Orbital Welding of Corrosion Resistant Materials for Bioprocess Piping Applications

Orbital Welding of Corrosion Resistant Materials for Bioprocess Piping Applications

ABSTRACT

Orbitally welded 316L stainless steel tubing has long been considered as an industry standard for pharmaceutical and bioprocess technology piping systems. Recently, there has been an increasing demand for an alloy having a greater resistance to corrosion than 316L stainless steel for use in bioprocess environments. To be of general use in the industry, it would be desirable if this hypothetical alloy could be installed by conventional methods without a loss of corrosion resistance. We have selected two corrosion resistant alloys, a super austenitic stainless steel, AL-6XN®, and a nickel-based alloy, Hastelloy® C-22 to determine whether these materials could be welded successfully with automatic orbital fusion (autogenous) welding equipment using the GTAW (TIG) process. These materials have been welded and subjected to accelerated corrosion tests to determine whether the initial corrosion resistance of these alloys would be retained after fusion welding. In addition, AL-6XN was fusion welded by the orbital TIG process using an insert ring of Hastelloy C-22 to determine whether the addition of filler material enriched in alloying elements would improve the corrosion resistance of the weld and heat-affected-zone of the parent material. Metallographic examination of the welds have been done to determine the effects of welding on the grain structure of the corrosion-resistant alloys.

INTRODUCTION

At last year's Bioprocess Technology Symposium at the ASME Winter Annual Meeting in Chicago in November, 1988, there was a great deal of interest and discussion about the selection and use of corrosion-resistant materials for bioprocess piping applications. It was pointed out that under the severe service environments encountered in biotechnology, an alloy that would provide greater corrosion resistance than the most commonly used corrosion-resistant alloy, 316L stainless steel, would be advantageous for some applications. In order to be of general use for the industry, this hypothetical material must be available, and it should be suitable for installation by conventional construction methods typically used by mechanical contractors to install pharmaceutical/ biotechnology piping systems. Autogenous (fusion) welding by the automatic orbital TIG process has been accepted as an industry standard because of the consistent high quality of the welds and because of the smooth inner weld bead that minimizes crevices where corrosion can begin and where bacteria could become established in the piping system.

Thus, it would be preferable to select an alloy that would retain its corrosion resistance following standard welding procedures, and one that is suitable for fabrication into fittings, valves, and other piping system components.

It must be recognized that the corrosion resistance and service lifetime of a piping system depends as much upon the corrosion resistance of the welds as on the corrosion resistance of the parent material. This has not been a serious problem with orbitally welded 316L stainless steel since, although there is inevitably some loss of corrosion resistance following welding, this effect is comparatively slight.7 It should be noted however, that with more highly alloyed material, there is a greater tendency for segregation of alloying elements to occur during welding.8 This leaves the heat-affected-zone (HAZ) adjacent to the weld vulnerable to corrosive attack.8 This is especially true for materials alloyed with molybdenum. While the corrosion resistance is increased for alloys high in molybdenum, there is also a greater difference between the corrosion resistance of the welded as compared to the unwelded materials.8,9

During manual TIG welding, the tendency of the alloying elements to segregate can be overcome, in some cases, by the addition of filler material enriched in alloying elements. However, manual TIG welding with filler wire is a less suitable welding process for the construction of high-purity bioprocess piping installations. Although the addition of filler material is possible with the automatic orbital TIG process, the equipment with this provision is more expensive and requires more training of welding personnel than fusion welding equipment. Alternatively, it would be far simpler and much less expensive to provide alloy enrichment to an orbital fusion weld by using an insert ring instead of filler wire.

While the corrosion resistance of most alloys is well documented, the data on the corrosion resistance of welds is less common. Where such data exists, it generally refers to manual welds. The purpose of this paper is to examine two examples representative of alloy categories8 having high resistance to corrosion, to determine whether these materials can be successfully installed using automatic orbital fusion TIG welding techniques. The use of an insert ring enriched with alloying elements has been explored to test its practicality as a method of retaining the corrosion resistance of automatic orbital fusion welds. Accelerated corrosion tests have been performed on the welds to determine the degree of difference in resistance to localized corrosion between the weld and heat-affected zone (HAZ), and the base metal. Metallographs have been made of some of the welds to show the effect of welding on the grain structure. It should be noted that alloy selection was not intended to be an endorsement of any particular candidate alloy, but rather to illustrate the kinds of considerations with respect to welding that must be addressed before adopting any new alloy for general use in the bioprocess industry.

