Synthesis Gas by Partial Oxidation

DuBOIS EASTMAN
The Texas Co., P.O. Box 300, Montebello, Calif.

M IXTURES of hydrogen and carbon monoxide are known

as synthesis gas. These can be used directly in various
catalytic processes t o produce hydrocarbons, methanol, or other
veloped to use natural gas as a fuel but was later applled to
light oil and to heavy bunker-type fuel oils.
The first application of this process was a t The Texas Co.
synthetic organics. The carbon monoxide present in synthesis Laboratory in Montebello, Calif., where from 1916 through
gas can be reacted almost quantitatively with steam via the 1954 it was used in a routine manner t o provide synthesis gas
water gas shift reaction t o produce a mixture of hydrogen and for a pilot unit on a modification of the Fischer-Tropsch process.
carbon dioxide. On the removal of this carbon dioxide, pure The process was also used a t the synthetic fuel and chemical
hydrogen is available for hydrogenation or ammonia synthesis. plant of Carthage Hydrocol. Since that time numerous instal-
Synthesis gas was first produced from coal. A two-stage proc- lations have been made a t synthetic ammonia plants xhere the
ess was used, first converting the coal t o coke in a coke oven flexibility of the process with respect to fuel source is particu-
and then reacting the coke with steam in a water gas apparatus larly attractive. The ability t o produce ammonia from heavy
to produce synthesis gas. This reaction of coke and steam is fuel oils by this process has made it possible to manufacture am-
highly endothermic. The necessary hcat is supplied by inter- monia from easily transported fuel oils a t locations where neither
rupting the gas-making process and blowing the hot coke bed coal nor natural gas is an economically practical raw material.
with air. This burns part of the coke and reheats the bed pre- It was immediately apparent that the noncatalytic reaction
paratory to the next gas-making cycle. Both the coke oven of methane and oxygen would proceed as a flame. It was also
and the water gas set are operated a t pressures near atmospheric. apparent from the general theory of flames that a detailed state-
For this reason the apparatus is large and bulky, and high costs ment of mechanism was not available.
are involved in compressing the synthesis gas to the pressures Experimental work soon showed, however, that free oxygen
normally required for use. was not a component of the product and that the partial oxidation
Another method, which has been widely used for synthesis reaction was characterized by the presence of methane, carbon
gas manufacture, is the methane-steam reaction. This reaction dioxide, and x a t e r vapor. Although the actual reactions are
is also highly endothermic, but is carried out as a continuous certainly much more complex, this behavior is consistent with a
process. This is done by using a catalyst packed in alloy two-stage reaction mechanism involving complete combustion
steel tubes which are arranged in a furnace. The equilibrium in the first stage via the highly exothermic reaction
of the methane-steam reaction requires such high temperatures
that, even with the best alloys available, pressures must be held CH, + 202 CO:! + 2HzO (1)
a t relatively low levels. These low pressures again result in Equilibrium is very far to the right of thiq reaction and no
bulky equipment and high compression costs. measurable oxygen remains. The quantity of oxygen fed is,
The exothermic reaction of methane and oxygen was studied however, much less than that required t o react with all the
as early as 1933 by Padovani ( 1 ) who carried out the reaction a t methane by Equation 1 so t h a t on completion of this reaction a
atmospheric pressure in the presence of a nickel catalyst. This large surplus of methane remains. This surplus methane can
same process w-as developed on a commercial scale by Schiller, then react with either of the products of the primary reaction
Bartholome, and Koch a t Opau in Germany. I n the Schiller, as follows:
Bartholome, and Koch process methane and oxygen were pre-
mixed and fed to a reactor through small openings in a refractory CHa + COP 2CO + 2H2 (2)
burner block, using feed velocities above the rate of flame propa-
gation. As in Padovanis work, a nickel catalyst was used and
CH, + HQO = CO + 3Hz (3)

