What Gas to Use for MIG Welding Stainless Steel (2026 Guide)

You finally got your stainless TIG-looking welds on a kitchen backsplash project, switched your MIG over to the cheap C25 bottle you had left from mild steel work — and the weld came out sugary, gray, and peppered with brown glass-like islands of slag. The puddle kept boiling and spitting, and grinding it back revealed sugary porosity underneath. Nine times out of ten, that failure isn’t your technique — it’s the bottle. Choosing the right shielding gas for stainless is the single cheapest upgrade you can make to your stainless MIG results.

Why Shielding Gas Matters So Much on Stainless Steel

The short answer: for MIG welding stainless steel you want an argon-rich gas with a small, controlled addition of CO2, oxygen, or helium. That means a tri-mix or a 98/2 argon/CO2 blend as your default, and 100% CO2 or a 75/25 C25 mix only as a last resort. Stainless earns its corrosion resistance from a thin, self-healing chromium oxide film, and the arc tears that film apart. The gas protects it while it reforms.

What the shielding gas actually does

In GMAW (also called MIG), an arc forms between a continuously fed filler wire and the workpiece. If air gets into that space, the weld pool picks up oxygen and nitrogen while molten. The gas you pay for displaces that air and keeps the pool chemically clean as it solidifies.

Think of it as environmental control, not cooling: the arc is far too hot for gas to protect the base metal directly. Protection comes from the gas column flowing around the wire and over the weld, sweeping air away from the hot pool. The TWI guide to shielding gas says it plainly: gas protects the molten pool and cooling weld from oxidation, and its composition sets the transfer mode, arc stability, and penetration.

Why stainless is far less forgiving

By definition, stainless steel contains a minimum of about 10.5% chromium. Exposed to oxygen, that chromium forms a passive oxide film only a few nanometers thick and roughly 100 times denser than the steel beneath, and that film is what stops corrosion. Nickel adds stability and toughness; molybdenum in 316 and super duplex grades improves pitting resistance.

Welding is destructive to that film. The pool, solidification zone, and heat-affected zone all oxidize, and the HAZ often cannot re-form the passive layer because of sensitization, where chromium ties up in carbides at grain boundaries.

Mild steel mainly wants to avoid oxygen pickup because it makes the weld brittle and hard. Stainless has a bigger problem: oxygen pulls chromium into the oxide film instead of leaving it available to reform the passive layer. That leaves less chromium for corrosion protection, and the trouble shows up months later as rust staining or pitting along the weld. Welding alloy 20 shows how sensitive high-alloy stainless families are to atmosphere control.

Heat tint and sugaring

A good stainless weld shows a narrow band of straw gold, then blue, then grey around the bead. That rainbow is heat tint, the normal oxidation of the surface at elevated temperature, and a thin, uniform band is acceptable on most work. Widen it, deepen the blue and brown, or let it spread into a rough crystalline crust, and you are looking at something different.

Sugaring is the welder’s word for heavy oxide formation on stainless: a matte, sandy, granular surface that looks like spilled sugar. That sugar is a thick layer of chromium and iron oxides that never melted out of the pool. It is porous, it traps moisture and chlorides, and it invites rust bleeding and corrosion undercut. The brown, blue, and purple colouring is the diagnostic clue: oxygen in the shielding envelope reached the chromium at welding temperature. TWI, Lincoln Electric, and ESAB all point at the same fix: an argon-based blend, tight shield coverage, and a short torch-to-work distance.

Carbon pickup

The other side of the coin is carbon. CO2 helps arc stability and controls the transfer mode, but it is also a carbon source, and too much of it drives carbon into the pool. On stainless that is a direct threat, because carbon ties up chromium as chromium carbide and reduces the chromium available for the passive layer.

That is why 100% CO2 is a poor fit for stainless: no argon to stabilise the arc, hard carbon pickup, and a spray-like transfer carrying far too much heat for thin sheet. It is also why the 75/25 C25 blend, excellent on mild steel, is a compromise most stainless authorities steer away from. AWS and Miller literature for stainless MIG consumables consistently recommends argon-based blends over carbon-dominant mixes.

Short-circuit transfer

Stainless fabrication is usually thinner material, often 18 gauge up to 1/4 inch, and often welded out of position: vertical, overhead, or on an angled fixture. That combination pushes you into short-circuit transfer, where the wire tip touches the pool and shorts the arc repeatedly, transferring small droplets at lower current and lower heat input than spray.

Short-circuit is the most common GMAW mode on stainless sheet, and it is unusually sensitive to gas composition. The arc extinguishes and re-strikes constantly, so the pool is repeatedly exposed wherever the gas column fails to protect it. Wrong gas gives you a boiling, spitting puddle that deposits sugar and porosity. Argon is why you want it in the blend: it holds the arc stable through those re-strikes in a way CO2-dominated mixes cannot, while pure argon alone tends to give an unstable short-circuit arc with poor wetting and a narrow, cold-looking bead.

What gas choice changes downstream

Gas choice shows up in several other measurable ways:

• Bead appearance. An argon-based blend gives the smooth, domed, clean silver bead people mistake for TIG work. A CO2-heavy mix leaves a rougher, darker bead with spatter fused into the surface.
• Heat input and the HAZ. Argon-helium runs a hotter, wider arc, raising deposition rate but widening the HAZ. Argon-CO2 keeps heat lower and the HAZ narrower, which is why shops like it for 16 to 20 gauge work.
• Cleanup and rework. Clean gas means less pickling and less grinding, and a sugared bead still has to be ground and passivated before it passes a sanitary or corrosion-critical inspection. Extra spatter means extra wheels and extra hours, and on a quoted job that labour decides whether you made money.

Bottom line: on mild steel a mediocre gas costs you cosmetics; on stainless the same gas costs corrosion performance, weld quality, and rework hours. That is why the right shielding gas is the cheapest upgrade available to a stainless MIG welder.

The Main Contenders: Common Gases for Stainless MIG

Only four gases really matter in stainless MIG, and once you know what each does, every bottle on the shelf makes sense. Argon is the base gas for nearly all stainless work because it is inert, will not oxidize your puddle, and gives a stable, clean arc. Helium is blended in to widen the arc and raise its temperature. CO2 is blended in for penetration and short-circuit stability. Oxygen, in tiny amounts, keeps the arc from wandering and helps the puddle wet out.

