WSRC-TR-2001-00364

Design, Construction and Testing of a Practical-Scale Palladium
Membrane Reactor for Cracking Tritiated Water

D. W. Howard and K. L. Sessions
Westinghouse Savannah River Company
Aiken, SC 29808

S. A. Birdsell
Los Alamos National Laboratory
Los Alamos, NM

This document was prepared in conjunction with work accomplished under Contract No. DE-AC09-96SR18500 with the U.S. Department of Energy.

DISCLAIMER

This report was prepared as an account 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.

This report has been reproduced directly from the best available copy.

Available for sale to the public, in paper, from:  U.S. Department of Commerce, National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161,  phone: (800) 553-6847,  fax: (703) 605-6900,  email:  orders@ntis.fedworld.gov   online ordering:  http://www.ntis.gov/support/index.html

Available electronically at  http://www.doe.gov/bridge

Available for a processing fee to U.S. Department of Energy and its contractors, in paper, from: U.S. Department of Energy, Office of Scientific and Technical Information, P.O. Box 62, Oak Ridge, TN 37831-0062,  phone: (865 ) 576-8401,  fax: (865) 576-5728,  email:  reports@adonis.osti.go

Abstract

A palladium membrane reactor (PMR) system has been successfully designed, constructed, tested, and put into service at the Savannah River Site for recovering elemental tritium from tritiated water on Disposable Molecular Sieve Beds (DMSB), filled during tests of the Tokamak Fusion Test Reactor. This PMR system is the first known large-scale application of the PMR to routine, production recovery of tritium from tritiated water.

This PMR incorporated several new design features to achieve high conversion efficiency of tritiated water to elemental tritium, while operating at the high flow rates and low pressure drops needed for a practical production-scale operation. The PMR demonstrated >98% conversion of water to elemental hydrogen during "cold" (non-tritium) tests.

During the first integrated system test with tritiated water, over 4.8 lbs. of tritiated water were processed in a period of about 150 hours of continuous operation, recovering approximately 2400 liters of mixed ‘hydrogen’ isotopes (H, D, & T). The system is now in routine operation recovering tritium from the legacy of DMSBs generated during the Princeton tests. Future plans for the PMR system are to process over a dozen DMSBs, containing about 75 lbs. of tritiated water, recovering about 38000 liters of mixed hydrogen isotopes (primarily D).

1. Introduction

A Palladium Membrane Reactor (PMR) system to recover tritium from tritiated water has been successfully demonstrated on a laboratory scale at the Tritium Systems Test Assembly (TSTA) at Los Alamos National Laboratory [1]. Large quantities of tritiated water waste exist world-wide because the predominant method of cleaning up tritiated streams is to oxidize the tritium to tritiated water, which can be collected with high efficiency onto molecular sieves and stored for subsequent processing and/or disposal.

The PMR is also under consideration for the tritium processing plant for the International Thermonuclear Experimental Reactor (ITER), which would contain tritiated impurities such as water and methane. The tritium will require recovery from these impurities for economic and environmental reasons.

The PMR is a combined permeator and catalytic reactor. Heterogeneous catalysts are used to foster reactions such as water-gas shift reaction,

Q2O + CO = Q2 + CO2

and methane steam reforming,

CQ4 + Q2O = 3 Q2 + CO

where Q represents any of the hydrogen isotopes H, D, and/or T. Thermodynamics limit these reactions to only partial completion. A Pd/Ag membrane, which is exclusively permeable to hydrogen isotopes, is incorporated into the catalyst bed of the reactor. A vacuum maintained on the permeate side of the membrane, continuously removes the product hydrogen isotopes, and enables the reactions to proceed toward completion.

Previous papers [2, 3, 4] have reported the results of various PMR laboratory-scale testing. Tests have been performed on simulated fusion fuels without tritium, simulated fusion fuel with tritium, and tritiated water, at flow rates up to (1100 sccm total gas; 500 sccm of Q2O; and/or 40 sccm of CQ4).