the pressure was near atmospheric. Whereas the primary reaction is highly exothermic, these reac-
Following World War 11, work was undertaken by Hydro- tions are both highly endothermic. It is therefore important
carbon Research on a pressurized version of this process. As in that the excess heat liberated in the primary reaction be made
the work of Schiller, Bartholome, and Koch the natural gas and available by some means of heat transfer t o the secondary endo-
oxygen were premixed and preheated a t velocities above the rate thermic reactions.
of flame propagation. The reactor itself was internally insulated The balance between these exothermic and endothermic reac-
and packed with a catalyst. tions is such that a departure from equilibrium in the secondary
The Texas Co. undertook vork on a noncatalytic process using stage has the effect of liberating heat. This increases the tem-
a specially designed burner whirh permitted all the mixing t o perature and tends to accelerate the secondary reactions. TO
be done within the reaction zone. This process was first de- this extent the reaction system is self regulating

1118 INDUSTRIAL AND ENGINEERING CHEMISTRY vci. 48, NO. 7

 SYNTHETIC FUELS AND CHEMICALS

I n addition to these reactions of methane, the products can Calculations of heat and material equilibrium such as those
react via the feebly exothermic water gas shift reaction discussed will give fair approximations of fuel and oxygen re-
quirements. They may be further refined through the use of
CO + HzO = COz f Hn (4) reaction rate data which measure the rate of approach to equi-
and carbon can be formed from carbon monoxide via librium in the slow secondary reactions. Experience has shown,
however, that more reliable predictions can be made by means
2co = con + c (5) of the thoroughly empirical relations discussed below.
Equilibrium calculations of temperature and effluent composi-
tion can be based on Equations 2,4, and 5 which include the prin- Equipment
cipal products: CH4, H,, CO, H,O, COS, and C. Values of the
equilibrium constants for these reactions have been translated Calculations of the type discussed in the previous section made
from Wagman ( 3 ) . it clear that equipment must be designed to withstand tempera-
It has been found that the methane and oxygen requirements tures above 2000' F. if the methane content of the product were
of this system can be fairly well approximated by a simultaneous t,o be reduced to a value on the order of 1%.
calculation of temperature by heat balance and composition by The very severe combination of a pressure of several hundred
equilibrium. In these calculations Equation 5 is used only to see pounds with a temperature above 2000' F. was met through the
that the product mixture is not in the carbon-forming range. use of internally insulated reactors. These are designed with
Actually the system is rate controlled by the slow secondary cylindrical carbon-steel pressure shells lined with successive
reactions, Equations 2 and 3. The temperature is, however, layers of specially selected refractories.
very sensitive to the proportions of methane to oxygen, and the I n this process it is possible, through a control failure, to subject
equilibrium is also very sensitive to the temperature. As a the inner wall of the reactor to temperatures approaching 6000" F.
result, the simultaneous solution for temperature and equilibrium However, in normal operations the refractories are subject only
composition offers a fair approximation of the methane and to temperatures that are allowable for firebrick.
oxygen demaad of the system. Stated in another way, a rela- Since the rate-controlling reactions are temperature sensitive,
tively large departure from the equilibrium concentration of there is a relation between allowable operating temperature and
methane in the product (1% as compared with 0.05%, for ex- the required residence time. I n practice residence times are
ample) may still correspond to a good approximation of the selected as required to bring the required temperature within the
methane and oxygen which must be fed to produce a given quan- safe working temperature of the refractory wall.
tity of hydrogen and carbon monoxide. It is also necessary to select values of internal diameter and
Since the desired over-all reaction length such as to provide adequate heat exchange between the
