Effects of Purge Gas Purity and Chelant Passivation on the Corrosion Resistance of Orbitally Welded 316L Stainless Steel Tubing

Effects of Purge Gas Purity and Chelant Passivation on the Corrosion Resistance of Orbitally Welded 316L Stainless Steel Tubing

The article was published in two parts. Part 1 describes the material, welding, and purging techniques, and the potentiodynamic polarization and auger electron spectroscopy methodology. The results of the electrochemical measurement of the pitting potentials of approximately 154 weldments of 316L stainless steel tubing, together with a visual description of the weldments, are presented. Part 2 details the results of auger surface line scans to show the surface element segregation and the depth profile for chromium, iron, oxygen and manganese. Chromium to iron (Cr/Fe) surface ratios and depth profiles will be presented. The discussion and summary sections describe the significance of findings and practical applications of this knowledge.

Part One

Introduction

It has long been known that welding of a stainless steel tube results in some loss of corrosion resistance relative to the unwelded condition of the base metal. When tubes are welded with a backing gas on the inside of the tube (ID) containing oxygen in the purge gas, a colored oxidation or "heat tint" is formed in the heat-affected-zone (HAZ) on either side of the welded joint. Several investigators have demonstrated that the intensity of the color is proportional to the amount of oxygen in the welding gas, and that the critical pitting temperature (CPT) which is an indication of corrosion resistance, is reduced by increased amounts of heat tint.1,2,3 [The CPT is the temperature at which significant corrosion is noted or measured after subjecting a sample to incrementally higher temperatures in a corrosive solution for a specified period. The higher the CPT, the higher the corrosion resistance of the sample.]

Critical pharmaceutical piping systems are typically placed in service in the as-welded condition. Grinding, pickling or other means of removing heat tint are not suitable for this application, but passivation is done to reverse the detrimental effects of welding and fabrication. Previous studies4 have shown that treatment of unwelded AISI 316L stainless steel tubing with various chelant mixtures results in a significant increase in pitting potential as determined by cyclic potentiodynamic polarization testing. Using our test procedures following the guidelines of ASTM G-61-86 with 0.1 M KCl as an electrolyte, unpassivated samples of unwelded 316L stainless steel have pitting potentials (Epit) in the 100 mV range with a tendency of electropolished (EP) tubing to have slightly higher pitting potentials than mechanically polished (MP) tubing. Pitting potentials in the range of 200-250 mV are characteristic of passivated samples showing marginal or low resistance to corrosion while fully passivated samples having good corrosion resistance have pitting potentials of 300 mV or higher. The resistance of 316L tubing to chloride pitting will vary according to chemical composition, surface finish, the number and type of inclusions as well as other factors.

The aim of the present study was to determine the relative efficacy of several mixed chelant solutions and their comparative effectiveness with respect to nitric and phosphoric acids. The mixed chelants were tested on three different degrees of heat tint produced by three different levels of ID purge gas purity on mechanically polished (180 grit) and electropolished (10Ra) tubing which had been orbitally welded.

The following levels of oxygen concentrations in argon were selected for purging the inside of the tubing during welding: 8 ppm, the upper limit of oxygen permitted in standard welding argon gas; 100 ppm which previous authors3 have shown is high enough to result in a noticeable drop in critical pitting temperature, and argon gas purified to the low parts per billion (ppb) level. Corrosion resistance of test samples was determined by potentiodynamic polarization testing.

While the semiconductor industry routinely specifies oxidation-free welds for process gas lines, the pharmaceutical industry usually permits welds to have either a light straw color or an amount of oxidation specified and approved by the owner. It is generally assumed that passivation will effectively remove light heat tint. Electropolished (EP) tubing is universally specified for process gas lines, while pharmaceutical piping may be EP or mechanically polished (MP).

Field observations have supported the assumption that it is more difficult to obtain an oxidation-free weld on mechanically polished tubing than on electropolished tubing. In this investigation, results were compared on EP and MP tubing from the same material heat of 316L stainless steel.

Auger electron spectroscopy (AES) and electron spectroscopy for chemical analysis (ESCA) studies5,6 have shown that welding changes the surface distribution of elements, including oxygen, chromium, nickel, iron and manganese, across the weld and HAZ with an adverse effect on the passivity of the surface layer and that these changes could be reversed by passivation with a mixed chelant solution.

In the present investigation, Auger line scans were performed across welds on as-welded EP and MP tubing as well as on similar samples passivated with a mixed chelant formulation which was shown to be the most effective based upon the results of the potentiodynamic polarization measurements of pitting potential (Epit). In addition, Auger depth profiles were performed to evaluate the Cr/Fe ratio, oxide, and manganese (Mn) levels from the ID surface to a depth of up to 450 Å. Measurements were taken in 4 Å increments beginning at the ID surface of the weld bead, the HAZ and an area away from the HAZ of welded samples of EP and MP tubing in both passivated and unpassivated conditions.

Methods

Materials and Preparation


Seamless 316L stainless steel tubing of Heat # A341855 was supplied as two twenty foot lengths each of ultra-high-purity (UHP) cleaned and packaged electropolished (10 Ra) and mechanically polished (180 grit) material. The elemental composition for this heat number is shown in Table I. Coupons were cut, without the use of cutting oil, into three inch lengths using a saw and the ends prepared for welding in a square butt configuration with a portable end-preparation machine.

Tubing was handled with gloves and coupons kept in individual plastic bags until ready for welding. A total of 154 coupons each of EP and MP tubing were prepared for welding. 


Welding and Purging Procedure

Welding equipment was an Arc Machines, Inc. Model 207 microprocessor-based Power Supply and a Model 9-2500 weld head. A weld program was developed based upon an electrode travel speed of five inches per minute. Weld parameters were established on test coupons and stored as a weld program or schedule in the power supply memory. The same program was used for all the samples virtually eliminating heat input as a variable.


Figure 1. ID purge set-up for welding of tube samples.
A separate bottle of argon was used to purge the weld head. Pressure gauge and O2 monitor were used during set-up but were not in-place during welding.

The Model 9-2500 is an enclosed type weld head which forms a chamber of inert gas that protects the outside of the weld from oxidation during welding. At the beginning of the weld, shrinkage may cause a gap to form between the two tube ends allowing gas in the weld head to mix with gas inside the tube. As the weld progresses the gap closes up. A weld done in an enclosed head is much cleaner than a weld done in an open frame head where a gap opens up the tube ID to atmosphere. The flow rate of purge gas to the weld head was 30 CFH (cubic feet per hour), while the ID purge flow rate was 20 CFH. Purge gas was supplied and certified to be 108 and eight ppm oxygen in argon. For the purpose of this experiment, an oxygen concentration in argon near the upper limit of the concentration in standard welding gas was requested. The supplier specifies seven ppm as the upper limit for oxygen in their standard purge gas and eight ppm was delivered. Similarly, an oxygen concentration near 100 ppm was requested and was certified to be 108 ppm. For high-purity purging of the tube ID, a standard bottle of argon, four ppm oxygen, was processed through a purifier which removes oxygen and moisture to the low ppb levels. For that experiment standard gas was used to purge the weld head and the purified gas used to purge the tube ID. A schematic of the ID purge set-up is shown in Figure 1.

An oxygen analyzer was used during purging to establish when the purge inside the tube reached the source level of oxygen. Plastic tubing was used to distribute argon from the regulator and flowmeters on the argon cylinder to the power supply and tube ID. An internally-placed expanded purge plug was used on the inlet side of the tube ID and a rubber stopper with a length of tubing to a tee connector was used on the exit end of the sample. Oxygen levels on the tube ID were verified prior to each weld.

Following welding, the weld was visually inspected on the OD and on the ID using a mirror and a flashlight. Welded samples were placed in plastic bags and later cut to 1.5 inch lengths prior to passivation.

Chemical Cleaning and Passivation

After the weld coupons were cut to the final 1.5 inch length, faced and deburred, all specimens were pre-cleaned using an alkaline phosphate/chelant wetting agent blend to ensure removal of any residual dirt, debris, oil, grease or surface contaminant which could interfere with efficient uniform contact of the passivation treatment with the bare metal and weldment areas of the tube ID. The time of exposure was two hours at 160°F - 180°F followed by DI water rinsing to a neutral pH.

