Purge TIG Welding: A Complete Guide to Back Purging

A fabricator finishes a flawless-looking TIG root pass on 316L tubing, packs up the TIG set, and moves to the next spool. Two days later the client’s inspector pushes a borescope through the joint and finds a root that looks like burnt sugar, dull gray and crusted. The outside of the weld was perfect. The inside was a corroding liability. What went wrong was not the welding — it was the air nobody thought to remove.

What Is Purge TIG Welding? The Basics Explained

Purge TIG welding means making a TIG root pass from one side of a joint while the back of the weld is flooded with an inert gas that pushes the air out and keeps it out. The dams, tape and analysers are just the machinery.

Your torch throws shielding gas over the front of the molten pool, keeping oxygen away from the puddle as it solidifies. The arc can only be in one place at a time, so the root gets nothing. Air is still sitting there, and at welding temperatures it oxidises hot stainless steel.

What the damage looks like

Heat tint is the rainbow or straw colouring that appears on hot stainless steel. Light straw and gold tints are mostly harmless, but a blue, brown or black tint means the oxidation has eaten into the metal. TWI explains heat tint as a visible indicator of the chromium oxide layer formed during welding, and the start of losing the passivating film that protects the alloy.

Sugaring is the uglier cousin: the rough, grainy, burnt-sugar texture you get when the root bead oxidises heavily and crystallises as it cools, so it crumbles like the sugar at the bottom of a pan. A sugared root is a corrosion initiation site and grounds for rejection in hygienic service.

Underneath the colour sitsfree iron. When chromium oxidises instead of the steel, the surface is left as bare iron with no passivation, and that iron rusts, pits and creeps under the weld. You cannot grind a sugared root back to a good state; the chromium is gone.

Linde’s shielding gas guidance puts it bluntly: failing to purge can significantly reduce the working life of a component, or in the worst case lead to its failure in service.

Three terms people mix up

Back purging, root purging and backing gas are related but not identical.

• Back purging floods the reverse side of the joint with gas during welding. This is what most people mean by “purge TIG welding”.
• Root purging is more specific, describing purging applied to protect the root pass itself, the first and most vulnerable layer. Root purging is the technique, back purging the usual way of achieving it.
• Backing or form gas sits behind the root to support the molten metal and control bead shape. It can add heat, since it supports the puddle rather than excluding air. On thin material it shapes the bead, a different job from protection.

One clarification people search for constantly: purging here means shielding a weld root from air. It has nothing to do with purging a gas cylinder or depressurising a vessel.

Where purging is required, and where it is not

Purging is usually carried out when the material will be used for its corrosion-resisting qualities, such as stainless steels and some nickel alloys, or when the material oxidises easily, such as aluminium and titanium alloys.

On the other hand, back purging is generally not required for carbon and many low-alloy steels for most applications. Mill scale on a carbon steel root will form some oxide, but carbon steel metallurgy does not depend on a preserved chromium passivation film, and the root is usually buried in a painted or coated joint. You are protecting a load path, not a corrosion barrier. “Generally not required” is not “never” though: purge if the root will be exposed in corrosive service, or if the specification calls for it.

Purge gas: almost always argon

Purge gas is typically argon because it is inert and forms no compounds. It dissolves cleanly, does not react with the molten root and does not change the microstructure of the deposit. Argon/hydrogen and nitrogen/hydrogen mixtures suit materials not susceptible to nitrogen or hydrogen attack, a more advanced choice covered later in this guide.

The three phases of any purge

Every successful purge, whether it takes ninety seconds or half an hour, has the same three-phase structure. Get one wrong and the root suffers.

1. Pre-purge. Gas flows before the arc is struck, displacing the air inside the enclosed volume. The joint is not ready to weld until the oxygen reading at the sample point falls to the level your procedure requires.
2. Weld purge. Gas continues to flow at the correct rate while the arc is on, protecting the root as it forms and cools enough to lose its susceptibility to oxidation.
3. Post-purge. Gas keeps flowing after the arc is off, holding the protective atmosphere over the hot root while it cools below roughly 250 °C. Killing the gas as the torch lifts turns a perfect root sugared in the last two minutes.

What this looks like in a real shop

Picture a shop running 316L tube for a dairy skid. The operator TIGs a butt joint on 50 mm stainless tube, with an inflatable purge dam one side, a vent the other, and argon fed through a diffuser so it does not jet straight at the root. Later a colleague fits a purge plug and shuts the argon off at the cylinder because the job is “done”. The joint cools with air on the root, and the borescope two days later shows a root that looks like burnt sugar. Same weld, same arc, same wire, same cylinder. The only mistake was treating the purge as something that ends when the arc does.

That is the heart of purge TIG welding. It is not a separate process. It is TIG welding with an atmosphere deliberately maintained on the far side of the joint for as long as the root is hot. Get it right and the joint lasts; get it wrong and you have a liability that looks fine from the outside.

Why Purging Matters: Heat Tint, Corrosion and Code Compliance

A clean weld bead is the visible half of a good joint. The half nobody sees decides whether the pipe still holds pressure in five years or fails a hygienic audit in five months. A dirty root shows up later, as rust bloom in a food-grade line or pitting around a chemical tank weld.

When stainless steel heats to welding temperatures in air, the chromium in the alloy oxidises. That chromium forms the invisible passive film that makes the alloy “stainless” in the first place. TWI’s guidance on heat tint in stainless welds is blunt: the tint forms a chromium-rich scale that is typically not very protective, leaving the surface chromium-depleted and potentially impairing corrosion resistance. Heat tint is not cosmetic. It is chemically altered metal with no passive film left where the process fluid sits.

The pitting temperature data that makes the case

The number that persuades engineers is the critical pitting temperature (CPT): the point above which stainless steel begins to pit in a chloride solution. TWI reports that the CPT of a type 316 weld in 0.1% NaCl at +300 mV SCE may fall from 60 °C to 40 °C when the surface is heat tinted.

Read that again. A fully passivated 316 weld fails at 60 °C; the same weld with a heat-tinted root fails at 40 °C. Many processes run warm, and clean-in-place cycles use hot caustic and hot acidic chemistry. Hot water at 45 °C is unremarkable in a food plant, and at 45 °C the heat-tinted weld is above its CPT while the parent metal beside it is not.