CHa + '/no* = CO + 2Hz exothermic primary reaction and the endothermic secondary
reaction. It is essential to avoid contact between hot, unreacted
proceeds with a doubling in volume, it is apparent that compres- oxygen and the reactor wall, as this is highly destructive.
sion costs can be greatly reduced by operating a t a high pressure. Care must also be exercised to provide for the differential ex-
If this is done it will be necessary to compress only about half pansion of the refractory and the external pressure shell and a t
as much gas as would be required if the operation were carried the same time to prevent the access of hot gases to the metal
out a t low pressure. Furthermore, the methane is usually avail- surface.
able a t a fairly substantial pressure, and the oxygen can be It is necessary to mix the reactants quickly and thoroughly as
pumped as a liquid before vaporizing. soon as they are introduced into the reactor in order to prevent
The disadvantages of pressure operation lie in the adverse thermal decomposition and coke formation. These reactions
effect on equilibrium and the severe mechanical problems that proceed very rapidly a t the temperatures used. The mixing
are introduced in this high temperature process. process is carried out by means of an alloy burner having carefully
Equilibrium calculations showed that a t pressures of about 200 controlled dimensions. It would be desirable to use an uncooled
to 1000 pounds/square inch gage the effect of pressure on equi- refractory burner in order to reduce heat loss and avoid dependence
lib1 ium could be offset by a moderate increase in temperature on coolant supply, but no satisfactory refractory has been
and that this increase in temperature could be obtained a t very available.
small cost in increased oxygen. When the product gases are t o be fed to a water gas shift
Energy balances are conveniently based on values of the heat reactor, it is convenient to quench directly with water, so that
content of the constituents expressed as B.t.u./pound mole the sensible heat content of the effluent gas is used to generate
relative to the elements a t 0" R. In'thermodynamic terms: the steam needed in the shift reaction.
When the hydrogen-carbon monoxide mixture is to be used
Fhere
(H? - Hg) + AH;
H? = heat content a t temperature, t, ' F.
directly without shifting the carbon monoxide, the heat of the
effluent is transferred to steam in a boiler. Care must be exer-
HI = heat content a t 0" R.
AH,: = heat of formation from elements a t 0' R. cised in the design of a boiler for this service t o avoid attack
through alternate oxidation and reduction. Provisions must
The 0 superscripts indicate that these values are exact only a t also be made to prevent fouling with soot or to remove such soot
infinite dilution. The errors due to their use a t several hundred as may be deposited. After the primary cooling step, any re-
pounds pressure are normally unimportant. maining traces of soot are removed by scrubbing.
Values of heat content as defined have been calculated from Because of the sensitivity of reaction temperature to the
Rossini ( 2 ) . fuel-oxygen ratio, unusual care is required in instrumentation.
When dealing with fuel oils, the heat content of the oil, also I t is, in fact, unlikely that the process could be used a t all without
referred to the elements a t 0' R., is evaluated in the usual way its unique arrangement of control instruments. In this service
from the heat of combustion. the usual ratio controllers are undesirable since they do not
As the ratio of carbon to hydrogen in the fuel is increased a provide a simultaneous record of both flows and both pressures
point is reached a t which the quantity of oxygen necessary to I n the event of an upset condition it is desirable to have this in-
react with the carbon is more than enough to bring the product formation immediately available so that the operator can de-
mixture to the necessary temperature. At this point steam can termine exactly and quickly where the upset originated. As a
be added to supply part of the total oxygen and to hold the reac- result of these considerations, separate pressure and flow con-
tion temperature a t the desired level. trollers are used on each feed stream, and a ratio alarm is in-

