FINAL REPORT Test of TDA's Direct Oxidation Process for Sulfur Recovery by Girish Srinivas, Ph.D. (P.I.) Steven C. Gebhard, Ph.D., P.E. Eugene Peeples TDA Research, Inc. 12345 W. 52nd Avenue Wheat Ridge, CO 80033 (303) 422-7819 Sandra Huzyk Huzyk Energy Management Randy Welch Whiting Petroleum Company

for Jesse Garcia U.S. Department of Energy Contract No. DE-AC26-01BC15333 Period of Performance: March 13, 2001 – February 29, 2004 January 2005

Disclaimer “This report was prepared as an of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.”

2

Abstract This project was a Phase III pilot plant test of TDA’s gas sweetening process done under realistic conditions. TDA Research Inc successfully completed the test at Whiting Petroleum’s Sable San Andreas Gas Plant. The feed was approximately 228,000 standard cubic feet per day (SCFD) of gas that contained approximately 60 vol% CO2, 20 vol% CH4 and 10 vol% C3+ and higher hydrocarbons. The feed was associated gas from CO2 flooding operations carried out on Whiting’s oil wells. The gas is collected and piped to the Sable gas plant where it is normally flared. We sited our pilot plant in line with the flare so that we could remove the hydrogen sulfide (H2S) prior to flaring. The average H2S concentration in the gas during the field test was 7341 ppm. The selectivity of our process for converting H2S into elemental sulfur was essentially 100% and the catalyst converted 90% of the H2S into sulfur and water (the remaining 10% of the H2S ed through unconverted). Importantly, no catalyst deactivation was observed for over the course of the 1000+ hour test. Minimal (ca. 10-15 ppm) of SO2 was formed during the test. Approximately 3.6 tons of elemental sulfur was recovered from a total inlet of 3.9 tons of sulfur (as H2S). The total amount of SO2 released from the plant (taking into flaring of the unconverted 10% H2S) was 0.86 tons. This amount of SO2 is much lower than the normal 8 tons that would have been emitted if all of the H2S were flared over the time of the pilot plant test. The pilot plant was simple to operate and required much less operator intervention than is typical for a new unit being commissioned. Our operator (Mr. Eugene Peeples) has more than 30 years of experience operating commercial scale liquid redox sulfur recovery processes and in his opinion, TDA’s Direct Oxidation pilot plant is easier to operate than liquid systems. The ease of use and low capital and operating costs of TDA’s Direct Oxidation process makes it an attractive technology to be used where traditional sulfur recovery technologies are too expensive (e.g. small to medium sized plants). Currently, TDA’s direct oxidation process has been exclusively licensed to SulfaTreat, and is being offered commercially under the trade name SulfaTreat-DO®. We anticipate that the first plant will be installed in 2005.

3

Table of Contents 1. 2. 3. 4.

Executive Summary Objectives Project Description and Background Description of the Pilot Plant 4.1. Overview 4.2. Catalytic Reactor 4.3. Sulfur Condenser 4.4. Sulfur Bins (“Pots”) 4.5. Modification of Flare Tie-In to Prevent Sulfur Vapor Plugging 5. Field Test Results 5.1. Feed Gas 5.2. Catalytic H2S Oxidation Test Results 5.3. Laboratory Testing of Fresh and Used Catalyst 6. Technology Transfer 6.1. Technology/Information Transfer 6.2. Public Relations and Marketing 7. Conclusions 8. References 9.

LOCATION: NPTO

4

5 5 6 7 7 11 12 14 14 18 18 19 22 23 23 23 24 25 26

1. Executive Summary This project was a Phase III pilot plant test of TDA’s gas sweetening process done under realistic conditions. TDA Research Inc successfully completed the test at Whiting Petroleum’s Sable San Andreas Gas Plant. The feed was approximately 228,000 standard cubic feet per day (SCFD) of gas that contained approximately 60 vol% CO2, 20 vol% CH4 and 10 vol% C3+ and higher hydrocarbons. The feed was associated gas from CO2 flooding operations carried out on Whiting’s oil wells. The gas is collected and piped to the Sable gas plant where it is normally flared. We sited our pilot plant in line with the flare so that we could remove the hydrogen sulfide (H2S) prior to flaring. The average H2S concentration in the gas during the field test was 7341 ppm. The selectivity of our process for converting H2S into elemental sulfur was essentially 100% and the catalyst converted 90% of the H2S into sulfur and water (the remaining 10% of the H2S ed through unconverted). Importantly, no catalyst deactivation was observed for over the course of the 1000+ hour test. Minimal (ca. 10-15 ppm) of SO2 was formed during the test. Approximately 3.6 tons of elemental sulfur was recovered from a total inlet of 3.9 tons of sulfur (as H2S). The total amount of SO2 released from the plant (taking into flaring of the unconverted 10% H2S) was 0.86 tons. This amount of SO2 is much lower than the normal 8 tons that would have been emitted if all of the H2S were flared over the time of the pilot plant test. The pilot plant was simple to operate and required much less operator intervention than is typical for a new unit being commissioned. Our operator (Mr. Eugene Peeples) has more than 30 years of experience operating commercial scale liquid redox sulfur recovery processes and in his opinion, TDA’s Direct Oxidation pilot plant is easier to operate than liquid systems. The ease of use and low capital and operating costs of TDA’s Direct Oxidation process makes it an attractive technology to be used where traditional sulfur recovery technologies are too expensive (e.g. small to medium sized plants). Currently, TDA’s direct oxidation process has been exclusively licensed to SulfaTreat, and is being offered commercially under the trade name SulfaTreat-DO®. We anticipate that the first plant will be installed in 2005.