MATERIALS

Alloy Selection


The two corrosion-resistant alloys selected for study were a 6% molybdenum super-austenitic alloy, AL-6XN® from the Allegheny Ludlum Corporation of Pittsburg, Pennsylvania, and a nickel-based alloy, Hastelloy® C-22 from Haynes International of Kokomo, Indiana. AL-6XN is alloyed with 20% Cr, 24% Ni, 6.5% Mo and 0.2% N. Hastelloy C-22, which is mostly nickel, contains 22% Cr, 3% Fe, 13% Mo, and 3% W. (See Table 1.) Thus, of the two, Hastelloy C-22 is more heavily enriched with alloying elements, is more corrosion-resistant, and more costly.

Tubing Sizes. The sizes of tubing selected for study were considered to be representative of tubing diameters and wall thicknesses appropriate for bioprocess piping applications. AL-6XN was available in two sizes: 1.000" outside diameter (OD) with a wall thickness of 0.060", and with an OD of 0.875 and a wall thickness of 0.028". The Hastelloy C-22 tubing was 0.875" OD with a wall thickness of 0.065. A 1-1/2 inch Hastelloy C-22 pipe, 1.900" OD with a wall of 0.109", was also fusion welded with orbital TIG equipment, but no corrosion testing of this size was done.

Insert Rings. Inverted-tee insert rings were machined from sections of 3/4 inch schedule 10 Hastelloy C-22 pipe (1.050" OD) as shown in Figure 1.

WELDING PROCEDURES

Welding Equipment


The welds were done with standard orbital tube welding equipment consisting of a solid-state DC power supply, associated cables, and an enclosed weld head. The weld head contains an internal rotor which holds a tungsten electrode, and which rotates around the work to do the weld. The portable power supply, which plugs into 115V VAC, controls the entire weld sequence including an inert-gas pre-purge, arc strike, rotation delay, rotational speed (RPM), and four timed levels of welding current with pulsation. This is followed by a downslope which gradually terminates the current, and a postpurge to prevent oxidation of the heated metal. These weld parameters are dialed into the power supply from a weld schedule sheet after determination of the proper parameters from test welds done on tubing samples. Fusion welds with automatic orbital TIG welding equipment is practical on tubing or small diameter pipe in sizes from 1/8 inch OD tubing to 6 inch schedule 10 pipe (6.625 inches OD, 0.134 inch wall), and on wall thicknesses up to 0.154 inches.

Weld Parameters and Weld Appearance

AL-6XN in both tubing sizes was easily weldable with weld parameters, including travel speed (RPM) and weld currents, comparable to those used to weld 316L stainless steel. Weld parameters including RPM, welding currents, arc voltage and time were recorded on a chart recorder during the welds. Chart records are shown in Figure 2. The weld appearance was excellent with a smooth, shiny, flat weld bead on both the OD and ID. For welds with inverted-tee insert rings, the inserts were simply placed between the two sections of tubing to be welded, and fusion welded as usual, except for a slight increase in welding current to compensate for the increased thickness of material contributed by the insert ring. These welds also had a pleasing appearance, with a slight crown on the OD and some inner-bead reinforcement.

Because the addition of nitrogen to the purge gas has been shown to increase the corrosion resistance of welds on AL-6XN, argon gas containing 5% nitrogen was used on some of the AL-6XN tubing, while pure argon gas was used on other samples.5

Hastelloy C-22 samples were welded with a mixture of 95% argon gas and 5% hydrogen. This gas mixture was used to provide a hotter arc, and thus more penetration, at the same weld current as a comparable weld done in pure argon. The Hastelloy weld had an acceptable appearance with a somewhat coarser weld bead than a comparable weld on stainless steel. There was a slight tendency towards concavity that was more apparent on the 1-1/2 inch pipe sample. While the tube sample could be welded at normal travel speeds (5 IPM), the pipe sample required the Synchro function of the power supply. In Synchro, the movement of the rotor is synchronized with the current pulses so that rotation is stopped during the high current pulse to achieve maximum penetration. The travel speed in Synchro is normally about half that of a conventional weld. When the wall thickness is over about 0.100" on a fusion weld done in the 5 G (horizontal) position, gravity affects the molten weld puddle so that at the 12:00 o'clock position (over the top) the weld bead tends to sag and become convex on the ID and concave on the OD. Pulsation and Synchro help to control the weld puddle and thus to minimize sagging, but some undercut was, nevertheless, apparent on the heavy-walled Hastelloy C-22 sample. This condition might be considered unacceptable for some applications.