The nitric acid passivation procedure was performed in accordance with the ASTM A-380-94 Standard Practice for Cleaning and Descaling Stainless Steel Parts, Equipment, and Systems, Table A2.1, Part 1, Code F. The phosphoric acid treatment was performed using a commercial blended product for sanitary stainless steel cleaning at the recommended five percent concentration at 120°F for two hours of treatment.

The various chelant formulations were prepared from commercially available food grade or reagent grade chelating agents compounded in combinations to evaluate efficiency in maximizing corrosion resistance and Cr/Fe ratio in the passive layer. The chelants employed were polyfunctional organic acids containing hydroxyl and/or amine substituents selected on the basis of their ability to dissolve and sequester free iron and the detrimental oxide products of welding. Actual chelant formulations and processing conditions are proprietary information.*


Figure 2.
Potentiodynamic polarization curves for chelant passivated and unpassivated MP orbitally welded samples (a) and EP chelant passivated and unpassivated orbitally welded samples (b) welded at the three different oxygen concentrations in the ID purge gas.

Potentiodynamic Polarization (PP) Measurements

Potentiodynamic polarization scans were performed using an EG & G Princeton Applied Research (PAR) Potentiostat/Galvanostat Model 173 connected through a PAR Model 276 interface to an IBM compatible PC. PAR software Model 342C was used to obtain the data.

Potentiodynamic pitting scans performed in accordance with a modified version of ASTM G 61-86 were conducted in O. 1 N KC1 electrolyte solution. The tests were initiated after equilibration for 50 minutes at the open circuit potential at a potential slightly more negative than the corrosion potential, Ecorr. The potential was ramped at a rate of 10mV per minute and the scan terminated when the current density reached 5000 µA/cm2. This abbreviated cycle was employed for economy of time and because of the primary interest in determining the Epit value.

Triplicate as-welded (unpassivated) coupons were tested for Epit for all three purge gas purities, for both EP and MP, with the exception of the ultra pure gas EP. For the latter, a total of 8 replicate welds were tested in order to obtain a better estimate of experimental precision. Duplicates of MP and EP samples purged with standard gas and passivated, and single MP and EP samples purged with high-purity gas and the lowest purity gas respectively, followed by passivation were scanned as described above.

Auger Electron Spectroscopy (AES) Surface Line Scan and Depth Profile

AES Surface Line Scan and Depth Profiles were performed at Photometrics, Incorporated, Materials Characterization Laboratory of Huntington Beach, California, employing a Perkin Elmer Model 590 Physical Electronics Spectrometer with interfacing and software upgraded. Operational parameters for the quantitative surface line scan across the weldment ID surface were as follows:

Depth profiles were obtained by sputtering the specimen with a 2.3kV argon ion beam at a pressure of 1 x 10-7 Torr. Sputtering rate was approximately 50 Ångstroms / minute referenced to Ta2O5.





Results

Potentiodynamic Polarization (PP) Tests
Mixed Chelant Solutions: The ASTM G61 Cyclic Potentiodynamic Polarization Standard Practice details the method for evaluating the pitting characteristics of stainless steel and provides for determination of corrosion potential (Ecorr), pitting potential (Epit), and repassivation potential (ERP). While Epit, and the potential difference between Epit and Ecorr (Epit - Ecorr) are widely used as a measure of susceptibility to pitting corrosion, this study focused on the value of Epit as the primary parameter to evaluate the passivity of the stainless steel surface. Good correlation has been reported,7,8,9 for Epit measured in 0.1M KCl and Cr/Fe atomic ratio as measured by AES.

A total of 77 samples including an equal number of welded samples as well as four unwelded tubes each of both electropolished and mechanically polished tubing were subjected to potentiodynamic polarization tests. Potentiodynamic polarization curves are shown in Figure 2 for MP samples (Figure 2a) and EP samples (Figure 2b), which were orbitally welded with the three different purge gas conditions shown in Table II. For each condition a polarization curve for an untreated and for a passivated sample are shown in Figure 2.

Results of Epit measurements are shown in Table II. A and B and plotted on the graphs in Figures 3-6. The improvements of the passive properties due to passivation in the chelant solution are evident by the decrease of the passive current density and the increase in Epit, which was more pronounced for EP samples (Figure 2). For unpassivated MP samples welded in the standard gas atmosphere or with the least pure gas, passive behavior was not observed (Figure 2a). The polarization curves resembled those of carbon steel in 0.1 M KC1. In Table II, these samples are indicated as "ACT". In some cases passivity was indicated, but Epit was quite low and could not be determined accurately. In Table II, these samples are listed as N/A and Epit was assumed to be more negative than -100 mV vs SCE.

If all of the Epit values from all levels of gas purity are averaged together, passivated samples of EP tubing have an average Epit of 385 mV (n=23). The MP passivated samples had an average Epit of 341 mV (n=18). For EP welded samples this represents an improvement of 268 mV over unpassivated samples, while for the MP welded samples an improvement of 224 mV for passivated samples was realized. Clearly there is a very significant improvement in the Epit of passivated welded samples over unpassivated samples. (Figure 3)

With the exception of the data for the MP tubing welded with 108 ppm oxygen in the purge which was N/A, the Epit values of welded samples were not significantly different from those of unwelded tubing samples obtained with a different heat of 316L and shown in Figure 4.4 The as-received orbital welds had similar Epit values to results obtained with unwelded, unpassivated tube samples. Epit values of passivated welded samples were comparable to passivated unwelded tubing. These results indicate that orbital welding of 316L stainless steel with standard or purified purge gas has no detrimental effect on the resistance to pitting. (Figures 4 and 5)

With the orbitally welded tube samples, the highest Epit values were obtained on EP tubing for mixed chelant solution BN with an Epit of 576 ± 71 mV (n=2), and chelant solution AN which had an Epit of 573 ± 17 mV (n=2). Both measurements were obtained from samples welded with the standard purge gas. For the EP tubing this represents an increase in 459mV over unpassivated samples, a 214 mV higher Epit than phosphoric acid, and a 225 mV improvement over passivation with nitric acid. For MP tubing, the mixed chelant, BN, had an Epit value of 510 ± 0 mV (n=2), which was 233 mV higher than the Epit for phosphoric acid. The AN chelant solution had an Epit 108 mV higher than that of the sample passivated with phosphoric acid, but no comparison was possible with nitric acid on MP tubing since the nitric acid sample had a pitting potential too low to measure. Orbitally welded samples treated with other chelant solutions had Epit values in the same range as samples passivated with nitric and phosphoric acids as shown in Figure 6.

Effects of Purge Gas Purity on Epit

The effects of purge gas purity on Epit appeared to be insignificant on as-welded unpassivated samples. It maybe significant that all three of the MP samples welded with the lowest purity gas were not passive, but all of the unwelded EP and MP samples (n=4) had Epit of less than 100 mV SCE (Table II A and Table II B).

A total of 5 MP samples were welded with argon containing 108 ppm oxygen, and only one of these gave a measurable result as shown in Table III. This was the sample treated with the mixed chelant solution, AN, which had an Epit value of 155 mV compared to a value of 265 mV for the AN sample welded with high purity gas. The AN samples welded with standard gas had an Epit of 385 ± 60 mV (n=2). No significant difference in Epit is seen between MP samples welded with standard gas and those welded with highly purified gas. Welding EP tubing with an argon purge containing 108 ppm oxygen resulted in no detrimental effect on Epit. However, distinct visual differences were observed in samples welded at the different gas purities after exposure to the PP testing as shown in Figures 7 and 8.


