An Overview of the Effects of Sulfur on the Orbital GTA Welding of AISI Type 316L Stainless Steel Tubing and Pipe

An Overview of the Effects of Sulfur on the Orbital GTA Welding of AISI Type 316L Stainless Steel Tubing and Pipe


Introduction

Heat-to-heat differences in the weldability of stainless steel have been recognized since 1967. Although other trace elements may be a contributing factor to these differences, weldability problems with AISI type 304 or 316 have often been found to be the result of low sulfur concentrations in the base metal, or may occur when welding low sulfur heats to heats that are in the mid-range or higher with respect to the sulfur concentration. The purpose of this presentation is to provide an overview of the effects of sulfur on the orbital welding of type 316L stainless steel tube and pipe for semiconductor applications. This overview is based on selected papers from the welding literature, observations and unpublished data, successful solutions to customers’ welding problems, and current semiconductor applications and practices. To understand this topic we need to consider the following questions:

  • What is the effect of sulfur (or lack of sulfur) on the weld
  • At what sulfur concentration level does the effect become apparent
  • What causes this effect
  • When does the sulfur concentration of 316L stainless steel become a welding problem
  • What approaches have been used to correct for weldability problems
  • How effective are these approaches
  • What sulfur ranges of type 316L stainless steel would be most appropriate for the different semiconductor applications
  • Where do we go from here?

What is the effect of sulfur on the weld?
The topic of sulfur concentrations in stainless steel was brought to our attention in 1984 by a paper written in 1982 by Fihey and Simoneau at Ontario Hydro in Canada. The paper described the phenomenon of arc deflection that may occur when welding a flat plate of stainless steel (ss) with a very low sulfur to one that was higher in sulfur. These investigators were trying to resolve difficulties in welding instrumentation tubing at a nuclear facility. It was known that autogenous welds on heats of 304 or 316 that were very low in sulfur had a very wide shallow weld bead with a width to depth ratio [W/D( Note: Some authors refer to the D/W ratio.)] of a fully penetrated weld much greater than that of welds on heats in the medium to high sulfur range. Fihey and Simoneau showed when welding heats of 304 or 316 stainless steel of unmatched sulfur concentrations, the arc would tend to deviate towards the low sulfur heat if one of the heats was below about 0.008% sulfur and the other was above this value. Oxygen was found to have a similar but weaker effect to sulfur.

Orbital Welding
Autogenous orbital welding of thin-walled tubing and pipe using enclosed weld heads is commonly done on sizes from 1/8 inch O.D. tubing to 6 inch schedule 5 pipe (O.D. 6.625 inches). The most common application of orbital welding is of 1/4”, 3/8”, and 1/2” O.D. tubing, fittings and other components of gas distribution systems. The practical wall thickness limitation for fusion butt welding of 300 series stainless steel in enclosed heads with a fixed arc gap is about 0.154” which is the wall thickness of 2 inch schedule 40 pipe. Autogenous orbital welding of pipe up to 12 inch schedule 10, with a 0.180” wall has been done for semiconductor oil-free air and general purpose nitrogen lines using open-frame type heads for diameters up to 6 inch pipe or full-function pipe weld heads that mount on a guide ring and crawl around the weld joint circumference for the larger diameters. The automatic arc voltage control (AVC) on these heads provides precise control of the arc gap and thus better control of the heat input allowing for uniform penetration on pipe that may be out-of-round or out of dimensional tolerances. These welds are generally done as a single-pass weld using a square-butt end preparation.

Effect of sulfur on orbital welds
To determine whether Fihey and Simoneau’s observations held true for orbital tube-to-tube welds, in 1984 the appearance and quality of orbital welds on various heats of 316L tubing for which mill test reports (MTRs) were available was evaluated by visual examination and measurements of the width to depth ratios of fully penetrated orbital welds. The size range of these materials collected at random from customers was from 1/8 inch O.D. tubing to 1 inch schedule 10 pipe (1.315 inches O.D.). Eleven of the first 13 of these heats had sulfur contents of 0.008% or less, while the AISI specification for Type 304 and 316 SS, (as well as 304L and 316L) has a maximum sulfur concentration of 0.030%. The low-sulfur materials (0.008% and less) had wide weld beads with W/D ratios of greater than 3 compared to a range of 1.5 to 2 for heats having more sulfur. The larger weld puddle is more affected by gravity, tends towards concavity and there is a finer line between full penetration and overpenetration which makes it more difficult to achieve a uniform, full-penetration weld. The affect of gravity is more pronounced on heavier-walled material. Autogenous 5G welds of pipe, with wall thicknesses greater than about 0.065 inches, which exhibit a wide weld bead configuration typically have some degree of concavity at the 12 o’clock position.

