CA1179622A - Scavenging hydrogen sulfide from hydrocarbon liquids - Google Patents

Abstract

SCAVENGING HYDROGEN SULFIDE FROM HYDROCARBON LIQUIDS
Abstract This invention relates to scavenging hydrogen sulfide from hydro-carbon liquids. For this purpose, the hydrogen sulfide in the liquid is contacted with iron oxide particles typically containing a crystalline phase of Fe3O4 or Fe2O3 and combinations thereof together with an amor-phous Fe2O3 portion, and having a surface area of at least 3.5m2/g.
Where for example, substantially anhydrous kerosene contains hydrogen sulfide in an amount in excess of 20 ppm, such hydrogen sulfide content is reduced to less than 10 ppm, thereby rendering it safe and salable.

Description

, 1~796Z~

SCAVENGING HYDROGEN SULFIDE FROM HYDROCARBON LIQUIDS
Techn;cal Field:
This invention relates to the use of particulate iron oxides for scavenging hydrogen sulfide from non-aqueous liquids such as l;quid 5 hydrocarbons.
Background Art:
The presence of hydrogen sulfide in liquid kerosene, for example, is unacceptable even in small amounts. Yet as well as this Applicant can determine, little has been accomplished in developing any simple 10 scavenging treatment for such non-aqueous liquids.
In U.S. Patent No. 4,008~775, issued February 22, 1977, there is described a process in which specific porous iron oxides are used in drilling muds~ primarily aqueous drilling muds to scavenge hydrogen sulfide (H2S) released from a well in the course of a drilliny opera-tion. These iron oxides are described as having an ideal composik~on of substantially Fe34, a particle size of about 1.5 to 60 microns and asurface area o~ at least ten times as great as magnetite (Fe304) part;-cles of equal size. It has been determined that the iron oxide de-scribed in said patent is Further characterized as having an amorphous Fe203 (non-crystalline) moiety together with an Fe304 crystalline phase.
It is described in U. S. Patent No. 4,089,809, issued May 16, 1978, that H2S can be removed from producer gas by passing the gas (at very - high temperatures) through a bed of pellets composed of silica and ~ Fe203; and U. S. Patent No. 4,201,751, issued May 6, 1980, describes ; 25 that H2S can be removed from coke oven gas by contacting the gas with a fluidized bed of particles of perllte containing steel-making dust comprising Fe203 and an alkaline material such as lime under very turbulent conditions. However, applicant is unaware of any published information which teaches that H2S can be scavenged from substantially non-aqueous liquid systems us;ng iron oxides, much less the iron oxides '~

Z;~

-2 which have been found to be useful in the practice of the present in-vention.
Disclosure of Invention:
The present invent;on provides a process for scavenging H2S from 5 a substantially anhydrous non-aqueous liquid such as a hydrocarbon liquid containing same, which comprises intimately contacting such liquid with dry or substantially dry particles of hydrogen sulf;de-reactive iron oxide particles having a surface area of at least 3.5m2/g and composed substantially of a crystalline phase of Fe203, Fe304 and 10 combinations thereof together with an amorphous Fe203 moiety or portion.
The contacting should be such that the H2S in the liquid is contacted with such iron oxide particles The amount of iron oxide particles used is generally such that the H2S content (generally above 20 ppm~ (parts per million) of the liquid is reduced to a predetermined level, for 15 example, less than 10 ppm and preferably to 4 ppm or less. The tem-perature of the liqu;d used is generally between 10C. and 50~C., but may be higher or lower depending primarily on the boiling point of the liquid at higher temperatures, and the ~iscosity of the liquid at lower temperatures.
A variety of stable reaction products are -formed in carrying out the process depending primarily on contact time, the properties of the specific iron oxide particles used and to some extent on the amount of water, if any, present in the liquid being treated and/or in the iron oxide particles. In general, if the iron oxide particles have an Fe304 25 crystalline phase and an Fe203 amorphous moiety, but no Fe203 crystalline phase, the crystalline reaction product will be stable and comprise a mixture of FeS2 (pyrite) and Fe3S4 (greigite), the ratio of one to the other be;ng primarily dependant on the contact time and the water con-tent of the liquid and/or particles. On the other hand, if the iron 30 oxide particles used have a crystalline phase of Fe203 and an amorphous Fe203 moiety, but no crystalline Fe304 phase, the crystalline reaction products will be stable and comprise primarily FeS2, although some Fe3S4 may also be formed. By "stable reaction products" is meant reaction products such as S, FeS2, Fe3S4 and other non-FeS species 35 which do not readily regenerate hydrogen sulfide, but are at least 70%
stable, in the presence of strong common acids.