July 1956 I N D U S T R I A L AND E N G I N E E R I N G CHEMISTRY 1119

 Natural Gas Operation
Table I. Generator Operation
Typical operating data for natural gas-oxygen, fuel oil-0x3-gen,
Natural Fuel Coal and coal tar-oxygen systems ale shown in Table I. The results
Gas Oil Tar with natural gas were obtained in a generator of commercial
Flow rates
Natural gas, thous. std. cu. ft./hr. 32.08 ... size, 11-hereasthe results of fuel oil and coal tar were obtained in
Oil, lb./hr.
Steam, lb./hr.
....
....
428.7
162.7
4i6:7
218.0
a smaller. pilot-size generator. Except for heat losses which
are a function of generator size, considerable experience has
Oxygen, thous. std. cu. ft./hr. 26.51 5.20 4.57
shown that there are no substantial differences in operation. in
Dry product gas, thous. std. cu.
ft./hr. 100.03 21.59 18.73 generators of different sizes.
Material balance and heat balance calculations for the test
Operating conditions period on natural gas are given in Table 11. I n this method of
Pressure, lb./sq. inch gage 219 342 347
Oxygen preheat, F. 209 69 64 calculation the volume of product gas is obtained by carbon bal-
Hydrocarbon preheat, ' F. 827 712 750 ance, and the water vapor content of the reactor effluent iE taken
Product composition, mole % as the average of the value indicated by hydrogen balance and
Carbon monoxide 38.02 47.97 54.36 the value indicated by oxygen balance.
Carbon dioxide 2.19 3.65 5.69 For the natuial gas operation, the d i y pioduct volume is
Hydrogen 59.54 47.45 38.87 obtained by dividing the total moles of carbon fed (106.278)
Nitrogen
Methane
0.15 0.22 0.82
by the molal fraction of carbon in the product (0.38025 +
Hydrogen sulfide
0.10
....
0.26
0.44
0.07
0.18 0.02193 + 0 00101 = 0 40319). The resulting product volume
Carbonyl sulfide .... 0.02 0.01 (263.593) multiplied by the molal composition of the product
gives the quantities of all constituents, except water vapor.
Performance data
Oxygen consumption, cu. ft./ T n o values for water vapor can be obtained. By subtracting the
1000 CU. It. Hz + CO 271.7 252.3 259.0 total atoms oi hydrogen as calculated (313.860 +
1.064 =
Cold gas efficiency (heating value 314.924) from the total atoms of hydrogen fed (375.888). a value
+
of HI CO), % heating value of 30.482 is obtained by hydrogen balance. Correspondinglg,
of HC Fed 81.86 83.89 86.41
by subtracting the total a t o m of oxygen (100 231 +
11.562 =
111.793) from the total atoms of oxygen fed (141.502), a value of
29.709 is obtained by oxygen balance. The average value
stalled to \Tarn of any deviation from the desired ratio. duto- (30.482 + 29.709)/2 or 30.095 is used.
matic shutoff devices are provided t o stop both flov-s if the ratio The heat balance is obtained by taking the sum of the products
departs more than a specified amount from the set value. of the feed constituents times their respective total heat contents
Platinum, platinum-rhodium thermocouples are used in re- a t the preheat temperature. I n the example given in Table I1
fractory protecting tubes. the total heat content of the natural gas stream is given as (69.243)
Temperature alarms are at,tached to the exterior of the pressure (-1640) + (5.926) (-11 22) + (5246) (-9.53), e t c , for a
shell. These consist of long, small-bore alloy tubes containing total of -1377.7 a t 800" F. The corresponding total a t 900" F.
two conductors separated by a temperature-sensitive insulating is -1234.9 so that the value for the actual preheat temperatuie
material. These tubes are stapled spiially to the outside of the of 827" F. is interpolated at -1339.1 thousand B.t.u. per hour.
pressure shell. A resistance-bridge potentiometer sounds an Correspondingly, heat content of the oxygen stieam at 208" F.
alarm if any part of the tube exceeds a safe temperature. is given as 325.5 thousand B.t.u. per hour so that the total
The methane content of the dry product gas, which is generally heat input in the feed streams is -1339.1 +325 5 or -1013.6
the control point in the system, is monitored by infrared analyzers. thousand B t.u. per hour. In thr samr Jvay, the hrat contrnt