Frames from an AMI in-house video showing lathe welds of 316L tubing using Arc Machine’s arc filtration system to view the weld pool. Leftr. Weld of 0.008% to 0.008% sulfur with symmetrical weld puddle. Right. Weld of 0.002% sulfur (top) to 0.014% sulfur. Note asymmetrical puddle. Electrode is centered on the weld joint, while the weld puddle is offset towards the low sulfur heat. Henon, unpublished data.
0.008% sulfur0.002% sulfur
0.008% sulfur0.014% sulfur


Results of AMI In-House Arc Deflection Test Welds April, 1992
In order to verify and document the arc deflection described by Fihey and Simoneau, tubing was obtained from several heats of type 316L, 3/4 “OD., 0.065” wall with sulfur contents of 0.002, 0.008, and 0.014 wt.%; and 1/2” OD tubing with sulfur contents of 0.001, 0.007, and 0.015 wt.% as shown in the table on page 14. The sulfur content was inscribed on each individual tube piece. A video showing arc deflection was made on June 11, 1992 using a direct view camera positioned on the torch. (See frames from the video on the previous page.) The welds for the video were not orbital welds, but were done on a lathe with the torch held stationary. The first weld was a 316L tube with 0.008 wt.% sulfur welded to another coupon from the same tube. The arc was symmetrical and was centered on the weld joint. A weld of tubing with 0.002 welded to another tube of 0.014 wt.% sulfur showed extreme deflection towards the low sulfur side. The weld puddle for this weld was asymmetrical with the trailing edge clearly on the low sulfur side, but the downslope trailed off towards the high sulfur side. Several heats of 316L in the mid and low ranges of sulfur content were orbitally welded and examined for arc deflection on welds between heats (See table). Not all of the low sulfur heats exhibited arc deflection when welded to heats higher in sulfur, probably because elements other than sulfur affect weld pool behavior.

Low Sulfur Memo - 1985
Based on information from the welding literature and our in-house results, a customer memo on the subject of sulfur was drafted in October of 1985. It was hoped that by controlling the sulfur content of type 316L it would be possible to guarantee consistency of weld results making automatic orbital welding a truly automatic process where the same weld schedule could be applied to any heat having the same OD and wall thickness. It was suggested that when ordering stainless steel tubing for autogenous orbital welding that customers specify sulfur concentrations between 0.005 and 0.016% in order to assure optimal weldability. However, weldability is not the only criterion for selection of alloy composition and in the industries in which autogenous orbital welding is used, there are other considerations.

Proposed mechanisms to explain the effects of sulfur on the weld pool

Above: Fihey and Simoneau thought that sulfur acted to constrict the anode spot producing a more focused arc. They showed that oxygen in the purge gas could constrict the arc and increase penetration. Below: Heiple and Roper (1982) attributed the effect of sulfur to the Marangoni effect on the weld puddle. When the sulfur concentration is greater than 0.008%, the temperature coefficient of surface tension is positive and penetration is deep. When the sulfur content is less than 0.008% the temperature coefficient of surface tension is negative and the weld profile is wide and shallow.


How does sulfur affect the weld pool?
The anode spot theory. Fihey and Simoneau suggested that the weld pool penetration pattern was linked to the anode spot as shown in the drawing on the previous page. The anode or anodic spot is the portion of the weld pool surface where electrons leaving the tungsten cathode enter the weld pool providing the heat for welding. Minor elements in the base metal affected the degree to which the anode spot was constricted. A diffused anode spot resulted in poor penetration while a constricted anode spot resulted in good penetration. Elements having a high electron affinity such as sulfur, oxygen, and halogens when present at certain concentrations produce a constricted anode spot. At low concentrations of these elements or combined with the presence of easily ionized elements such as Cs, K, Al, Ca, Cr, Mn, etc. produce a stable diffuse anode spot and poor penetration welds. Aluminum which is used in the processing of stainless steel, combines with oxygen and thus has a negative effect on penetration. Fihey and Simoneau mentioned a blue ionized metallic vapor emitted from the surface of the weld pool which they identified as manganese. They also described the smoother weld bead surface associated with the lower sulfur material.

Marangoni convection, the surface tension driven fluid-flow model. Heiple and Roper (1981) proposed that elements affecting weld pool shape do so by altering the surface tension. According to this model, fluid flow in the weld pool determines GTA weld pool shape and the force driving the fluid flow is the surface tension gradient. This is called Marangoni convection since Marangoni investigated this phenomenon in the 19th century and discovered that when a surface tension gradient exists on the surface of a liquid, then fluid will be drawn along the surface from the region(s) of lower surface tension to the region(s) of higher surface tension. High speed motion pictures of the weld pool showed that sulfur added to the weld caused AlO2 particles at the surface to move from a ring at the edge of the weld pool towards the center and disappear, while aluminum itself had no effect on movement of the particles.

Pure metals (and low sulfur materials) have a negative temperature coefficient of surface tension. When the area under the arc is heated, the surface tension drops in the center of the weld pool and the direction of fluid flow is from the center towards the edges of the pool. When sulfur, oxygen, selenium, or other surface active elements are introduced to the weld pool at very low levels they cause the surface tension to rise with an increase in temperature. In this case, the direction of fluid flow is from the outside of the pool, where the surface tension is lower, towards the center, thus increasing penetration. Burgardt and Heiple (1985) of Rockwell International proposed that elements changed the GTA weld shape by changing the temperature gradient on the weld pool surface through changes in input power density. Magnetohydrodynamic flow was also considered to explain the effect of trace elements on the weld pool.

Effects of other trace elements. Pollard, 1988. The heat-to-heat variability observed in automating the welding of strips of stainless steel into tubing at tube mills led Pollard (1988) to test the effects of various trace elements on penetration. He found that sulfur had a large effect on the penetration of production heats which were low in oxygen, but little effect on laboratory heats containing 150-200 ppm oxygen. With these higher oxygen heats, good penetration was possible at sulfur concentrations less than 0.005%. Pollard recommended that the sulfur contents of type 304 stainless steel be held to between 0.010 and 0.015% as a good compromise for weldability, hot workability, and formability. He did not consider the effect of sulfur on surface finish or on weld bead smoothness which is of particular interest to the semiconductor industry. Pollard’s results were consistent with the surface tension driven fluid flow model of Heiple and Roper.