2. Objectives The objective of the project was to perform a field test of TDA’s gas sweetening process under realistic conditions. In TDA’s process, hydrogen sulfide gas (H2S) is catalytically oxidized (using air) to form elemental sulfur and water. The elemental sulfur is formed in the vapor phase and is subsequently condensed downstream of the catalytic reactor. The small amount of water vapor produced by the process remains in the sweet gas exiting the sulfur condenser. Prior to the field test, all testing had been done at the laboratory scale where synthetic mixtures of feed gas were tested. While laboratory scale testing is sufficient for catalyst development and initial studies, field-testing at the pilot plant scale is necessary to determine how the catalyst will perform when exposed to the wide variety of hydrocarbons, mercaptans and other impurities commonly found in natural gas and associated gas streams. Also, by testing at the pilot scale, temperature gradients within the catalyst bed, non-uniformities in gas flow, and engineering scale issues are addressed that cannot be investigated at the much smaller laboratory scale. This report describes the results from a successful field test of TDA’s process to remove H2S from an associated gas stream that contained approximately 8000 ppm of H2S, 60% CO2, 20% CH4, and 20% ethane, propane, butane and traces of higher hydrocarbons. We also report an 5

economic analysis based on the field test data that shows that TDA’s process (which we refer to as Direct Oxidation or DO) is much less complex and lower in cost than competing commercial technologies.

3. Project Description and Background Direct Oxidation (DO) uses a proprietary patented catalyst that was developed by TDA. The catalyst oxidizes hydrogen sulfide (H2S) to elemental sulfur and water, with minimal if any SO2 formation. DO catalyst development and demonstration at the bench scale was done with funding from the U.S. Department of Energy under Phase I and Phase II SBIR grants. After the laboratory scale demonstration of DO, the Gas Technology Institute (GTI) and Unocal funded the design and fabrication of a $1,300,000 pilot plant to test the process at close to commercial scale. The pilot plant was originally scheduled Figure 1. Location of Plains, TX. to be field tested on a Unocal offshore platform in Mobile Bay, located off the coasts of Mississippi and Alabama. Unocal anticipated that the offshore platform would be subjected to more stringent SO2 emissions regulations and saw the Direct Oxidation process as a way to achieve increased gas production while reducing SO2 emissions. The environmental regulations on the platform however, were not increased and Unocal decided not to spend the additional money needed to install and operate the plant. After extensive negotiations and exploring possible sites for the field test, TDA completed construction of the pilot plant and installed it at a new site in Texas (Figure 1) with funding from the Department of Energy (this project). The plant was installed at Whiting Petroleum Company’s Sable San Andreas gas plant outside Plains, TX and the field test was completed in 2003. The Whiting gas plant is located about 5 miles northwest of Plains, TX, which is about equidistant between Midland, TX, Lubbock, TX and Hobbs, NM (Figure 1). The gas that was used as the feed for the pilot plant contained about 8000 ppm of H2S, 20% methane and about 60% CO2 with the balance being ethane, propane and butanes. This gas is associated gas obtained after CO2 flooding of Whiting oil wells. The pilot plant test was completely successful and the data have enabled us to evaluate the economics of the process and compare our process to currently available small-sulfur recovery technologies such as various liquid redox processes. Having successfully demonstrated the TDA DO process, potential customers will be much more confident about the reliability of the technology. As a result of the field test and favorable economics (discussed later), TDA has licensed the process to SulfaTreat, a business unit of M-I. The process is currently being offered commercially under the trade name SulfaTreat-DO®. Westfield Engineering is the primary contractor for detailed engineering design and for subcontracting the construction of commercial SulfaTreat®-DO units. Westfield is a well-established company with a long and successful history in deg and building sulfur recovery units, including Claus and SulFerox units. The first commercial SulfaTreat-DO® plant is expected to be installed some time in 2005.

6

4. Description of the Pilot Plant 4.1. Overview The pilot plant tests were conducted at the Whiting Petroleum Sable San Andreas Gas Plant located near Plains, TX. Startup and shakedown of the plant was done during the first week of October in 2002. This included installation of the main electric gas heater, tuning controllers and establishing the flow of process gas. Figure 2 is a photograph of the site where the pilot plant was eventually located, and Figure 3 is a photograph of the plant during its initial installation (before being heat traced and insulated).

Shed Pipe rack with flare and feed gas lines

Pad area for pilot plant

Figure 2. Photograph of Whiting Sable San Andreas gas plant before installation of the TDA sulfur pilot plant. The pilot plant has five main “subsystems:” 1) the feed gas system which includes the gas manifold, a liquid knockout (KO) drum and a reactor by; 2) the catalytic reactor where H2S oxidation into sulfur vapor and water takes place; 3) a sulfur condenser; 4) a sulfur u-trap and heat traced liquid sulfur handling piping; and 5) sulfur storage bins. The liquid knockout (KO) drum is located upstream of the catalytic reactors and its function is to remove and hydrocarbon liquids that may be present in the gas stream. This minimizes the amount of hydrocarbon vapors to which the catalyst can be exposed. The KO drum is equipped with level controls, automatic drainage, a flare by etc. The P&ID for the inlet piping and KO are shown in Figure 4.

7

Electric heater attaches here Reactor Sulfur storage pots Sulfur condenser

Flare line Gas inlet not attached in this photo

Figure 3. Photograph of pilot plant during installation at the Whiting site.

Figure 4. P&ID of the knockout drum section upstream of the catalytic reactors.

8

Figure 5 is the process and instrumentation diagram for the reactor and condenser sections of the plant. Radian International did the detailed engineering (and made the drawings) early in the history of the project under subcontract to TDA. Two fixed bed reactor vessels are shown in Figure 5 because in the original plan was to place the plant on a Unocal offshore platform in Mobile Bay. In this application, Unocal wanted the option of drying the gas over a desiccant prior to desulfurization by catalytic oxidation. The second “reactor” was to be used to contain the desiccant bed. In the field test at Plains, this vessel was empty and was valved off.