Table II. Heat and Material Balances

Mole Moles/ Total Heat Content, Thous.
Feed streams % Hour Elemental Balances B.t.u./Mole
Natural gas C H 0 800' F. 900' F. Total
Methane 81.91 69.243 69.243 276.972 ... - 16.40 - 14.98
Ethane 7.01 5.926 11.852 35.556 ... -11.22 -8.80
Propane 6.20 5.246 15.738 41.968 ... -9.53 -6.03
Butane 1.69 1.424 5.696 14.240 ... -8.35 -3.79
Pentane 0.71 0.596 2.980 7.152 ... -7.64 -2.03
Carbon dioxide 0.91 0.769 0.769 ... 1.538 -157.38 - 156.18
Oxygen 0.14 0.123 ... ... 0.246 +9.10 +9.89
Nitrogen 1.43 1.209 ... ... ... _ _ _ _ +8.86 i-9.60
Total 100.00 84.536 - 1377.70 -1234.90 - 1339.1 at 827' F.
200' F. 300' F.
Oxygen a .
_ _ - _ _..._ _ _ _..._
_69.859 139.718 f4.59 f5.31 = f 325.5 at 209'F.
Total Input 154.395 106.278 375.888 141.502 - 1013.6
Products 2500' F.
~- 2600' F.
___
Carbon monoxide 38.02 100.231 100.231 ... 100.231 -26.30 -25.45
Carbon dioxide 2.19 5.781 5.781 ... 11.562 -134.78 -133.36
Water vapor
Hydrogen
..
59.54
30.095
156.930
...
...
60.190
313.860
30.095
...
-74.78
+21.27
-73.63
122.05
Nitrogen 0.15 0.385 ... ... ... 122.43 +23.27
Methane
Total, Wet
- 0.10 _0.266
293.688
_0.266
106.278
_1.064
375.114
- ...
141.888
~ +14.63 ~
-2315.3
~
+16.73 ~ _
-2064.0 = -2184.6 at 2552O F.
_
Dry 100.00 263.593

1120 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 48, No. 7

 SYNTHETIC FUELS A N D CHEMICALS

of the products is calculated to be -2315.3 thousand B.t.u. per

hour a t 2500" F. and -2064.0 a t 2600' F., giving an interpo-
lated value of -2184.6 thousand l3.t.u. per hour a t a temperature
of 2552" F. The difference between input and output (-1013.6
and -2184.6 or -1171.0) represents the heat loss from the
reaction system. Since the heating value of the natural gas fed
was 38,444 thousand B.t.u. per hour, the loss is 3.08% and the
hot gas efficiency is 96.92y0.
The over-all heat balance is therefore:

Thousand
B.t.u./
Hour Per Cent
Heating value of natural gas feed 38,444 100.00
Heatjng value of HBand CO in product 31 ,470 81.86
Heating value of methane in product 102 0.27
Heat loss from reactor (as determined
Rhove) 1.171 3.08
Heat aqailable for steam production 5;701 14.79
Total 38,444 100.00

The heating value of hydrogen and carbon monoxide in the cold

product, expressed as a percentage of the heating value of the
fuel fed (81.86Yo in this case) is identified as "Cold Gas Effi-
ciency" in Table I.