~7~;Z~
~3 Best Mode for Carrying Out the Invention:
In carrying out the processes of the present invention, any sub-stantially anhydrous non-aqueous liquid containing H2S may be used al-though it is preferred to employ a substantially anhydrous hydrocarbon liquid, that is a liquid composed of one or more compounds consisting of carbon and hydrogen atoms, and such compound or compounds may be satura-ted or unsaturated, that is, contain -c=c or -c-c bondsO Hydrocarbon liquids such as those obtained from producing wells or earth formatîons often contain H2S ;n concentrations in excess of 20 ppm, and the present processes are des;gned to reduce the H2S content thereof to an accepta-ble level so that the liquid can be used ln refining operations wlthout po;soning catalytic materials or in combustion processes without emit-ting deleterious quantit;es of noxious S02 fumes.
The terms "substantially anhydrous non-aqueous liquids" or "sub-stantially anhydrous hydrocarbon liquids" as used herein are intended toidentify liquids which are free of water or which only contain the normal amounts of water present in the liquid as a result of low or high atmospheric humidity conditions existing during the storage of the liquid, and, therefore, such amounts will vary to some extent depending on the hydrophilic or hydrophobic characteristics of the liquid. In general, the water content of the liquid will seldom exceed 2% by weight, and in most cases will be in the range of from about 0.001% to about 0.05% by weight of water. In accordance with the present lnvention, H2S
can be successfully scavenged from non-aqueous llquids or hydrocarbon liquids even though they are essentially anhydrous or contain such small amounts of water.
The liquid is preferably intimately contacted wlth the iron oxide part;cles by incorporating the particles into the liquid and subjecting the resulting mixture to high shear mixing or agitation so as to expose the surface area of the particles to as much solid-liquid interface as possible thereby enabling the particles to contact the H2S in the liquid.
By operating in this manner, it is also possible to scavenge the H2S
from the liquid as quickly and as efficiently as the reactivity of the parkicles permit. Generally, the contact time will be less than four hours, and usually satisfactory scavenging of the H2S will occur wi~hin a period of 30-150 minutes.
As noted previously herein, the iron oxide particles employed are dry or substantially dry prior to incorporation in the liquid, by which is meant that the particles only conta1n an amount of water such as the particles will have after manufacture and/or after storage in a low or high humidity atmosphere prior to use~ In any event, even though the particles may contain some water they are free-flowing, and appear and 5 feel to be dry particulates of matter. In general, the water content of the particles will seldom be ;n excess of 8% by weight, and ;n most cases will be in the range of from about 0.001 to about 1% by weight.
Usually, the water content of the iron ox;de particles w;ll be partially what may be termed "free" water; that is, water wh;ch ;s picked up from 10 the atmosphere under storage cond;tions, and part;ally what may be termed "bound" water; that ;s, water contained in the particles after rather v;gorous drying, for example, drying under a vacuum at 100-110C.
for 24 hours. The amount of "free" water is usually from about 1.5 to about 5 t;mes greater than the amount of "bound" water.
It has been found that two classes of ;ron oxide particles have unexpectedly higher H2S react;ve capac;ty and rates of react;on than other iron oxides within the scope of th;s invent;on, and, accordingly~
are preferred. These preferred classes of ;ron ox;de particles are as follows:
(1) A class of iron ox;de particles having a surface area of at least 3.5 m /9, a kinetic "K" value of at least 1000 and which are composed of an Fe304 crystalline phase, substantially free of crystall;ne Fe203, and an amorphous Fe203 moiety. Said particles are descr;bed ;n sa;d U. S. Patent No. 4,008,775. Particularly preferred are spec;fic ;ron ox~de particles, hereinafter des;gnated Compound A, the pert~nent properties of which are disclosed in the specific examples herein. Th;s class of partlcles has the theoret-;cal capac;ty to react w;th more H2S than any other ;ron oxides of which I am aware; and (2) A class of iron ox;de particles having a surface area of at least 3.5m /9, a kinet;c "K" value of at least 2000 and wh;ch are composed of an Fe203 crystalline phase, substan~ially free of crystalline Fe304, and an amorphous Fe203 moiety. Such particles are prepared by the conventional high temperature oxidation of ferrous sulfate. Particularly preferred are iron oxide particles, hereinafter designated Compound D, the pertinent properties of which are described in the specific examples herein.
The temperature of the liquid during contact;ng of w;th the ;ron .. .. . .. . .