Figure 5. P&ID for reactor and sulfur condenser section of the pilot plant. One of the reactor was used and was filled with catalyst. The reactor is made of carbon steel and is lined with refractory (described in detail in Section 4.2). The sulfur condenser (described bed in Section 4.3) is a heat exchanger with a half-height tube sheet that is half filled with boiling water. The process gas exiting the catalytic reactor (that contains the elemental sulfur vapor) flows tubeside. The temperature of the sulfur condenser tubes is maintained by controlling the steam pressure on the shell side (boiling water) of the condenser. Because the saturation pressure of the shellside steam determines the saturation steam temperature, controlling the steam pressure controls the temperature of the sulfur condenser. The condenser is mounted slightly off horizontal so that liquid sulfur exiting the condenser will flow downhill and into a U-tube type trap and into sulfur bins (Figure 6). The gas exiting the sulfur condenser was at the sulfur condensation temperature, which was 250-260°F in the field test. Liquid sulfur exits the bottom of the bonnet of the condenser and sweet gas exits at the top. Because this gas is at 250-260°F and is in equilibrium with the liquid sulfur leaving the condenser, the sweet gas contains about 75 ppm of sulfur vapor. Unless this vapor is removed, it will plug downstream piping. To prevent plugging of Whiting’s flare header, 9

we installed about 75 ft of 6 inch SCH40 carbon steel pipe between the pilot plant and Whiting’s flare header. This pipe was unheated so the 75 ppm sulfur vapor would condense in it leaving only clean sulfur free gas to into the flare header. This is discussed in detail in Section 4.5. Commercially available technologies exist for condensing sulfur vapor, and because the focus of the pilot plant test was proving the catalytic oxidation of H2S to sulfur with a real gas stream, we installed the long pipe because this option was quick and inexpensive. In a commercial application, we will use a heated coalescing filter (as recommended by Westfield Engineering).

Figure 6. P&ID for sulfur collection “pots.” The sulfur condenser and sulfur storage vessels are located on the bottom level of the unit (the unit has two levels to save space as per the original off-shore requirement). Liquid sulfur exiting the sulfur condenser ed through a U-trap that acts as a sulfur seal and into collection bins. The bins are purged with air to degas the sulfur (Figure 6). Unfortunately, during the field test, we found that the bins badly leaked gas and were therefore not used. In addition, the heat tracing on the P-trap was not sufficient to maintain the proper liquid level. Therefore, we removed sulfur directly from the heat exchanger bonnet, storing the sulfur in 55 gal drums. Several contractors were involved in the installation and start up of the unit. The primary contractors were Huzyk Energy Management, Denver, CO; Wilbanks Engineering, Houston, TX; and Ref-Chem, Odessa, TX. Sandra Huzyk of Huzyk Energy Management was the project manager for installing the plant and found the test site.

10

Ref-Chem is a heavy industrial construction, maintenance and engineering company. They provide roots facilities, plant expansions, revamps & retrofits, and in-plant contract maintenance services for the petrochemical, chemical, oil & gas, refining, air separation, minerals processing, power, cement and manufacturing industries. Ref-Chem did the physical installation of the pilot plant at the Whiting site (Figure 7). Wilbanks installed the process control system and the heat tracing. Figure 7. Pilot plant installation (photo courtesy of Huzyk Energy Mgmt.). The catalytic reactor is made from carbon steel, lined with refractory, and designed to operated at 25 psig. Figure 8 is a detailed drawing of the reactor and how it was loaded with inerts and catalyst prior to the field test. The reactor was operated adiabatically and contained about 1700 lb of catalyst in the form of 1/8 in x ¼ in pellets. 4.2. Catalytic Reactor

Solids are retained in the reactor by means of a cylindrical screen that is located at the bottom of the reactor. The cylindrical screen also has a cap. The first layer was a 6-inch deep layer of 1-inch diameter ceramic balls (NorPro Denstone). A 6-inch layer of ½ inch ceramic balls and then a 6-inch thick layer of ¼ in ceramic balls followed this. A 10 x 10 mesh stainless steel screen as placed over the Denstone balls and the catalyst (1700 lb of 1/8 inch x ¼ inch cylindrical pellets) was placed on top of the screen. A sheet of 10 x 10 mesh stainless steel screen was then placed on top of the catalyst and a final single 6-inch thick layer of ½ inch ceramic balls was place on top. The balls act as a flow distributor for the feed and product gases as a for the catalyst pellets.

Figure 8. Detailed drawing of the pilot plant catalytic reactor.

11

4.3. Sulfur Condenser The sulfur condenser is essentially a Steam to air-fin cooler thermosiphon “reboiler” type of heat exchanger that operates at about 265°F. Hot gas from reactor At this temperature the saturation Sweetened product gas pressure of steam is about 38 psia (26 Water level psig at the altitude of Plains, TX). By controlling the steam pressure (back pressure on the steam line from the condenser to the air fin cooler (see P&ID in Figure 5), the temperature of the boiling water inside the condenser can be controlled. Figure 9 is a schematic of how the condenser functions and Figure 10 is Liquid sulfur out Water returns from air-fin cooler a close up photograph of the sulfur condenser (the gas outlet is difficult to see Figure 9. Schematic of sulfur condenser. in the photograph – it is located on top of the bonnet at the end of the condenser just left of the water level gauge). Water level gauge Gas outlet Hot gas inlet

Liquid sulfur outlet

Water return

Flare line

Figure 10. Photograph of sulfur condenser. It is important to keep the sulfur condenser temperature between about 240°F (the melting point of sulfur) and 320°F (where the viscosity rises dramatically) because of the peculiar viscosity versus temperature behavior of liquid sulfur. Our typical operating temperature range is 250°F to 280°F. At these temperatures, liquid sulfur has its minimum viscosity (Figure 11). The viscosity data shown in Figure 11 below 240°F (left side) are for the supercooled liquid, which we never have in our application. Above 320°F, the viscosity of sulfur rapidly increases, eventually reaching enormous values (right side of Figure 11) when the so-called plastic or µ−sulfur exists (Tuller 1954). For example, Figure 11 shows that the viscosity of sulfur at 375°F is about 60 lb/ft-sec, which is equal to 89,000 centipoise. For comparison, the viscosity of water at room temperature is about 1 centipoise. The huge increase in viscosity of liquid sulfur with

12

increasing temperature above 320°F is due to the rupture of the normal eight-atom sulfur rings (Figure 12) followed by polymerization of the now linear S8 chains. The resulting material, (called µ-sulfur) is extremely viscous. The phase change is irreversible, and therefore once µ-sulfur is formed, the equipment will plug and will require shutdown for cleaning.