Figure 3. Average pitting potentials (Epit) of all passivated and unpassivated orbitally welded samples of electropolished and mechanically polished tubing. Pitting potentials of passivated weld samples were significantly higher than those of unpassivated welded samples. Electropolished tubing has somewhat higher Epit than mechanically polished tubing. Data from MP tubing welded with 108 ppm oxygen in the purge gas were not included since they failed to provide meaningful results. Data from orbitally welded samples purged with standard or purified gas show similar pitting potentials to those of unwelded passivated and unpassivated tubing from another heat of 316L.
(See Figures 4 and 5)


Figure 4. Pitting potentials of passivated and unpassivated unwelded 316L tubing (Lin, 1995). Data from the two chelant formulations with the highest pitting potentials were averaged and compared to unpassivated samples and samples passivated with nitric acid and phosphoric acid. Mechanically polished tubing was 180 grit polished and electropolished tubing had a 10 Ra µinch finish. Note: Data shown in Figures 4 and 5 not from same heat of 316L.


Figure 5. Pitting potentials of unpassivated and passivated EP AND MP 316L welded tube samples. All available data from the best two mixed chelant solutions were averaged. Passivation with the best mixed chelants improved Epit of welded MP tubing by 281 mV and EP tubing by 430 mV over as-welded unpassivated samples. The data show no detrimental effects of orbital welding on Epit with the exception of MP tubing welded with 108 ppm oxygen in the purge.


Figure 6. Pitting potentials of orbitally welded 316L tubing samples. Pitting potentials of mechanically polished and electropolished tube samples welded with standard gas (8 ppm oxygen) are plotted on the graph above.


Effects of Electropolishing and Mechanical Polishing on Epit

Electropolished tubing had somewhat higher pitting potentials than mechanically polished for both welded and unwelded tube samples (Figure 3-5). In addition EP tubing seemed more resistant to the detrimental effects of welding at all purge gas oxygen levels but particularly at oxygen levels greater than 100 ppm. Four out of five (80 percent) of the MP samples purged with 108 ppm oxygen had active polarization curves (shown as ACT in Table II B) while all five of the EP tubing samples, both as-welded and passivated, had measurable pitting potentials.

For samples welded with standard purge gas purity, two EP samples out of 19 (11 percent) were shown to be N/A with pitting potentials too low to be measured with accuracy, while of the MP weld samples four out of 18 were found to be N/A and one was ACT for a total of 28 percent. For UHP purge gas, zero N/A's were recorded for either EP or MP tubing. In every category, and most dramatically in the case of the lowest purity purge gas, orbitally welded EP tubing resisted failure. The results with the very low pitting potentials and samples without passivity (ACT) show that both electropolishing and increasing purge gas purity significantly reduce or prevent the loss of corrosion resistance associated with welding.

Visual Evidence of Oxidation (Heat Tint)

As expected, the visible amount of oxidation on the welded samples was proportional to purge gas purity. Purge gas with 108 ppm oxygen produced brownish rust-colored heat tint on the weld bead extending into the HAZ on both sides of the weld. The color was much more pronounced on MP tubing. EP tubing purged with purified gas had barely visible straw color. Some of the samples had a slight blue color at the start of the weld, probably as a result of mixing of gas on the tube ID with gas from the weld head which was standard gas. MP tubing, even with the highly purified gas showed light heat tint in the HAZ of the weld.

Passivation was not effective in removing heat tint discoloration from any of the weldments with the exception of the very lightest heat tint found on the EP welded samples purged with UHP gas. This is to be expected as chemical passivation is known to be effective only to a depth of 30Å to 50Å,5 while the depth of colored oxide bands in the HAZ of welds have been measured to depths exceeding 1000Å.1

Effects of Purge Gas Purity - ELECTROPOLISHED TUBING

<0.1 PPM OXYGEN

8 PPM OXYGEN

108 PPM OXYGEN
Figure 7. Visual observations of corrosion testing. Welds done on unpassivated electropolished tubing purged at three different oxygen concentrations in the ID argon purge after PP testing. Surface corrosion bands on either side of the weld bead are visible on the sample welded with the lowest purity gas. The surface corrosion band is only visible in the overlap and downslope area of the weld on the sample purged with 8 ppm oxygen in the argon. No surface corrosion band is seen on the sample purged with less than 0.1 ppm oxygen in the purge. Pits are visible on all samples, but the density of pits/cm2 increased with increasing oxygen concentration.


 
   Effects of Purge Gas Purity- MECHANICALLY POLISHED TUBING
 
<0.1 PPM OXYGEN
 
108 PPM OXYGEN
 
108 PPM OXYGEN

   Figure 8. Visual observations of corrosion testing. Weld samples on unpassivated mechanically polished tubing purged at three different oxygen concentrations in the ID argon purge. Note bands of surface corrosion on either side of the weld bead which are proportional to the amount of oxygen in the purge. The heaviest concentration of surface corrosion is in the area of weld overlap and downslope.


Effects of Chelant Passivation - ELECTROPOLISHED
 
108 PPM OXYGEN - AS WELDED
Epit = 178 mV
 
108 PPM OXYGEN - CHELANT
PASSIVATED, Epit = 392 mV
 
   Figure 9. Visual observations of chelant passivation. Weld samples on electropolished tubing purged with 108 ppm oxygen in the ID argon purge and submitted to corrosion testing. The top sample was unpassivated, while the lower sample was passivated with the mixed chelant which gave the highest Epit value. Surface corrosion bands on either side of the weld bead are visible on the top sample, but are considerably reduced after chelant passivation. Light straw and blue colored heat tint are still visible on the passivated sample. Note pitting distribution across base metal, HAZ and weld bead.



Visual Evidence of Corrosion


As the PP test parameters are set so as to cause pitting corrosion, i.e., reach Epit, every welded tubing sample that had been subjected to testing showed definite visible evidence of corrosion. Localized corrosion in the form of pits was found in the weld bead, HAZ and base metal of all of the samples. The pits were distributed differently on the EP and MP tubing. Pits were concentrated at a high density in the HAZ of the MP tubing, while they were more uniformly distributed in a less-dense pattern on the base metal, weld bead and HAZ of the EP tubing as shown in Figures 7-9.

In addition to the pitting, a generalized surface "etching" band which appeared to be a type of intergranular corrosion, was found in varying amounts in the HAZ on both sides of the weld bead. The amount of surface etching was found to be proportional to the amount of heat tint present on the welded samples, and like the heat tint, was much more pronounced on the MP samples welded at similar purge gas oxygen levels. The amount of surface etching was greatly reduced by passivation with the mixed chelant solutions as shown in Figure 9, but even after passivation, the amount of surface etching was proportional to the level of purge gas purity.


Part One References


1. Turner, S. and P.F.A. Robinson. The effect of surface oxides produced during welding on the corrosion resistance of stainless steels. Science. Volume 45, Number 9, 1989.

2. Kearns, J.R. and J.R. Maurer Welding guidelines to minimize service degradation of stainless alloys in bioprocess systems. Bioprocess Engineering Svmposium, ASME WAM, Atlanta, Georgia, pp. 109-117, 1991.

3. Hansen, J.V., Nielsen, T.S., and P. Aastrup. Root surface quality requirements- high efficiency purging or pickling? Paper 46, Conference on Duplex Stainless Steels, Glasgow, Scotland 13-16 November, 1994.

4. Lin, S.H. Corrosion protection of high copper aluminum alloys and stainless steels by surface modification. Ph.D. Thesis, University of Southern California, August, 1995.

5. Balmer, K.B. Auger surface line scan to evaluate element segregation in and chelant passivation of a 316L weld. Presented at the Bioprocess Engineering Symposium, ASME International Conference and Exposition, Chicago, IL., November, 1994.

6. Shirai, Y., Kojima, T., Miyoshi, S., Narazaki, M., and Ohmi, T. Specialty gas distribution system free from corrosion, gas decomposition and reaction- Perfect Cr2O3 treated tubing system. MICROCONTAMINATION Conference Proceedings, 1994.

7. Gramlich, C.T., O'Keefe, T.J., and Hsien-Tsung Liao. Characteristics of stainless steel surfaces after passivation with nitric acid solutions. Abstr. BED. Vol. 23 p. 63. Bioprocess Engineering Symposium, ASME WAM, 1992.