~7~

oxide particles is nok believed to be critical but is preferably such that a satlsfactory reaction rate between the H2S and the particles ~s obtained and the react;on products are stab1e, that 1s, do not read;ly decompose to form H2S, and so that the liquld is not so viscous as to preclude intima~e mixing of the liquid and particles. A preferred tem-perature range for admixture of the liquid and particles is from about 15C. to about 40C.
The following specific examples are given to illustrate the present invention and the practice thereof, but are not to be construed to limit the scope thereof.
EXAMPLE I
A. Description of Materials Used:
~1) Hydrocarbon liquid used was kerosene having a boiling point range of 80-165C. at 760 mm of pressure and a flash point range of 65-85C.
(2) Dry kerosene is the kerosene of (1) dr;ed with molecular sieve material to a water content of 48 mg per kilogram of kerosene or 0.0048% water by weight.

(3) Wet kerosene is the kerosene of ~1) mixed with deionized water and allowed to separate into two layers from which the water layer was removed giving a kerosene containing 89 mg water per kilogram of kerosene or 0.0089% water by weight.

(4) Compound A is composed of iron oxide particles having a particle size of 6-8 mm, a surface area of 10 m2/g~ a klnetic "K"
value (determined as hereinafter described) of 2000 and containing a crystalline phase of Fe30~, substantially free of Fe203 crystalline phase, and an amorphous Fe203 mo~ety. This material had a water content of 2940 mg per kilogram of iron oxide or 0~29%
water by weight.

(5) Dry Compound A is the same as Compound A except that it has been dried in a circulating oven at 100-110C. for 48 hours to a moisture content of 60 mg per kilogram of iron oxide or 0.006% by weight of water.
B. Description of Equipment Used:
Slurry Reactor: .
All reactions described herein were performed in a water-cooled, glass blender reactor (Lab Glass, Inc., Vinelane, NJ) attached to a Waring blender. The reactor had an internal diameter of 3 1/8" and was il79~Z~

-6-10 1/2" hlgh. The blender reactor was fltted w1th a gasketed l~d sup-plied with various ground glass flttings. The bl~nder contents were mixed by a four blade turbine assembly, located at the base of the reactor, at 10000 rpm. Hydrogen sulfide gas was supplied to the reactor s from a lecture bottle regulated to 0.42 kg/cm2 (6 psig). A flowmeter was used to control the delivery of the gas to the reactor discharge tube. The tube outlet was located below the surface of the blender contents. Gas eXiting the reactor was measured by a ~lowmeter and excess hydrogen sulfide was absorbed in a caustic trap.
Infusion rates of hydrogen sulfide gas were controlled to allow a min;mum of 4.0 9 H2S/hour to enter the reactor. The reactor contained 500 ml of kerosene and 28.4 9 of iron oxide material (20 lbs/bbl), and the contents were maintained at a temperature of 22 C. to 25C.
Several comb;nations of ;ron oxide mater;als were used with the dry and wet kerosene. All test cond;t;ons were carried out in duplicate.
To determ;ne the completion of each run, the gas ex;t;ng the reac-tor was measured by a flowmeter. The reaction was stopped when the exit hydrogen sulfide flow was equal to the ;nlet hydrogen sulfide flow. At this point, the react;on was considered to have reached capacity.
At capacity, a sample of slurry was removed from the reactor and diluted ten-fold with dry kerosene, and the sample was placed ;n a sealed conta;ner. This dilution lowered the H2S partial pressure in the sample container and provided a convenient method to reta;n the sample ~or analysis. The delay between sampl;ng and analysis was less than 30 m;nutes.
Three samples were analyzed for hydrogen sulflde. One sample contained slurr;ed solids and hydrocarbon from the reactor. A second sample was centrifuged and the supernatent was analyzed. A third sample was taken from the caustic trap. These analyses were used to obtain a material balance and determine the quantity of hydrogen sulfide reacted with the iron oxide.
After the react;on, a large port;on of the reacted solids were separated by centrifugation and rinsed three times with acetone to remove residual kerosene. The solids were then dried at 100-110C.
for 24 hours and analyzed for reaction products by x-ray crystallography.
Table I, which follows, shows the conditiuns used in the equipment for scaven~ing H2S employing various combinations of dry ker~sene, wet kerosene, Compound A and dry Compound A, and the H2S reactive capacities of Compound A under these conditions.