Crevice corrosion follows the same logic. A heat-tinted root is often rougher and more undercut than a clean one, and roughness plus a depleted surface plus a narrow crevice traps chloride. A sugared root in a 316 line is a reliable precursor to a crevice leak.

Heat tint as an indicator, not a decoration

Because colour correlates with oxidation, it works as a proxy, and that is the logic behind AWS D18.2. The industry agreed a reference scale of colour samples, each tied to a purge gas oxygen content, and the AWS C5.10 and D18 committee work cover the GTAW and PAW processes used here.

Sample number	Purge gas oxygen level (ppm)	Typical interpretation
1	10	No visible tint; fully passivated
2	25	Very light tint, minimal depletion
3	50	Light tint; generally acceptable
4	100	Threshold of concern; unacceptable
5	200	Clear amber to brown band
6	500	Blue/brown band, clearly oxidised
7	1,000	Heavy blue, corrosion resistance compromised
8	5,000	Dark grey to black, sugaring possible
9	12,500	Severe oxidation, unserviceable root
10	25,000	Air contamination, effectively unpurged

Industry practice is that sample 4 and above is typically unacceptable unless the parties agree otherwise. ASME BPE has agreed to refer to the D18.2 guide for acceptance levels, which is why D18.2 is cited inside BPE documents and why the AWS D18 committee matters to anyone welding stainless tube.

The standards that actually bind you

AWS D18.1/D18.1M:2020 carries most of the weight in sanitary stainless work, covering gas tungsten arc welding and plasma arc welding of stainless steel tube and pipe 1/4 in (6 mm) diameter and larger for sanitary processing systems. It sets out the visual examination requirements for as-welded and finished surfaces, and that is where the D18.2 colour chart enters acceptance: a root reading sample 2 or 3 is generally fine, sample 5 or higher is a reject.

The BPE and the ISPE baseline guide lean on this language for bioprocess equipment, so pharmaceutical and dairy fabricators specify D18.1 welds with documented purge records. When an inspector asks for the purge log, they want proof you had low oxygen at the joint.

Why the outside can pass and the inside still fail

The torch and shielding gas cover the front of the pool; nothing covers the back unless you deliberately flood it. The bead a welder, a QC person and even a client inspector can see from outside is almost always the good side. Root oxidation develops in the gap behind the joint, where nobody looks, and only shows later as a rust streak in a product batch or a failed clean-in-place validation. Borescope inspection is standard on sanitary and high-purity work for that reason: it records the root bead, so acceptance is judged on the surface that touches the product. Where the purge record and the borescope image disagree, the borescope wins.

For shops building capability, pair this with the base process. If your shielding gas fundamentals are shaky, the air gas welding explained guide covers how shielding gases behave, while the aerospace TIG welding standards and techniques reference shows how tight the same acceptance logic gets in safety-critical work.

Purge gas is not a consumable you buy for convenience. It is the only thing protecting the corrosion-resistant surface of your stainless, and the colour chart exists because that surface is graded, measurable and enforceable. Get the oxygen down, match the tint to the sample scale, and the weld still passes inspection in five years.

Choosing the Right Purge Gas: Argon, Nitrogen and Hydrogen Mixtures

Argon is the workhorse of purge TIG welding, and for most stainless steel and nickel alloy work it is the only gas you need. But argon is not always the right answer, and choosing badly can quietly ruin a joint that looks perfect from the outside. Here is how the practical options compare.

Argon: the default for stainless and nickel alloys

Argon is inert. It does not react with the molten pool, does not dissolve into the weld metal, and forms no compounds with chromium, nickel, or molybdenum. That simple chemistry is why argon became the industry-standard backing gas, and why TWI continues to recommend pure argon as a backing gas for most stainless applications. It keeps the root clean without upsetting the balance of the alloy.

One practical detail catches people out. Argon is roughly 1.4 times as dense as air, so it sinks. Feed it into the low point of the purged volume and vent it at the high point, and the argon pushes air upward and out of the way. That geometry is why most setups put the gas inlet at the bottom and the vent at the top. Helium is lighter than air and wants the opposite arrangement. Get the orientation wrong and you end up fighting the gas instead of using it.

On oxygen targets, Linde’s shielding gas guidance notes that for many stainless applications an oxygen level of 0.01 percent, or 100 ppm, is satisfactory. For more critical service, 20 ppm may be needed to prevent pitting corrosion or subsequent stress corrosion cracking. Nickel alloys typically need 25 to 50 ppm O2, while reactive metals such as titanium and zirconium sit at 5 to 10 ppm.

Nitrogen: useful, and dangerous in the wrong hands

Nitrogen is cheap, widely available, and does a good job of displacing air. The problem is what it does once it gets into the weld.

Nitrogen is an austenite stabiliser. Absorbed into stainless weld metal, it changes solidification behaviour and can leave a brittle, untestable martensitic structure in the root of the joint. It also raises the risk of nitrogen-induced porosity in some alloys. It is not truly inert for titanium or aluminium either, since both can pick up nitrogen from the atmosphere.

That said, nitrogen is not the enemy. TWI notes that argon/nitrogen backing gases produce good results on highly alloyed stainless steels such as duplex and super-austenitic grades, where the intent is deliberately to stabilise austenite rather than fight it. Specialty argon/nitrogen blends also suit joints where travel speed or penetration matters more than inertness. The rule of thumb is simple: if you cannot explain why the joint needs nitrogen in it, use argon.

Hydrogen-bearing mixtures: scavenging the last traces of oxygen

Hydrogen does something argon cannot. It reacts with residual oxygen in the purged atmosphere to form water vapour, pulling the oxygen level down to levels argon alone will not reach. That is why small amounts of hydrogen are blended into argon or nitrogen for the most demanding applications.

According to Linde, hydrogen is added to argon to improve oxygen-scavenging characteristics. Commercial mixes such as Stainshield TIG and Specshield 5% H2 are used where only very low levels of oxidation can be tolerated, and only on materials not susceptible to hydrogen attack. That last condition matters enormously. Hydrogen can cause embrittlement in high-strength steels and can be troublesome in some nickel alloys and titanium. For a stainless sanitary joint, a hydrogen-bearing mix can be a smart choice. For a high-strength, pressure-retaining nickel system, it needs engineering sign-off first.

Helium: for heat and conductivity, not for chemistry

Helium is chemically inert too, but its reason for being on the list is thermal. It conducts heat far better than argon, changing the arc behaviour and letting you run higher amperage for a given electrode size. That makes it useful for high-conductivity metals and high heat input work, including thick-section aluminium and copper. It costs more, and it is lighter than air, so it feeds from the top of the purged volume with the vent low, the reverse of argon practice. For most stainless pipe purging, argon remains the sensible default.