Figure 11. Sulfur viscosity at low temperatures (left) and high temperatures (right); data from the Sulfur Data Book (Tuller 1954). The narrow temperature window (ca. 60°F) for operating the sulfur condenser ∆ requires that there be a minimal T > 320°F temperature gradient along the length of the condenser tubes (which are about 15 S8 µ-S ft long). In a conventional cold-fluid-in, Figure 12. S8 ring breaking and polymerization. hot-fluid-out, counter-current type of heat exchanger, it is difficult to guarantee enough a uniform enough temperature along the length of the tubes so that overheating (making µ-sulfur) or too much cooling (sulfur solidification) can be prevented. Either situation would plug the condenser. To avoid this, the constant temperature/boiling water design was used. Figure 10 is a close-up of the condenser with the various lines labeled. The sulfur condenser is a standard design and it functions essentially as a boiler with a thermosiphon. Steam generated on the shell side of the condenser (by heat transfer from the hot product gas in the tubes) flows up to an air-fin cooler where it is condensed and returned through a standpipe (see Figure 10). The height of the standpipe determines the hydraulic head that drives the water back into the shellside of the condenser. In our case this was about 15 ft of water (approx 2 psi). One problem we observed with this design during operation was that the volume of water in the standpipe was too small. This was because the return pipe was 2 inch SCH40 and there was no water reservoir. As a result, establishing natural circulation was very sensitive to the operation of the steam control valve. Too much steam pressure would blow the standpipe dry and with too little steam pressure, the steam in the condenser would drive water backwards up the standpipe into the air-fin cooler. We were able to control the loop and operate with continuous circulation, but in commercial models we will use a larger diameter standpipe so that the standpipe volume is greater.

13

4.4. Sulfur Bins (“Pots”) The sulfur bins were not used in the field test because of an operational difficulty in the lines that transport liquid sulfur from the outlet of the sulfur condenser to the bins. During the test, we found several cold spots between the exit of the sulfur condenser and the bins that would plug with solid sulfur. Worse, the bins were not gas tight, which would have made using them dangerous. Rather than shut down the plant and re-heat-trace the entire downstream liquid piping, we modified the system so that we could withdraw the liquid sulfur directly from the sulfur condenser. While not a permanent solution, this enabled us to continue the field test and was acceptable because the test was focused on the performance of the Direct Oxidation catalyst, and not how to handle liquid sulfur (which is existing technology that can be easily engineered by Westfield Engineering for commercial units). 4.5. Modification of Flare Tie-In to Prevent Sulfur Vapor Plugging

Vapor Pressure (psia)

The sulfur condenser was operated at 0.6 about 260°F. While the vapor pressure 0.5 of sulfur at this temperature is relatively -3 low (Pvapor = 0.063 Torr = 1.22x10 psia; 0.4 see Figure 13) the volume of gas 0.3 ing over the liquid sulfur is large enough that considerable sulfur will exit 0.2 the condenser as vapor. The average sour gas flow rate was about 200 0.1 SCFM. Assuming that this gas is in 0.0 equilibrium with the liquid sulfur in the 200 250 300 350 400 450 500 550 condenser when it exits the condenser, Temperature (F) then the sulfur vapor concentration in the gas would be approximately 75 ppm. Figure 13. Sulfur vapor pressure (Daubert et al. At this concentration and 200 SCFM, 1997). about 100 lb of sulfur would deposit in the piping downstream of the condenser during the test. If not removed, this sulfur would plug Whiting’s flare lines, which was unacceptable. The goal of the pilot plant test was to demonstrate that the catalytic technology; methods and equipment for handling liquid and vapor phase sulfur is commercially available. Thus, we needed a quick and importantly, inexpensive way to prevent sulfur vapor from condensing and depositing in Whiting’s flare line while permitting us to finish the field test on budget. We investigated several options to remove the sulfur vapor before the gas reached Whiting’s flare header. Among the options considered were: 1) treating the gas with a cold water spray; 2) scrubbing the gas with a solvent; 3) plumbing the gas through a packed tower scrubber (using water) – there was an unused tower located on-site, and 4) adding a long section of unheated, large diameter, pipe section upstream of the flare heater to allow the sulfur to deposit in that line before reaching the flare header.

14

File: C:\STEVE GEBHARD\SG_ PROJECTS\SULFUR proje We chose to protect Whiting’s flare line by 0.008 kmol S installing 75 ft of 6 inch SCH40 pipe between the vapor outlet of the sulfur 0.007 condenser and Whiting’s flare header. 0.006 With this length of pipe, the gas exiting the sulfur condenser was cooled to below 0.005 200°F before it entered the flare header. 0.004 In addition, the pipe was not a straight 0.003 run; several tees were installed (instead of elbows) to permit both cleaning (if 0.002 necessary) and to create turbulence 0.001 S8(g) S6(g) S2(g) where sulfur would preferentially deposit. S7(g) S5(g) Temper 0.000 Several thermocouples were installed C 0 50 100 150 200 250 along the length of pipe so that we could check the gas temperature. In a Figure 14. Composition and dew point of for 8000 commercial application, we would install a ppm of sulfur vapor. commercially available heated sulfur coalescing filter, but time and cost constraints prevented us from using such a device during the field test.