100% CO2: Why It Is Rarely the Right Answer

CO2 is cheap, available everywhere, and gives genuinely deep penetration on mild steel. That is exactly why it is a trap on stainless. Carbon dioxide is reactive: when it contacts hot stainless it oxidizes the weld and drives carbon into the metal. You get a heavy oxide layer, lots of spatter, and a dull, dirty bead. On corrosion-critical work or anything with a cosmetic finish, it is avoided.

75/25 and 90/10 Argon/CO2 (C25 and the Universal Blend)

These are the workhorse blends for mild steel, found in nearly every shop, and they are the bottles most people already own, which is why they end up on stainless by accident. C25 (75% argon, 25% CO2) is designed for short-circuit and spray transfer on carbon steel; that much CO2 on stainless causes spatter, oxide coloring, and carbon pickup. The gentler 90/10 can produce a workable bead on thin stainless, but it is a compromise, not a stainless gas.

98/2 Argon/CO2: The Budget Stainless Choice

Drop the CO2 to 2% and the behavior changes. At that level, CO2 stabilizes the short-circuit arc without dominating the weld chemistry, which is why 98/2 is the common low-cost alternative to helium tri-mix and is widely recommended for short-circuit welding on thin stainless sheet and tubing.

The trade-offs are predictable: the arc is narrower and cooler than a tri-mix, the bead is flatter and less smooth, the toes tie in less gracefully, and you see more spatter. On a 16-gauge backsplash panel or a length of 304 tubing, it gets the job done.

The Oxygen-Containing Blends: 98/2 Ar/O2 and 90/8/2 Ar/CO2/O2

A small amount of oxygen, roughly 1% to 2%, does something counterintuitive: it helps the arc. Oxygen sharpens the arc and reduces the wandering pure argon is known for on stainless. It changes how the puddle behaves too, giving better wetting and a bead profile that flows rather than balls up.

The catch is that it must be controlled tightly. Too much oxygen and you simply oxidize the weld, the exact failure you were trying to avoid. These blends turn up more in automated, high-production stainless work like tube welding and food equipment fabrication than in a typical shop. The ESAB stainless welding guidance makes the same point: oxygen improves arc characteristics in small, carefully controlled amounts. The 90/8/2 argon/CO2/oxygen blend is the family hybrid: CO2 for penetration and stability, argon for the inert base, and a touch of oxygen for arc shape and wetting. It behaves much like 98/2 Ar/CO2, but with a cleaner, better-wetting bead.

Helium Tri-Mix: 90% He / 7.5% Ar / 2.5% CO2

This is the industry-standard tri-mix for short-circuit GMAW on stainless, often sold simply as “Stainless Tri-Mix” or under a vendor name like Tri-Mix 90/7.5/2.5. It is also the gas most Lincoln Electric and Miller recommend for common stainless work.

Helium is the star. It raises the arc voltage and widens the arc footprint, producing a hotter, broader, smoother arc with much better tie-in of the toes. The bead comes out noticeably rounder and more “wet looking” than an Ar/CO2 weld, and on a decorative railing that difference is everything. The argon keeps the base gas inert, while the 2.5% CO2 stabilizes short-circuit transfer and helps the puddle flow without wandering.

85/10/5 and the Helium-Heavier Variants

As you increase helium and reduce argon, the arc gets hotter and the bead smoother, but the cost climbs and short-circuit transfer gets less stable. The 85% helium blends are pushed for spray or pulsed transfer on thicker stainless, where extra heat matters more than arc stability at low current.

Blend	Arc	Best For	Watch Out
100% CO2	Deep, harsh	Rarely	Spatter, oxidation, carbon pickup
75/25 Ar/CO2	Punchy	Thin sheet only	Oxide color, rough bead
98/2 Ar/CO2	Clean, easy	Budget short-circuit	Flat bead, tight tie-in
98/2 Ar/O2	Sharp, wets well	Production, tube work	Oxidation if mix off
90/8/2 Ar/CO2/O2	Stable, wets well	Sheet metal, food equipment	Clean gas, steady flow
90/7.5/2.5 He/Ar/CO2	Hot, broad, smooth	Standard for MIG	Cost, more flow
85/10/5 He/Ar/CO2	Very hot, less stable	Thick, spray or pulsed	Cost, low-current

One practical note before you buy: check the cylinder label rather than trusting color bands or memory. Suppliers mix to order, and cylinder blends drift as bottles age, so an old tri-mix may not deliver the mix on the label. Cylinder handling and labeling notes on gas supplier resources are worth a read, and the next question comes down to cost, flow rate, and bead appearance.

I’ll rewrite this section to hit the 600–900 word range while keeping the heading, facts, and all links intact.

Helium Tri-Mix vs. Argon/CO2 Blends: Real-World Trade-offs

Two families of shielding gas cover nearly every MIG welding job on stainless steel in North America. The first is the helium-bearing tri-mix, typically 90% helium, 7.5% argon and 2.5% CO2, with an 85/10/5 version showing up in heavier shops. The second is the argon/CO2 blend, most often 98% argon with 2% CO2 or the old 75/25. Both will weld 304 and 316 stainless. Both can produce a perfectly serviceable fillet weld. They are not interchangeable, though, and the differences show up in three places: your cylinder bill, your gas flow settings, and the bead coming off the plate.

What Helium Actually Changes in the Arc

Helium is the single biggest technical difference between the two approaches. Because helium raises arc voltage at a given wire feed speed, the arc runs longer and hotter for the same amperage. The practical result is deeper, wider fusion, better wetting, and a smoother, flatter bead profile. That extra heat input is a blessing on heavy plate and a liability on thin sheet, which is a large part of why the decision splits so cleanly by thickness and position.

Shops that regularly weld 3/8 inch and up, or work out of position on stainless tanks and pipe, lean on that behaviour. The small CO2 fraction in an argon blend stabilises the arc, adds a touch of penetration, and holds arc voltage lower, so the heat stays more manageable on light gauge. General guidance on how each gas component affects the arc is well documented by TWI’s shielding gas explainer, and the practical parametrising guidance from ESAB and Lincoln Electric is worth reading before you dial anything in.

Cost and Availability: The Part Nobody Mentions

Argon/2% CO2 blends are markedly cheaper than tri-mix and are stocked by essentially every rental supplier, welding supply house and local gas dealer. A 98/2 cylinder is an impulse purchase. A 330 CF tri-mix cylinder is a planned purchase, frequently a special order with lead time, and it costs meaningfully more per unit of gas.

Then there is the consumption factor. Because tri-mix runs at higher flow rates, you burn through the bottle faster. The cost per foot of finished weld goes up twice: once because the gas is dearer, and again because you need more of it. If you are quoting a job rather than building a weekend project, price the cylinder count before you price the wire. The Airgas welding supplies guide and the AllGas welding supply guide both walk through cylinder sizes and what distributors typically keep on the rack, which is useful for planning ahead.

