Performance Testing of a Prototype Pd-Ag Diffuser

Gregg A. Morgan

Brittany J. Hodge

37th Tritium Focus Group, Rochester, NY October 25-27, 2016

SRNL-STI-2016-00551-S

2

Outline

• Background and Introduction

• Design of Prototype Pd-Ag Diffuser

• Experimental Design and Set-Up

• Results Bleed Flow Rate vs. Total Feed Flow Rate Curves Amount of H2 permeating vs. Total Feed Flow Rate Curves Amount of H2 in Bleed Stream vs. Total Feed Flow Rate Curves Average Shell Pressure vs. Total Feed Flow Rate Curves Tube Pressure Drop vs. Total Feed Flow Rate Curve Residual Gas Analysis Spectrum

• Summary and Conclusion

• Acknowledgments

3

Introduction

• The Savannah River Site Tritium Facilities currently uses Johnson-Matthey (JMI) Pd-Ag diffusers for the separation of hydrogen isotopes from various inert species on the process stream

• Several years ago (circa 2012), JMI decided to exit the business of manufacturing Pd-Ag diffusers

• Over the last several years, SRNL has been heavily involved in the identification of potential vendors for the manufacture of tritium compatible Pd-Ag diffusers

• Potential Challenges of Finding a Vendor: • Limited number of domestic vendors • Pd-Ag diffusers are typically custom components based on flow rates and

pressures • Long Lead times • Tritium service requirements

4

Background

• Most vendors manufacture hydrogen purifiers instead of diffusers/permeators

Hydrogen Purifier

Designed to purify hydrogen

Not typically designed for

tritium service

Hydrogen isotopes remain in the bleed stream and would require additional

processing

Hydrogen Diffuser/Permeator

Typically designed for once through application where nearly all of the hydrogen

isotopes are removed from the process stream.

5

Operation of a Pd-Ag Diffuser

• The H2 in the feed gas permeates through the wall of the Pd-Ag tubing into the lower pressure cavity of the diffuser shell and is known as the permeate stream. The non-hydrogen gases of the feed stream continue through the tubes and discharge to the bleed or raffinate stream.

6

Design of Prototype Pd-Ag Diffuser

• The prototype Pd-Ag diffuser uses a micro-channel membrane technology

Pd-Ag membrane

Stainless Steel

Images taken from http://www.powerandenergy.com/

7

Experimental Manifold

2TS test system

• Mass flow controllers up to 20 SLPM

• Pressure transducers

• Back pressure regulators

• Dycor Dymaxion DM200M Residual gas Analyzer (RGA)

• Inficon 3000 micro gas chromatograph

8

Gas Composition Pump Type Tube Pressure (Torr)

2% H2 / 98% N2 Normetex 15 700

50% H2 / 50% N2 Normetex 15 700

96% H2 / 4% N2 Normetex 15 700

2% H2 / 98% N2 Edwards nXDS15iR 700

50% H2 / 50% N2 Edwards nXDS15iR 700

96% H2 / 4% N2 Edwards nXDS15iR 700

2% H2 / 98% N2 Normetex 15 1140

50% H2 / 50% N2 Normetex 15 1140

96% H2 / 4% N2 Normetex 15 1140

2% H2 / 98% N2 Edwards nXDS15iR 1140

50% H2 / 50% N2 Edwards nXDS15iR 1140

96% H2 / 4% N2 Edwards nXDS15iR 1140

Experimental Parameters

• Characterization testing of the diffuser considered the effects of various feed gas compositions, feed flow rates, internal tube pressures and pump configurations on the permeation of H2 through the Pd-Ag tubes.

9

Bleed Flow Rate vs. Total Feed Flow Rate Curves – Tube Pressure and Pumps

• Each of the bleed versus feed curves for the three compositions at 700 Torr overlap almost exactly with the curves at 1140 Torr using both pump combinations. The increase in tube pressure has a negligible effect on the H2 permeation.

10

Amount of H2 permeating vs. Total Feed Flow Rate Curves

• The amount of H2 permeating through the membrane is nearly 100% for the 96% H2 and 50% H2 cases based on the calculation using the flow controllers and flow meters.