8. Wu, Chyang-Jen, Ph.D. Dissertation. Effect of surface treatment on localized corrosion of stainless steel. University of Missouri, 1989.

9. Seo, M and M. Sato. Epit shifts in a more noble direction with increasing chromium content. Trans Japan Institute of Metals, Vol. 21, No. 12, 1980.

 

Part Two

Effects of Purge Gas Purity and Chelant Passivation on the Corrosion Resistance of Orbitally Welded 316L Stainless Steel

AES Surface Line Scans

Distinct differences were observed in surface element segregation between the unpassivated weldments of mechanically polished (MP) as compared to electropolished (EP) tubing welded with standard purge gas (8 ppm oxygen). The scans were taken across a length of approximately 30 mm with the weld bead at the center in order to traverse all or most of the HAZ. Similar to the results previously published,5,6 the MP specimen displayed an iron peak at a concentration of 32-35 atomic percent across the weld bead (total length of approximately 6 mm) with little or no chromium or nickel present at the surface ( Figure 10a). A chromium peak of 18-19 percent occurs at a distance of six to eight mm from the centerline on either side of the weld corresponding to an iron concentration valley of eight to 10 percent. A distinct manganese peak of about 4.5 percent occurs at four to six mm at either side of the weld bead.

Figures 10. Auger quantitative line scans across orbital welds and HAZ of 316L tubing.

Figure 10a. Mechanically polished tubing - unpassivated weld. Auger line scan across an unpassivated weld on mechanically polished tubing. Note the increase in iron and decrease in chromium on the weld bead and chromium and manganese peaks in the HAZ.

Figure 10b. Mechanically polished tubing - mixed chelant passivated weld. Elements are more similar to unwelded tubing with favorable Cr/Fe ratio.



Figure 10c. Electropolished tubing - unpassivated weld. Note high iron and low chromium levels. Little or no manganese peaks in HAZ.

Figure 10d. Electropolished tubing - mixed chelant passivated weld. Passivation reduces iron and restores favorable Cr/Fe ratio.


In contrast, the element distribution across the surface of the weld and HAZ of the unpassivated EP (Figure 10c) weld does not display the peaks and valleys to the degree shown by the MP welds. The iron concentration across the entire 30 mm scan ranges from 26 percent to 34 percent and remains well above the other elements. Again chromium is absent from the weld bead and peaks slightly at about seven percent at five to six mm distance from the weld center line. It may be significant that the manganese peaks observed for the MP specimen do not appear on the EP Auger surface scan.

After passivation by the mixed chelant that gave the highest pitting potential, surface iron was reduced with attendent enrichment in chromium for both MP and EP welded specimens (Figures 10b and 10d). Surface Cr/Fe atomic ratios were calculated for these samples (Figures 11a and 11b). The magnitude of the Cr/Fe ratios for the MP sample before passivation ranges from zero at the weld bead to 3 to 5 in the HAZ at a distance of five to eight mm from the weld bead centerline. After passivation of the MP sample, the Cr/Fe ratio fluctuated from 2.0 at the weld bead to as high as 19 in the HAZ. Similar extreme Cr/Fe values have been seen in welded MP tube samples received from various construction sites. It should be realized that very high Cr/Fe values in the HAZ in the data from AES surface scans or depth profiles are not necessarily indicative of a passive condition.

For the EP sample (Figure 11b), the Cr/Fe ratio before passivation ranged from zero at the weld to 0.3 at the seven mm heat tint band to 0.7 at 13 to 14 mm as the scan exits the HAZ towards parent material. After passivation the Cr/Fe ratios ranged from 0.6 to a high of 2.7.


Figures 11a-b. Cr/Fe ratios across welds. Data from Auger surface line scans across welds to show the Cr/Fe ratios on orbitally welded electropolished tube samples comparing as-welded samples to samples passivated with a mixed chelant formulation. Figure 11a (top) shows data from mechanically polished tubing while Figure 11b (bottom) shows data from electropolished samples. Note difference in vertical scale.

Auger Depth profiles
Cr/Fe Depth Profiles

After each Auger line scan was concluded, an Auger depth profile was performed at three locations on the sample: 1) weld bead 2) darkest area of HAZ and 3) away from the HAZ. Figure 12a presents the Cr/Fe ratio depth profiles for the MP as-welded and BN chelant passivated welds and figure 12b shows the Cr/Fe ratio depth profiles for the EP as-welded and BN chelant-passivated welds.

As-Welded MP. The Cr/Fe ratio for the as-welded MP weld bead starts at the surface at essentially zero, i.e., little or no chromium, and increases slowly to the parent metal Cr/Fe ratio of 0.25 to 0.3 at a depth of ~70Å to 90Å (Figures 12 a and b).

The location analyzed in the HAZ of the as-welded MP sample shows a surface Cr/Fe ratio of 1.9 progressing to a peak Cr/Fe value of 4.8 at a depth of 50Å. The ratio drops towards the bulk alloy value beyond 200Å depth. The Cr/Fe ratio away from the HAZ still shows some enrichment at the surface with a Cr/Fe ratio of ~1.3 approaching the parent Cr/Fe value beyond 200Å. (Figure 12a).

BN Chelant-Passivated MP Weld Sample. The chelant BN process established a more "normal" passivation in the area of the weld bead (Cr/Fe = 2) and away from the HAZ (Cr/Fe = 1.4). However, in the HAZ, the Cr/Fe ratio attains extreme levels of 12.5 surface to 18.5 peak (10Å) with the removal of iron oxide as shown in Figure 12b.


Figure 12a. Cr/Fe ratio depth profiles - as-welded. Auger depth profiles were taken of the Cr/Fe ratio in the HAZ, weld bead and area outside of the HAZ of an MP unpassivated welded sample from the surface to a depth of 200Å. Note: Data shown in Figures 12a and 12b taken from different weld samples.

Figure 12b. Cr/Fe ratio depth profiles - passivated. Auger depth profiles were taken of the Cr/Fe ratio in the HAZ, weld bead and area outside of the HAZ of a passivated MP welded sample from the surface to a depth of 200Å. Note: Data shown in Figures 12a and 12b taken from different weld samples.

 

As-Welded EP
At the surface of the EP weld bead the Cr/Fe ratio is close to zero, and like the MP sample, the ratio increases to the parent metal value, 0.25 to 0.3 at ~70Å depth. In the HAZ the Cr/Fe ratio starts at 0.3, gradually increases to a maximum of 0.59 at 65Å to 80Å and decreases to parent value beyond 170Å . Away from the HAZ, more typical EP Cr/Fe ratios of 1.1 surface and 1.3 peak (at 15Å) are observed. (Figure 13a.)

BN Chelant-Passivated EP Weld Sample
The chelant passivated EP weld presented more typical Cr/Fe ratio depth profiles at each of the three locations, as might be expected in unwelded tubing. The only location with lower than expected values was that away from the HAZ. It is noteworthy that the two EP samples did not display the extremely high Cr/Fe ratios observed in the HAZ of the weld on MP tubing. (Figure 13b).


Figure 13a. Cr/Fe ratio depth profiles. Auger depth profiles were taken of the Cr/Fe ratio in the HAZ, weld bead and area outside of the HAZ of an EP unpassivated welded sample from the surface to a depth of 200Å. Note: Data shown in Figures 13a and 13b taken from different weld samples.

Figure 13b. Cr/Fe ratio depth profiles. Auger depth profiles were taken of the Cr/Fe ratio in the HAZ, weld bead and area outside of the HAZ of a passivated EP welded sample from the surface to a depth of 200Å. Note: Data shown in Figures 13a and 13b taken from different weld samples.

Figure 14a. Manganese Auger surface line scans. Data from Auger line scans shown in Figure 10 was plotted on an enlarged scale to show the distribution of manganese across an orbital weld and HAZ. Mn is present as peaks of 4.5 percent in the HAZ on both sides of an unpassivated weld and is reduced to 2.3 percent in the passivated weld sample. Data shown were from different weld samples. Manganese peaks were not seen on orbital welds on EP tubing.