Where aluminium is the subject rather than the purge, remember that oxide removal drives quality far more than gas choice. Aluminum cleaner for welding covers surface prep properly, and the practical guide to aluminum MIG welding covers the process side. Inert gas cannot rescue a dirty surface.

Selection table

Material and service	Recommended purge gas	Target oxygen level
General stainless fabrication, structural service	Argon	100 ppm (0.01%)
Stainless in hygienic, food, dairy or pharma service	Argon, or argon with 5% H2 where oxidation must be minimal	20 ppm
Nickel alloys such as Hastelloy and Inconel	Argon	25 to 50 ppm
Duplex and super-austenitic stainless grades	Argon, or argon/nitrogen per approved procedure	25 to 50 ppm
Titanium and zirconium	Argon or helium, chamber purged	5 to 10 ppm
Aluminium, tubular and thin wall	Argon, or helium for high heat input	50 to 100 ppm
High-strength steels susceptible to hydrogen attack	Argon only, no hydrogen-bearing mix	Per material specification

Bottom line: start with argon unless there is a documented reason not to. Reach for nitrogen only where austenite stabilisation is wanted or the code permits it, and reach for hydrogen mixes only where oxygen levels of a few ppm matter and the material can tolerate hydrogen. Source your cylinder and regulator from a supplier who understands shielding gases, such as AllGas welding supply, and check the gas label every time, because a mislabelled or switched cylinder is one of the fastest ways to ruin a purge.

Setting Up the Purge: Dams, Plugs, Diffusers and Shrinking the Volume

Getting inert gas to the root is half the job. The other half is deciding what to seal, where the gas enters, where it leaves, and how much space you are asking it to fill. Get those four things right and the analyser will read 20 ppm. Get them wrong and you can still hit 20 ppm on the gauge while the root is quietly oxidising, because the gas you measured never actually bathed the weld.

Step 1: Clean the bore before you seal it

Everything inside the joint has to come off before you close it up. Grinding scale, rust, mill oxide, oil and gasket residue all hold oxygen against the metal and, worse, they hold moisture that comes out of the joint the moment you heat it. On stainless, the scale is usually the enemy. On aluminium, it is the tenacious oxide layer that even a dedicated aluminum cleaner for welding needs to fully remove. Scrub the bore, then wipe it, then seal it. You will not get another chance once the dam is in.

Step 2: Install the dams deep enough to survive

Purge dams must be installed deep enough inside the pipe section that the heat of welding will not damage them. That sounds obvious until you consider a 3-inch stainless spool with a closure disc 6 inches from the joint. The heat-affected zone, the spatter, the fume and the radiant heat all land well beyond the edge of the bead, and a thin plastic disc or a water-soluble blocker placed too close will melt, collapse or blow out right in the middle of the weld.

The closure options described in the welding literature cover most bases: water-soluble dams that dissolve away with hot water afterwards, inflatable bladder dams that press against the bore wall, collapsible disc dams, and thermally disposable dams that are designed to be burnt or purged out of the line later. Pick the type that suits the wall condition, the bore diameter and the cleanup your site allows.

Step 3: Seal everything except one opening at the top

The joint should be sealed around the circumference except for a small opening at the top position of the pipe, with gas entering at the bottom and exiting at the top. That single design decision is what makes the purge work. Gas fed in low displaces air upward in a predictable sweep, and the vent at the high point lets the displaced air escape instead of pressurising the volume until it finds an unplanned leak.

“Sealed” means every branch, every nozzle, every instrument connection, every open end. A purge is only as good as its weakest opening, and a single 6 mm vent stub on a header can keep oxygen at 8,000 ppm no matter how long you purge.

Step 4: Fit a diffuser so the gas does not jet at the root

Purge gas arriving as a hard jet straight at the molten root will scour it, blow it concave, or drag oxygen back toward the weld from the far end of the bore. A diffuser, a porous plug, or simply a short length of perforated tube taped to the inlet spreads the flow into the volume instead of blasting it. The gas should feel like a gentle breeze across the root, not a nozzle.

Component quality matters here. Intercon specifies that all components used to make purging units should be halogen-free, heat-resistant and free of dissimilar metals, and that a purging unit should provide a well-sealed chamber, reduce purge volume and introduce no contaminants. Those are sensible criteria for any setup, whether you buy a commercial unit or build one from fittings and tape.

The hierarchy of purge methods

Method	Best suited to	Watch out for
Whole-length purging with tape and bored plugs	Straight runs, open piping, larger diameters where the bore is easy to access	Highest gas use and longest purge times; every untaped joint is a leak
Local purge chamber with inflatable or water-soluble dams	Production work, repair jobs, anything where the joint is accessible	Dams must sit far enough from the heat; chambers cost money and take time to set
Purge plugs in small-diameter tube	Instrument tubing, capillary and fine tube under 1 inch OD	Plug must not sit in the heat zone; small volumes purge fast
Full chamber purging	Reactive metals such as titanium, zirconium and some nickel alloys	Heavy gas consumption, longer cycle times, needs a sealed rigid chamber and trained setup

Shrink the volume and everything gets easier

Here is the single biggest cost and time saver in purge TIG: reduce the enclosed volume. Haynes International points out that while purging the entire volume of a long pipe run is often suggested, local purging of the immediate volume around the weld zone is an alternative and is generally the preferred technique. Local purging means less gas, a faster cycle, a smaller sealed chamber and fewer things to leak.

The number that tells you how long to wait comes from orbital welding guidance: as a starting point, exchange the purge zone volume at least 5 to 7 times. Scale that to your actual sealed volume and your measured flow rate, then verify with an analyser rather than a timer. For small-diameter tubing under 1 inch OD with plugs set close to the joint, 5 to 7 volume exchanges can take only 30 to 60 seconds. For a 12-inch line with dams 20 feet apart, the same rule means minutes and a very large cylinder.

Work through the setup checklist in order: clean bore, dam or plug, seal all openings, gas in low, vent high, diffuser fitted, analyser sample taken from a point that represents the atmosphere right at the joint. If you are building a procedure for reactive metals, aerospace TIG welding practices cover the chamber discipline those applications demand. And when the joint itself is the deliverable rather than a structural weld, many fabricators lean on a certified procedure instead of shop intuition. For reference, AWS D18.1/D18.1M covers welding of stainless steel in sanitary applications, including the cleaning and purging expectations that make a root pass acceptable in food and pharmaceutical service.