Flow Rate: Where Setup Gets Fussy

Helium tri-mix commonly runs at 35 to 45 CFH, while argon/CO2 blends sit around 15 to 25 CFH. The reason is density. Helium is far lighter than argon or CO2 and escapes the shielding envelope quickly, so you need more volume of gas to keep the same coverage over the puddle. Reference flow ranges are tabulated across the industry on Engineering ToolBox‘s shielding gas tables.

Higher flow is not free. Push it too far and the jet turbulences, entrains room air, and you are back to the oxidised, sugary bead you bought the good gas to avoid. That means the tri-mix setup needs a properly sized gas lens and nozzle, a shorter contact tip-to-work distance, and more attention to drafts. Argon/CO2 is far more forgiving on all three counts.

Head-to-Head Comparison

Factor	Helium Tri-Mix (90/7.5/2.5)	Argon/CO2 Blend (98/2 or 75/25)
Typical flow rate	35 to 45 CFH	15 to 25 CFH
Arc voltage / heat input	Higher voltage, higher heat input at given wire speed	Lower voltage, more moderate heat input
Penetration profile	Deeper, wider fusion zone	Shallower, more stacked profile
Bead appearance	Clean, well-wet, brighter surface with visible colour on 304 and 316	Clean on 98/2; 75/25 can show slight oxidation
Spatter	Noticeably less spatter	More spatter, especially with higher CO2
Best positions	Vertical up, overhead, outside pipe and tube joints	Flat and horizontal, short-circuit on sheet
Material thickness	3/16 inch and up, or thin wall where heat must be spread	16 gauge up to 1/4 inch
Cost and availability	Higher cost per cylinder, often special order, 330 CF typical	Cheap, widely stocked, 250 CF common
Setup tolerance	Demanding: gas lens, nozzle and stickout all matter	Forgiving, even with a mediocre gas cup

Scenario Guidance

• Kitchen and food service equipment. 304 and 316 sheet at 16 to 20 gauge. Use 98/2 argon/CO2. The bead is clean, the heat stays low enough to limit distortion on cabinet panels and sink bodies, and passivation handles the corrosion question, not the gas. Tri-mix adds cost and burn-through risk for little visible gain. Where the weld is fully exposed and cosmetic, a light tri-mix touch-up pass is sometimes used.
• Brewery and food tank seams. 304 and 304L at 3/16 to 1/4 inch, frequently welded vertical up. This is tri-mix territory. The higher arc voltage and better wetting let you hold a consistent, well-wetted bead up a vertical seam without chasing spatter, and the deeper penetration profile supports the pressure and sanitation requirements.
• Automotive exhaust work. 304 and 409 tubing in 16 to 20 gauge. 98/2 is plenty for most tube. The bigger issue is back-purging on 409, since the ferritic grade picks up nitrogen and carbon unless the inside is shielded with low-flow argon. Tri-mix earns its cost here mainly on the outside, where appearance and spatter cleanup on visible pipe matter.
• Decorative rails and architectural work. 316 tube, often 1.2 to 1.5 mm wall. Tri-mix delivers the most consistent surface finish with the least cleanup, and the bright, well-wetted bead photographs better under showroom lighting. 98/2 will work, but expect more spatter on the floor and a little more time with the grinder. For structural work with an AWS D1.6 acceptance requirement, pick the gas that gives you the most repeatable profile.

The bottom line is less about which gas is better and more about what you are welding. Helium tri-mix buys you heat, penetration and a clean bead, and it bills you for both the gas and the setups that go with it. Argon/CO2 blends are cheaper, easier to find, easier to run and completely adequate for thin sheet and general fabrication. Buy the bottle that matches the thickest metal and the worst position in the job, and use the cheaper blend for everything else.

Matching Gas to Transfer Mode: Short Circuit, Pulsed, and Spray

Gas choice and transfer mode are one decision split in two. In GMAW the wire feeds continuously into the arc, and the shielding gas controls how the arc behaves, how the droplet detaches from the tip, and how much heat lands on the workpiece. Pick a gas that fights your transfer mode and you will spend the job fighting spatter and an unstable arc.

Most shops run three modes on stainless: short circuit for thin material, spray for heavier sections, pulsed in between. For a refresher on the hardware, our Air Gas Welding Explained guide covers cylinders, regulators, and flowmeters.

Short-Circuit Transfer: Thin Stainless

Short circuiting is the default for thin-gauge stainless, and it is where most food-equipment and sheet-metal work ends up. The wire touches the pool, the arc shorts, a droplet transfers, the wire feeds forward again. It is forgiving and limits heat, but the arc is repeatedly extinguished and restarted, which is where gas chemistry matters most.

The usual answer is 98/2 argon/CO2. That 2% carbon dioxide stabilizes the arc and gives the puddle decent clean-up without the heavy spatter and hard oxide of 75/25 (C25). A tri-mix also works, though most fabricators save expensive helium blends for thicker plate or out-of-position work.

What to avoid in short circuit is pure argon. On stainless a pure argon arc is notoriously unstable here: it wanders, hunts for a grounding point, and leaves a stiff puddle that fights the wire. Welders who reach for argon “because it is stainless” end up with a gray bead and a sore throat.

More CO2 is not the fix either. Going from 2% to 10% (90/10 Ar/CO2) buys extra penetration but adds spatter and oxide film, the last thing you want on a corrosion-resistant stainless-family joint.

Pulsed GMAW: The Middle Ground

Pulsed GMAW is the most versatile mode for stainless. The machine modulates current rapidly so droplets transfer in a controlled, spray-like fashion at an average current low enough to limit burn-through on thin material. Because the underlying transfer is essentially spray, the gas must support a spray arc.

That means helium-rich blends. Commercial stainless mixes for pulsed work commonly run 60 to 70% helium with the balance argon plus 1 to 3% CO2 or oxygen, and heavier helium tri-mixes in the 75 to 95% helium range give the flattest possible profiles. Helium lifts arc voltage, producing a longer arc, a wider bead, and more heat to the joint. Argon/oxygen mixes also work in pulsed mode at higher currents.

The trade-off is that helium’s extra heat widens the HAZ and adds distortion on thin or stressed parts. On nickel alloys and other stainless-family materials where heat input control is critical, follow the alloy-specific filler and procedure guidance instead of copying a carbon steel chart.