Table 111. Comparison of Observed and Predicted Oxygen

OXYGEN CONSUMPTION, C U F T / U . CU F T Hz f CO
Consumptions
Predicted Figure 1. Natural gas operat;on-relation of oxygen
Observed Oxygen consumption
Methane Oxygen Con-
Content of Con- sumption
Net Heat Dry sumption, (Figure I),
Input, Thous. Product Cu. Ft./1000 Cu. Ft./1000
B.t.u./Mole Gas, Cu. Ft. Cu. Ft. The correlation shown in Figure 1 may be expressed by the
Carbon Fed Mole 70 H2f CO H a + CO equation
-13.4
- 13.4
0.07
0.10
261
253
260
257
O ~ ( C Uft./1000
. +
CU. f t . Hz CO) =
-13.0 0.15 256 260 196.5 - (17j(log % CHdj - (3.25) (net heat input)
- 13.8 0.28 252 251
-13.6 0.30 252 250
- 17.5 0.06 275 275 The process is also sensitive to the ratio of oxygen to natural
-17.7 0.10 271 271 gas. For an increase of about 5% in this ratio, the temperature
-18.0 0.16 269 269 increases about 200" F., and the methane content of the product
-22.5 0.53 273 274 gas declines sixfold. It is this sensitivity of methane content
- 17.1 0.98 2 53 253
to temperature and feed ratio that has made it desirable to install
-22.4 2.29 260 264
-22.4 2.47 261 263 infrared spectrometers to monitor product streams. Although
-22.4 2.51 265 263 these instruments could probably be used for control, they have
-22.4 2.57 265 263 actually been installed as multipoint recorders, following several
-22.8 2.85 264 263
streams intermittently.
Although subject to some control through burner design, there
is always some net production of carbon black. This material
is produced by the thermal cracking of incompletely mixed gas
There is a relationship between conversion (as measured by adjacent to the burner and is largely destroyed by reaction with
the methane content of the product) and operating tempera- carbon dioxide and water vapor as the mixture passes through
ture. Over the normal operating temperature range, the methane the reactor. It is thermodynamically possible to form free
content of the dry product gas changes by a factor of two for carbon by the back reaction of carbon monoxide to produce car-
temperature changes of about 50" F. This relationship is sensi- bon and carbon dioxide as the reaction mixture is cooled, but
tive to residence time, changing about 50' F. for each twofold this can be avoided by using a rapid initial cooling rate as the
change in time. effluent leaves the reaction zone.
The relation of methane content and net heat input to oxygen Residual carbon is effectively scrubbed out of the product gas
consumption, shown in Figure 1, is much more general in applica- with water. It is not practical to measure the carbon remaining
tion. The net heat input is defined as the algebraic sum of the in the gas after this scrubber, because of the very small quantities
heat contents of the input oxygen and gas and the heat loss, and the small particle size. The best evidence of its complete
expressed as thousand B.t.u./mole of carbon in the natural gas. removal is found in the long continued operation of a fixed-bed
This completely empirical correlation has been applied to a very shift converter on the scrubbed gas. After a year of operation
wide variety of operating temperatures, pressures, reactor vol- such a converter was found to be free of any appreciable deposits
umes, and methane contents and predicts values for oxygen of carbon black. The carbon produced in the gas generation
consumption within 1% of the experimental values. Typical operation is similar in particle size to commercial carbon black,
comparisons are shown in Table 111. but has a very high oil absorption index.