Figure 14b. Manganese depth profiles. Auger depth profiles on passivated and unpassivated MP orbital welds showing distribution of manganese from the surface to a depth of 200Å. Manganese levels on these samples were not appreciably greater than the manganese content of the base metal level of 1.28 percent. Passivated and unpassivated data were taken from different welded samples.

Manganese Depth Profiles
The highest Mn concentration in the HAZ of the orbital welds done for this study was 4.7 atomic percent on the MP unpassivated weld sample (Figure 14a). The Mn peaks can also be seen on the Auger surface line scans in Figure 10a. Depth profiles taken in the darkest part of the HAZ of MP tubing from the ID surface to a depth of 450 Å did not show Mn levels higher than 1.5 percent as shown in Figure 14b. The Mn peaks observed in the surface scan were very narrow with a width of about 2 mm. The depth profile was taken in the darkest (most heavily oxidized) part of the HAZ and did not show increased levels of Mn significantly beyond the 1.28 percent Mn value listed for the heat of 316L parent material used in this study. Thus, if Mn was increased at deeper levels in the HAZ, it was in a narrow band missed by the depth profile.

The highest Mn value on the as-welded EP surface scan was 1.8 percent 5.5 mm from the weld centerline. EP depth profiles in the darkest part of the HAZ showed surface Mn levels of 1.5 -1.8 percent increasing to a peak of 3.0 percent at a depth of 198 Å which is not indicative of material deposited on the surface.

Mn has a lower boiling point (2150°C/3900°F) than iron or other alloying elements. It is present in the blue vapor which appears over the weld pool during the welding of 304 or 316 stainless steel10 and it precipitates in the HAZ where it has been implicated in loss of corrosion resistance.6

Previous unpublished work on a weld taken from a pharmaceutical construction site indicates that Mn may be deposited in the HAZ of welds at much higher levels than on the orbital welds described above. (See Figure 15). An Auger surface line scan was conducted on that unpassivated MP construction site weldment, after initially sputtering away the top 10Å layer and starting at the weld bead fusion boundary, traversing through the HAZ in one direction only. Mn in that sample displayed a peak concentration of 22 percent while the maximum concentration of Mn permitted in 316 base metal is only 2 percent. An Auger depth profile of the same sample was taken at ~3mm from the fusion boundary. Starting at about 17 percent at a depth of 10Å, the Mn decreased gradually to 10 percent at 60Å, and approached parent metal concentration at 100Å to 120Å. Since the deposition of Mn in the HAZ is detrimental to corrosion resistance, it will be important to determine the extent to which the Mn concentration of the base metal, surface finish and welding techniques, especially heat input, influence the amount of Mn precipitation during welding. This is particularly important because Mn may persist beneath the surface at levels inaccessible to passivation.


Figure 15a. Manganese surface scan of construction site weld. Manganese distribution from weld centerline across HAZ of a construction site weld as shown by an Auger surface line scan. Manganese reached a peak of 22 percent between 2 and 2.5 mm from the weld center. Compare to 4.5 percent Mn peak on orbital welds done for this study.

Figure 15b. Manganese depth profile of construction site weld. Manganese levels were 17 percent at the surface gradually returning to base metal levels at a depth of just over 100Å from the surface.
Oxide Depth Profiles in Heat Tint Band
of Orbital Welds - 8 ppm O2


Figures 16a and 16b. Oxide depth profiles. Auger data taken in the HAZ of orbital welds welded with 8 ppm oxygen in the purge. Graphs show atomic percentage of oxygen from the surfaces to a depth of 450Å. Top samples (16a) were unpassivated and bottom samples (16b) were passivated with the most effective chelant formulation. In both cases there is significantly more oxide which extends to a greater depth in mechanically polished compared to electropolished tubing.

Oxide Depth Profiles

Oxide depth profiles were determined for the as-welded (unpassivated) MP and EP samples and for the BN chelant processed MP and EP welds (Figures 16a and 16b) in the darkest heat tint band of the HAZ to a depth of 450Å. The atomic percent oxygen for the as-welded samples were 48 percent for the MP sample and 45 percent for the EP sample at the surface, peaking at 60 percent for the MP samples and 58 percent for the EP sample at 25Å for the EP and at 48Å for the MP sample. The oxide concentration is higher for the MP sample at every depth beyond 50Å.

The difference in surface roughness between EP and MP tubing may explain the greater amount of heat tint adhered to the MP surface and the relative depth of the oxide layers. The fact that the oxide layers extend to greater depth on the MP tubing may account for the increased amount of surface corrosion or "etching" observed on the MP welds compared to welds on EP tubing.


DISCUSSION



The heating and cooling which occurs during welding of austenitic stainless steels produces changes in the microstructure and surface condition of the welds which make the weldment more vulnerable to corrosion than the unwelded base metal. Improper purging resulting in excessive heat tint is a well-documented cause of corrosion in welds, with the amount of oxidation proportional to the loss of corrosion resistance.11 Rouging in WFI (water for injection) systems, after a year or so of service, has been reported to be associated with the HAZ of welds having visible heat tint. The HAZ of welds on 304L and 316L has been reported to be more susceptible to crevice corrosion in microorganism-containing water than the parent material, but similar welds from which the heat tint was removed by pickling, grinding, or electropolishing did not display the same vulnerability.12

Although treatments such as annealing, grinding away heat tint and the underlying chromium-depleted layer, or pickling with a solution of nitric and hydrofluoric acids, have been used successfully on stainless steel to undo the changes produced by welding,13 none of these treatments are practical for welds in pharmaceutical piping systems which are typically put into service in the "as-welded" condition.

Most welds in piping systems are inaccessible. Furthermore, pickling and grinding would produce unacceptable surface roughness which would be detrimental to the goals of cleanability and sterilizability. These treatments would be particularly detrimental to a smooth electropolished finish. The only postweld treatment available for welds in critical biopharmaceutical piping systems is passivation. Therefore, it is important to understand the beneficial effects of passivation and to determine the most efficacious passivating agents. It is also necessary to understand the limitations of the passivation process and to work within these limits by controlling the welding process in order to minimize the harmful effects produced by welding.

AISI type 316L, which contains 2-3 percent molybdenum, experiences a comparatively small loss of corrosion resistance during welding. Alloys which contain more molybdenum, which tend to have higher corrosion resistance to begin with, lose proportionately greater percentages of their corrosion resistance during autogenous welding and are usually welded with filler material overalloyed with molybdenum to counteract the effects of molybdenum segregation.14

The fact that 316L provides acceptable corrosion resistance to a broad spectrum of corrodents and can be welded autogenously without the need for postweld heat treatment has made it the standard material for use in critical biopharmaceutical and semiconductor piping systems. However, changes which occur in the HAZ of welds make this area particularly vulnerable. A typical weld-related corrosion problem is "sensitization" in which carbon precipitates out of the grains combining with chromium in grain boundaries as chromium carbide. This leaves a chromium depleted layer along the grain boundaries which is "sensitized" and lower in corrosion resistance, particularly in service environments where chlorides are present. The "L" grade of type 316 was designed to have less carbon, 0.03 percent compared to 0.08 percent in the standard 316. Reducing the carbon content limits the amount of chromium carbide formed but does not entirely prevent it. Carbide precipitation is time and temperature dependent in the range of 800-1500° F which is reached in the HAZ during welding. The precise control of the welding thermal cycle provided by orbital welding technology, in combination with the low carbon content of the L grade of 316, have kept carbide precipitation and the accompanying chromium depletion to a level at which it is no longer a problem.

Stainless steel at room temperature is highly resistant to oxidation and corrosion due to the presence of a passive surface layer 5-50Å in depth, however, when heated to temperatures above 400° C in the presence of atmospheric oxygen, surface oxidation occurs which may extend to a depth of 1000 Å or more. Different color heat tints are produced at different temperatures, with the dark blue oxide tint being most susceptible to corrosion.2 Heat tint originates when vaporized elements from the base metal become oxidized and precipitate in the HAZ.