Spray Transfer: Thicker Stainless

Spray transfer needs higher current and wire speed than short circuit, plus an argon base with a small oxidizing addition, roughly 1 to 3% CO2 or oxygen. That addition stabilizes the arc and improves puddle fluidity and clean-up, which matters on stainless because you want residual oxide to float out rather than trap in the bead.

More CO2 gives deeper, narrower penetration like mild steel practice, plus more spatter. Oxygen instead gives a cleaner, more fluid puddle with less spatter but a heavier oxide film.

Why Oxygen Must Stay Tight

Push oxygen past roughly 2 to 3% and you start paying for it. The puddle gets fluid and cleans up beautifully, but oxide films get heavier, the weld picks up more surface oxidation, and there is a real risk of reduced corrosion resistance on stainless, defeating the point of using stainless.

The TWI guidance on shielding gas selection and AWS filler metal and gas combinations both stress matching the blend to the transfer mode and base metal rather than swapping cylinders mid-project. Stainless filler metal and process limits are usually detailed under AWS A5.18 and A5.9, via aws.org.

How Gas Shapes Penetration and the HAZ

Gas blend	Typical mode	Penetration	HAZ and heat effect
98/2 Ar/CO2	Short circuit, thin sheet	Moderate, rounded	Low heat, narrow HAZ
90/10 Ar/CO2	Short circuit to light spray	Deeper, narrower	More spatter, wider HAZ
Ar with 1 to 3% O2	Spray, medium plate	Moderate, fluid puddle	Cleaner puddle, heavier oxide
He-rich tri-mix (60 to 70% He)	Pulsed, spray, out-of-position	Wide and shallow	Higher voltage, broader HAZ, distortion

For machine-specific blend and parameter charts, see Miller, Lincoln Electric, and ESAB. Before you buy, make sure the cylinder you order from a supplier such as AllGas Welding Supply is a stainless-rated blend and that the label matches the transfer mode you plan to run.

Gas Parameters That Make or Break Your Weld: Flow Rate, Stickout, and Nozzle Care

Pick the perfect tri-mix and the weld can still go wrong. Gas, wire, and settings all correct, and you still grind off a gray, oxidized bead because of what happened in the ten seconds around the arc. Stainless is more gas-sensitive than mild steel, since its chromium oxidizes rapidly at welding temperature.

Flow Rate: Less Is Usually More

For most stainless short-circuit MIG work, start at 20 to 35 CFH depending on blend and nozzle size, then confirm with an actual flow meter on the outlet side of the regulator, because cylinder pressure and regulator condition skew what reaches the torch.

Argon blends want the lower end. Helium tri-mixes take more, since helium is lighter and shields less per unit volume. C25 and other CO2-heavy mixes run richer, one of several reasons they perform badly on stainless.

More gas is not better shielding. Excessive flow creates turbulence at the nozzle exit, and turbulent gas entrains atmospheric oxygen into the weld zone, so you pay for shielding while destroying it. Back the flow down and retest.

Gas blend	Starting flow	Notes
98/2 Ar/CO2	20–30 CFH	Stable, economical, thin sheet
Tri-mix (90/7.5/2.5 Ar/He/CO2)	25–35 CFH	Run higher; verify with a meter
Helium-rich tri-mix (85/10/5)	30–40 CFH	Hotter arc, needs more flow
100% CO2 or C25	Not recommended	Unsuitable for stainless

Contact Tip-to-Work Distance: The Single Biggest Lever

If you change one thing, change contact tip-to-work distance (CTWD). Manufacturers generally recommend 3/8 to 1/2 inch (10–13 mm) on stainless, shorter than most welders use on mild steel.

The gas column protects the hot puddle and the solidifying bead. Hold the torch longer and the gas travels farther while hot stainless passes through more air on the way out of the nozzle, and that exposure lets oxygen in, causing heat tint, sugar, and lost corrosion resistance. If joint position or torch geometry will not hold a short CTWD, many fabricators lean on gas shielded MIG with a good nozzle setup rather than a long stickout.

Nozzle Size, Gas Lens, and Nozzle Care

Nozzle diameter shapes the shielding stream. Too small starves the weld of coverage; too large and the gas spreads and lets turbulence set in. The gas lens is the internal diffuser that turns a swirl into a column, and it must match nozzle size and flow.

• Match the nozzle to amperage and flow, not to what came in the box. A 3/4 inch nozzle at 30 CFH of tri-mix guarantees turbulence.
• Check the gas lens for blockage; spatter inside the nozzle distorts flow and shows up as erratic bead color.
• Replace nozzles on stainless. Brass nozzles used on mild steel pick up copper, and copper contamination discolors the bead.
• On thin sheet, a smaller nozzle with adequate flow beats a big nozzle at reduced flow.

Drafts, Torch Angle, and Coverage Checks

Stainless welds are extremely sensitive to ambient airflow. An open bay door or a shop fan aimed at the booth can strip argon out of the shielding column and oxidize the weld in front of you.

• Use a curtain or screen to block cross drafts at the workstation.
• Turn off or redirect fans and air conditioning aimed at the booth, and point the nozzle slightly into any remaining draft.
• On closed vessels, cylinder-shaped sections, and food equipment, back-purge with an inert shielding gas to keep the root side clean. Non-negotiable for 409 exhaust tubing and sanitary welds you plan to passivate.

Hold the nozzle so the gas column reaches the joint by the shortest, symmetrical path; a wandering nozzle leaves one side clean and the other sugar-coated. Check coverage before striking an arc too: with the torch unlit at your intended CTWD, a healthy setup gives a straight, coherent stream, while a swirling plume means your flow is too high or the gas lens is wrong.

Pre-Flow and Post-Flow

Pre-flow purges the hose, regulator, and nozzle of air before the arc strikes, typically a second or two. Post-flow gets skipped more often and matters just as much: three to five seconds after the arc dies protects the still-hot weld and crater from oxidation as it cools. Crater cracking and the brown sugar ring around the bead are usually post-flow problems, not wire or voltage problems. If your machine lacks a post-flow timer, find it in the setup menu and learn to keep gas flowing after the arc dies; Lincoln Electric and ESAB document these settings in their manuals, and the AWS Welding Handbook covers shielding gas fundamentals.

Regulator, Hose, and Clean Consumables

Use a regulator rated for the gas and pressure. Argon blends and helium tri-mixes do not need high delivery pressure; most GMAW setups run best at 15 to 25 psi at the machine, and more pressure does not deliver more shielding. Inspect the hose for contamination, a swollen liner, or moisture, since oil and moisture in the gas path cause porosity. Secure the cylinder upright and out of walkways and keep the area ventilated, as OSHA hot work guidance requires.