The sulfur vapor recovery line was sized by assuming that free convection around the pipe was the only mechanism that would cool the gas. Most sections of the pipe would be cooled by radiation (especially at night when the air was cool); however during the day, solar radiation would slightly heat the pipe. Ignoring radiation heat transfer simplified the analysis and more importantly, gave a pipe length that was longer than actually needed because natural convection is less efficient at cooling in this situation; ”too much cooling” is not a problem. To make the calculations even more conservative, the temperature of the air was assumed to be rather warm (80°F). The properties of air at this temperature were used in the free convection calculations. The gas inlet temperature was assumed to be 266°F and the outlet was assumed to be 230°F (10°F below the sulfur melting point). The temperature of the surface of pipe was assumed to be the average of these temperatures (248°F). For free convection from a horizontal pipe in air, the Rayleigh number (Ra) is given by Equation 1. This dimensionless number expresses the balance between thermal expansion, temperature, thermal diffusivity, viscosity, and the thickness of the convection layer of the gas in which the body being cooled is immersed (in this case air). Natural convection begins near Ra = 2000 and becomes turbulent for Ra ≥ 106.

Ra =

3 gβ(Tsurface − Tair )d pipe

ν air α air Equation 1. Rayleigh number for free convection (Incropera and Dewitt 1996). The definitions of the in Equation 1 are as follows: αair is the thermal diffusivity of air [k/(ρ∗)], β is the coefficient of thermal expansion of air (2.725x10-3 K-1), g is the gravity constant (32.2 ft/s2), and νair is the kinematic viscosity of air (taken at the average pipe wall temperature, νair = 1.74x10-4 ft2/sec). The value of Ra calculated from Equation 1 was found to be Ra = 5.8x107 which is in the turbulent flow regime.

15

The Nusselt number is used to calculate a heat transfer coefficient from which the required heat transfer area can be calculated. The Nusselt number is given by Equation 2 when 10-5 ≤ Ra ≤ 1012 and in this case was Nu = 47.9 (using a Prandtl number for air of Pr = 0.707). The heat transfer coefficient under these conditions was calculated from Nu using Equation 3 and was hfc = 0.97 Btu/(ft2 hr °F). This low value of heat transfer coefficient is typical for free convection situations where values are typically 1-2 Btu/(ft2 hr °F). 2

      1   0.387Ra 6   Nu = 0.6 + 8   9   27    1 +  0.559  16      Pr        Equation 2. Nusselt number for horizontal pipe in free convection (Incropera and Dewitt 1996).

h fc =

k air Nu dpipe

Equation 3. Heat transfer coefficient Because the inlet and outlet temperatures are different, a log-mean temperature difference (LMTD) was used to calculate the heat transfer area required for cooling. The LMTD is given by Equation 4 and the calculated value was LMTD = 168°F.

LMTD =

(ThotIN

− Tcold ) − (ThotOUT − Tcold )

 (T − Tcold )  ln hotIN   (ThotOUT − Tcold )  Equation 4. LMTD calculation

The amount of heat to be removed is the heat needed to cool the gas from 266°F to 230°F and was q = m*gas*(ThotIN-ThotOUT) = 12,190 Btu/hr. There is so little sulfur in the vapor (75 ppm) that its contribution to the total heat duty (via the heat of fusion) is miniscule and essentially all of cooling duty is used in removing sensible heat from the gas. The heat transfer surface area required to remove 12,190 Btu/hr of heat from the gas given a LMTD = 168°F and hfc = 0.97 Btu/(ft2 hr °F) is A = q/(h*LMTD) = 75 ft2. This corresponds to 43 feet of 6 inch SCH40 pipe (the schedule number is relevant because it determines the outside diameter of the pipe for a standard pipe size). To be conservative we installed 75 ft of pipe, an overdesign factor of 1.7. The flow of gas inside the pipe is turbulent (Reynolds number = Re = 1.1x105) which improves the collection efficiency of sulfur depositing on the walls of the pipe rather than being carried along as tiny solid particles in the streamlines of cool gas (as would occur if the flow were

16

laminar). The calculated pressure drop (assuming incompressible at these low pressures) was less than 1 psi and in practice, we observed a pressure drop of about 0.6 psi. Figure 15 is a photograph of the pilot unit with an overlay of how the new line was attached (the actual configuration was somewhat different when installed, but the basic concept was unchanged). In the original configuration, the pilot plant was attached to Whiting’s flare header (on the left side of the photograph) using 6 inch SCH40 uninsulated pipe that served as both the inlet gas by as well as the exit from the sulfur condenser (red line in Figure 15). If this line were to plug with condensed sulfur vapor, then there would be no way to by the sour gas feed at the pilot plant. In addition, this line was much shorter than 75 ft cooling line making it more likely that sulfur would reach Whiting’s flare system. More alarming, sulfur powder was observed in the original line at a small sampling port located about 15 ft upstream of the header entrance, so sulfur was making it to the flare piping in the original layout. In the new configuration, the original tie-in to the flare header remained connected but was only used for bying and flaring the feed gas. The tee that connected the original flare line to the exit of the sulfur condenser was disconnected and blinded, and the independent, 75 ft x 6 inch SCH40 line was attached between the exit of the sulfur condenser and the flare header (yellow line in Figure 15). Blocking valves were placed at both ends so that the line so that it could be taken out of service and cleaned if necessary. Flanged tees were used rather than elbows to facilitate cleaning and to increase turbulence in the piping to help with sulfur collection.

Use minimum number of tees w/blinds that will still allow for easy cleaning

Old flare connection blanked off

Point where inlet gas can be byed to flare remains unchanged

New flare line has its own tie into the flare header (manual block valve just before header)

New vapor line to flare heater 75 ft of 8 inch SCH40 pipe tied into vapor outlet of sulfur condenser

Sulfur Condenser

Existing flare line

Skid Slope line down hill to a tee so that line can drain after cleaning

Block valve

Figure 15. Photograph of pilot plant showing original flare line (highlighted in red) and planned location of new flare line (yellow).