A persistent "blue halo" found next to the weld zone has frequently been reported in high-purity applications even when oxygen levels in the purge gas were in the low ppm range. Evidence cited in the welding literature suggests that the blue halo may result from manganese which vaporizes during welding and precipitates in the weld HAZ. Spectroscopic analysis of gases in the arc during GTA welding of stainless steel characterized a blue vapor consisting of ionized manganese, chromium and iron.15 It was reported that vaporized Mn which precipitates and adheres in bands on either side of the weld bead is corroded preferentially.6 Evidence for the volatility of the blue material is provided by instances in which a strong ID purge during welding results in the blue halo appearing on the downstream side of the weld only. However, Auger analysis indicated that the blue discoloration in the HAZ was an optical effect resulting from a thicker oxide layer in which manganese was present but at a lower concentration than in adjacent clear areas.16

An Auger analysis of an unpassivated weld on 316 MP tubing in a previous study revealed a manganese peak about seven mm from the weld centerline on both sides of the weld which was removed by passivation.5 The larger manganese peak and manganese depth profile of the construction site weld shown in Figure 15 suggest that a zone of manganese enrichment may be encountered beneath the passive layer at a level too deep to be affected by chemical passivation. The smaller manganese peaks observed on the orbital welds (Figure 14a) are in the same area of the MP HAZ which exhibits the "frosted" or "etched" band of corrosion which suggests that manganese may be a contributing factor in this type of corrosion. The corrosion bands appear to be wider than the blue halo which can be seen adjacent to the weld bead in samples 9E and 32E.

At oxygen levels of 108 ppm, the weld and HAZ were covered with a deep brown rust-colored heat tint. Other investigators have identified the composition of this heat tint as oxides of chromium and iron. This is presumably formed when chromium and iron are vaporized from the metal and redeposited on the surface leaving a chromium-depleted underlayer. The surface oxide presents a poor barrier to corrodents and the chromium-depleted layer is subject to localized corrosion including pitting. In fact, corrosion of the heat-tint oxide may activate adjacent metal initiating corrosion in much the same way as free iron.2

Accelerated corrosion tests are used to predict the performance of a material or welds in a particular service environment. The potentiodynamic polarization test was selected for the present study because it had been shown to give good discrimination between passivated and unpassivated unwelded tubing in previous tests. Our results have shown that orbitally welded samples of 316L tubing, both EP and MP, have similar pitting potentials (Epit) to unwelded, unpassivated tubing.4 Passivated welded samples were shown to have pitting potentials in the same range or higher as those of passivated unwelded EP or MP 316L tubing from a different material heat. Thus, based on Epit measurements by potentiodynamic polarization, while orbital welding with standard or high purity argon seems to have little effect on Epit, passivation, especially with the optimized mixed chelant formulations, offers significant corrosion protection.

The results of this study clearly demonstrate that passivation improved the corrosion resistance of welded 316L tube samples both by increasing Epit and by reducing the amount of visible surface etching that occurred during corrosion testing. Of the six mixed chelant formulations tested, four gave results in the same range as nitric and phosphoric acid, while two formulations provided superior protection. Thus chelant formulations which utilize safer, less toxic passivating agents than nitric acid, have been shown to be at least as effective, while the Epit values obtained for the most effective chelant mixtures suggest that superior results are possible. The fact that chelants are safer to handle, and present fewer concerns for disposal while being similar in efficacy to nitric acid, should lead to more widespread use of these agents in the future.

Although passivation, especially with the optimized mixed chelants, unquestionably improves the corrosion resistance of both welded and unwelded stainless steel tubing, our results have also shown that passivation was not effective in removing all of the heat tint produced by welding, even at gas purity levels of eight ppm, and significant blue and brown discoloration remained after passivation on samples purged with gas containing 108 ppm oxygen (Figure 9). This was further demonstrated by the Auger line scans in which Cr, Fe, oxygen, etc. did not completely return to unwelded metal levels after passivation, and in the depth profiles which showed that changes in elemental composition as a result of welding extended well below surface levels. Since chemical passivation is effective in removing only the light heat tint from the surface (5-50 Å) but not deeper discoloration which may extend to 1000 Å or more, it must not be considered as a means to undo the effects of improper purging.

Clear indication of the benefits of electropolishing for providing increased corrosion resistance, particularly corrosion resulting from welding, have been demonstrated. The pitting potentials obtained on welds with EP tubing were slightly higher than those of MP tubing on both as-welded and passivated samples. The failures with the MP samples welded with the lowest purity gas also indicate that MP tubing is less resistant than EP tubing to the effects of oxygen in the purge gas. At the oxygen concentrations in the argon purge used in the present study, it was not possible to eliminate all evidence of oxidation from the welds on MP tubing. Light straw-colored bands were still present on both sides of the weld with ID purge oxygen levels below 0.1 ppm, while with the EP tubing with the same purge gas, some samples had a barely visible blue tint adjacent to the weld bead near the start of the weld, while other samples were free of visible oxidation with a mirror-like finish. Thus, welding specifications calling for no visible oxidation on MP tubing may be unrealistic.

Further evidence for the superior protection offered by electropolishing was provided by the data from the Auger line scans and depth profiles. These results suggest that the changes in surface element distribution produced by welding are less severe for EP tubing than for MP tubing and are more easily reversed by passivation. The Auger oxygen and Cr/Fe ratio depth profiles showed that changes brought about by welding do not extend as far beneath the surface on EP tubing. The Auger results, combined with the occurrence of darker heat tint discoloration, and increased surface etching after corrosion testing on MP welded samples purged with argon of comparable purity support the hypothesis that EP tubing provides greater protection than MP from oxidation and other detrimental changes leading to the loss of corrosion resistance after welding.

Our results on MP tubing are consistent with those who found that the corrosion resistance of AISI 316L stainless steel purged with argon was reduced at oxygen levels between 50 - 200 ppm, with the greatest drop occurring at oxygen levels over 100 ppm.3 These authors suggest limits of 25 - 50 ppm of oxygen for AISI 304 and 316L stainless steels, equivalent to a light straw color, as the maximum permissible levels which would result in critical pitting temperatures (CPTs) close to base metal levels. They recommend a value of less than five ppm oxygen for minimal reduction of CPT of 316L or for acceptable results with duplex stainless steel. They did not study these effects on EP tubing.

The present investigation has shown visible differences in localized corrosion of the HAZ of welds even between low ppm and low ppb levels of oxygen. The benefits of passivation, particularly chelant passivation, in overcoming the detrimental effects of heat tint produced by welding have been clearly demonstrated. Orbital welding which accurately regulates the heat input into the weld, is much more effective than manual welding in maintaining the corrosion resistance of the base metal and was shown in this study to have no significant effect on Epit when purge gas of standard purity levels or better were used.17 Nevertheless, even the combined use of orbital welding, EP tubing and the best chelant passivation can not completely eliminate the detrimental effects of welding 316L stainless steel tubing when purge gas oxygen levels exceed 100 ppm. This was shown by the etched appearance of weldments after being subjected to potentiodynamic polarization testing.

The observation that the pattern of corrosion was heaviest in the overlap and downslope areas of the weldments suggests that heat input affects the corrosion resistance to 316L stainless steel. Whether the increased corrosion in these areas which receive double exposure to the heat of welding results from additional oxidation or from other effects of heat, weld procedures which minimize the overlap area, and provide precise control of heat input would be expected to minimize the loss of corrosion resistance produced by welding. The practice of permitting a second pass to repair a weld in cases where the weld joint is found to be incompletely penetrated or, especially the practice of permitting an additional pass to correct concavity which results from excessive heat input in the first place, should be discouraged in systems where corrosion resistance is a concern.

Welding defects such as arc strikes, weld spatter, flux and grind marks associated with loss of corrosion resistance are virtually eliminated by orbital welding.13 A manual GTA welder will ordinarily start and stop the arc several times during the course of manual welds such as the two-inch OD tube welds done for this study. Manual GTA welding does not provide the precise control of heat input that is provided by orbital welding with the result that to assure a completely penetrated weld, excessive heat is introduced into the weldment. The resulting weld is unlikely to be as smooth as an orbital weld, and the quality less consistent. It should be realized that if manual welds had been the subject of this study, the pitting potentials obtained would most likely have been considerably lower. Thus orbital welding has become the preferred method of joining for industries requiring a high degree of weld quality and cleanliness.