The Savannah River Site (SRS) currently has in storage a large number of Disposable Molecular Sieve Beds (DMSBs) containing tritiated water recovered from the experimental runs at the Princeton Tokamak Fusion Test Reactor. These DMSBs are being processed to recover the tritium. The previous processing method required the desorbed water, driven off the DMSB by heating, to be reacted with heated Mg metal to convert the Q2O to Q2, which was then isotopically separated. The use of heated Mg to convert the Q2O to Q2, has multiple disadvantages: generation of contaminated solid waste, periodic shutdown (with hands-on, contaminated maintenance) to replace spent Mg, and low flow rate.

The present paper reports on the design, construction, and testing of a Palladium Membrane Reactor, now installed at the Savannah River Site and used for production-scale recovery of tritium from tritiated water. This PMR system has now eliminated the generation of solid waste, eliminated the need for replacement of spent Mg, and has shortened the processing time for each DMSB. This PMR was the product of a collaborative effort between the Savannah River Site and Los Alamos National Laboratory. The PMR was designed by LANL personnel, constructed by Johnson Matthey Inc., and tested by LANL and SRS personnel at LANL. SRS personnel designed the instrumentation, auxiliary equipment, and support equipment. The PMR was installed, started, and put into production operation by SRS personnel.

2. PMR Design

Figure 1 shows a schematic drawing of the SRS PMR. This PMR was designed to be similar to the 1st stage PMR described by Birdsell & Willms [1] used for tritiated water processing, but having increased Pd/Ag tube permeation area and an increased ratio of permeation area to catalyst volume. The SRS PMR contains 16 tubes (vs. 6 for the LANL unit), each 28.25 inches long by 0.25 inch OD, but has a higher ratio of tube surface area to catalyst volume than the 1st stage of the LANL PMR system. In regards to the ratio of tube surface area to catalyst volume, the SRS PMR was designed to be more similar to the 1st stage PMR of LANL’s cold test bench PMR system, described by Birdsell, Willms, and Wilhelm [3].

The reactor body is approximately 30.5 inches long by 3.5 inch OD, constructed of 347 stainless steel. The SRS PMR contains approximately 4000 gms of Pt/a-Al2O3 catalyst (Engelhard A-16825). The catalyst pellets are approximately 0.125 inch diameter by 0.125 inch long. The catalyst surrounds the Pd/Ag tubes within the reactor body.

The sixteen Pd/Ag tubes are arranged on two concentric circles. The inner circle contains five tubes and the outer circle contains eleven. The tube arrangement was designed so that the majority of points in the catalyst bed would less than 0.3 inches from the nearest tube surface, and that no point would be more than 0.5 inch from the nearest tube surface. This was intended to minimize concentration gradients within the catalyst bed. The top end of each tube is sealed, and the bottom end passes through the bleed end cap assembly and is open to the vacuum manifold that routes to the product pumping system.

The bleed end cap is attached to the bottom of the reactor body. A ten-micron stainless steel filter separates the catalyst from the bleed end cap. The Pd/Ag tubes penetrate through the filter and the entire bleed end cap assembly and are welded to the outer surface of the bleed end cap. The byproduct gas (unreacted CO and H2O, unpermeated H2, and CO2) flows out of the catalyst bed, through the ten micron filter, and into the bleed end cap where it is pumped away by the by-product exhaust pump train. The ten-micron filter was intended to provide sufficient pressure drop to prevent channeling of the gas through the catalyst bed near the by-product exhaust port.

The vacuum manifold is welded to the outer portion of the bleed end cap assembly. The open end of each Pd/Ag tube connects to this vacuum manifold allowing the hydrogen that permeates into the tube to be removed from the reactor via the vacuum manifold and product pumping system.

A heating and insulating jacket surrounds the reactor body. The jacket has three separate heating elements. The first (top) element or zone is designed to heat the inlet portion of the reactor and is approximately six inches long. The second zone is approximately 18 inches long, and the bottom zone is approximately 6 inches long. There are two thermocouples associated with each heating zone; one thermocouple measures the temperature of the heating element while the other measures the skin temperature of the reactor body. Additionally 4 thermocouples measure temperatures within the catalyst bed: one on the centerline of the bed near the feed inlet, and 3 others 0.1 inch in from the ID of the reactor body spaced at 120° angular positions and approximately equal axial intervals. These thermocouples were intended to allow precise adjustment of a uniform temperature in the PMR.