I I~ ~ I o ~ ~o ~o ll ~ ~ ~ oo :

~ ~ N o ~ ID ~ ~, r~ d- z : m~- _ .
_~ ~ If'> ') N U'l CO ~ D ~ N

J~ n ~ ~ g ~ O
E ~ O ~ ~ ~i N N N ~i N C~ N r7 N ~') N ~ t~ N O
_ C~ I _ It~ ) C~ O ~

IJ, L ~ ~ 0;~ It1 1~ 0 U:l I O CO O ~r D
, V)~ _ ~ ;
~ oO O 0~ D .~ Co . ~ o~ ~ z ~ o ¦ o ~ ¦ 3 3 3 3 O

v~¦ o o 3 3 3 3 3 1~ O Q O
N ~ ~ 0 Cn ` ~'79~Z~

~ o o ~, o ~ o o o .

V 1~ ~ I ~ ~
C L ~ O

~3 e o ,~ ~ e~ *
~ ~ ~ O ' N ~ ~ a~ o (11 5_ C~ ~ .0 ~ N o~ 1~) L~7 0 n O If O L

~ ~ ~ ~ ~ C

S ~ IJ_ ~ ~ r~ ~ ~ 0 N ~_ O ~

3 C = . D
S~ ~ ~ 7 l ~a~ - lo~
~E x ol . ~ 0 ~ u~ ~D 0 o c~

Z~

It will be noted from the experimental results of TABLES 1 and 2 that the H2S reactive capacity o~ iron oxide containing free water (Experiments 13) is higher than that of the dried oxide (Experiments 4 and 8), with the exception of Experiment 10 in which a larger amoun~ of H2S was infused, whereas an increase in water content of the kerosene actually lowers the reactive capac;ty of such oxide (compare Experiment 1 and Experiments 2 and 3~. On the other hand, the H2S reactive capac-ity of either Compound A or dry Compound A is of somewhat the same magnitude whether the kerosene contains more or less water (Experiments 2 and 6). The analytical results indicate that both FeS2 and Fe3S4 are formed in varying amounts depending on the reactant combinations used.
However, both of ~hese products are stable and do not release H2S in the presence of air and are also "acid stable".
EXAMPLE II
A. Description of Materials Used:
(1) The kerosene used was the sa~e as described in items (1), (2) and (3) of Example I.
~2) Compound D is composed of iron oxide particles having a parti-cle si~e somewhat smaller than Compound A (of Example I), a surface area of 4m2/g~ a kinetic 'IK'' value of 4000 and containing a crystalline phase of Fe203, substantially free of crystalline Fe304, and an amorphous Fe203 rnoiety. Thls material has a water content of 250 mg per kilogram of iron oxide or 0.025% by we;ght of water.
(3) Dry Compound D is the same as Compound D except that ~t was dried prior to use at 105C. for 24 hours to a water content of 170 mg per kilogram of iron oxide or 0.017% by weight of water.
B. Descript;on of Equipment Used:
The equipment used was the same as described in Example I.
Table 3, which follows, shows the cond;tions used in the equipment for scavenging H2S employing various combinations of dry kerosene, wet kerosene7 dry Compound D, and the H2S reactive capaci~ies of Compound D
under these condit;ons.

7~3~

Exp. Reactant Infused Unreacted Reacted Time Infusion Capacity No. Combina- H2S M2S in H2S at Rate gH2S
tions Slurry and capacity (yF~d~ ) (g) Trap (g) (g) (hours) (g/h) Kero- Compound sene D
. .
11. D(l) D(3) 19.6 1.4118.2 4.89 4.79 0.64 12. W(2) D 22.2 1.5 20.7 4.43 5.01 0.73 (1) D is dry kerosene (2) W is wet kerosene (3) D is dry Oompound D
Note: Although no experiment was run using Compound D prior to drying,an experiment essentially identical to Experiment 11 but utilizing 5% of water, based on kerosenel gave a capacity of 0.62 indicating that free moisture in the particles would not influence reactivity significantly.