CORROSION TESTING

The most reliable method to determine the suitability of an alloy, or a type of weldment, for a particular service environment would be to test it under the conditions of use over a period of months or years. Since this is usually not practical, accelerated corrosion tests have been developed. By raising the temperature and the strength of corrodents, the corrosion rate can be raised to produce measurable corrosion over a period of hours or days instead of years, thus providing an indication of a material's corrosion resistance in a particular environment.

Accelerated Corrosion Test

The major corrodents in bioprocess piping environments are chlorides. Chlorides are known to initiate localized corrosive attack, particularly pitting and crevice corrosion. Accelerated corrosion tests designed to predict the resistance to pitting or crevice corrosion in chloride environments are outlined in ASTM G-48. The material being tested is immersed in a solution of 10% FeCl3 . 6H20 for a specified period and temperature. The sample is weighed and observed after each testing period and re-exposed to the test solution at increasingly higher temperatures until failure occurs. Failure is defined as a weight loss of greater than 0.0001 gram /centimeter2 during any single test period. If the test is done without a crevice (Practice A), the temperature at which failure occurs is the critical pitting temperature for the material being tested. If a crevice is used (Practice B) the temperature at which failure occurs is the critical crevice corrosion temperature. Test procedures and periods vary somewhat with the testing center. Test samples must be run together if comparative results are desired.

AL-6XN Fusion Welds on Thin-Walled Tubing. Crevice corrosion testing was done on thin-walled (0.028") AL-6XN fusion weld samples according to ASTM G-48 Practice B. The exposure time was 72 hours for each temperature beginning at 100º F (37.8º C). The shield gas for two of the samples was pure argon, while two other samples were welded in the presence of 95% argon with 5% nitrogen. The tube samples were sectioned lengthwise and a teflon crevice was placed across the weld and held in place with rubber bands. Although some surface etching was apparent at lower temperatures (see Table 2.), failure of 3 of the 4 samples occurred at 125º F (51.7º C). The fourth sample, which had been welded in argon gas, had measurable corrosion, but no attack of the weld or HAZ at temperatures up to 140º F (60º C). The critical crevice corrosion temperature for AL-6XN base metal has been reported to be 43º C (110º F).1,3 Thus, with thin-walled tubing of AL-6XN, there appears to be little or no loss of corrosion resistance following orbital fusion butt welding.

Exposure of Heavier-Walled AL-6XN Tube Samples to 10% Ferric Chloride ASTM G-48 Practice A. Five welded samples of 1.000" OD, 0.065 wall AL-6XN tubing were subjected to corrosion tests in ferric chloride according to ASTM G-48 Practice A (without a crevice) to determine the relative corrosion resistance of various TIG welding techniques. The welds included an automatic orbital fusion tube-to-tube weld, a manual TIG weld using Inconel 625 wire as filler, and a manual TIG fusion weld. In addition, two automatic orbital welds using insert rings of Hastelloy C-22 were tested. On one of the welds with the insert rings, both the shield gas and ID purge were pure argon. The second insert ring weld was purged with a mixture of 95% argon and 5% nitrogen. The results of this test are shown in Table 3. At 25º, all of the samples were acceptable with no visible attack (NVA). The weld sample with the insert ring purged with the argon/nitrogen mix showed some visible, but non-measurable attack at 35º C, but showed no further attack until 65º. At 45º C, the manual TIG fusion weld failed with visible weld attack. At 55º, the only welds surviving were the two with the Hastelloy C-22 insert rings. Both fusion welds and the manual TIG welds with the Inconel 625 filler showed attack of the weld and HAZ. The remaining welds failed at 65º with corrosion noted mainly in the weld and HAZ and slight attack of the parent material. A different group of similar tube samples was immersed in boiling ferric chloride solution for 24 hours. Exposure to this extreme environment resulted in severe corrosive attack to all regions of all of the samples except for the weld with the Hastelloy C-22 insert that was purged with the argon/nitrogen mixture. On this sample, the parent material was severely corroded with the weld nearly separated from the tube on one side, but the weld was still intact. This demonstrates a definite advantage to the use of the insert ring and suggests an advantage from the addition of nitrogen to the purge gas for welding AL-6XN.