3. PMR System

Figure 2 shows the overall process flow diagram of the PMR system.

The PMR system is exhausted through three sets of pumps. Three Normatex scroll pumps (Model PV-12) arranged in parallel exhaust the product gas (hydrogen isotopes) from the vacuum manifold. Based on the PV-12’s performance curve, the three pumps in parallel were intended to achieve a suction pressure of approximately 1 torr at a flow rate of 0.5 std. liter per minute. The three scroll pumps discharge into a train of 4 Metal Bellows pumps (Model MB-151) arranged in series. The discharge from this pump train is collected into storage tanks for later isotopic separation. The byproduct gas (H2O, CO2, CO and residual H2) is exhausted by a train of 4 Metal Bellows pumps (Model MB-151) arranged in series and configured to pull the gas through a molecular sieve bed. The molecular sieve bed is intended to remove the unreacted water from the byproduct stream; the remaining gas is compressed by the pump train and discharged into storage tanks for further processing.

The carbon monoxide (research grade; minimum purity 99.99%) is supplied to the PMR process from a gas cylinder, through a pressure regulator, check valve, and a set of three automatic valves in a Block-and-Bleed configuration. The carbon monoxide supply can either be routed to the PMR through the DMSB (sweep mode) where the flowing CO helps to sweep the desorbed water vapor out of the DMSB, or be routed directly to the reactor inlet (bypass mode) to mix there with the stream of desorbed water.

Flow rates of both water vapor and carbon monoxide are measured and controlled by mass flow controllers and a PLC that adjusts the flow rate set points to maintain an excess of approximately 5%-20% CO in the PMR feed.

The PMR’s normal operating conditions will be between 400 –500 C, and 600 – 1000 torr, and flow rates from 100 – 3000 sccm.

Instrumentation for monitoring the performance of the PMR system includes a mass spectrometer sample point for the byproduct gas downstream of the byproduct molecular sieve bed, and two moisture meters on the byproduct stream, one upstream and one downstream of the byproduct molecular sieve bed. These two moisture meters provide the control of the residual moisture in the gas exiting the PMR, and indicate when the byproduct molecular sieve bed has become saturated and requires changeout.

3. Testing

3.1 Initial Testing

After fabrication by Johnson Matthey Inc., the PMR was installed in support and auxiliary equipment at LANL similar to the final equipment at SRS, and put through a "cold" (non-tritium) test regimen to check out the equipment, to condition the catalyst and to verify design performance. Diagnostic instrumentation at LANL used a gas chromatograph instead of a mass spectrometer.

The PMR was leak tested, and heated to operating temperature. The catalyst was dried by flowing dry argon through the PMR at operating temperature. The catalyst was conditioned by flowing a 5% oxygen in argon mixture through the PMR, followed by a 5% hydrogen in argon mixture, both at operating temperature. The permeation rate of hydrogen through the PD/Ag tubes was determined by supplying a mixture of 50% hydrogen in argon to the feed and measuring the flow rate of hydrogen gas in the product stream.

CO and H2O feeds that were similar to the design operating conditions were used for performance testing of the PMR. Water was sorbed onto a spare DMSB, which was subsequently used to supply the water vapor to the PMR system. This permitted demonstration of the ability to control the delivery of water vapor to the PMR. Both the sweep and bypass modes of mixing the CO and water vapor were tested and proved satisfactory. Better than 98% conversion of the water vapor to molecular hydrogen and removal from the fed stream was demonstrated at flow rates up to 500 sccm of water vapor.

3.2 Startup Testing at SRS

After initial testing at LANL, the PMR was shipped to SRS and installed into the pre-fabricated, permanent support and auxiliary equipment, in the tritium-processing production environment. The system was then put through a period of initial testing and checkout, which included leak-checking and electrical functionality.