Stainless needs clean, dedicated consumables: copper liner, dedicated gun, and a filler spool that never shares a stand with mild steel wire, or contamination reaches the puddle and causes porosity and discoloration. Store wire dry and sealed, wipe the spool face before loading, and never handle it with oily gloves, since oil and silicone break down under arc heat into gases that cause porosity.

Troubleshooting Stainless MIG Problems Rooted in Gas Choice

Most shops blame the machine first: bump wire feed, trim a volt, swap the contact tip, then suspect the cylinder an hour later. On stainless that order is backwards. A gray, sugared, spitting bead is more often a gas or cleanliness problem than a settings problem.

Here is a symptom-to-fix guide. Change one variable at a time.

60-Second Triage Before You Touch the Machine

1. Is the oxide uniform or random? Even gray-brown coverage across the whole bead points to gas or shielding coverage. Random streaks and islands mean contamination or a draft.
2. Does the problem follow the cylinder? If the same wire, voltage and joint give a clean bead on a different bottle, stop tuning the machine.
3. What changed in the last week? New spool, new bottle, new hose, an open shop door, an aerosol sprayed nearby. Most stainless MIG mysteries come down to one of those.

Gas deserves this attention because the blend governs arc stability, penetration and bead appearance more than most first-time TIG converts expect. The TWI guide to shielding gas for steel treats mix selection as a core process variable, not a consumable detail.

Symptom, Likely Cause, Fix

Symptom	Likely gas-related cause	Fix
Sugary, gray, crumbly oxide	Oxygen contamination: wrong gas (C25 or 100% CO2), excessive stickout, air drawn into the shield	Move to 98/2 Ar/CO2 or a tri-mix, cut stickout to 3/8 to 1/2 inch, block drafts
Heavy spatter off the bead	CO2 above roughly 3 to 5 percent, or voltage set for a different mix	Switch to 98/2 Ar/CO2 or Ar/O2, re-tune voltage and travel speed
Pinholes, wormholes, no root reinforcement	Low flow, a leaky draft, moisture or oils on the wire	Check flow at the nozzle, shield the work, degrease and dry the spool
Dull gray bead, no sheen	Hydrocarbon or silicone contamination, or wrong polarity	Degrease base and wire, clear aerosols, confirm DCEP (electrode positive)
Dark blue or brown heat tint	Too much heat input or poor trailing coverage	Lower heat, boost coverage, trail-shield thin work
Burn-through on thin sheet	Rarely the gas; usually excess heat and arc length	Cut wire feed and travel speed, add backing, consider pulsed GMAW

Reading the Symptoms in Detail

• Excessive spatter. Spatter climbs once CO2 passes about 3 to 5 percent in a stainless blend. CO2 makes the arc hotter and more fluid, so droplets get larger and less stable. A C25 bottle bought for mild steel will spit and leave a brownish-gray bead, which is expected, not bad wire. Lincoln Electric and ESAB note that a blend change needs a parameter reset, so treat the swap as a fresh setup.
• Porosity. Porosity in stainless MIG is often traced to wire drawing lubricants, oil or moisture rather than gas alone. Round, clustered holes point to the spool; holes at the start and end of a run point to the gas shield starving as you accelerate or as the arc breaks. Cleanliness of base and filler metal is essential, so wipe the spool before you swap bottles.
• Gray, dull beads. A flat, lifeless bead with no reinforcement is usually contamination. Cutting oil, a dirty joint, or silicone from an aerosol sprayed near the booth break down at welding temperature and contaminate the weld. Check polarity too: stainless MIG runs DCEP, meaning the electrode is positive. DCEN gives a stiff, dirty bead that mimics a gas problem but is not.
• Heat tint. Brown or blue tint does not always mean a defective weld, and light straw or gold on a thin section is usually fine. Heavy dark oxide suggests too much heat input or poor shielding on a chromium-bearing alloy. Because stainless relies on chromium to rebuild its protective oxide layer, heavy oxidation is not just cosmetic.
• Burn-through on thin sheet. Rarely a gas fault, but gas still shapes it: longer stickout and higher arc voltage raise heat input, which is how a 26-gauge panel ends up with a hole in it. If burn-through follows a gas change, re-trim arc length and travel speed.

Gas Problem, Technique Problem, or Material Problem?

A defect appearing only at the top of the bead, at the ends of a run, or in one position is technique: travel angle, speed, heat input. A defect that is random and intermittent, with clean metal between bad spots, is contamination or drafts. A defect that is consistent on every joint is the gas, the polarity, or the material. That triage saves a lot of knob-turning.

Keep the Wire Clean and the Bottle Honest

Stainless filler wire must be stored as it came: sealed in its original packaging or a plastic bag, out of standing moisture, off a concrete floor. Wipe the spool before welding, and never use a carbon-steel brush or a rag that has touched mild steel. Oil breaks down into gases during the arc, creating porosity and gray deposits that look exactly like a shielding problem.

Give the cylinder label and the mix percentage the same respect. welding gas supply and cylinder labeling is worth a read before you buy, since aged or mismatched bottles are a common, expensive surprise.

And if the root cause is airborne contamination, no gas blend on the market will save you. Fix the environment, clean the metal, then re-evaluate the bottle. Ventilation and fume control matter here too: OSHA hot work guidance covers the general requirements for a safe welding area.

Gas Selection by Application: Food Service, Exhaust, Decorative, and Structural

No single gas wins everywhere. A tri-mix that gives a mirror finish on a 1.5 mm handrail burns through a 0.6 mm exhaust tube, while 98/2 that runs a brewery tank all afternoon leaves you grinding heat tint off a showcase railing. Match the blend to the metal, thickness, position, and the finish standard the customer actually paid for.

Food, Dairy, and Pharmaceutical Equipment

Sanitary stainless welding is a cleanability problem, not a strength one: a rough profile, heat tint band, or loose oxide island traps food residue, and an inspector finds it. The accepted standards are AWS D18.1 for stainless and D18.2 for hygienic welding, and D18.2 includes a chart for evaluating heat-tint discoloration, where a lower rating means better shielding. Standards sit at aws.org.