It would be tempting to think that all of the detrimental effects of welding are preventable, but this is probably not the case, since it was found that the HAZ of even well-purged welds that had been pickled and passivated were more susceptible to microbially-induced corrosion (MIC) than unwelded base metal.18 Nevertheless, while these changes can not be prevented entirely, they can certainly be minimized to a level that is correctable, or nearly so, by passivation. In fact, in this study, when orbital welds were done on electropolished tubing purged on the ID during welding with purified argon gas, and passivated with the most effective mixed chelant passivation, no detectable visual or electrochemical difference in corrosion resistance between welded and unwelded tubing was found.

It also must be realized that corrosion in pharmaceutical piping systems can originate from sources other than welding. These sources include contamination, especially by iron from tools, particularly wire brushes used for cleaning the tubing, weldments, or surfaces in the fabrication area. Susceptibility to corrosion has been associated with non-metallic inclusions which become corrosion initiation sites.l9 One explanation for the low Epit values could be inclusions in the base material that lowered the corrosion resistance. All four of the unwelded tube samples were shown as N/A. However, the MP tube samples welded with greater than 100 ppm of oxygen in the purge most likely reflect loss of corrosion resistance due to excessive heat tint. If surface anomalies are responsible for some of the low pitting potentials, then electropolishing appears to provide some protection and also to minimize the harmful effects of welding with greater than 100 ppm oxygen in the purge gas. The rougher surface of MP tubing may permit more outgassing of contaminants at high temperatures and adsorb more oxides onto the surface during welding. This would account for the greater amount of discoloration observed on MP samples and the larger surface corrosion bands after PP testing.

Other sources of corrosion are dissimilar metals such as carbon steel, or improper passivation or electropolishing. However, for the more critical biopharmaceutical applications where orbital welding is specified, life cycle considerations would probably justify the additional expense of effective purge gas purification, attention to detail in purging and fabrication techniques, as well as the use of electropolished tubing coupled with the most effective chelant passivation.


Summary


The results of this study provide some practical implications for the installation of pharmaceutical piping systems of 316L stainless steel to obtain the optimal corrosion resistance of the material. Passivation should always be done to welded piping systems before they are put into service. Our results indicate that all of the experimental mixed chelants tested provided an improvement in pitting potential at least comparable to that of nitric and phosphoric acid, while the two optimized chelant formulations were found to provide higher pitting potentials and thus provide better corrosion resistance. However, passivation affects only the outer 50Å of the stainless steel surface while the heat tint produced during welding can extend to deeper levels. To obtain the maximum benefits of passivation, it is necessary to control the amount of heat tint produced during welding. The elimination of heat tint requires that purge gas of the highest purity be used for purging the ID of the weld joint during welding. Our results indicate there may be benefits to limiting oxygen and moisture in the purge gas to as low as the low ppb range.

For more critical systems, electropolished tubing appears to offer somewhat better corrosion resistance than mechanically polished tubing and appears to be somewhat more resistant to the loss of corrosion resistance during welding. The heat input during welding should be precisely regulated so that only the amount of heat required for a full-penetration uniform weld bead is applied. Excessive heat input may result in loss of corrosion resistance in the weld HAZ. Weld parameters which provide the optimal heat input for a particular heat number, diameter and wall thickness of tube or fitting can be easily achieved with orbital welding for every weld joint in the piping system.


Conclusions



Orbital welding was found to have little effect on Epit values at ID purge gas oxygen levels as high as 108 ppm for EP tubing and as high as eight ppm for MP tubing when compared to unwelded untreated samples and chelant-passivated samples of a different heat of 316L tubing.

Much greater differences in corrosion resistance as measured by Epit were seen in passivated compared to unpassivated samples than between welded and unwelded samples.

Samples treated with mixed chelants showed a significant difference in Epit values between passivated and unpassivated values indicating superior protection from corrosion. While all chelants gave results at least as good as nitric and phosphoric acids, optimized chelant formulations gave consistently higher Epit values.

EP welded, unpassivated tubing showed somewhat higher Epit than MP welded unpassivated tubing; differences in Epit were not significant after passivation except for the lack of passivity for MP tubing purged with 108 ppm oxygen in the purge gas.

The oxygen concentration of the argon purge at the levels tested had little apparent effect on Epit in either passivated or untreated samples with exception of MP tubing at purge gas oxygen levels of 108 ppm for which passivity was not observed.

Visual examination of samples after corrosion testing showed clear differences between levels of oxygen in the purge gas, visible oxidation, and the extent of corrosion, with the purest gas resulting in the least amount of etching corrosion. The amount of visible surface corrosion was markedly reduced in all cases by mixed chelant passivation. However, while passivation removed light heat tint produced by standard gas, it did not completely remove deep heat tint.

MP tubing appeared much more susceptible to differences in gas purity as shown by the lack of passivity in Epit tests, by the amount of visible heat tint in welded samples, and the extent of "etching" in HAZ in samples subjected to corrosion tests.

Auger line scans showed clear differences in element distribution between EP and MP tubing after welding.

Auger depth profiles indicated that changes in oxygen concentration and Cr/Fe ratios extended to greater depths beneath the surface on MP than on EP welded samples. These changes, both at the surface and at deeper levels, were more pronounced in MP tubing, and most likely have a negative impact on corrosion resistance, especially on surface corrosion or "etching".

Detrimental changes in surface element distribution across the welds were restored to near unwelded metal levels by mixed chelant passivation as evidenced by pitting potential measurements and Auger analysis.


By Arnie Grant, Barbara K. Henon, Ph.D., Arc Machines, Inc.
and Florian Mansfeld, Ph.D




Acknowledgments

1. Potentiodynamic Polarization tests were done by L.T. Han at the Corrosion Environmental Effects Laboratory of the Department of Materials Science and Engineering at the University of Southern California, Los Angeles, California.

2. Tubing was supplied by Valex Corp., Ventura, California as two 20 foot lengths each of mechanically polished (180 grit) and electropolished (10 Ra) type 316L stainless steel, heat number A341855.

3. A Nanochem® Model 1400 Purifier System from Semi Gas, Inc., Santa Clara, California was used to purify standard argon gas to the low ppb range of oxygen concentration.

4. The Auger surface line scans and depth profiles were performed by Photometrics, Inc., Huntington Beach, California.

5. Weld photos (Macrographs) were taken by Kars' Advanced Materials, Inc., Garden Grove, California.


References


1. Turner, S. and P.F.A. Robinson. The effect of surface oxides produced during welding on the corrosion resistance of stainless steels. Science. Volume 45, Number 9, 1989.

2. Kearns, J.R. and J.R. Maurer. Welding guidelines to minimize service degradation of stainless alloys in bioprocess systems. Bioprocess Engineering Symposium, ASME WAM, Atlanta, Georgia, pp.109-117,1991.

3. Hansen, J.V., Nielsen, T.S., and P. Aastrup. Root surface quality requirements- high efficiency purging or pickling? Paper 46, Conference on Duplex Stainless Steels, Glasgow, Scotland 13-16 November, 1994.

4. Lin, S.H. Corrosion protection of high copper aluminum alloys and stainless steels by surface modification. Ph.D. Thesis, University of Southern California, August, 1995.

5. Balmer, K.B. Auger surface line scan to evaluate element segregation in and chelant passivation of a 316L weld. Presented at the Bioprocess Engineering Symposium, ASME International Conference and Exposition, Chicago, IL., November, 1994.

6. Shirai, Y., Kojima, T., Miyoshi, S., Narazaki, M., and Ohmi, T. Specialty gas distribution system free from corrosion, gas decomposition and reaction- Perfect Cr2O3 treated tubing system. MICROCONTAMINATION Conference Proceedings, 1994.