The negative effect of aluminium on penetration was explained by its interaction with oxygen to form aluminum oxide thus depleting oxygen in the base metal. Manganese and silicon were found to increase penetration at low concentrations, while penetration was decreased at higher concentrations. While not contradictory to the surface tension model, the effects of these elements are thought to result from additional factors such as the effect of silicon on viscosity. Silicon and aluminum were shown to increase the amount of slag inclusions and the ratio of silicon to aluminum to influence the type of slag inclusions formed during welding.

Above: Data from Pollard (1988) comparing the depth to width ratio of welds on heats of 316, 316L, and laboratory heats as a function of sulfur concentration. In the laboratory heats containing 150 to 200 ppm of oxygen, the effects of sulfur were minimal. Lack of sulfur had a greater effect on 316L than on 316. Below: Plot of weld d/w ratio versus sulfur content for approximately 200 heats of type 304L stainless steel. Each point is an average of multiple sulfur analyses and weld d/w ratio measurements. If single values are used, the scatter is greater. From Burghardt and Campbell, 1992.

 

Burgardt and Campbell, 1992.
Burgardt and Campbell presented data for over 200 heats of 304L stainless steel and a few heats of 316L. Their results showed that, although there was a clear increase in weld D/W ratios with increasing sulfur concentrations, there was a great deal of variability for a given sulfur concentration. The variability appeared to increase at the lowest sulfur concentrations between 0.001 wt.% and 0.005 wt.%. They attributed the variability to interactions of the surface active elements sulfur and oxygen with other elements in the steel. The effect of trace elements that are relatively weak in the presence of sulfur may be increased as sulfur is reduced.

Calcium interacts with oxygen and also with sulfur to form stable compounds that are unlikely to become surface active. Aluminum and silicon also interact with oxygen to form stable compounds. This leaves the amount of sulfur and oxygen available to segregate to the weld pool surface a complicated function of the total weld pool chemistry. The implications of this study clearly demonstrate that it is impractical to attempt to predict the weldability of a particular heat of stainless steel based on sulfur levels alone. It appears that at the lower end of the sulfur range that other trace elements may have a greater effect than at higher sulfur concentrations either by acting alone or in combination.

Approaches to welding low sulfur heats
Addition of 0.1% oxygen or SO2 to the shielding gas. One approach to obtaining a narrow weld bead on heats of stainless steel with unacceptably wide beads was to compensate for the lack of oxygen or sulfur in the base metal by adding these elements to the shielding gas. Fihey and Simoneau were able to reduce the weld bead width of low sulfur heats of 304 stainless steel by adding 0.1% oxygen to the shield gas to restrict the “anodic spot”. This technique resulted in an acceptable weld profile but they found that increasing the oxygen to 1.0% reversed the positive effect. Similarly, attempts to make up for the lack of sulfur by adding SO2 to the shield gas were tried by Heiple and Burgardt (1985). The SO2 is toxic at concentrations used, and the 0.1% oxygen concentration in the shield gas which was found to be effective is equal to 1,000 ppm which would not be acceptable for high-purity applications.

Fihey and Simoneau (1982) suggested the use of 0.1% oxygen in the argon shield gas to compensate for lack of oxygen in the base metal to constrict the anode spot and increase penetration. The positive effect of oxygen on penetration was reversed at oxygen concentrations of 1%. The photos above show that when 0.3% oxygen was used as a shield gas a fairly narrow weld bead was obtained. However, oxygen at these levels in the shield gas is disallowed for high purity applications because of discoloration and the possibility of particulate generation. Henon, unpublished data.



ID pressurization.
Fihey and Simoneau used the application of pressure to the ID weld bead to correct for the overpenetration that resulted from increasing the amperage to overcome lack-ofpenetration defects associated with arc deflection. This technique is widely used today in fabrication of components for HP and UHP gas distribution systems. For thin-wall tubing, 1/4”, 3/8” and 1/2” OD a pressure gauge is used with a restrictor at the tube outlet to obtain a predetermined I.D. purge pressure during the weld. This allows the welding operator to use weld parameters that would otherwise be slightly too hot, and keep the I.D. flat rather than over penetrating. The use of this technique provides consistent welds with fewer minor adjustments to the weld program. The correct application of ID pressurization on a very low sulfur material can produce orbital welds that are flat on the ID and OD and virtually invisible to the naked eye. However, with single pass welds the weld bead will typically be wider for low sulfur heats than for heats in the mid to high range of sulfur.

Henon and Overton (1988) and Henon and White (1992) discussed the use of pressure gauges to prevent weld blow-out in UHP field installations. The liquid weld puddle can be pushed outwards by the application of I.D. pressure, which if excessive, can result in blowing out the weld. Pressure gauges are used in the field to determine the correct flow rate for each individual weld which will vary depending on the location of the weld with respect to the length and configuration of the system being installed. To set the correct flow rate, a tee is inserted at the weld joint and connected by a length of tubing to a pressure gauge. The flow rate is set to give a predetermined pressure reading, then the tee is replaced by the weld head and the joint is welded.

Orbital welding procedures for UHP piping
It was recognized at an early stage that orbital welding for UHP semiconductor applications required a systematic approach with control and documentation of weld-related procedures such as cleaning, cutting and facing, surface finish, purging techniques including I.D. pressurization, and awareness of material composition including sulfur content (Henon and Overton, 1988).