Most food equipment and tank work calls for full-penetration welds ground flush and passivated, so two choices dominate:

• Tri-mix (90% helium / 7.5% argon / 2.5% CO2, or 85/10/5 variants) gives the cleanest, most oxide-free bead and the rounded stainless profile that grinds flat without chatter. Helium raises arc voltage, spreading heat and slowing the puddle, so you avoid melting through and gain time to clean the root between passes, at the cost of higher heat input on the same amperage. Slow the wire speed and control travel on thin sanitary sheet.
• 98/2 argon/CO2 is the budget option, widely used where the joint gets ground anyway. Argon keeps the bead clean and the arc stable, and the small CO2 keeps short-circuit from going harsh. Expect more heat tint and a flatter, spatter-prone bead, but on 2 mm to 3 mm 304 and 316L it is workable and economical.

One caution: 98/2 is not “food grade” gas in the sense of touching the product. Food safety comes from steel grade, grind, passivation, and cleaning. But gas drives post-weld cleanup, and cleanup is a real cost line.

Decorative and Architectural Stainless

Railings, paneling, handrails, elevator trim, storefront work, and signage. Here the bead is the product: nobody grinds down a decorative handrail, and a 5 mm 316 tube with a splotchy, heavily oxidized bead gets rejected on sight, however sound the structure. That is why helium tri-mix is the favored choice for decorative and architectural stainless. The small CO2 gives arc stability and puddle control on vertical runs and overheads, while argon and helium keep the bead bright, smooth, and free of heavy brown oxidation. Flow consistency matters more here than anywhere else, because a decorative job is judged as a set of visible welds and shielding variation shows up as color differences between posts. On thin ornamental sheet and tube, run flow conservatively and keep a tight CTWD, since the underside gets scrutinized too. Aluminum is a different rulebook: the Aluminum MIG welding guide covers why it needs pure argon and dedicated equipment.

Automotive Exhaust: 304 and 409 Tubing

Exhaust fabrication is a short-circuit game. Automotive systems built from 304 or 409 stainless generally use short-circuit GMAW with argon/CO2 blends at moderate parameters to avoid burn-through on thin-wall tubing: 0.6 mm to 1.2 mm walls, lap and butt joints, often fitted in a driveway or a cramped bay. For this work 98/2 argon/CO2 is the sweet spot, with a stable short-circuit arc, low spatter, easy cleanup, and a bead that will not melt a thin wall. Tri-mix works too and looks cleaner, but its extra heat and higher voltage mean more burn-through risk on 0.6 mm wall. Plain C25 or 100% CO2 is a poor match: carbon oxidizes the bead, the arc is harsh and spattery, and the results get dirty fast.

• Back-purge on 409. Ferritic 409 has lower chromium and nickel content than 304, so the backside of the joint oxidizes and rusts quickly at welding temperature. Low-flow argon behind the joint, held by a dam or foil dam, keeps the root bright and slows rust bloom, which is the difference between a clean-looking weld and one weeping orange within a week.
• Full penetration. Exhaust tubing carries heat, vibration, road salt, and debris, and a partial joint on a lap or butt repair will eventually crack. V-groove the joint, run a root, and fill.

Thin-wall tube work of this kind shares the same short-circuit discipline, as covered in alloy wheel welding and similar techniques.

Heavy Plate, Tanks, and Structural Stainless

Past about 6 mm, or on large vessels, the calculus changes. You need pulsed GMAW and a gas that carries heat and penetration, which is where helium tri-mix earns its cost. Helium’s higher arc voltage delivers a deeper penetration profile, a wider bead, and better out-of-position control on a 10 mm 304 or 316L tank shell or a heavy structural bracket. Producers such as Lincoln Electric and ESAB both document pulsed GMAW with helium-bearing mixes for stainless.

On very thick structural stainless, or where volume and shop-floor robustness matter more than cosmetics, flux-cored wire is a legitimate alternative: gas-shielded stainless flux-cored wires deposit faster with good penetration and tolerate poor fit-up better than solid wire, at the cost of slag, a rougher bead, and grinding. If the welds will be hidden in a structural frame, a flux-cored process may beat chasing a perfect MIG bead.

Why Passivation and Pickling Change Your Gas Decision

Weld heat leaves the stainless surface with a chromium-depleted, oxide-covered layer that is not the same material as the surrounding sheet. Passivation, typically a nitric or citric acid treatment, strips that heat-affected oxide and restores a uniform passive film; pickling is the harsher cousin, dissolving the oxide and some base metal with it.

Here gas choice becomes a money conversation. Every layer of heat tint and embedded oxide must come off before the part can be passivated, and that removal is chemical or mechanical work you charge for. A gas that produces a clean, lightly tinted bead needs less pickling time, less grinding, and less acid; a sugary, heavily oxidized weld needs a full pickling cycle, often with a grinding pass first. Better shielding, less cleanup, and an easy argument for tri-mix on sanitary work.

Citric acid is the gentler, more common modern passivation chemistry for finished food equipment, chosen because it is less aggressive and leaves less residue than nitric. Nitric-based pickling remains standard for heavier fabrication, heat tint removal, and weld oxide, typically followed by a neutralization step. Follow the spec for the alloy and end use, not a generic rule.

A Quick Application Summary

Application	Recommended Gas	Why
Food, dairy, pharmaceutical equipment	Helium tri-mix, or 98/2 Ar/CO2 on a budget	Minimal oxide, easiest passivation, cleanest grind-back
Decorative railing, paneling, trim	Helium tri-mix	Bead appearance is the deliverable; bright, low-tint surface
Automotive exhaust, 304/409 tubing	98/2 Ar/CO2	Stable short-circuit arc, low burn-through risk; back-purge 409
Heavy plate, tanks, structural	Tri-mix with pulsed GMAW, or flux-cored	Deeper penetration, wider bead, better out-of-position control

The through-line: match the gas to the finish standard you are paid to hit. Bright, grindable, oxide-free beads for sanitary and decorative work. Stable, low-spatter arcs for thin tubing. Higher heat and deeper penetration for plate. Everything else on the machine, from wire size to voltage to stickout, should support the gas you chose rather than fight it.

Common Mistakes and Misconceptions About Stainless MIG Gas

Most stainless MIG problems get blamed on shielding gas. Sometimes gas is the culprit, but more often it is not. Here are the mistakes that keep repeating, including some with nothing to do with gas at all.

Wrong Gas and Wrong Flow

Mistake 1: reusing the C25 or 100% CO2 bottle from mild steel work. Burning a dedicated bottle on a couple of hours of stainless feels wasteful, so the 25/75 mix gets used instead. CO2 oxidizes the molten pool, and stainless has no tolerance for it. Expect a dull, grainy or “sugary” bead, brown glassy slag islands, a boiling puddle, and porosity underneath. The bead may still look fine, but the corrosion resistance you paid for is gone in the heat-affected zone.