7. Gramlich, C.T., O'Keefe, T.J., and Hsien-Tsung Liao. Characteristics of stainless steel surfaces after passivation with nitric acid solutions. Abstr. BED. Vol. 23 p. 63. Bioprocess Engineering Symposium, ASME WAM, 1992.

8. Wu, Chyang-Jen, Ph.D. Dissertation. Effect of surface treatment on localized corrosion of stainless steel. University of Missouri, 1989.

9. Seo, M and M. Sato. Epit shifts in a more noble direction with increasing chromium content. Trans Japan Institute of Metals, Vol. 21, No. 12, 1980.

10. Fihey, J.L. and R. Simoneau. Weld penetration variation in GTA welding of some 304L stainless steels. Presented at the American Welding Society Conference, Kansas City, Missouri, May, 1982.

11. Kearns, J.R. and G.E. Moller. Reducing heat tint effects on the corrosion resistance of austenitic stainless alloys. Materials Selection and Design. NACE, May,1994.

12. Ping Li, R.A. Buchanan, C.D. Lundin, D.E. Sachs, A.H. Tuthill, R.E. Avery, and J. Bullen. Corrosion studies of stainless steel weldments in a microorganism containing Arizona ground water system. 1995 INTERNATIONAL CONFERENCE ON MICROBIALLY INFLUENCED CORROSION, New Orleans, Louisiana, May 8-12, 1995.

13. Tuthill, A.H. Stainless steel: Surface cleanliness. Pharmaceutical Engineering. pp. 35-44, Nov/Dec, 1994.

14. Rahoi, D.W. Technical considerations in selection of nickel alloys and stainless steels for the bioprocess industry. Bioprocess Engineering Symposium. ASME Winter Annual Meeting, Chicago, Illinois, 1988.

15. P. Kieffer, "Spectroscopie de l'arc de soudure", IREQ Report 2513, 1982.

16. Mehalko, T. The science of welding for bio-processing application. Bioprocess Engineering Symposium, ASME WAM, Chicago, Illinois, 1988.

17. Henon, B.K. Orbital welding of corrosion resistant materials for bioprocess piping applications. ASME Bioprocess Engineering Symposium, ASME WAM, San Francisco California, 1989.

18. Punter, A. Influence of weld discoloration on the susceptibility of stainless steel weldments to microbiologically influenced corrosion. Stainless Steel Europe, pp. 38-43, Oct., 1994.

19. Suter, T. and H. Bohni. New microchemical investigations to study the initiation mechanisms of pitting. Conference on Critical Factors in Localized Corrosion. Chicago, Illinois 9-11 October, 1995.

20. Banes, P.H. Passivation: Understanding and performing procedures on austenitic stainless steel systems. Bioprocess Engineering Symposium, ASME Winter Annual Meeting, New Orleans, LA, BED Vol.27. pp.41-46, November, 1993.

21. Butters, J. and A. Reynolds. Stainless steel surfaces in bio-processing. Construction materials, pp. 65-69, Process Engineering, February, 1990.

22. Electrochemistry and corrosion overview and techniques. EG&G Princeton Applied Research, Electrochemical instruments division. Application note corr-4.

23. Cormia, R.D., Schiefelbein, B. and P.A. Olsen. Electropolishing of stainless steel: Meeting the materials requirement for today's demanding applications. Presented at the Photo Chemical Machining Institute (PCMI) annual meeting in Anaheim, California, February 26, 1990.

24. Geipl, H. Orbital TIG (GTA) welding of electropolished high alloy steel tubes. Stainless Steel Europe, pp. 4855, May, 1992.

25. Grant, A. Passivation of welds on stainless steel tubing. Presented at the ASME Bioprocess Engineering Symposium, San Francisco California, October, 1995.

26. Henon, B.K. and Grant, A. Orbital welding for hygienic pharmaceutical and biotechnology applications. 15th Annual Pharmtech Conference, Oxford, UK, March, 1996.

27. Henon, B.K. Corrosion effects of oxidation resulting from welding on 316L stainless steel tubing. Presented at the ASME Bioprocess Engineering Symposium, San Francisco, California, October, 1995.

28. Mangan, D. Metallurgical, manufacturing and surface finish requirements for purity stainless steel system components. Bioprocess Engineering Symposium, ASME WAM, San Francisco, California, pp. 29-38, 1989.

29. Mansfeld, F. Corrosion testing of welds on stainless steel. Presented at the ASME Bioprocess Engineering Symposium, San Francisco California, 1995.

30. Marshall, P.I. and T.G. Gooch. Effect of composition on corrosion resistance of high alloy austenitic stainless steel weld metals. Paper No. 296. Corrosion '92 The NACE Annual Conference and Corrosion Show, 1992.

31. McWane, O. Current fabrication, welding and quality practices for the construction of biotech facilities. Pharmaceutical Engineering, Vol. 9, No. 3., pp. 20-23, May / June, 1989.


About the Authors


Arnie Grant is Director of Research at Cal-Chem Corp. Grant received his BS in chemistry from Fairleigh Dickinson University and pursued post-graduate courses in analytical chemistry, instrumentation, biochemistry, quality assurance and reliability at California State University at Long Beach and the University of California at Los Angeles. He has more than 35 years of experience in analytical chemistry methods development, corrosion prevention and control programs, parts-materials processes evaluation and standardization and contamination control and prevention. He has held both laboratories management and project management positions. He was past chairman of the Infared Committee for the ICRPG and is currently very active in the ISPE and ASME authoring and presenting papers and conducting seminars in passivation and the causes and prevention of corrosion.
Cal-Chem Corp., 2102 Merced Ave., So. El Monte, CA 91773.

Barbara K. Henon, PhD is Manager of Technical Publications at Arc Machines, Inc. Dr. Henon has 12 years of experience in the on-site training of welding personnel and development of welding procedures for applications in the pharmaceutical and bioprocess industry as well as the semiconductor, and other critical industries. She has authored numerous technical articles on welding issues for industry publications including Pharmaceutical Engineering, presented seminars on orbital welding topics at the Interphex Pharmaceutical Show and Conference, at the ASME Bioprocess Technology Symposia, the American Welding Society Conference and the National Association of Corrosion Engineers Conference. Dr. Henon received her PhD from the Department of Biological Sciences at the University of Southern California and conducted research in neurophysiology at the Beckman Research Institute at the City of Hope in Duarte, California. She is the past Chairman of the ASME Bioprocess Engineering Subdivision, and a member of the ASME Bioprocess Equipment Main Committee and the Subcommittee for Materials Joining which developed the new ASME standard for the fabrication of bioprocess equipment and facilities in the United States (BPE '97).
Arc Machines, Inc.,10500 Orbital Way, Pacoima, CA 91331.

Florian Mansfeld, PhD is joint Professor of Materials Science and Engineering and of Chemical Engineering at USC. He was department manager, Interface Phenomena, at Rockwell International Science Center, for seven years prior to joining USC, and before that served for nine years as a member of the technical staff in the Fracture & Metal Physics and Physical Chemistry groups. He received the 1993 W.W. Horner Award from the American Society of Civil Engineers, NACE's 1988 W.R. Whitney Award (in recognition of contributions to the science of corrosion), the ASTM's Sam Tour Award of 1984, the Second Best Paper published in Plating and Surface Finishing (1979), the US Senior Scientist (Humboldt) Award in 1979, the NASA Technology Utilization Award in 1975, and a National Research Council Postdoctoral Fellowship in 1968. Professor Mansfeld was named a Fellow of the National Association of Corrosion Engineers in 1994 and a Fellow of the Electrochemical Society in 1995. He has authored or coauthored over 300 papers in the field of electrochemistry including some on the deposition of nickel and copper, and brings with him an extensive background in the theory and practical aspects of electrochemical deposition. Indeed, his team at Rockwell investigated the deposition of copper and nickel from a common electrolyte that is central to EFAB. Professor Mansfeld holds a BSc, MSc, and PhD degrees, all from the University of Münich, Germany.
University of Southern California, Rm.714, Vivian Hall of Engineering, Los Angeles, CA 90089-0241.