Technique of using a pressure gauge to achieve correct ID pressure and a flat inner weld bead. From Henon and White, 1994.



Ohmi technology. Ohmi in a number of publications in the early 1990s proposed an integrated approach to orbital welding for UHP piping. His recommendations included using a shield and I.D. purge gas mixture of 95%argon/5% hydrogen in combination with high speed, multi-pass autogenous welds which were not pulsed in order to achieve a narrow, smooth weld bead. Internal pressurization was used to control over penetration on specially-formulated very low sulfur materials. A higher travel speed at the same welding current will result in a narrower weld bead, but produce less penetration, so more than one pass was needed to achieve a completely penetrated weld using this technique. A narrow weld bead is considered desirable for HP and UHP semiconductor applications because the surface roughness of the I.D. weld bead is greater than that of the electropolished tubing. A narrower weld bead reduces the area of greater surface roughness to which the ultraclean process gases are exposed. Although the appearance of these welds on the electropolished low-sulfur material was generally good, Ohmi soon recognized that making several passes significantly increases the heat input into the weld increasing the size of the heataffected zone and reducing the corrosion resistance in aggressive service environments. Ohmi modified his welding procedure in order to reduce the excessive heat input by recommending a high travel speed and higher amperages to achieve a narrow weld in a single pass. However, with this technique he noted excessive chromium depletion in austenitic materials and began to recommend the use of ferritic stainless steels for UHP piping.

Use of mixed gas. On heats of 304 or 316 stainless, the addition of 5% hydrogen to the arc shield gas will usually constrict the arc resulting in deeper penetration with a narrower weld bead. On heats of low sulfur material with high W/D ratio weld beads, the 95%argon/5% hydrogen mixture does not typically reduce the weld bead width. It does reduce the amount of amperage required to weld them, it produces a smoother weld pool and its reducing action removes trace amounts of oxygen in the purge gas producing a cleaner looking weld. However, using a mixed gas produces a less stable arc and arc wander may occur. It is essential that the composition of the mixed gas remain constant since small differences can affect penetration.

Pulsation. Ohmi tried to eliminate the appearance of ripples in the weld bead by omitting the pulsed current. While the unpulsed welds had a smooth visual appearance without magnification, SEMs of pulsed welds compared to unpulsed welds were similar microscopically with respect to surface roughness (Burton, personal communication, 1997).

Above: Unacceptable orbital weld (BOP) on a type 316 stainless steel 3 inch schedule 10 pipe with 0.130 inch wall. The weld bead width was 0.47 inches, the W/D was 3.6. The shield gas was 95%argon/5% hydrogen. Below: Orbital weld on the same pipe using an insert ring made from another 316 ss pipe. The full-penetration weld had an O.D. bead width of 0.25” and a W/D ratio of 1.9. The shield gas was argon. Henon, unpublished data.

STEP procedure. The use of a STEP welding procedure during which rotational travel is stopped or slowed during the primary current pulse and resumed during the background current pulse has been found to be effective on some poor weldability heats. This technique allows for greater penetration at the same amperage than welds done with a continuous welding speed. It is particularly effective for tube-to-fitting welds with less than ideal fit up or for joining materials of differing weldabilities.

Use of insert ring for pipe-to-pipe welds.
Autogenous orbital welding of thin-wall pipe up to about 0.154” wall is normally very practical and easily accomplished. However, occasional heats of stainless steel may be problematic. In 1990 a customer was having difficulty obtaining an acceptable orbital fusion butt weld on a 3 inch schedule 10 stainless steel pipe. The wall thickness of the pipe was about 0.130”. The weld bead was nearly 1/2 inch wide with concavity over about 30% of the weld and had a concave downslope. The sulfur content of the pipe was listed as 0.007%, but the welding characteristics were typical of the low-sulfur profile. It is common industry practice to add filler which is chemically enriched in elements which are lacking in the parent material or may segregate non-uniformly in the weld metal. For pipe welding the filler material is usually in the form of wire, but since at that time the cost of orbital welding equipment with wire-feed capabilities was significantly more than fusion welding equipment, the customer wanted to do the weld autogenously if at all possible. Washer shaped inserts were cut from a type 316 three inch schedule 10 stainless pipe that had a very narrow weld bead. When welded with the inserts the weld bead on the problem pipe was reduced from 0.47 inches without an insert to 0.25 inches with the insert when argon was used as a shield gas and to 0.21 inches when the shield gas was 95%argon/5% hydrogen. The mixed gas was used to increase the amount of penetration at the same current setting which it did. However, it did not reduce the weld bead width on the pipe with the wide weld bead, but it did reduce the weld bead width when an insert ring was placed in the weld joint. No mill test report (MTR) was available for the pipe from which the insert ring was made. However, it was clear that this technique worked very well as shown on the photos above.

A similar solution was employed on an application in Taiwan with autogenous welds on 4 inch schedule 10 pipe (0.109” wall) which was difficult to weld. The pipe had a sulfur content of 0.001%. As a control, an insert ring from another low sulfur heat was tried, but this ring had no effect, so higher sulfur rings were used in production. Prior to the actual weld, a pass was made at low amperage to blend the materials in the pipe and ring on the joint surface. The customer made over 500 welds with the 0.009% sulfur insert rings. One out of 20 of these welds failed X-ray due to lack of fusion of the ring, but they were able to reweld these successfully. When an insert ring is used it is very important to make certain that all surfaces are completely clean to assure fusion. Today, since the cost of orbital welding equipment with wire feed capability has come down appreciably, the use of a pipe weld head to add wire of a known higher sulfur concentration into the weld puddle would be recommended for pipe welding if the specifications permitted the addition of wire. In the case cited above, a low sulfur material was used to manufacture the pipe because it was softer and thus easier to draw over a mandrel in forming. The contractor was able to save money on the purchase price of the pipe, but the unexpected delay and increase in the complexity of the welding procedure drove up the cost of the installation.