July 1956 INDUSTRIAL A N D ENGINEERING CHE.MISTRY 1122

 operations in which the oxygen rate is too low, and efficiency
suffers because of poor conversion. Substantial quantit,ies of
unconverted carbon are produced in this operating range. On
the upper limb of the curve, hovever, the conversion of carbon is
substantially complete. I n this range a further increase in oxygen
rate serves only t,o convert, hydrogen and carbon monoxide to
unnanted water vapor and carbon dioxide. Optimum operation
with this oil feed then obviously lies nithin 250 to 270 cubic feet
of nsygen/1000 cubic feet of Hz +CO. From an economic
standpoint it is generally desirable to operate in the neighbor-
hood of 250 cubic feet/1000 cubic feet of oxygen consumption
and 847, efficiency, although this is subject to some variation
n-ith the local costs of oxygen relative to fuel.
-4s \vould be expected, the amount of unconverted carbon in
an oil generator operation is a function of the oxygen feed rate,
the amount of unconverted carbon increasing practically linearly
n i t h decreasing oxygen feed rate. Unconverted methane de-
creases a t a decreasing rate as temperature is increased, the re-
lation being very similar to t'hat described for natural gas. Es-
perience has sholm that m-ith differences in fuel oil quality and
differences in steam ratio, the various relationships betxeen tein-
perature, unconverted carbon, oxygen consumption, thermal
f 30
80 81 82 83 e4 OS efficiency, and met,hane content of the product gas remain fairly
THEflUAL C f f l C I E N C Y - HEArINB VALUE OF U , + C O , PEfl C E N T O P OIL constant,. The last column of Table I shows, for comparative
purposes, the results obtained when charging to the reactor a
Figure 2. Fuel oil operation-oxygen and fuel efficiency stripped coal tar of very high carbon content. Except for
a n even higher rat'io of carbon monoxide to hydrogen, the results
are substantially the same as those obtained when charging fuel
oil. The oxygen consumption and cold gas efficiency figures are
Fuel Oil Operation quite comparable.
I n most residual fuel oil operations some unconverted carbon
.Llthough similar in many respects to the natural gas operation, is produced, and it is necessary t o install filters to remove this
the application of this process t o fuel oils is complicated by an carbon from the circulating water stream. Although this carbon
additional process variable (steam rate) and by the wide range of is produced in a particle size comparable to carbon black, it is
fuel oil quality. I n addition sulfur and ash are normal constitu- easily %-etby water, flocculates rapidly, and can be settled com-
ents of fuel oils, and their presence needs to be considered. pletely in about 20 minutes. The settled floc has a loading of
A typical set of data on fuel oil operation is given in Table I. about 1 weight 7ocarbon, and it is therefore necessary to filter
The fuel oil used in this run had the following properties: the stream in most cases. The usual method is t o cool the blow-
down stream by exchange wit,h the return n-ater and remove the
Gravity, ' API 13.1 carbon with a continuous vacuum filter.
l'iscosity, se:. Furol a t 122" F. 405 ?\lention was previously made that most of the sulfur present
Pour point, F. d0 in the oil feed appears in the product as hydrogen sulfide. Ana-
Carbon residue (Conradson), % 10.2% lytical studies show that the distribution of sulfur compounds is
-Ish content, % 0 06%
Gross heat of combustion, B.t.u. /lb. 18011 very similar to t h a t calculated for equilibrium. One study
Ultimate analysis,. 5%
.. ehowd, for example:
Carbon 55 59
Hydrogen 11 38
Nitrogen 0 72 Distribution of Sulfur,
___- yo-
Sulfur 1 96 Observed Equilibrium
Oxygen 0.35 H d r o g e n sulfide 89 93
Carbonyl sulfide 11 7
Carbon disulfide 0 03 0 05
It is evident from Table I that the general level of performance,
as nieasured by oxygen consumption and cold gas efficiency, is
very similar t o that for natural gas. The principal points of Acknowledgment
difference are the higher ratio of carbon monoxide t o hj-drogen
for the fuel oil operation and the presence of sulfur in the prod- .Is is commonly the case in n-ork of this nature, many people
ucts. The sulfur compounds present in the oil feed are con- have been concerned in the pilot plant development of this
verted primarily to hydrogen sulfide n-hirh can readily be re- process. Those most directly insTolved have included L. P.
moved from the product gas by any one of a number of conven- Gaucher, F. E. Guptill, Gordon Kiddoo, J. B. Nalin, C. P.
tional operations, Only very minor amounts of carbonyl sul- Marion, TV. L. Glater, and D. XI. Strasser.
fide and a trace of carbon disulfide are found in the product
gases.
literature Cited
Figure 2 s h o w the oxygen consumption (in cubic ieet/1000
+
cubic feet of HZ CO produced) plotted against cold gas effi- (1) Padovani, C., Francbetti, P., Giorn. C h ~ m .i n d . app2icnta 15,
+
ciency (heating values of Hz CO, as per cent of heating value 429 (1933).
of the oil feed) for typical fuel oil operations. Although all (2) Rossini, F. D., others, U. S.Bur. Standards, Circ. C-500, Fehru-
ary 1952. and C-461, November 1947.
these data fall within the range 81 to 847, thermal efficiency,
(3) Wagman, D. D., others, U. S. Bur. Standards, Research Paper
it is evident that a maximum was reached in the neighborhood of R P 1634 1.7. Research 34 (February 1945)l.
84'3 with about 260 cubic feet/1000 cubic feet of oxygen con-
sumption. The lowcr limb of this curve is representative of RECEIVED
f o r review December 10, 1933. ACCEPTED
March 29, 1056.

1122 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 48, No. 7