Get the setup right before you worry about the settings

Most of the purge failures we see are not gas problems at all. They are geometry problems: an unsealed branch, a vent hole at the wrong end, a jet pointed straight at the root, or a purge zone six times larger than it needs to be. Fix the setup and the flow rate, the oxygen reading and the heat tint usually follow without any other change.

Before you strike an arc, walk the joint one more time and confirm six things: the bore is clean, every dam or plug is deep enough to survive the heat, every opening is sealed except the top vent, gas enters at the low point, the diffuser is fitted and pointing along the bore rather than at the root, and the analyser sample line comes from a location that reflects the atmosphere at the joint.

If your setup is sound, move on to the numbers. Flow rate, purge time and volume are what turn a good setup into a good weld, and that is exactly what the next section covers.

Flow Rate, Purge Time and Volume: The Numbers That Matter

Most purging failures trace back to a number nobody wrote down: how much gas, for how long, through how much space. Get those three right and the root comes out clean almost by default. Get them wrong and even a perfect arc leaves a sugared root.

The volume displacement rule

Work out the enclosed volume: the space inside the pipe or tube your dams, plugs or tape have sealed off. The standard approach is to displace that volume by a factor of about 6, so about six times that volume of gas passes through before you strike an arc.

Six is enough for most applications, stainless included, but not for titanium, zirconium and other reactive metals, where the oxygen limit is far tighter. Intercon, a purge equipment manufacturer, recommends rest oxygen content below 100 ppm as highly desirable on stainless work, and going from roughly 21% oxygen in air to 100 ppm is a long way in one pass of gas.

A quick calculation shows the cost of a big purge. A 4 in nominal pipe with a 10 ft purge length holds roughly 8 to 9 cubic feet of internal volume, so at six volume changes you need around 55 cubic feet of argon before the arc is struck.

Why more gas is not better

The instinct on a bad joint is to turn the flow up. Resist it. Excessive flow turns the purge volume into a turbulent path, and turbulent gas does not sweep air out in a controlled front: it stirs air against the walls and creates suction that draws air back in through every gap in the sealing, so the analyser stalls at 400 ppm however long you flow. That plateau is almost always a leak or bad seal, not a flow problem.

There is also a physical limit. Purge pressure that gets too high pushes against the molten root from behind, leaving a concave root surface, and push it harder and the pool blows through. That is the classic back purge blowout, caused almost always by a flow rate set by guesswork rather than by joint size.

Starting flow rates by pipe size

Use these as starting points and confirm with an analyser on the actual joint.

Pipe / tube OD	Starting back purge flow
1/4 to 1/2 in	5 to 10 CFH
1/2 to 1 in	10 to 15 CFH
1 to 2 in	15 to 25 CFH
2 to 4 in	20 to 35 CFH
4 to 6 in	30 to 50 CFH
6 to 10 in	40 to 70 CFH
10 in and above	50 to 100+ CFH

These figures assume you have already reduced the volume as far as you can; if you cannot shrink the purge zone, flow rate and purge time both go up with it. Long runs with branches need more flow and more time, and tight-radius elbows and valves trap air.

Back purge flow versus torch shielding flow

Back purge and torch shielding gas are separate systems with separate regulators, and they must be set relative to each other. Haynes International found that a ratio of approximately 4 to 1 between back purge flow rate and welding torch flow rate is necessary to achieve a sound root-pass weld, working from 40 CFH back purge against a 10 CFH torch on HASTELLOY alloy pipe. That 4:1 anchor balances the pressures: enough back purge to keep the root clean, not so much that it deforms the root.

Set the back purge to four times the torch flow, then adjust from there: concave root surface, reduce back purge flow; oxidation creeping onto the root, increase it. A dedicated regulator and flow meter for the purge side, rather than tapping off the torch line, makes that far easier to hold steady. For regulator and flow meter selection, see All Gas Welding Supply’s guide to gases, cylinders and safety.

Reducing the volume: the highest-leverage number

If you take one action from this section, make it this one: shrink the volume and every other number gets easier. An AWS forum rule of thumb shows the gain. A U-groove preparation with a small root face, a 1/4 in bevel radius and no root opening reduces purge volume compared with a square butt, and a 1/16 in vent hole at the high point lets displaced air escape instead of pressurising the cavity. The same joint then needs less gas and less time to reach the same oxygen level.

Preheating and its effect on oxides

One more factor changes what you are purging against. Intercon warns that preheating pipe at high temperatures can encourage chrome oxide formation. Those oxides form faster and are harder to displace than room-temperature air, and on stainless they are the direct route to a browned or sugared root that will not clean up no matter how good the weld looks.

So get the oxygen down before the metal gets hot, keep the pre-purge flowing as the part comes up to temperature, and use a flow rate that clears the volume without the turbulence that pulls air back through leaks. On high-specification work with a tightly controlled oxygen level, read the requirements before setting any numbers: the aerospace TIG welding standards and techniques guide shows how those tolerances shape procedures.

Seal the joint, seal every branch, vent the high point, feed gas at the low point, then weld once the analyser settles below target.

Oxygen Analyzers: Choosing, Calibrating and Reading Them Correctly

You cannot see a purge and you cannot time it. Nothing tells you the air has gone from inside a pipe, and a timer assumes the volume, leak tightness and flow rate repeat on every job. Two pipes of the same diameter can need pre-purge times an order of magnitude apart if one has a leaky branch stub and the other does not.

That leaves measurement. An oxygen analyser is the only instrument that tells you what the atmosphere at the root of the joint looks like while you weld it.

Why eyeballing and timers fail

Welders purge until it feels right, or for a fixed two minutes. Both habits produce a root that looks fine outside and crumbles on the inside, and worse, one that reads acceptable on a coarse indicator, giving false confidence downstream. As the Intercon purging guidance explains, most precision welding happens well below 70 ppm, so an indicator resolving to 0.1% (1,000 ppm) or 0.01% (100 ppm) may not be sensitive enough.

Two very different instruments

Shops lump “oxygen meters” into one category, and that causes real problems on the floor. There are two distinct families of device.