Correcting for Arc Deflection
The semiconductor industry soon discovered that sulfur in the base metal combined with manganese to produce non-metallic inclusions called manganese sulfide “stringers”. These “stringers” are removed from the stainless steel surface during electropolishing leaving microscopic voids which may become initiation sites for pitting corrosion. This lead to industry specifications for much lower levels of sulfur to achieve acceptable surface finish and corrosion resistance. However, the presence of the manganese sulfide “stringers” improves machinability, and some fitting manufacturers specified sulfur contents in the mid to high range of sulfur concentrations for type 316L. As a result of mixing materials with widely different levels of sulfur, the incidence of arc deflection problems was increased. A prominent mechanical contractor reported at an Interphex West workshop that arc deflection that occurred on welds between tubing and fittings cost his company $180,000 per year until they discovered the cause of the problem and began to track heat numbers (MacLaren, 1995). Another mechanical contractor complained that the arc in his weld head was coming off to one side. When coupons from the identical heat number were welded together in the head, the problem went away. Heiple and Burgardt explained the arc deflection phenomenon as resulting from a surface tension gradient in the weld pool towards the region of higher surface tension. Since the temperature coefficient of surface tension is negative in low sulfur heats, when the heat of the arc is applied to the center of the weld pool, the surface tension is reduced on the low sulfur side and the direction of fluid flow is from the center outwards, but on the high sulfur side, fluid flow is from the edges of the weld pool towards the center. Logically one would expect that since the low sulfur material typically requires higher amperage to penetrate than higher sulfur heats, that penetration would be greater on the high sulfur side, but this is not the case. When arc deflection occurs the penetration is almost always on the low sulfur side and this deflection may be serious enough to shift the penetration entirely to the low sulfur side completely missing the inside of the weld joint.

One approach to welding mismatched heats when arc deflection occurs is to make a preheat pass at a lower amperage in an attempt to blend the dissimilar materials and even out the sulfur gradient prior to welding. This approach does yield some improvement, but is not entirely effective. When the cause of the arc deflection problem became more widely known, some manufacturers recommended making fittings and tubing from the same heat of material to prevent arc deflection. This is not always practical, but heat traceability for both tubing and fittings is important. A better approach is to track material heats and try to match the sulfur contents on the job as closely as possible. In our experience differences in sulfur do not always result in arc deflection. While welding a heat of 0.001 to 0.007 wt.% no arc deflection was observed, but welds between a heat of 0.002 and 0.008 wt.% did show arc deflection. (See Table).

Restricted sulfur ranges.
AISI type 316L stainless steel, joined by orbital fusion butt welding, is widely used in both the semiconductor and biopharmaceutical industries, and to an increasing extent, in sanitary process piping applications in the food, dairy and cosmetic industries. Type 316L is an austenitic stainless steel containing chromium, nickel, manganese, and silicon in addditon to iron. The carbon content is limited to 0.030% compared to 0.08% in AISI type 316 to reduce the incidence of carbide precipitation that can occur during welding. Sulfur is present in steel as an impurity, but affects weldability to a greater extent than other trace elements. When problems with weldability occur on a jobsite it can result in unacceptable delays and cost overruns. Attempts to weld the material(s) in question may be less than satisfactory. When the AMI customer memo suggested in 1985 that users of orbital welding equipment select materials with a limited sulfur range, this idea met with considerable resistance in the industry, particularly from manufacturers and suppliers of stainless steel tubing.

However, the rapidly developing semiconductor industry soon discovered the need for a smooth inner surface for tubing to be used in gas distribution systems. The need to be able to quickly achieve acceptable levels for moisture in gases passed through a system demanded an electropolished interior surface of lower and lower average roughness (Ra) with similar demands on weld bead smoothness. Manganese sulfide inclusions as well as other types of inclusions containing oxides of calcium, aluminum, silicon and others have been associated with pitting corrosion and make it difficult for manufacturers of tubing and fittings to achieve specified surface finish requirements. The weld bead itself is smoother with the low inclusion materials. Thus it would be desirable for reasons of surface finish and corrosion resistance to severely limit the amount of sulfur present in type 316L.

SEMs of orbital welds on 316L stainless steel tubing. Above: Orbital weld on VOD melt at 50X magnification. Below: Orbital weld on EB melt at 100X magnification. The EB melt which has fewer impurities in the base metal has a much smoother weld bead. Courtesy of Valex Corp.



Stainless steel melt methods for high purity applications have become more and more effective in eliminating inclusion-forming impurities. Older melt methods such as electric arc were replaced by AOD (argon oxygen decarburization) or VOD (vacuum oxygen decarburization) which remove a significant amount of trace elements and make it possible to limit sulfur to concentrations to as low as 0.001 wt.%. Double refining techniques such as AOD/VAR (argon oxygen decarburization/ vacuum arc remelt), and VIM/VAR (vacuum induction melt/vacuum arc remelt) remove even more impurities. Electron beam (EB) refining technology involves the use of several stages of increasingly high vacuum and the resulting product has virtually no sulfur or slag-forming elements. The SEMs shown above clearly show the relationship between impurities in the base material and weld bead smoothness.