Mistake 2: believing more flow is always better. Under-shielding is real, but the shielding envelope works only over a narrow window. Push past it and the nozzle gas turns turbulent, entraining room air instead of shielding, so 40 CFH still hides an oxidised weld. The same logic governs MIG welding aluminum, where contamination shows faster. For 0.030 and 0.035 wire, 15 to 25 CFH is the practical window for argon/CO2 blends, with tri-mix needing a bit more. TWI’s shielding gas guide explains why.

Equipment, Cylinders and Labels

Mistake 3: reusing a hose and regulator without checking them. CO2 is a more aggressive sealing agent than argon, so a regulator that spent a year on CO2-rich gas can have degraded seals and internal parts. A contaminated regulator can also trap moisture and hydrocarbons, which show up as porosity and get misdiagnosed as purity.

• Hose: cracking, chafing, or a bulge near the fitting means replace, not patch.
• Seals and diaphragms: flattening, hardening, or a whistling leak.
• Purge the line with a few minutes of shielding gas, valve open.
• CGA fitting and threads must match the regulator; argon and CO2 use different conventions.

Cylinder handling and ventilation rules sit in our AllGas welding supply guide. Review OSHA hot work requirements first.

Mistake 4: assuming tri-mix beats 98/2. Tri-mix costs more, empties faster, and on very thin sheet the extra helium heat makes it harder to control. 98/2 argon/CO2 handles short-circuit work on thin to medium material and is cheaper and better stocked. Tri-mix earns its place out of position, on heavier sections, and where bead appearance justifies it.

Consumption scales with flow rate and arc-on time. A 330 cubic foot tri-mix cylinder may last roughly half to two-thirds as long as a 98/2 blend of the same size, so shops should cost it in.

Mistake 5: ignoring cylinder labels and mix drift. Blends drift, so a cylinder that has sat through heat or cold for months can deliver a composition that no longer matches its label. Nitrogen can creep in. Read the label every time: percentage breakdown and fill date. If a cylinder is unlabelled or temperature-cycled, keep it off stainless. Our Air Gas welding supply guide covers cylinder types.

Store cylinders upright and capped; one on its side carries liquid into the regulator, showing erratic flow and a sputtering arc.

Trailing Shielding on Thin Stainless

Mistake 6: skipping trailing shielding. Trailing shielding keeps a small gas flow running after the arc stops so the hot metal is covered as it cools to a golden or straw tint. Skip it on thin gauge and the last inch of bead oxidises, leaving brown or blue heat tint that becomes a rust stain. Set a post-flow timer of a few seconds, or hold the trigger to the end of the pass. If the machine has no post-flow setting, add one; it is the cheapest fix for clean stainless.

Gas Is Only One of Several Variables

Mistake 8: believing gas is the only factor. Gas is one variable in a system with at least seven others. When a weld fails, check the rest in order:

• Polarity: DCEP, electrode negative. DCEN ruins the result; Lincoln Electric and ESAB both specify reverse polarity, and torch lead and clamp positions differ by polarity on many MIG welders.
• Filler metal: AWS ER308L for 304 and 308, ER316L for 316 and 316L. ER308L on a 316L base works for fill passes, not as a universal answer.
• Wire size, voltage and feed: 0.030 and 0.035 are the workhorses, 0.045 for heavier plate. Voltage too high for the wire speed burns the tip.
• Joint prep: a bevel, proper root gap, and a clean land matter more than most expect.
• Cleanliness: degrease with a dedicated stainless cleaner, then brush to bright metal. Contamination is the top cause of porosity and no gas blend fixes it.

What Each Mistake Looks Like

What you see	Likely mistake	First change
Sugary bead, brown slag	CO2 or C25	98/2 or tri-mix
Oxidised bead, flow wide open	Air entrainment	Cut flow to 15-25 CFH
Porosity, spitting at start	Moisture in hose or wire	Purge, replace hose
Bead ends brown, not straw	No trailing shielding	Set post-flow timer

The cylinder gets blamed because it is the easiest thing to change. But no shielding gas rescues a dirty joint, wrong polarity, too high a voltage, or a cylinder that drifted out of spec. Get the fundamentals right, then pick the gas for the job.

For process background, see our guide to air arc welding and carbon arc gouging. The Airgas welding supplies buyer guide helps when building a gas setup.

How to Choose: A Decision Framework, Setup Checklist, and Next Steps

You know what each blend does and where the failures live. What remains decides whether your next weld looks like stainless or looks like cold roll.

Step 1: Identify the base metal and the thickness

Grade matters because of chromium and nickel content. 304 and 316L are the workhorses and weld with the same gases. 409 (exhaust) is lower nickel, rusts readily, and needs cleaner shielding plus a backing purge. 430 and other ferritic grades are nickel-free and unforgiving. Check the mill markings, and pull a harder alloy’s procedure reference the way you would for alloy 20 work, since the same gas logic holds.

Thickness drives everything: under about 1/8 inch (3 mm) you short-circuit, 1/8 to 3/16 inch mixes short-circuit and pulsed transitions, and above 3/16 inch spray transfer becomes realistic, pushing thicker plate toward pulsing or flux-cored.

Step 2: Pick the transfer mode

Gas and transfer mode are linked: pick the mode first and let it push you toward a gas.

• Short-circuit (thin sheet, most shops): 98/2 argon/CO2 or a helium tri-mix, where most fabricators live.
• Pulsed GMAW (stainless, 1/8 inch and up): 75 to 95 percent helium blends or argon/oxygen mixes. Pulsing needs a stable, hot arc that pure argon and CO2-heavy mixes cannot deliver.
• Spray (3/16 inch and up): argon with 1 to 2 percent CO2, or a small oxygen addition, at higher voltage and current.

Step 3: Match gas to mode, appearance, and position

Rule of thumb worth memorizing: choose 98 percent argon / 2 percent CO2 for thin-sheet short-circuit work on a budget, and 90/7.5/2.5 helium tri-mix when appearance, penetration, or out-of-position welding matters. A 16-gauge 304 backsplash? 98/2 is fine and cheap. A sanitary tank, a visible railing, or a tube held in a fixture all day? Tri-mix earns its extra cost.

Step 4: Check budget and availability

Tri-mix costs more per cubic foot than 98/2 and burns faster at the nozzle, so a cylinder that lasts a month on C25 might last three weeks on tri-mix. Ask about cylinder size, price per fill, and swap programs. Local gas suppliers and welding supply houses often bundle mixes with cylinder rental, and tri-mix is nearly universal in stainless shops. Buying your own supply? Read a buyer’s guide to welding gases and cylinders first.