Instrument	What it does	What it cannot do
Trace oxygen analyser	Measures oxygen in the sampled gas down to low ppm and drives the arc-strike decision	Protects nobody and will not warn of a hazardous atmosphere
Personal oxygen-depletion alarm	Warns the operator that ambient oxygen has fallen to a dangerous level	Not a substitute for measuring low oxygen levels for purging

The distinction matters because a personal oxygen-depletion alarm is a life-safety device. Purge gases are inert, but they still displace breathable air, and inside a vessel or long pipe the atmosphere can become unbreathable without anyone noticing. Keep both instruments on site, and follow the OSHA welding safety fact sheet on ventilation and fume control.

Range, accuracy and what to buy

For structural stainless work a low-ppm analyser is plenty. For hygienic service, pharmaceutical work and titanium the spec tightens: accurate to 1 ppm, with enough resolution to confirm 20 or 30 ppm rather than only knowing it is under 100.

• Low range capability. A dual-range instrument beats a single-range percent meter, but confirm the low range is a true low-ppm measurement, not an extrapolated scale.
• Sensor type. Electrochemical cells are the usual choice for trace oxygen: quick to respond and inexpensive. Paramagnetic and thermal-conductivity cells cover higher ranges but are the wrong tool for judging a 30 ppm purge.

Also consider how the instrument lives on a shop floor: battery power, a readable display, a case that survives being dropped. Where the weld sits in a pressure boundary the discipline is formalised, as the aerospace TIG welding guide shows.

Calibration

Analysers drift, so treat them like any other measuring instrument. Zero with nitrogen, which contains no measurable oxygen, then span against a certified gas of known oxygen content. The Process Sensing manual sets the interval between span calibrations at no more than three months and notes that output can drift about 0.11% of reading for each °C of change from the calibration temperature. Left in a hot van or a freezing yard, it is not giving you a number you can trust.

Check it against the calibration gas at the start of every shift and write the result down, not only at the formal three-month interval. Where work is audited, the reading belongs on the job file, since AWS D18.1/D18.1M treats documented condition as part of acceptance.

Sample flow, pressure and where to sample

Two settings decide whether the display means anything. Sample flow: 1 to 5 SCFH has no significant effect on accuracy, but 1 to 2 SCFH is recommended for optimum performance. Sample pressure: regulate the inlet between 5 and 30 psig. Outside that window the sensor can give false readings. The sample must also represent the joint atmosphere, so draw it from the far end of the purged volume, not the analyser inlet.

When the reading will not come down

If the oxygen value refuses to fall, the sensor is almost certainly not the problem. Work in this order:

1. Leaks in the purge. Every unsealed branch, poorly taped joint and open valve feeds fresh air back in.
2. Leaks in the sample system. If the O2 reading changes inversely with sample flow rate, the leak is in the sample system, not the purge.
3. Sample location. A probe in a dead leg or the wrong end of the pipe misreports the volume.
4. Sensor last. Only when the plumbing is right should you suspect the cell. A second analyser qualifying the sample gas settles it.

Maintaining the Purge Through Welding and Cooling: The Part Most People Miss

Ask a shop what goes wrong with back purging and you will hear about seals, flow rates, and gas quality. Few will blame the last five minutes of the job. Yet the root pass is the easiest part of a purge to get right and the easiest to undo, because the metal keeps working long after the arc goes out.

Picture the classic failure: the welder holds a clean 10 ppm oxygen reading through the root pass, lays down a textbook bead, taps the valve closed at the regulator because the weld “looks finished,” and moves on. Outside, the joint is beautiful. Inside, a hot, bare, unshielded root sits in air for the next twenty minutes while somebody grabs lunch. Two days later the client’s inspector pushes a borescope through and finds a dull grey, crusted root. That root is now a defect to be cut out and re-welded, and none of it happened during the “good” weld.

The three-stage discipline

Back purging is not a single action at the start of a weld. It is a sequence, and AWS D18.1/D18.1M and most shop specifications treat it that way. Air Products’ welder’s handbook breaks purging into three stages, and every good field procedure follows the same logic:

1. Pre-purge. Displace the air slowly and deliberately before striking an arc, then confirm with an analyser. Air Products’ guidance is to get oxygen below 50 ppm before starting. Above that, you are welding in air with extra steps.
2. Weld purge. Maintain the purge through every pass, not just the root. Every interruption gives air a route back into the bore.
3. Post-purge. Keep the gas flowing after the arc is off, until the root has cooled below roughly 250 °C.

Why the cooling threshold matters

That 250 °C figure is not a guess. TWI’s guidance on heat tint notes that most specifications require the back purge to be held until the root and hot pass, and sometimes the first few fill passes, are completed, or until the root region temperature falls below about 250 °C.

Orbital welding guidance from Intercon expresses the same rule in Fahrenheit: keep the gas flowing until the weld zone is below 500 °F (about 260 °C), and treat post-purge as non-negotiable, because the root can still be sitting near 800 °F when the arc is struck.

Back purge does not stop the heat. It only controls what the hot root is breathing while it cools. Switch the gas off early and the bore atmosphere changes from inert to oxidising while the root is still above 250 °C, which is exactly the worst possible combination.

Thick walls, hot passes and the flow adjustment

On heavier wall thickness the problem gets longer, not shorter. Air Products recommends using backing gas for the first, second and third passes as well on wall thicknesses greater than 10 mm, because the root stays hot across the interpass interval rather than just for the moment the root bead is made.

There is a second, subtler point in the same stage. TWI notes it may be beneficial to reduce the gas flow pressure slightly after the root pass to stop an over-pressure build-up, which matters especially where the root bead is thin and nearly re-melted by the second bead. A back purge that is correct for the root can become too aggressive once a second bead is laid. Excess back-purge pressure physically pushes the molten root toward the torch, and a root that gets pushed can come out concave, badly shaped, or blown out entirely. Blowing the root out on the second pass means grinding it back and redoing it, with the same purge discipline needed to do it right.

The routine is simple: strong enough flow to hold the atmosphere during the root, then a small reduction once the second pass is coming, keeping the inert atmosphere without loading the pool.

Tack welds and the second pass

If a second pass will re-melt the root, the post-purge has to survive the whole interpass period, not just the arc-off moment. Haynes International recommends that back purge gas generally be maintained for the first two fill pass welds, and that tack welds be ground back to a feather edge so the closure weld ties in without detectable defects.