Semiconductor sulfur range. There is general agreement that type 316L stainless steel has adequate corrosion resistance for most semiconductor process gas line applications. Gases such as HBr and HCl are not very corrosive unless moisture is present thus corrosion damage tends to occur when the sytem is opened such as for changing gas bottles. While it is not considered necessary in most cases to specifiy a more corrosion resistant alloy for these applications, it is important to optimize the corrosion resistance of type 316L. The semiconductor industry, i.e., SEMI, has chosen to restrict the sulfur content of type 316L for the most critical applications. For high purity gas distribution systems (SEMI E49.8-96) the sulfur content includes the entire range of sulfur specified in AISI type 316L which has a maximum of 0.030 wt.%. Sulfur contents for UHP gas distribution systems (SEMI E49.9-96) have different specifications depending on the diameter of the tubing. For tubing over 1/2 inch, sulfur is limited to less- than- or- equal- to 0.005 wt.% to 0.015 wt.%, while the sulfur content of tubing of 1/2 inch or under is limited to less- than- or- equal- to 0.003 wt.% to 0.012 wt.%. Tubing of 1/2 inch OD and under are commonly welded with some type of ID back pressure and typically have thinner walls which facilitate the welding of the lower sulfur materials. SEMI F2-94 Specification for 316L Stainless Steel Tubing for General Purpose Semiconductor Manufacturing Applications limits sulfur for type 316L stainless steel to between 0.003 and 0.010% for all nominal sizes. This is somewhat lower than would be considered to be ideal for welding but avoids the lowest sulfur levels. However, an even lower sulfur concentration (0.001 to 0.004%) is recommended by SEMI F20-95 for 316L stainless steel bar which is used for componenets for high purity chemical (gas or liquid) distribution systems. While some of the SEMI standards recognize that the lowest sulfur content materials may present difficulties for welding, particularly in the larger sizes, SEMI F20 specifies sulfur concentrations to provide superior surface finish and corrosion resistance for which the ideal sulfur concentration would probably be zero.

Above: Arrows show location of orbital welds on a low sulfur (maximum of 0.005 wt.%) 316L electropolished stainless steel valve. Left: ID of an orbital weld viewed through a Sight-Pipe®. Welds are inspected for full penetration, absence of discoloration, good alignment and workmanship.
Above and below: Orbital tube-to-tube and tube-to-fitting welds on 1/4" OD 316L cut open to show inner weld beads.

Specification for the pharmaceutical industry. The ASME Bioprocessing Equipment Standard (BPE - 97) has recently addressed considerations for the design and construction of bioprocessing equipment and piping in order to achieve systems that can be maintained in a clean and sterile condition [CIP/SIP (clean-in-place/sterilize or steam-in-place)]. The industry requires smooth, crevice-free corrosion resistant welds in order to limit the growth of microorganisms. This standard limits the sulfur content of 316L to between 0.005 and 0.017% in order to eliminate the lowest end of range which may be difficult to weld and the highest end of range which typically has a rougher weld bead. The ASTM A270 Standard has adopted the recommendations of ASME BPE-97 and added a supplement, S2 for Pharmaceutical Grade Tubing to this standard. This industry commonly uses 316L tubing in sizes of 1 to 4 inches OD with 0.065 inch wall thicknesses and the recommendations for sulfur are comparable to those of the larger diameter tubing specified by SEMI E49.9-96 for UHP gas distribution systems.

However, all specifications are subject to change and Art Tuthill, an expert on corrosion and a consultant to the Nickel Development Institute, citing concern for higher corrosion rates of materials having significant amounts of manganese sulfide inclusions has proposed at a recent ASME Seminar that 316L stainless steel strips used to make welded tubing for bioprocess applications be limited to a maximum of 0.005% sulfur. For tanks, vessels and similar equipment which are welded by different processes he recommends that sulfur concentrations be limited to 0.001 wt.%.

Other restrictions in chemical composition
Restricting sulfur content would not solve all problems associated with poor weldability. In an attempt to eliminate problem heats and to define the ideal composition for weldability Cohen (1997) examined 5 “trouble” heats of 316L stainless steel that had presented difficulties on a large installation. He found that higher copper levels seemed to promote arc wander, tin was present in higher than usual concentrations in 2 of the “trouble” heats, and another “trouble” heat was higher in manganese.

Collins et al. (1999) demonstrated a strong relationship between the cosmetic quality of the welds on various heats of 316L and the base metal chemistry as expressed by the ratio of the chrome equivalent (Cr eq) to the nickel equivalent (Ni eq) where chromium, silicon, niobium and titanium contribute to the Cr eq, and nickel, manganese, carbon, nitrogen and copper contribute to the Ni eq. Collins proposes the use of insert rings to adjust the chemistry of weld components to achieve a favorable Cr eq/Ni eq which will produce cosmetically attractive welds with good corrosion resistance.