Metallographs of Welds in this group were done to determine the effects of welding with and without insert rings on the weld-metal microstructure.

Ferric Sulfate Environment. Welded samples of AL-6XN tubing, similar to the ones tested in ferric chloride, were exposed to a ferric sulfate environment according to ASTM A 262 Practice B. This is a strongly oxidizing environment. The results are shown in Table 4.

Corrosion Testing of Heavier-Walled AL-6XN and Hastelloy C-22 Samples. The heavier-walled tube samples (1.000" OD, 0.065" wall) of AL-6XN as well as the fusion welded sample of Hastelloy C-22 were subjected to corrosion testing in a solution containing 11% H2SO4, 1.2% HCl, 1% FeCl3, and 1% CuCl2, also known as "green death". This test is designed to indicate corrosion resistance in low pH, oxidizing acid environments.

DISCUSSION

The oxidizing chloride environments of bioprocess piping systems present a formidable challenge to materials. Although materials highly enriched in molybdenum have demonstrated superior corrosion resistance in these environments, previous welding research with manual TIG welding indicates that fusion welds are considerably less corrosion resistant than the parent materials. Our results suggest that automatic orbital fusion welding, with the use of an alloy-enriched insert ring, may be an excellent method for joining these materials, making both the materials and the process a practical as well as economical choice for bioprocess piping applications.

Crevice-corrosion tests of orbital fusion welds on thin-walled tubing of AL-6XN gave satisfactory results. it would be difficult to weld very thin-walled tubing such as this using filler material with the manual TIG process. It could probably be done with orbital TIG with an insert ring, but may not be necessary. On the thin-walled samples that were tested, there was no apparent advantage in corrosion resistance provided by the addition of 5% nitrogen to the shield gas.

The corrosion tests of the various samples of AL-6XN in ferric chloride indicate that the best-performing weld was the orbital fusion weld with the Hastelloy C-22 insert ring welded in the presence of 5% nitrogen added to the argon purge. This is consistent with the findings of Kearns et al., 1987 who found that nitrogen enhanced the corrosion resistance of FeCrNiMo alloys used in the Chemical Process Industries. Presumably, nitrogen in the purge gas helps to prevent the loss of nitrogen from the parent material to reduce the incidence of pitting of the base metal. This result is indicative, but with this limited sample size, must be considered preliminary until further testing can be done. The next best weld with respect to corrosion resistance was the fusion weld with the C-22 insert ring done with argon gas. This failed at 60º C compared to the manual TIG weld with the Inconel filler material which failed at 55º. The superior performance of fusion welds with C-22 insert rings is consistent with the hypothesis that enriched insert rings are effective in counteracting the effects of molybdenum segregation that are bound to occur during welding.

Both manual and orbital fusion welds failed at temperatures below those welded with filler. However, the orbital fusion weld of the heavy-walled AL-6XN material appeared to be somewhat better than the manual fusion TIG weld. This might be expected on the basis of the more controlled heat input of orbital welding compared to manual welding. In addition, the starting and stopping of the arc done in manual welding may increase the tendency towards molybdenum segregation with detrimental effects on corrosion resistance of the weld.

Our results with the ferric sulfate environment did not discriminate between the different kinds of welds done on the AL-6XN. Ferric sulfate provides an oxidizing environment without chlorides and is not particularly useful as an indicator of molybdenum segregation.

For Hastelloy C-22, no alloy or insert material exists that is more highly enriched in alloying elements than the base metal. Therefore, the chemistry of the fusion weld metal would not be expected to differ significantly from that of a weld done with an insert or filler metal of C-22. (Lee Flasche, Haynes International, personal communication). Hastelloy C-22 tubing, joined by orbital fusion buttwelds, have been successfully used for the transport of HCI gas in the semiconductor industry. While 316L is not attacked if this gas remains free of moisture, in the presence of very low concentrations of moisture, hydrochloric acid is formed and aggressive corrosive attack occurs. Since some moisture contamination of these process gas lines is inevitable, semiconductor manufacturers have sought an alloy that could resist this severe attack. Fusion-butt-welds of thin-walled Hastelloy C-22 tubing has been highly successful in this application.