A tack weld is a leak path. A crater at the end of a tack, an unfilled start-stop, or a tack that was ground roughly and left with a gap under it all give oxygen a route into the bore. On critical work, where procedure documentation and aerospace-grade TIG procedures are the benchmark, tacks are treated as part of the purge system rather than a shortcut around it. Grinding tacks to a feather edge, cleaning between passes with a stainless brush, and keeping the purge running right up to the start of the next pass are what separate a procedure that passes a borescope from one that does not.

The routine that works

Stage	What you do	Hold point
Pre-purge	Slow displacement of air, sample taken at the joint	Below 50 ppm oxygen before striking
Weld purge	Maintain through the root and all hot and fill passes	No interruption between passes
Flow adjustment	Ease back pressure after the root pass	Avoids root distortion on the second bead
Post-purge	Keep gas on after the arc is off	Until below about 250 °C (500 °F)

The habit that separates a good purge from a lucky one is unglamorous: a timer. Set a clock for the post-purge at the moment you kill the arc, and do not walk away from the job while it runs. On material such as Alloy 20 or a thin-wall pharmaceutical tube, the post-purge is the entire difference between a root that passes AWS D18 acceptance criteria and a root that has to be scrapped. Nobody has ever regretted a weld for staying under purge a few minutes longer. Plenty of people have regretted one for staying under purge a few minutes too short.

Common Purge TIG Mistakes and How to Troubleshoot Them

Most purge failures are not exotic. The same handful of problems repeat in every shop that back-purges stainless: short pre-purges, excessive flow, unsealed branches, dirty regulators, plugs welded over by mistake. Each leaves a signature, and once you read it you fix the joint instead of guessing.

Work symptom first, then likely causes, then what to change. It all comes from real purge and back-purging practice, including Intercon’s guide to purging explained and Haynes International’s back-purging requirements.

The Three Warning Signs to Recognise

Three patterns cover most of what goes wrong.

1. The reading plateaus. Oxygen falls, then stops falling at, say, 800 ppm however long you purge. Something feeds air in steadily, or the gas is swirled hard enough to pull fresh air in through every gap.
2. The reading falls, then rises. Numbers climbing back mid-purge or mid-weld mean the seal was disturbed, the flow changed, the vent blocked, or air is entering the sample line rather than the joint.
3. A good reading with a sugared root. Oxygen at 20 ppm before welding, yet the root comes out dull and crusted. Protection was lost during welding or cooling, or the sample point never represented the joint.

Symptom-to-Cause Troubleshooting Table

Symptom	Likely cause	What to change
Reading stays high	Pre-purge too short; unsealed branch; porous tape	Extend purge, seal openings, re-tape
Reading plateaus	Excess flow causing turbulence	Reduce flow, re-test, hunt leaks
Reading falls, then rises	Disturbed seal, blocked vent, air in sample line	Check dams, regulator, vent, sample line
Good reading, sugared root	Purge removed early; unrepresentative sample	Hold purge through cooling, move sample point
Concave or blown-out root	Over-pressurisation from back purge	Lower back-purge flow, open the vent
Weak or erratic flow	Dirty hose or regulator, uncapped cylinder	Purge regulator, replace hoses, cap cylinders

Working Through the Usual Suspects

• Pre-purge too short. A welder strikes the arc after thirty seconds while trapped air is still in the pipe and the analyser samples near the vent, not the worst corner. Haynes notes that back-purging must continue until the oxygen in the purge zone reaches the specified level before the root is welded, verified with an analyser rather than by eye or feel.
• Flow too high. Cranking the flow does not clear a volume faster once air is being displaced; it creates turbulence that pulls atmospheric air in through hairline gaps at a dam edge or tape seam. Back the flow down, let the volume settle, re-read, then hunt leaks.
• Removing the purge too early. A root that turns to sugar after the arc goes off is a common field failure: the hot root keeps oxidising long after the pass. TWI identifies heat tint as a direct result of oxidation, so the root must stay shielded until it cools below roughly 250 °C, a couple of minutes on thin wall and ten or more on a heavy section.
• Unsealed branches and porous tape. Close every branch, tapping, instrument nozzle and open end, and give tape enough overlap, because one poor seam can dominate the result. Where a permanent seal is impossible, use an inflatable or water-soluble plug. Intercon’s purging guide makes the point: leaks and open branches keep replenishing oxygen, so extra flow will never fix the reading.
• Contaminated gas. A regulator never purged, a hose with moisture or oil in it, or an uncapped storage cylinder all put water, oil or air into the purge. The tell-tale sign is erratic flow plus an oxygen value that climbs on its own. Grease inside the bore outgasses too, so degrease the bore, extend the pre-purge and add absorbent.
• Purge zone too large. Purging ten feet of pipe when twelve inches will do wastes gas, time and money for no benefit. Keep the enclosed volume as compact as practical; local purge chambers and thorough surface preparation habits matter more than people expect.
• Welding through a purge plug. It happens when nobody marks where the plug is. Arc heat melts plug material into the weld, and you end up grinding out filler metal to find it. Mark plug positions on the outside of the pipe before welding.
• Over-pressurisation. Too much back-purge pressure pushes the molten root toward the torch, giving a concave, blown-out or undercut root. Lower the flow and keep the vent at the high point clear.

A Diagnostic Order That Works

1. Confirm the instrument: zero it on nitrogen, span it on certified gas. A bad sensor is rarer than a bad seal, but check anyway.
2. Check the sample point: in the purge zone, away from vent and torch, and representative of the joint?
3. Check the enclosure: dams seated, branches sealed, vent open, plug intact.
4. Check the gas supply: regulator purged, hose clean, correct gas, cylinder not exhausted.
5. Reduce flow and re-read: if the plateau was turbulence, the number will fall.
6. Extend pre-purge and repeat: if it still misses target, the seal is leaking somewhere you have not found.

The pattern: oxygen that will not come down is usually a leak, a turbulence problem or a bad sample point. Oxygen that falls then climbs means something changed mid-job, so check dams, gas supply, vent and sample line. A clean reading with a dirty root means protection was lost after the measurement, which points straight back to purge duration and sample placement. More on the surrounding process sits in the welding and fabrication guides on Weld Gallery.

Purge TIG by Material and Application: Stainless, Titanium, Aluminium and Nickel Alloys

Every question in this guide, from flow rates to dam placement, comes back to one thing: how clean must the root of this joint be? An amber tint on a stainless handrail is acceptable; the same amber on a titanium pressure vessel is a rejection.