Table 4, which follows, shows the analytical results obtained with regard to unreacted iron oxide and x-ray diffraction analyses of the reaction products obtained in the experimental runs of Table 3.

__ Exp. Remaining FeS2 Fe3S4 Ratio Total Capacity No. Iron Oxide Fe3S4 Sulfur (gH S
(%) (%) (%) S2 (%) g-cpdD) 11. 20 46~5 4-5* 10/90 22.4 0~64 12. 25-40 44-5 6* 10/90 26 3 0.73 *About '3~Z'~

It will be noted from Table 3 and Table 1 that Compound D has a higher H2S reactive capaci~y than Co~pound A even though the theoretical reactive capacity of Compound A is higher than that o~ Compound D. Also the experimental runs indicate that the H2S reactive capacity of Com~
S pound D is not significantly affected by the water content thereln, and that, in contrast to Compound A, the capacity of Compound D is increased somewhat if the water content of the kerosene is higher. Further, Table 4 shows that the reaction product formed when us;ng Compound D under various wa~er content conditions is substantially FeS2 whereas the reaction products obtained with Compound A (Table 2) contain varying amounts of FeS2 and Fe3S4 under such conditions.
' EXAMPLE III
A. Description of Materials Used:
(1) The kerosene used was the same as described in items (1), (2) lS and (3) of Example I.
(2) Compound C is composed of iron oxide particles having a parti-cle size somewhat smaller than Compound A (of Example I), a surface area of 4 m2/g~ a kinetic "K" value of 100 and containing an Fe293 and Fe304 crystalline phase and an amorphous Fe203 moiety. This material has a water content of 676 mg per kilogram of iron oxide or Q.068% by weight of water. These iron oxide particles are iron oxide waste dusts from open hearth or basic oxygen furnace steel making operations.
(3) Dry Compound C is the same as Compound C except that it was dried at 100-110C. for Z4 hours prior to use, and has a water content of 560 mg per kilogram of oxide or 0.056% by weight of water.
(4) Magnetite is magnetite ore particles which are composed sub-stantially of an Fe304 crystalline phase and an amorpllous Fe304 moiety and have a surface area of about I m2/g and a kinetic "K" value of less than 1Ø This material was received as a water slurry, centrifuged to - 30 remove as much water as possible and dried in vacuum at 100-110C. for 24 hours prior to use. It has a water content of 100 mg per kilogram of magnetite or 0.001% by weight of water.
B Description of Equipment Used:
.
The equipment used was the same as that described in Example I.
Table 5, which follows, shows the conditions used in the equipment for scavenging H2S employing various combinations of wet kerosene, dry kerosene, dry Compound C and magnetite, and the H2S reactive capacities of these iron oxides under these conditions.

7~6 Exp. Reactant Infused Unreacted H2S Reacted Time Infus~on Capacity No. Combina- H2S in Slurry and H2S at Rate ( ~ S
nat;ons Trap Capacity g ~ron (g) (9~ (g) (hours) (g/h~ oxide Kero- Iron sene Oxide 13. D(l) Dry 8.5 4.17 4.33 1.67 5.1 0.15 Cpd.C
14.D Magnetite 15.1 2.76 10.78 2.54 5.93 0.38 15.W~2) Magne- 3.5 2.27 1.28 0.84 4.2 0.05 tite (1) D represents dry kerosene (2) W represents wet kerosene .
An experiment was also run as in Experiment 13 but adding 5% by weight of water to the reactants with the result khat the capacity of Compound C was increased to 0.34. These experiments indicate that Compound C has considerably less capacity than Compounds A and D under substantially dry conditions (compare with Tables 1 and 3) whereas the capacity of magnetite was substankially reduced when the kerosene con-tained more water.
Table 6, which follows, shows the analytical results obtained with regard to unreacted iron oxide and x-ray diffraction analyses of the reaction products obtained in the experimental runs of Table 5.