11

Amount of H2 in Bleed Stream vs. Total Feed Flow Rate Curves

• The amount of hydrogen in the bleed stream as calculated from the mass flow controllers and meters never meets the diffuser’s design criteria of 0.03%. This indicates that this diffuser design may not be suitable for a tritium purification system without an additional means to remove the tritium remaining in the bleed stream.

12

Comparison of Micro-GC analysis

• The micro-GC analysis of the bleed stream provided confirmation that the amount of H2 in the bleed stream is above the design criterion.

13

Average Shell Pressure vs. Total Feed Flow Rate Curves

• For all three gas compositions at both 700 torr and 1140 torr, the shell pressure is lower when the permeate stream is pumped by the Edwards nXDS15iC compared to the Normetex Model 15 scroll pump. This indicates the Edwards could be a potential replacement option for the Normetex pump.

14

Tube Pressure Drop vs. Total Feed Flow Rate Curve

• The relatively small pressure drop across the diffuser assembly indicates that multiple diffuser stages could be placed in series to remove hydrogen isotopes from the bleed stream without the use of additional process equipment.

15

RGA of Bleed Stream for a 96% H2/4% N2 gas mixture

• The hydrogen signal is very low (~1 x 10-9) for both of the cases indicating that most of the available hydrogen in the feed stream is permeating through the Pd-Ag membrane into the shell space of the diffuser.

16

RGA of Bleed Stream for a 50% H2/50% N2 gas mixture

• For the 50% H2/50%N2 gas mixture, nearly 100% permeation is only achieved up to 1500 sccm. This may be attributed to a blanketing effect caused by the nitrogen or to the small residence time through the diffuser.

17

RGA of Bleed Stream for a 2% H2/98% N2 gas mixture

• The hydrogen in the bleed stream is barely detectable in the RGA, indicating that very little hydrogen is present in the bleed stream. With a composition of only 2% H2 in the feed stream, it would be expected that only a small amount of hydrogen would be evident in the bleed stream.

18

Summary and Conclusions • Characterization tests on a new Pd-Ag diffuser design have been completed to

gain an understanding into how the permeation of H2 is affected by varying experimental test conditions. − Tests were completed at 96% H2, 50%H2, and 2% H2 with the balance N2

− Test pressures of 700 and 1140 torr

− Normetex and Edwards scroll pumps

• The results presented indicate that:

− This prototype Pd-Ag removes nearly 100% of the hydrogen isotopes at the flow rates for which the Pd-Ag diffuser was designed.

− This prototype Pd-Ag diffuser does not meet the initial design criteria of 0.03% hydrogen in the .bleed stream

− The residence time through the Pd-Ag tubes ranges from 5.3 seconds at 100 sccm to 0.17 seconds at 3000 sccm

− The remaining hydrogen isotopes present in the bleed stream will need to be processed using additional techniques in order to maximize the tritium recovery and minimize the environmental discharges if used for tritium service.

19

Path Forward • SRNL recommends continuing the search for additional manufacturers of

commercially-sourced Pd-Ag diffusers for testing over a range of hydrogen isotopes

• Optimize the test parameters and the experimental set up to possibly improve the permeation of the H2 and decrease the amount of H2 remaining in the bleed stream using the prototype Pd-Ag diffuser

• Three production Pd-Ag diffusers have been received from Power and Energy. These are similar in design to the prototype unit but contain more Pd-Ag tubes than the prototype unit. • Complete testing of the production units at various flow rates and gas

compositions • Compete testing of multiple Pd-Ag diffusers in series.

• 3-D radiographs of prototype Pd-Ag diffuser will be completed to obtain a better

understanding of the design and identify potential opportunities for improvement

20

Acknowledgements

• The authors acknowledges the contributions of – Anita Poore – Benton Randall – Henry Sessions – Kipplin Neikirk

• The authors gratefully acknowledge the financial support of the National Nuclear Security Administration NA-231 Convert Mo-99 Program Office.

21

Thank you for your attention!!!

Questions?