The system was then subjected to a period of non-heated, non-tritium functional testing. This regimen verified operability of all valves, mass flow controllers, instrumentation, etc.

3.3 Integrated System Tests with Tritium

The first test with tritium used a DMSB containing approximately 4.8 lbs. of tritiated water, and was expected to yield approximately 2400 STP liters of hydrogen molecular isotopes if fully desorbed and converted. The first test was run with the CO feed in the bypass mode, and with an 8% excess of CO. Flow rates up to 500 sccm of water were used. A total of 1608 STP liters of molecular hydrogen isotopes were collected in the product tank, and the average water feed rate was 220 sccm. A total of 1584 liters of byproduct gas were collected, with an average composition of 9.5% CO, 88.3% CO2, and 0.3% molecular hydrogen isotopes. During this test no attempt was made to completely purge the feed DMSB of water.

Subsequent tests have operated the PMR with the CO feed in the sweep mode, which permits a more complete purging of the last water out of the DMSB. This mode of operation has been successfully demonstrated.

4. Routine Production Operations

The PMR system is now in routine production operation recovering hydrogen isotopes from the tritiated water on the DMSBs. Maximum sustained feed rates seem to be limited to about 390 sccm of water vapor, which seems to be limited by the pumping capacity of the product pump train, specifically the Metal Bellows pumps. There are indications that one or more of the Metal Bellows pumps may be malfunctioning. Currently, the PMR system is being used to recover hydrogen isotopes from the inventory of 14 DMSBs from the Princeton Tokamak Fusion Test Reactor. These DMSBs are estimated to have a total of about 70 lbs. of tritiated water, and are expected to yield about 35,700 STP liters of molecular hydrogen isotopes (primarily D).

In the future, the PMR system can be modified to process tritiated water from other offsite and onsite sources. The modifications required are mainly small piping configuration changes to physically connect other geometries of molecular sieve beds to the feed inlet.

5. Conclusions

This PMR system is the first known large-scale application of the PMR process to a routine, production recovery of tritium from tritiated water. This system achieves a high throughput rate of water, and produces no solid waste, thus reducing environmental impacts. It requires less contact maintenance than alternative processes, thereby reducing radiological exposure. The PMR offers a means to process the tritiated water in a more cost effective and environmentally friendly manner than alternative processes.

The design and construction of this PMR were straightforward. No unexpected difficulties in testing or operation have been found. The design appears to be scalable to higher flow rates and/or lower concentrations of hydrogen in the byproduct. The main limitation in achieving either of these goals may lie in the ability of tritium-compatible vacuum pumps to achieve lower pressures in the vacuum manifold at higher throughputs.

6. Acknowledgment

This document was prepared in connection with work done under Contract No. DE-AC09-96SR18500 with the U.S. Department of Energy.

7. References

  1. S. A. Birdsell and R. S. Willms, Tritium Recovery from Tritiated Water with a Two-Stage Palladium Membrane Reactor, Fusion Engineering and Design, Vol. 39-40, ISFNT-4; 4th International Symposium on Fusion Nuclear Technology, Tokyo (Japan), 6-11 Apr. 1997.
  2. R. S. Willms, S. A. Birdsell, and R. C. Wilhelm, Recent Palladium Membrane Reactor Development at the Tritium Systems Test Assembly, Fusion Technology, Vol. 28, No. 3, (1995) pp. 772-777.
  3. S. A. Birdsell, R. S. Willms, and R. C. Wilhelm, Ultra-High Tritium Decontamination of Simulated Fusion Fuel Exhaust using a 2-Stage Palladium Membrane Reactor, Fusion Technology, Vol. 30, No. 3, Part 2A, (1996) pp. 905-910.
  4. S. A. Birdsell, R. S. Willms, Effect of Inlet Conditions on the Performance of a Palladium Membrane Reactor, Proceedings of the 17th IEEE/NPS symposium on furion engineering, San Diego, CA, (Oct. 6-10, 1997) pp. 301-303.

 

Figure 1. Palladium Membrane Reactor (not to scale)

 

Figure 2. PMR System Process Flow Diagram