Step 5: Set the parameters

Parameter	Stainless starting point
Wire	.030 for 16 to 20 gauge, .035 for 1/8 in and up
Polarity	DCEP (reverse polarity), always
Voltage	18 to 22 V on .030 short-circuit; tune by ear
Wire speed	Set to feed, adjust voltage for a stable arc
Stickout (CTWD)	3/8 to 1/2 in, shorter than mild steel
Flow rate	15 to 20 CFH on 98/2; 25 to 35 CFH on tri-mix
Pre-flow / post-flow	2 to 5 s pre-flow, 8 to 15 s post-flow on thin material

Tune voltage by watching the bead, not by chasing a chart. A quiet crackle with a slightly convex bead is the target; add voltage in small steps, and back it down if the arc sounds harsh. Lincoln Electric, Miller, ESAB, and AWS publish recommended settings for their own products, and those tables belong on the bench.

At-the-machine setup checklist

1. Cylinder upright, chained, away from heat and traffic
2. Regulator and hose rated for the gas service, contamination-free
3. Nozzle and gas lens matched to the mix, clean, not spattered
4. Flow meter verified with an unlit torch, full flow, no turbulence
5. Draft shield in place, or the job moved out of the wind
6. Grinder wheel dedicated to stainless, ideally labeled
7. Dedicated stainless liner and contact tip, wiped and stored dry
8. Base metal cleaned and dry, joint prep matched to thickness

Item three catches a surprising number of failures: a dirty gas lens lets atmosphere into the shielding cone, and you get oxides no gas blend can fix.

Quick FAQ

• Is 98/2 safe for all stainless? For most thin-gauge and general fabrication work, yes. It struggles out of position and lacks tri-mix penetration.
• Can I use tri-mix on mild steel? You can, but it is wasteful on thin mild steel and gives a wider, less penetrating arc.
• Do I need a shielding blanket or backing gas? On 304/316L, usually not, if post-flow is long enough. On 409 exhaust tubing, back-purging is strongly recommended.
• Why is my bead gray and dull? Almost always contamination: oil on the wire, dirty nozzle, or dirty liner. Change the liner and tip before blaming the bottle.
• Is argon flammable? No, but it is an asphyxiant. See the safety notes below.

Safety notes

Argon, helium, and CO2 are colorless, odorless, and all displace oxygen. Confined-space welding can quietly knock you down with no alarm. OSHA’s ventilation standard for welding and the agency’s hot work guidance call for adequate ventilation and, where oxygen levels can fall, monitoring. Never weld stainless in a closed tank, hopper, or duct without forced exhaust and a spotter outside. Keep a NIOSH-rated respirator and fume extraction close, since stainless fume contains hexavalent chromium.

Secure cylinders upright with a chain or strap, cap the valve when the regulator is off, and never move a cylinder with the regulator attached and the valve open. Confirm the regulator matches the gas service, and check mixed-gas cylinders for certification date and mix accuracy, since blends drift in storage.

Where to go deeper

For wire handling, feed speed, and machine setup, see practical MIG welding guidance and process fundamentals. Many “gas problems” are actually cleaning problems, and a dedicated surface prep step explains more failures than any blend change. For stainless that must look beautiful, TIG remains the benchmark, and TIG procedure fundamentals are worth reading even if you never lift a TIG torch.

The takeaway

Most people buy a bigger machine, then wonder why the stainless still looks like gray sugar. A capable inverter will not fix a bad bottle. Shielding gas is the highest-return modification you can make on an otherwise fine welder, and a modest gas upgrade often beats a large machine upgrade on finished quality alone.

And if you remember one habit from this guide, make it this: keep a dedicated stainless setup. Separate gun, liner, contact tip, regulator, and cylinder from the mild steel side of the shop. Half of all mysterious stainless failures trace back to a crossover that never should have happened.

Frequently Asked Questions

What is the best gas for MIG welding stainless steel?

For most short-circuit stainless MIG work, 90% helium / 7.5% argon / 2.5% CO2 tri-mix gives the best combination of arc stability, penetration, and bead appearance. A cheaper and still very capable alternative is 98% argon / 2% CO2 for thin sheet and general fabrication.

Can I use regular C25 (75/25 argon/CO2) gas for stainless MIG?

You can run a weld with it, but it is not recommended. The higher CO2 content causes more spatter, heavier heat tint, and greater oxidation of the chromium in stainless, which can compromise corrosion resistance and require more cleanup. Reserve C25 for mild steel.

Why does my stainless MIG weld look sugary or gray?

That appearance is heavy oxide formation, usually caused by oxygen contamination — from the wrong shielding gas, too long a contact-tip-to-work distance, drafts in the work area, or contaminated filler wire. Switching to a proper stainless gas and shortening stickout typically resolves it.

What flow rate should I use for MIG welding stainless steel?

For argon/CO2 blends on stainless, start around 20–25 CFH; for helium tri-mix, start around 35–45 CFH and verify with a flow meter. Higher is not always better — excessive flow creates turbulence that pulls air into the weld zone.

Is tri-mix gas worth the extra cost over 98/2 argon/CO2?

For decorative, sanitary, or critical work where bead appearance and fusion matter, yes — tri-mix noticeably reduces spatter and produces a smoother, better-wetting bead. For everyday thin-sheet repairs on a budget, 98/2 is usually adequate and much cheaper to run.

Can I use 100% CO2 for MIG welding stainless steel?

It will technically arc and deposit metal, but 100% CO2 produces heavy spatter and aggressive oxidation on stainless, and it is generally unsuitable for anything beyond crude, non-critical work. It is not used in professional stainless fabrication.

Do I need to purge the back side of a stainless MIG weld?

For full-penetration root passes on sanitary, food-grade, or corrosive-service work, back purging with argon is standard practice to prevent sugary oxidation on the root side. For partial-penetration cosmetic or non-critical welds, purging is usually unnecessary.

What filler wire pairs with these shielding gases on stainless?

Match filler to base grade: ER308L for 304 stainless, ER316L for 316, and so on. All common stainless MIG gases work with these wires on DCEP polarity; the wire and gas should be selected together as part of a coherent procedure.

How do I know if my gas mix is the problem and not my technique?

A quick test is to run a bead on scrap of the same thickness with a shorter stickout, a draft shield, and a known-correct gas. If the weld quality improves dramatically and immediately, the gas or coverage was the culprit. Persistent porosity or spatter under good conditions points to wire cleanliness or parameter setup instead.