CONCLUSIONS

While preliminary, our results indicate that automatic orbital welding is a very promising joining technology for use with highly-alloyed corrosion resistant materials in bioprocess piping applications. While orbital fusion butt welding alone may be satisfactory for some alloys or for thin-walled tubing of other alloys, it appears likely that the use of an insert ring for fusion welding of alloys with high molybdenum content will have industry-wide applications.

By Barbara K. Henon, Ph.D., Arc Machines, Inc.

Reprinted From Bioprocess Engineering Symposium. Editors: T. E. Diller, R. M. Hochmuth, and Y. 1. Cho Mechanical Engineers Book No. H00579 - 1989


ACKNOWLEDGEMENTS

We gratefully acknowledge the contributions and assistance of the following individuals and companies:

Materials: Tubing of AL-6XN was supplied by Jack Mauer, Sales and Marketing Manager of the Allegheny Ludlum Steel Corporation, Brackenridge, Pennsylvania.

Corrosion Testing: Crevice-corrosion testing of thin-walled samples of AL-6XN were done by the Allegheny Ludlum Steel Corporation, Brackenridge, Pennsylvania.

Corrosion testing with ferric chloride and ferric sulfate solutions of the heavy-walled samples of AL-6XN were done by the Engineering Test Center, Engineering Service Division of E.I. Du Pont De Memours & Company, Newark, Delaware. Gordon Gaesser coordinated the testing, and was advised by Dr. Joe Dimo.

Additional corrosion tests on heavy-walled weld samples of AL-6XN and Hastelloy C-22 were done by Dr. Narafi Sridhar and Test Engineer, Lee Flasche of the Corrosion Service Center, Haynes, International, Inc., Kokomo, Indiana.

Metallography: Metallographs of some of the heavy-walled AL-6XN welds were done by the Engineering Test Center, Engineering Service Division of E.I. Du Pont De Memours & Company, Newark, Delaware.

REFERENCES

1. Franson, I.A., and Maurer, J.R., An Advanced Fe-Ni-Cr-Mo-N Austenitic Alloy for Utility Heat Exchanger Service. The International Conference on Advances in Material Technology for Fossil Power Plants. Chicago, Illinois, 1-3 September, 1987.

2. Haynes International, Inc. Hastelloy Alloy C-22. Haynes International Corrosion-Resistant Alloys, pp. 1-23, copyright, Kokomo, Indiana, 1988.

3. Johnson, M.J. The Beneficial Effects of Increasing Cr, Ni, Mo, and N on the Corrosion Resistance of Stainless Steels. Bioprocess Engineering Symposium pp. 53-58, ASME Winter Annual Meeting, Chicago, 1988.

5. Kearns, J.R. and Deverell, H.E. The Use of Nitrogen to Improve the Corrosion Resistance of FeCrNiMo alloys for the Chemical Process Industries. Presented during CORROSION/86, Paper No. 188 NACE, Houston, TX, 1986, Reprinted from MP, Vol. 26, No. 6, pp. 18-28, June, 1987 Copyright 1987 by the National Association of Corrosion Engineers, P.O. Box 218340, Houston, Texas 77218

6. Kiikvee, W. and Kaufman, J. Resistance of Alloys to Corrosion by Hydrogen Chloride with Reference to Gas Control Components. Advanced Pressure Technologies, San Rafael, California, 1988.

7. Mehalko, T. The Science of Welding for Bio-Processing Application. Bioprocess Engineering Symposium, pp. 71-87, ASME Winter Annual Meeting, Chicago, 1988.

8. Rahoi, D.W. Technical Considerations in the Selection of Nickel Alloys and Stainless Steels for the Bioprocess Industry. Bioprocess Engineering Symposium pp.41-48, ASME Winter Annual Meeting, Chicago, 1988.

9. Whitcraft, P.K. Considerations in Selecting a Stainless Steel for Resistance to Chloride Pitting and Crevice Corrosion. Bioprocess Engineering Symposium pp. 59-62, ASME Winter Annual Meeting, Chicago, 1988.

TABLES

 


 



Figure 1. Sketch of insert ring machined from a 1-1/2 inch pipe of Hastelloy C-22 and used as filler material in fusion welds of AL-6XN tubing.


Figure 2. Chart records of welds on 1" OD AL-6XN tubing. Top lines are pulsed welding currents (4 levels in A). Center trace indicates arc voltage, and bottom trace (A only) indicates RPM. A. automatic orbital fusion TIG weld. B. Manual fusion TIG weld. Both welds were done in a single pass. Note starting and stopping in the manual weld.