17

We found that the long cooling pipe was necessary and functioned as designed. At the end of the run after the pilot plant had been shut down, about 100 lb of sulfur was removed from the flare extension when it was opened for inspection (we predicted that about 102 lb should have been deposited). Most of the sulfur was deposited close to the first tee where the concentration of sulfur vapor was highest and where the flow first impinges on a pipe tee at a right angle. Importantly, smaller amounts of sulfur were found farther downstream, as close as 20 ft upstream of the flare header; but fortunately, no sulfur made it to the flare header. Therefore, our solution for removing entrained and vapor phase sulfur from the product gas exiting the sulfur condenser worked quite well and was inexpensive to implement. For a commercial application of course, a more permanent solution such as a heated sulfur coalescing filter would be used.

5. Field Test Results As mentioned previously, the field test was very successful. In this section we describe the results in detail in Section 5.1 we discuss the composition of the feed stream; in Section 5.2 we discuss the catalytic oxidation results and finally in Section 5.3 we compare the results of laboratory testing the catalyst used in the field test with fresh catalyst. Whiting’s Sable San Andreas Gas Plant is used to collect and flare associated gas from CO2 oil well flooding operations. The total gas flow varied somewhat but averaged about 300,000 ft3/day (ca. 200 SCFM) and contained about 6000 to 8000 ppm of H2S. The H2S concentration varied from day to day, as did the total feed flow rate when different oil wells went on and off line. The main components of the gas were methane (ca. 20%), CO2 (ca. 60%), ethane (910%), propane (6-7%) and butanes (ca. 3%). The gas also contained about 3000 ppm of hydrocarbon liquids as vapor in equilibrium with the liquid in the knockout drum located upstream of the feed preheater.

5.1. Feed Gas A detailed analysis of the feed and product gases was performed on grab samples by an outside laboratory. The results of those analyses are given in Table 1. Most of the C5+ hydrocarbons in the gas are removed when then enter the Whiting plant using a knockout drum. In addition there is a second liquid knockout located on the pilot plant skid (Figure 4). Still the inlet gas analysis in Table 1 shows that the gas contained almost 2.8 vol% of C5+ hydrocarbons. Under these conditions, the catalyst was continuously exposed to these hydrocarbons at the reaction temperature of 360°F (182°C) for the entire 1000+ hours of operation. After 1000 hours at sour gas flow rate of 200 SCFM, more than 13 million standard cubic feet of gas were processed and assuming that the total concentration of hydrocarbons with molecular weights higher than butane (i.e. C5 and up) was a constant 2.8 vol%, the catalyst (1700 lb) was exposed to over 35 short tons of C5+ hydrocarbons over the course of the test. This represents over 40 times the weight of the catalyst in higher molecular weight hydrocarbons. If these hydrocarbons had been deleterious to catalyst performance, we would have observed severe catalyst deactivation and we did not; no catalyst deactivation was observed over the 1000+ hour course of the test.

18

Table 1. Gas analysis from Laboratory Services, Hobbs, New Mexico (only compounds that were detected are listed). Mol% (inlet) Mol% (outlet) Component Mol% (inlet) Mol% (outlet) Component H2S N2 CH4 CO2 Ethane Propane isobutane n-butane isopentane n-pentane cyclopentane 2-methylpentane 3-methylpentane

0.8000 1.5972 17.6750 58.6264 8.7006 6.4640 0.9798 2.3841 0.7385 0.7715 0.0105 0.1512 0.1345

0.095 2.8794 17.7853 58.6625 8.5607 6.2643 0.9228 2.3154 0.7115 0.7453 0.012 0.1354 0.1203

n-hexane methylcyclopentane benzene

0.2658 0.1375 0.2107

0.2353 0.1144 0.1692

cyclohexane n-heptane methylcyclohexane toluene n-octane ethyl benzene p and m xylene o-xylene n-nonane

0.1698 0.0976 0.0517 0.0218 0.0053 0.0029 0.0029 0.0007

0.1387 0.0703 0.0378 0.0136 0.0032 0.0024 0.0027 0.0004 0.0022

Mercaptans (ppm)

101.00

20

Specific gravity (air =1 ) Gross BTU/CF dry Gross BTU/CF wet

0.45 731.00 718.00

5.2. Catalytic H2S Oxidation Test Results Figure 16 shows the sour gas flow rate and the inlet H2S concentration during the 1000+ hour test. During the first 150 hours of the test (6 days), the flow rate was about 280 SCFM but the H2S concentration was low (6000 ppm). There was a significant change in the sour gas flow rate during the pilot plant test as different wells were brought on and off line. The H2S concentration gradually increased over the first 300 hrs (12.5 days), but after that the concentration was essentially constant at about 8000 ppm. The inlet and outlet H2S, SO2 and O2 concentrations were measured using GasTech stain tubes. The stain tubes provided a simple and reliable method to determine the H2S, SO2 and O2 concentrations without the problems we encountered with gas chromatography (GC). The hydrocarbon vapors present in the gas (few thousand ppm) contaminated the GC column making the GC gas analysis results erratic. We also analyzed for O2 to determine if there was any oxygen slipping through the catalyst bed; none was detected. The oxygen concentration was also measured using an electrochemical analyzer. Because H2S will damage the electrochemical cell, this instrument was protected from H2S and hydrocarbons by a series of molecular sieve, and Purafil (KMnO4 impregnated alumina) beds. The molecular sieve traps hydrocarbons and the potassium permanganate (KMnO4) oxidized H2S to sulfate (which remains adsorbed on the Purafil beads). Figure 17 shows the H2S conversion as a function of time on stream during the pilot plant direct oxidation test. The H2S conversion was approximately 90% for the duration of the test and did not decrease with time. Because we kept the catalyst temperature constant during this time, the constant H2S conversion indicates that no catalyst deactivation had occurred in 1000+ hours of the test (this was confirmed by testing the spent catalyst from the pilot reactor in the laboratory with fresh catalyst as discussed in Section 5.3). The reason that the conversion remains constant even through the flowrate of feed gas is changing, is because the reaction is mass transfer limited. The lack of deactivation is especially significant considering that the catalyst was exposed to over 40 tons of C5+ hydrocarbon vapors and even larger quantities of methane,

19

ethane and butanes during the test without any adverse affects. The resistance of the catalyst to deactivation when exposed to hydrocarbon vapors is consistent with our earlier laboratory investigations where we contaminated our synthetic gas with toluene, xylene and hexane. No deactivation was observed in the laboratory studies and the hydrocarbons simply ed over the catalyst bed unreacted without coking or fouling the catalyst. 300

250 8000 200

150

7000

100 6000

Sour gas flow rate (CFM)

H2S inlet concentration (ppm)

9000

50

5000 0

200

400

600

800

1000

0 1200

Time on stream (hours)

H2S inlet (ppm)

Sour gas flow CFM

Figure 16. Sour gas flow rate and inlet H2S concentration during field test.