1~L7 -l3-_ . _ _ _ Exp.Remaining FeS2 Fe3S~ Ratio total Capacity No.Iron Oxide Sulfur (%) (%) (%) Fe ~ _ (%~ (gH ~ ) Fe ~ g ron oxide .. ..
l3. 95 5~ 2.7 0.l5 14. 85 14~ -- 7.5 0.38 l~. 85 5-l -- -- 2.7 0.05 _ -As to the kinetic "K" value in the above description in ~he pH
range 8-lO, the derived rate law ~or Compound A is as follows:
d[S~ = K x [St]2 x [H ] l.06 x [A~
dt~
wherein ~St] is sulfide concentration in ppm, t is time in ~inutes, d~St~ dt is the instantaneous rate of change of dissolved sulfide concen-trations, [H ] is hydrogen iron concentration and [A] is iron oxide concentration (lb./bbl~. K is the rate constant in min.l ppm lcm2 x ltmole and equal to approximately 2000. At Ph 8-lO, the derived rate law agrees closely with Rickard's analysis of the reaction of hydrated iron oxide (ferric hydroxide) and hydrogen sulfide ~Am. J. Sci., 2l4:941 (1974)]. When [Stl and [H~ are measured intermittently during the course of continuous acid reaction the substitution oF observed ~S~]
and CH ] values into the rate law equation above yields apparent K
values which define relative differences in reaction rates among different iron oxides.
For purpose of this application, the rate constant K, measured as min l ppm l cm 2 x liters/mole, is calculated for iron oxide concentra-tions measured in kg/m3, and is as appears in the fol7Owing derived rate law:
d[St~ = -K x [H~l-06 x [A]
dt where:
[S~] is sulfide concentration in ppm.
t is time in minutes.
dCSt~/dt is the ;nstantaneous rate of change of dissolved sulfide t ~t ~J ~
i ~ 7~3 concentrations.
~H+~ is hydrogen ion concentration~
~A~ is iron oxide concentration kg/m3.
Industrial Applicability:
Benefits of the present invention include the followiny: as applied to hydrocarbon liquids, for example kerosene, the liquid is rendered sub-stantially free of hydrogen sulfide; thus previously contaminated liquid products, theretofore incapable of meeting commercial specifications and which may be corrosive in use, are by the present process rendered fully acceptable.

Claims (4)

Claims The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. The process for scavenging hydrogen sulfide from a hydrocarbon liquid containing said sulfide which comprises contacting the hydrogen sulfide in said liquid with dry or substantially dry iron oxide particles having a surface area of at least 0.5m2/g and composed of a crystalline phase of Fe2O3, Fe3O4 and combinations together with an amorphous Fe2O3 moiety, said particles being used in an amount sufficient to reduce the hydrogen sulfide concentration in said liquid to a predetermined level.

2. The process according to Claim 1, wherein the contacting is carried out for a period of time sufficient to form stable reaction products including FeSz and Fe3S4.

3. The process for scavenging hydrogen sulfide from a substan-tially anhydrous hydrocarbon liquid containing said sulfide in amounts in excess of 20 ppm which comprises contacting the hydrogen sulfide in said liquid with iron oxide particles having a surface area of at least 3.5 m2/g, a kinetic "K" value of at least 1000 and composed of a crystal-line phase of Fe3O4, substantially free of a crystalline Fe2O3 phase, and an amorphous Fe2O3 moiety and containing from about 0.001 to about 0.5% by weight of water, said particles being used in an amount sufficient to reduce said hydrogen sulfide content to less than 10 ppm and said contacting being carried out for a period of time sufficient to form stable reaction products including FeS2 and Fe3S4.

4. The process for scavenging hydrogen sulfide from a substantially anhydrous hydrocarbon liquid containing said sulfide in amount in excess of 20 ppm which comprises contacting the hydrogen sulfide in said liquid with iron oxide particles having a surface area of at least 3.5 m2/g, a kinetic "K" value of at least 1000, a water content of from about 0.001 to about 0.5% by weight and composed of a crystalline phase of Fe2O3, substantially free of an Fe3O4 crystalline phase, and an amorphous Fe2O3 moiety, said particles being used in an amount sufficient to reduce said hydrogen sulfide content to less than 10 ppm and said contact-ing being carried out for a period of time sufficient to form stable reaction products including FeS2 and F3S4.