C LTIO;": O F EXTESDED ETHYL B L E N D I N G CHART

ference in compression ratio becomes smaller as the compression ratio increases. These two factors prevent an accurate extension of the chart by this method. ANOTHER approach is evident if it is considered that the extension would have a tetraethyllead scale on both the horizontal and the vertical axis of the chart, hence, if only the slope of the line representing isooctane mere known, the intersections of that line with the abscissas representing lead concentration would determine the ordinate scale. Figure 6 shows the lead susceptibility lines for isooctane-n-heptane blends containing 0, 70, 85, 90, and 95 volume per cent isooctane, respectively, and haring slopes of 2.5, 2.7, 2.8, 3.0, and 3.2. By extrapolation a slope of 3.5 seems most likely for pure isooctane. On Figure 7 the isooctane line has been drawn with this slope through the 100-octane-number clear point of the chart, and the isooctane-plus-lead ordinates have been drawn horizontally through the intersections of that line with the abscissas of lead concentration. The final verification of the extended chart must come from its accuracy in application. The several fuels plotted on Figure 8 show that good alignment is obtained and hence that the extension is valid. It is surprising to note how large are the gains from small amounts of tetraethyllead. The difference between 100 octane number and isooctane plus 0.2 cc. tetraethyllead is about equal to that between 95 and 100 octane number. The gain

November, 1941

INDUSTRIAL AND ENGINEERING CHEMISTRY

from unleaded isooctane to isooctane plus 3 cc. tetraethyllead is about equal to that from 75 to 100, or from 0 to 75 octane number. On the supercharged C. F. R. engine the indicated power output is substantially proportional to the manifold pressure a t constant-mixture strength, and therefore the power output at incipient detonation for any given fuel is proportional to its allowable boost ratio. If we call the power output possible on isooctane 100 per cent, then under a certain set of

1373

engine conditions isooctane plus 3 cc. tetraethyllead permits, say, 140 per cent power, and a 75 octane number fuel, 65 per cent (Figure 4).

Literature Cited (1) Boerlage, G.D., Peletier, L. A., and Tops, J. L., Aircraft E w . , 7, 306-8 (1935). (2) Hebl, L. E., Rendel, T.B., and Garton, F. L., IND. ENG.CHEM., 25, 187-91 (1933). (3) Ibid.. 31, 862-5 (1939).

Catalytic Poisoning in Liquid-Phase Hydrogenation Effect of Sulfur Compounds of Various Degrees of Oxidation Catalytic hydrogenation is an established process in the vegetable oil, petroleum, and chemical industries. In many cases process designs, costs, and operating schedules are greatly affected by the presence of catalyst poisons. Results are presented on the poisoning effect of sulfur compounds of various degrees of oxidation.

HE poisoning effect of sulfur compounds on nickel catalysts has been observed in many instances Ellis and Wells ( 2 ) showed that bromine, iodine, antimony bromide, sulfur, phosphorus, sulfur chloride, arsenic, mercury, and lead have a distinct poisoning effect on the hydrogenation of cottonseed oil with a nickel catalyst. Hydrogen sulfide, sulfur dioxide, and chlorine were also shown by Moore, Richter, and Van Arsdel (6) to destroy the catalytic activity of nickel. However, sodium sulfate, sodium chloride, sodium nitrate, and nickel chloride were without effect. Sodium sulfide gave a gradual poisoning effect. Ueno (8)in an exhaustive investigation of the hydrogenation of oil with a nickel catalyst found a retarding effect by soaps of potassium, sodium, lithium, magnesium, barium, beryllium, iron, chromium, zinc, cadmium, lead, mercury, bismuth, tin, uranium, and gold, while the soaps of calcium, strontium, aluminum, cerium, nickel, manganese, copper, silver, vanadium, thorium, and platinum had no effect upon the catalytic action. Nickel acetate, butyrate, stearate, lactate, oxalate, and succinate had also no influence on the catalytic hydrogenation. Fatty acids, such as acetic, lauric, stearic, and oleic, had no influence on catalytic action, but glycolic and lactic acids, hydroxystearic acids, oxalic, succinic, and fumaric acids, and hydroxy acids such as malic, citric, and tartaric acted as catalyst poisons. Sodium taurocholate had a restrictive influence while nucleic acids had no effect. Proteins, blood albumin,

T

A. GARRELL DEEM AND JOSEPH E. KAVECKIS University of Illinois, Urbana, Ill.

blood fibrin, and gelatin showed a restrictive influence, but hemoglobin was inert. Glycerol and lecithin had considerable action, but cholesterol and squalene were ineffective. Carbohydrates such as sucrose, dextrose, mannitol, and starch behaved as negative catalysts, but glycogen had no influence. Alkaloids such as morphine and strychnine were pronounced poisons. Kelber (8) found that nickel catalysts made in various ways behaved differently toward poisons such as hydrocyanic acid or potassium cyanide, hydrogen sulfide, and carbon disulfide when used for the hydrogenation of sodium cinnamate in aqueous solution. The effect of ethyl mercaptan on the kinetics of the hydrogenation of ethyl cinnamate was investigated by Schwab and Brennecke (7). Maxted and Evans (6)were able to show that the poisoning effect of hydrogensulfide, carbon disulfide, thiophene, and cysteine varied directly as the concentration of the poison. Kubota and Yoshikawa (4) found that, while thiophene inhibited the reaction of benzene to cyclohexane,the nickel catalyst so poisoned was still active for the hydrogenation of phorone. Later Yoshikawa (9) studied the effect of thiophene on the relative activities of nickel and nickel-copper catalysts for the hydrogenation of benzene. The nickel-copper catalyst was found to be poisoned to a lesser degree than the plain nickel catalysts. From these investigations the qualitative generality may be drawn that the poisoning effect of acids, halogens, soaps, m d sulfide sulfur compounds are dependent on the catalyst used and the material to be hydrogenated. No work, however, has shown the effect of sulfur of intermediate stages of oxidation between sulfide sulfur and sulfate or sulfonate sulfur. Since the progress of hydrogenation is of fundamental importance to the vegetable oil, petroleum, and chemical industries, more complete information of the effect of sulfur compounds of all types, especially sulfides, sulfoxides, sulfones, and sulfonic acid salts and esters, would be of value'. For this investigation Raney nickel was selected as the catalyst, because of ease of preparing batches with reproducible activity. Phenol, naphthalene, and quinoline were chosen as the materials to be hydrogenated because of the ease with