Implications of Sulfur Content for Successful Orbital Welding

Maintaining repeatability of the welding process. Orbital welding is a highly repeatable process with the capability of producing a large number of high-quality welds at a high level of consistency. For orbital welding to be most effective, certain factors must be controlled. These include dimensional tolerances for tubing OD and wall thickness, shield gas composition, tungsten length which determines the arc gap length, tungsten geometry, gas purity, gas flow rates, etc. Changes in weld procedures from autogenous welding to procedures requiring the addition of filler material or the use of an insert ring, or a change in arc gas composition from pure argon to a mixed gas are considered to be essential variables according to ASME Section IX of the Boiler and Pressure Vessel Code and require requalification of the welding procedure. Since requalification is both time consuming and costly, it is important to be able to use the same welding procedure for all material heats for a particular application. While minor changes in amperage are easily accommodated, a major change in the weld procedure from heat to heat may result in weld quality that is not acceptable for the application and a loss in productivity.

While the use of insert rings is certainly an effective method for achieving desirable weld quality on heats with excessively wide weld beads, it adds a degree of complexity to fusion butt welding procedures. Where the use of an insert ring (or filler wire) is necessary, for example for welding an alloy such as AL-6XN (Henon, 1989), a welding procedure using the insert ring or filler wire of specified composition must be qualified to ASME Section IX. This is a somewhat more complex procedure than fusion butt welding, but produces satisfactory, repeatable results.Thus it is not the procedure itself, which if planned for can be factored into the cost of a project, but the unexpected change in procedure and failure to achieve the expected level of weld quality that is problematic.

Successful Fabrication of Low Sulfur 316L. Orbital welding of very low sulfur 316L stainless steel need not be a problem when done in a systematic way. High-production orbital welding on small diameter tubing and components has become routine for the Veriflo Division of Parker Hannifin, which manufactures valves and regulators for UHP semiconductor applications. Dan Miller reports that they do approximately 100,000 orbital welds each month on welds of tube or fitting-to-valve or regulator body with a reject rate of only 0.05%. Not all of the rejected parts are due to welding defects but include operator errors such as welding a male nut to a part instead of a female nut. This extremely low reject rate is being achieved on materials with a maximum sulfur concentration of 0.005%.

Veriflo’s fabrication area includes 22 orbital welding stations operating on 3 shifts a day with a total of 55 people making an average of about 200 welds each per shift. The welds are done in a single pass with a 15 second downslope. The OD weld bead widths are presently about 3 T or 3 times the wall thickness of the joint. Miller has found that a 95%argon/ 5% hydrogen mix does help to make the weld bead narrower even on these low sulfur heats and intends to incorporate the mixed gas into his welding procedures in the near future. This procedural change will also help to maintain the high levels of cleanliness required of these components.

Chip Doherty and Roger Ames of Amescor also reported success with orbital welding of small diameter, low sulfur materials. They routinely weld materials conforming to SEMI E49.9 and have not experienced any changes in weld acceptance rates with variations in sulfur content. They pressurize the weld during welding by restricting the flow at the weld outlet and have an acceptance level of 1-1/2 T for weld bead width.

Even though UHP semiconductor fabricators are consistently welding materials in the 0.003 to 0.005 wt.% range of sulfur concentrations shown by Burghardt and Campbell to show the greatest variability in D/W as well as weldability, this does not seem to be a problem, especially for the smaller tubing sizes. The success of fabricators today in dealing with the lower sulfur materials may be the result of newer melt methods and improved control of base metal chemistry, or it may simply be due to improved standard orbital welding practices such as high purity cleaning procedures, better end preparation, improved shield and purge gas purification, and control of weld bead pressurization during welding.


Some fabricators of components for UHP gas distribution systems routinely weld thousands of joints a week using only very low sulfur materials. Photo courtesy of Advanced Micropolish , Inc.

Sulfur Concentrations of 316L for Orbital Welding. For more critical applications using small diameter thin-wall tubing, sulfur concentrations of 0.001 to 0.005 wt.% sulfur are routinely and successfuly orbitally welded but may require more complex welding procedures and allowances made for higher weld bead W/D ratios. Cost considerations of achieving an excellent surface finish and optimal corrosion resistance for orbitally welded piping systems must include the cost of more complex welding procedures, including training of welding personnel, that may be needed to weld these materials. For less critical applications where weldability is of foremost consideration, a sulfur range of 0.005 to 0.017 wt.% may be appropriate.

On installations where there may be heats of 316 with a large range of sulfur concentrations it is essential that installers carefully track the heats of tubing, fittings, and other system components being welded and attempt to keep the sulfur contents of materials being joined as similar as is practical.

In so far as possible, it is necessary to aim for consistency of orbital welding procedures and to avoid unnecessary changes in procedures. An understanding of the effects of sulfur concentrations in stainless steel is essential for making the best choice of materials for a particular application, but be aware that sulfur is not the only element controlling the weldability of 304 and 316 materials and that the effects of other trace elements and combinations of elements may also have a significant effect. Limiting the sulfur range of 316L in order to optimize orbital fusion butt welding would eliminate some, but not all of the weldability problems associated with sulfur. Inaccuracies of the MTRs, lack of reporting of some trace elements such as aluminum, selenium, manganese, iodine, and others that may affect weldability, and the unpredictability of interactions between the various trace elements make it impossible to predict from the MTRs whether or not a particular heat of stainless steel will be weldable using conventional orbital welding techniques.

Although the incidence of problem heats is very low, at least one major end user of 316L semiconductor grade stainless steel tubing has their materials prescreened for weldability prior to purchase to assure that orbital welds on every heat they purchase will meet their stringent weld quality requirements.

Presented by Barbara K. Henon, Ph.D., Arc Machines, Inc. at the SEMI Workshop on Stainless Steel, Semicon Southwest Austin, Texas October 16, 2000.