Stainless Steel: The Bread and Butter

Stainless is where most purge TIG happens, and the margin is widest. Type 304 and 316 rely on a thin, adherent chromium oxide passive film for corrosion protection, so purging keeps oxygen and free iron out of the root. TWI’s heat tint work shows a visible tint on a 316 weld corresponds to a measurable drop in critical pitting temperature. The colour is a symptom, not a cosmetic issue. Material sets the oxygen tolerance, service sets the margin.

For general fabrication a clean silver to light straw root is the goal. For hygienic and high-purity service the bar rises: AWS D18.1/D18.1M addresses new construction and may also be applied to maintenance and repair of food processing equipment, while AWS D18.2 is the visual discoloration acceptance reference most inspectors quote. For austenitic pipe and tube, AWS D10.4/D10.4M covers manual, semiautomatic, automated, mechanized and orbital processes, and AWS D10.11 covers root pass welding of pipe without backing, the coded procedure for what purge TIG does by hand.

Titanium and Zirconium: The Tightest Limits in Commercial Welding

Reactive metals work differently. Titanium and zirconium form no protective oxide, so oxygen reaching the molten root is consumed and the result is a brittle intermetallic. The Linde shielding gas guide puts it plainly: as little as 5 or 10 ppm O₂ is normally required to prevent embrittlement, and the component may need a purge gas chamber rather than a back purge alone.

Hydrogen is the second problem. Titanium absorbs hydrogen at welding temperatures and can embrittle the joint after cooling, long after the weld looks perfect. High-purity argon, clean dry regulators and hoses, and no carbon steel contamination in the setup are non-negotiable. A shop then moves to a purge TIG procedure built for aerospace work: a chamber, an analyser reading single-digit ppm, documented gas quality.

Nickel Alloys and Duplex Stainless

Nickel alloys vary widely: some Hastelloy grades are purged with argon, others accept or require an argon/nitrogen blend, and back purge requirements depend on joint condition, service temperature and code. Haynes’ guidance on back purging Hastelloy pipe systems helps you match gas choice to alloy and service. Duplex stainless adds nitrogen to the shielding gas to stabilise the austenite phase, so nitrogen in the root must be controlled too. In chemical service, shops lean on written procedures such as Alloy 20 welding procedures, not shop-floor memory.

Aluminium

Aluminium is purged with argon, but the harder half of the job is the surface. The oxide film is tenacious, melts at a different temperature to the base metal, and if it survives into the pool it rolls into the root as dross and inclusions. A clean, oxide-free surface is non-negotiable, so prepping aluminium properly before welding matters as much as the gas. On aluminium tube, purge TIG is usually thin-wall hydraulic, marine or heat-exchanger work, and the cleaning logic in our aluminium MIG guide applies.

Where Purge TIG Dominates

Industry	Typical root oxygen target	Dominant requirement
General stainless fabrication	Below 50 ppm	Corrosion resistance, code compliance
Food and dairy	Below 50 ppm	Cleanable surface, no crevice corrosion
Pharmaceutical, general	Below 50 ppm	Validated cleaning regime
Pharmaceutical WFI and clean steam	Below 10 ppm	Product contact surface integrity
Semiconductor and UHP	Below 1 ppm	Particle and outgassing control
Chemical processing and offshore	Per code and service	Pressure integrity, corrosion allowance

For chemical processing and offshore work, economics often favour purging over backing rings: a backing ring needs internal access and costs money per joint, while a purge needs only a seal and a gas supply.

Manual TIG, Orbital TIG and Automated Purge Systems

	Manual TIG	Orbital TIG	Automated purge system
Purge control	Welder’s judgement	Programmed flow and timing	Closed-loop monitoring
Repeatability	Operator dependent	High	Very high
Purge volume	Full bore, often oversized	Local chamber, small volume	Smallest, tightly sealed
Cost	Lowest	Medium to high	Highest, plus fixtures
Best fit	One-off repairs, small bore	Production pipe spools	UHP, semiconductor, high volume

Manual TIG wins on flexibility and capital cost, orbital on repeatability and purge volume, automated systems where every weld must be documented. Purge unit manufacturers describe the same progression, from dam-and-tape setups to instrumented systems with analysers and flow control built in.

Disclosure: some of the guides linked in this section are affiliate links, and we may earn a commission if you buy through them.

Safety, Cost and a Practical Purge TIG Checklist

Asphyxiation: the risk that lives inside the pipe

Purging fills a space you cannot see with a gas that cannot support life. OSHA FS-3647 notes that gases such as helium, argon and carbon dioxide displace oxygen in the air and can lead to suffocation, particularly in confined or enclosed spaces. The UK’s HSE welding fume guidance agrees: shielding gases can cause asphyxiation, usually through accumulation in confined spaces and/or displacement of oxygen.

Gas density decides where it collects. Argon is denser than air, so it settles into low points: the bottom of a horizontal pipe, a vessel sump, a floor pit. Nitrogen sits near air density. Helium rises and gathers overhead. The rule never changes: never enter or work in a purged vessel or bore until it has been vented and the atmosphere tested.

Fume, hexavalent chromium and ventilation

Purge gas is not a fume control plan. Welding stainless still generates fume containing hexavalent chromium, the compound most associated with stainless weld fume and its respiratory effects. Both OSHA and the HSE stress two controls: clean coatings, paint and solvent residue off the surface before you light the arc, and position local exhaust ventilation close to the plume source rather than across the shop. AWS D10.4M sets out recommended welding and cutting safety practices, and AWS C5.10 covers the properties, distribution, safe handling and use of shielding gases, including mixtures. Keep cylinders secured upright and stored in ventilated areas away from heat.

What purging actually costs

Purging is rarely the largest line on a stainless job, but it is the most commonly underestimated. The drivers are gas consumption, elapsed time, consumables and rework risk.

Cost driver	What drives it	How to cut it
Shielding gas	Purge volume multiplied by the number of volume changes	Shrink the enclosed volume; use a purge chamber instead of a whole length
Time	Pre-purge dwell plus post-purge cooling hold	Monitor with an analyser and purge to an oxygen target, not a fixed timer
Consumables	Tape, dams, plugs, sealant, spill kits	Reusable dams, a dedicated purge kit, fewer re-taped joints
Rework and failure	Rejected roots, borescope re-inspection, scrapped assemblies	Proper sealing and a verified low-oxygen reading before striking the arc

Intercon’s purging guidance puts it plainly: a good purging unit reduces purge volume, provides a well-sealed chamber, introduces no contaminants, and distributes purge gas slowly enough to keep turbulence to a minimum. Every one of those points lowers cost. Buying gas and fittings for the first time? A supplier overview such as AllGas Welding Supply’s guide to gases and safety is a sensible place to compare cylinder sizes and regulators.