100%

10

90% 80% H2S conversion

60% 50%

5

40% 30%

Air flow (cfm)

7.5

70%

2.5

20% 10% 0% 0

200

400

600

800

1000

0 1200

Time on stream (hours) H2S conversion

AIR FLOW (cfm)

Figure 17. H2S conversion and air flow rate during field test Figure 17 also shows the airflow as a function of time during the pilot plant test. The airflow is included in Figure 17 to show how the fluctuations observed in the H2S conversion (although small) were due to our intentionally varying the airflow (and consequently the O2/H2S ratio). This was done to determine the affect of changing O2/H2S ratio on H2S conversion and the catalyst’s selectivity to sulfur and SO2. When the airflow was held constant, the H2S conversion was constant (see for example the data between 400 and 500 hours in Figure 17).

20

100

6.0

5.0

4.0

50

3.0

2.0

Air Flow (CFM)

SO2 outlet (ppm)

75

25 1.0

0

0.0 0

200

400

600

800

1000

1200

Time on stream (hours) SO2 outlet (ppm)

AIR FLOW (cfm)

Figure 18. SO2 outlet concentration and air flow during field test. Figure 18 shows the concentration of SO2 in the outlet (product gas) during the test. The airflow is shown again to illustrate that by varying the airflow, we can control the concentration of SO2 in the outlet gas. Note that the SO2 concentration tracks with the airflow quite well. In any case, the SO2 concentration never exceeded 75 ppm and was typically about 12 ppm when operating under conditions that produced a high selectivity for elemental sulfur (e.g. the time between 200 and 700 hours in Figure 18. Table 2 compares the sour gas flow rates, sulfur production and SO2 emissions when the pilot plant is removing an average of 89% of the sulfur in the gas. Because the pilot plant processes Table 2. Production summary for field test at time = 1000 hours. Field Test Case No Sulfur Recovery Case Average sour gas flow rate 228 228 (SCFM) Average air inlet flow rate 3.5 0 (SCFM) Average H2S inlet (ppm) 7341 7341 Average H2S outlet (ppm) 790 0 Selectivity to elemental sulfur 100% N/A Average sulfur yield 89% 0 Total sulfur inlet (lbs) 7953 7953 Total sulfur inlet (tons) 3.9 3.9 Total sulfur recovered (lbs) 7097 0 Total sulfur recovered (tons) 3.6 0 Total sulfur flared as SO2 (tons) 0.86 7.95 all of the gas that would normally be flared by Whiting, and the selectivity to elemental sulfur is practically 100%, the reduction on SO2 emissions from the Whiting flare is equal to the mass of sulfur (as H2S) converted. For example, operation for over 1000 hours with an average H2S concentration of 7341 ppm, a total of 3.9 tons of sulfur enters the plant. The DO process converts 89% of that into elemental sulfur. The unconverted (11%) H2S is burned in the flare in the usual manner producing SO2. Over the course of 1000 hours, our process discharged about 0.8 tons of SO2 including the amount of unconverted H2S that is flared. For comparison, if

21

the feed gas were directly flared, then almost 8 tons of SO2 would have been emitted. This comparison illustrates one advantage of TDA’s direct oxidation process, that even at modest low sulfur recovery efficiencies (ca 90%), very large reductions in SO2 emissions are possible. The Direct Oxidation process also can be used if higher sulfur recovery efficiencies are required. In that case, a tail gas treatment unit (TGTU) such as liquid redox, amine with recycle or scavengers (e.g. SulfaTreat®) is placed downstream of the DO unit to remove the unconverted H2S.

5.3. Laboratory Testing of Fresh and Used Catalyst On concern we had was that because the H2S conversion remained constant at about 90%, but that over time the flow rate of feed gas slowly decreased (Figure 16), was that this effect might have masked catalyst deactivation. To test this possibility, we tested used catalyst taken from the pilot plant reactor after the end of the pilot run, with fresh catalyst. The tests were done in the laboratory under carefully controlled conditions of temperature, gas composition and gas flow rate. Figure 19 shows the results from the laboratory testing of fresh (blue bars) and used (red bars) catalyst. The used catalyst was removed from the pilot plant reactor after the 1000+ hour test. The samples were taken from three radial locations at three depths in the catalyst bed. The used catalyst tested in the laboratory was an average sample made by grinding and sieving a portion of each sample from the reactor and then mixing equal amounts of these samples together. 95

Percent Conversion

90 85

T = 356oF P = 1 atm 8000 ppm H2S 3200 - 4400ppm O2 Balance = N2

No SO2 Detected

80 75 70 65 0.8

0.9 O2 Concentration (Fraction of Stoichiometric) Used Catalyst

1.0

Fresh Catalyst

Figure 19. Laboratory testing of fresh (dark bars) and used (light bars) catalyst. The reaction conditions for the laboratory test were essentially identical to those used during the field test. The temperature was 356°F (180°C), the space velocity was approximately 600 ft3gas/ft3catalyst/hr, and the gas contained 8000 ppm of H2S. The oxygen concentration was varied between 3200 ppm and 4400 ppm and data were obtained at O2/H2S ratios that corresponded to 80%, 90% and 100 % of the stoichiometric amount of O2 need for the reaction H2S + ½ O2 Æ S + H2O (i.e. O2/H2S = 0.40, 0.45 and 0.5 respectively). Fresh catalyst (from the same batch that was used when filling the reactor on the pilot plant) was tested under identical conditions. Both the fresh and the used catalysts exhibited essentially identical performance indicating that

22

there was no deactivation of the catalyst while it was in service in the pilot plant test. The reason that the conversion remained at 90% while the gas flow dropped was that, at least at the lower flow rates, the H2S oxidation reaction was mass transfer limited.