References

1. Ames, R. and C. Doherty, Amescor, Presentation at the Semi Workshop on Stainless Steel, Semicon Southwest, Austin, Texas, October 16, 2000.

2. ASME Bioprocessing Equipment Standard (BPE -97). American Society of Mechanical Engineers, 3 Park Avenue, New York, New York. 1997.

3. ASTM A270 American Society for Testing and Materials, 106 Barr Harbor Drive, West Conshohoken, Pennsylvania 19428-2959.

4. Burgardt, P. and R.D. Campbell. Chemistry effects on stainless steel weld penetration, Ferrous Alloy Weldments. Trans Tech Publications. Switzerland, 1992. pp. 379-415

5. Burgardt, P. and R.D. Campbell. The effects of chemistry on weld profiles in partial and full penetration welds. International Trends in Welding Science and Technology. Proceedings of the 3rd International Conference on Trends in Welding Research, Gatlinburg, Tennessee, USA, June 1-5, 1992.

6. Burton, C. Effect of stainless steel melt methods on the corrosion resistance of welds in UHP gas lines. Presented at the MICROCONTAMINATION Conference, San Francisco, CA, 1997.

7. Burton, C. Theory behind improved corrosion resistance after welding. Dendrite correlation to solute content. Valex Corp., 1997.

8. Campbell, R.D., C.R. Heiple, P.L. Sturgill, A.M. Robertson, and R. Jamsay. Surface preparation effects on GTA weld shape in JBK-75 stainless steel. Welding Journal, Weld. Res. Suppl. 67- s-77-s, February, 1993.

9. Cohen, R.M. Identifying trace elements affecting the weldability of 316L stainless steel tubing. Future Fab International Issue 2, Vol.1. p.271-282, 1997.

10. Collins, S.R., P.C. Williams, P.C., and B. DeForce. Weldability and corrosion studies of AISI 316L electropolished tubing. Stainless Steel World 99 Conference, 1999.

11. 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, MO, May 16-19, 1982.

12. Heiple, C.R. and P. Burghardt. Effects of SO2 shielding gas additions on GTA weld shape. Welding Journal, 64(6): 159-s to 162-s, 1985.

13. Heiple, C.R. and J.R. Roper. Mechanism for minor element effect on GTA fusion zone geometry. Welding Journal 61(4): 97-s to 102-s, 1982.

14. Heiple, C.R. and J.R. Roper. Effects of minor elements on GTAW fusion zone shape. Rockwell Int. RFP 3224, 1981.

15. Henon, B.K. Customer advisory on fusion welding of low sulfur heats of stainless steel. Arc Machines, Inc., 1985.

16. Henon, B.K. Orbital welding of 316L Stainless Steel Tubing. The Fabricator, May, 1997.

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

18. Henon, B.K., and J.S. Overton. Part 1. Design planning for class 1 UHP stainless-steel process- gas piping system. Part 2. Constructing a class 1 UHP stainless-steel process-gas piping system. MICROCONTAMINATION, February, 1988 and March, 1988.

19. Henon, B. and J. White. Use of ID pressurization to control weld profile in semiconductor ultra-high-purity process gas lines. Arc Machines, Inc. in-house publication, 1994.

20. MacLaren, Ian. Process Piping: Construction issues. Presented at the ISPE Conference at Interphex West, Anaheim, California, 1995.

21. Miller, D. Veriflo Division of Parker Hannifin Corp. Presentation at the Semi Workshop on Stainless Steel, Semicon Southwest, Austin, Texas, October 16, 2000.

22. Miyoshi, S., K. Kawada, and T. Ohmi. High performance, high speed one-revolution welding technology for piping construction - Corrosion free, non-catalytic and fume free. Tohoku University, Japan. 1993.

23. Mizuguchi, Y., T. Ohmi, and K. Sugiyama. Butt welding technology for ultra high purity gas delivery system. Department of Electronics, Faculty of Engineering, Tohoku University, Japan, 1989.

24. Pollard, B. The effects of minor elements on the welding characteristics of stainless steel. Welding Journal, September, 1988.

25. SEMI E49.8-96 Guide for High Purity Gas Distribution Systems in Semiconductor Manufacturing Equipment. Semiconductor Equipment and Materials International, 805 East Middlefield Road, Mountain View, California 94043.

26. SEMI E49.9-96 guide for Ultrahigh Purity Gas Distribution Systems in Semiconductor Manufacturing Equipment. Semiconductor Equipment and Materials International, 805 East Middlefield Road, Mountain View, California 94043.

27. SEMI F2-94 Specification for 316L Stainless Steel Tubing for General Purpose Semiconductor Manufacturing Applications. Semiconductor Equipment and Materials International, 805 East Middlefield Road, Mountain View, California 94043.

28. SEMI F3-94 Guide for Welding Stainless Steel Tubing for Semiconductor Manufacturing Applications. Semiconductor Equipment and Materials International, 805 East Middlefield Road, Mountain View, California 94043.

29. SEMI F20-95 Specification for 316L Stainless Steel Bar, Extruded Shapes, and Plate for Components used in High Purity Semiconductor Manufacturing Applications.

30. Tuthill, Arthur. Corrosion Control. Presentation at the ASME Bioprocessing Seminars, Bioprocess Equipment Design, September, 2000.

31. Walsh, D.W. and W.F. Savage. Technical Note: Bead shape variance in AISI 8630 steel GTAW weldments. Welding Journal, Weld. Res. Suppl. pp. 137s-139s, May, 1985.