The field-ready purge TIG checklist

1. Preparation. Clean the bore and joint faces: degrease, remove oxides and heat tint. Confirm the material and purge gas match the WPS.
2. Sealing. Fit closure dams or plugs, seal every branch and tape joint. A film of leak-detection fluid on the fittings shows where a reading refuses to come down.
3. Gas selection. Verify the cylinder label and purity. Fit a dedicated regulator and flowmeter; confirm the seal is clean and the hose free of oil and moisture.
4. Vent and feed. Vent at the high point, feed gas at the low point, and fit a diffuser so gas does not jet onto the root.
5. Flow. Start from the flow table for your pipe diameter and reduce it if the reading plateaus high or the root starts to look concave.
6. Monitoring. Run the analyser until the reading holds at or below target, with the sample line drawing from the joint atmosphere. Never start on a timer.
7. Welding. Keep the purge running through every pass, and hold it after the arc is off until the root cools below roughly 250 °C.
8. Post-purge and close-out. Release the gas slowly, remove dams and plugs, and check the root colour before the joint is released for inspection.

Records to keep

• Welding procedure, joint identification and welder or operator qualification.
• Purge gas type, cylinder or lot number, and regulator details.
• Flow rate, pre-purge duration and the oxygen reading achieved before welding.
• Readings taken during welding and at the end of the post-purge hold.
• Analyser identity and calibration date.
• For hygienic service, the documentation required by AWS D18.1/D18.1M.

Key takeaways and where to go next

Purge TIG is a measurement problem wrapped in a sealing problem wrapped in a safety problem. Shrink the volume, seal the joints, get an honest oxygen reading, and hold the gas on until the root is cold. Treat the bore as a confined space, treat stainless fume as a respiratory hazard, and keep the records. For the standard-practice side, read AWS D18.1/D18.1M and the D18 committee’s sanitary applications guidance for hygienic work, and AWS C5.10 for gas handling. For equipment selection, an airgas welding equipment overview and aerospace TIG welding standards show how purge control and weld traceability are handled where quality is audited.

Frequently Asked Questions

What is purge TIG welding and when do you actually need it?

Purge TIG welding is gas tungsten arc welding where the back side of the joint is flooded with an inert gas that displaces air while the root is welded from one side. You need it whenever the back of the weld is going to be in service and the base material is oxidation-sensitive — stainless steel, nickel alloys, titanium, zirconium and aluminium. Carbon and low-alloy steels generally do not need back purging for most applications.

How much oxygen is acceptable in a purge for stainless steel?

For many stainless applications, 0.01% (100 ppm) is considered satisfactory, though some surface oxidation will still occur. For extreme service conditions where pitting or stress corrosion cracking is a concern, 20 ppm may be required. In hygienic and food-grade work, most fabricators target below 50 ppm, and AWS D18.2 sample number 4 or greater is generally considered unacceptable.

How long should I purge before starting to weld?

Time alone is not a reliable measure — the answer is the oxygen reading. A common rule of thumb is to exchange roughly 5 to 7 times the enclosed volume of the purge zone, and to keep purging until the analyser shows you are at target, typically 70 ppm or below for stainless and 50 ppm or below for titanium. In practice that often takes two to four minutes with a properly sealed purge unit.

Can I use nitrogen instead of argon for purging stainless steel?

Argon is the default because it is fully inert and forms no compounds. Nitrogen is used in some applications, but it is an austenite stabiliser and can be absorbed into the weld metal, producing an unacceptable microstructure, and it is a poor choice for aluminium and titanium. Argon/nitrogen mixtures are sometimes specified for duplex and super-austenitic stainless steels, and nitrogen/hydrogen mixes are used where faster oxygen scavenging is needed.

Why is my root pass sugared even though my oxygen reading was good?

It usually means protection was lost after the analyser reading was taken, or the sample point did not represent the joint. Common culprits are stopping the purge before the root cooled below about 250 °C, a leak that opened up mid-weld, or a gas jet disturbing the pool. Keep the analyser connected during welding, watch for a reading that climbs, and maintain the purge through the hot pass and into cooling.

Should I keep purging until the whole pipe is cool?

No. Most specifications only require the back purge to be maintained until the root and hot pass, and sometimes the first few fill passes, are complete, or until the root region falls below roughly 250 °C. For wall thickness over 10 mm, backing gas is generally recommended for the first three passes. Holding the purge for the entire cooldown is usually unnecessary and wastes gas.

What is the difference between an oxygen analyser and an oxygen-deficiency alarm?

They solve different problems. A trace oxygen analyser measures the low ppm levels inside the purge zone so you know when it is safe to weld; a personal oxygen-deficiency alarm warns you that the ambient air near the pipe is being depleted. A personal alarm is not a substitute for measuring purge quality, and you need both when working inside a pipe or in a confined space.

Does the root bead of a purged stainless weld still need pickling?

Usually not to the same degree. A properly purged root is bright and clean, with no scale to remove, and post-weld pickling passivation of the ID is often unnecessary. If the root was heat tinted or oxidised, the scale and the chromium-depleted layer must be removed by post-weld cleaning to restore corrosion resistance, which means pickling or mechanical cleaning followed by passivation.

Is purging dangerous for the welder?

Yes, and it is underappreciated. Argon, helium, nitrogen and carbon dioxide displace oxygen and can cause asphyxiation, especially in confined spaces and inside pipe, which is why OSHA and HSE both warn against welding in confined spaces without proper ventilation. Purging also adds fume from stainless welding, which contains hexavalent chromium, so local exhaust ventilation and fume control are still required.

Do I need to purge when welding stainless steel with GTAW on a flat plate?

Usually not. On a flat plate or an open joint, the root is exposed to air only momentarily and the heat tint that forms is thin and usually removed by pickling. Purging becomes necessary once the joint is closed enough to trap air, or whenever the inside surface will remain visible in service, or when the code, client or WPS calls for it — which is almost always in hygienic, pharmaceutical and chemical process work.