6. Technology Transfer 6.1. Technology/Information Transfer TDA has successfully completed the Phase III pilot plant testing project and has shown that the TDA catalyst developed for the direct oxidation of H2S into sulfur and water is active, selective and long-lived. The test lasted for more than 1000 hours (42 days) desulfurizing 100,000– 300,000 SCFD of CO2-rich, associated gas that contained 8000 ppm of H2S. The overall H2S conversion was approximately 90% during the course of the test, and minimal SO2 was formed. Currently, TDA’s direct oxidation process has been exclusively licensed to SulfaTreat, a business unit of MI and is being offered commercially under the brand name SulfaTreat®-DO. The original pilot plant (now in Plains, TX) will be disassembled and shipped back to GTI in Chicago, IL by June 2005.

6.2. Public Relations and Marketing TDA’s H2S oxidation process has been licensed to SulfaTreat as the SulfaTreat-DO® process (SulfaTreat®-“Direct Oxidation”). SulfaTreat is a business unit of M-I, LLC. M-I is a SmithInternational/Schlumberger Company. SulfaTreat is actively marketing the technology, and the first commercial unit will be installed in 2005. SulfaTreat is in final negotiations with the customer. That unit will process approximately 1 million SCFD of natural gas containing about 1% H2S. The SulfaTreat-DO® process is being d in appropriate trade magazines and has generated many inquires. Figure 20 is a copy of the ment published in Oil, Gas and

Figure 20. ment in Oil, Gas and Petrochemical Equipment for the SulfaTreatDO® process. 23

Petrochemical Equipment and Figure 21 is a copy of the ment for the process that appeared in the March 2004 issue of Hydrocarbon Processing (Gulf Publishing Co).

Figure 21. Ad for direct oxidation in March 2004 issue of Hydrocarbon Processing (scanning and slight rotation of the image caused blurring of the text – the original is crisp and clear).

7. Conclusions The objective of the project was to perform a field test of TDA’s gas sweetening process under realistic conditions. TDA Research Inc has successfully completed this Phase III pilot plant test at Whiting Petroleum’s Sable San Andreas Gas Plant. The average gas flow during the test was 228,000 SCFD. The gas contained significant quantities of CO2 (ca 60%), CH4 (ca 20%) and C3+ hydrocarbons (ca 10%). The average H2S concentration was 7341 ppm. No catalyst 24

deactivation was observed for over the course of the 1000+ hour test. Minimal (ca 10-15 ppm) of SO2 was formed during the test. The selectivity to sulfur was essentially 100% and the catalyst converted 90% of the H2S into sulfur and water. Approximately 3.6 tons of sulfur (in the elemental form) were recovered from a total inlet of 3.9 tons of sulfur (as H2S). The total SO2 emissions from the plant (taking into flaring of the unconverted H2S) were 0.86 tons. This much lower than the normal 8 tons of SO2 that would have been generated with straight flaring of the feed gas over the same period of time. The pilot plant was simple to operate and required much less operator intervention than is typical for a new unit coming on line. It was also considerably easier to operate than a typical liquid redox system. The chief plant operator during the test was Mr. Eugene Peeples. Mr. Peeples has more than 30 years of experience operating commercial scale liquid redox processes, and thus he is highly qualified to judge the ease of operation of the Direct Oxidation process compared to existing commercial liquid redox sulfur recovery processes. Mr. Peeples indicated that the Direct Oxidation process was considerably easier to run from an operator’s standpoint. The ease of use and low capital and operating costs of the TDA Direct Oxidation process makes this an attractive technology to be used where traditional sulfur recovery technologies are too expensive. As a result, the Direct Oxidation process is well suited for sulfur recovery in small to medium sized plants. Currently, TDA’s Direct Oxidation process has been exclusively licensed to SulfaTreat, and is being offered commercially under the trade name SulfaTreat-DO®. SulfaTreat is a division of M-I, which is a t venture between Schlumberger and Smith International. SulfaTreat’s flagship H2S scavenging technology is the market leader and is currently used to treat nearly 3 trillion cubic feet of gas annually from more than 1,000 applications in more than 20 countries worldwide. The SulfaTreat Company has become a widely recognized and trusted leader in the industry serving international markets of Energy, Environmental and Water Treatment. A new SulfaTreat system is installed somewhere in the world nearly every business day. With SulfaTreat’s background and expertise, they are ideally suited for marketing the Direct Oxidation process. In SulfaTreat-DO, the Direct Oxidation process can be operated as a stand alone unit, or used upstream of a gas polishing step such as the SulfaTreat® scavenger, a liquid redox process or other sulfur recovery technologies. In addition to stand-alone units, the SulfaTreat-DO® process can be used to “unload” existing sulfur recovery systems, thus providing customers with a low cost simple alternative to increasing the sulfur handling capacity of their equipment.

8. References Daubert, T.E.; Danner, R.P.; Sibul, H.M. and Stebbins, C.C. (1997) Physical and Thermodynamic Properties of Pure Chemicals, Design Institute for Physical Property Data (DIPPR), American Institute of Chemical Engineers, Taylor and Francis. Incropera, F.P. and DeWitt, D.P.(1996), Fundamentals of Heat and Mass Transfer, 4th ed., John Wiley & Sons, New York (1996). Tuller, W.N. (1954) Sulfur Data Book, Freeport Sulfur Company.

Updated By:

Date:

Girish Srinivas

January 2005

25