Dissimilar Metal Welding: Guide, Challenges & Techniques

A refinery crew once spent three days trying to weld a stainless spool to a carbon steel header, only to watch hairline cracks creep along the fusion line hours after the weld cooled. The welder’s technique was fine — the problem was that two metals that expanded, contracted and reacted to heat in completely different ways were being forced into a single joint. That is dissimilar metal welding in a nutshell: the most unforgiving, most fascinating, and most failure-prone corner of the trade.

What Is Dissimilar Metal Welding? Definition and Why It Matters

Dissimilar metal welding, often shortened to DMW, is any weld in which the metal on one side of the joint is not the same as the metal on the other side. That covers two different base metals joined together, a filler metal deliberately chosen to differ from the base metal it melts into, and cladding or overlay work where a layer of one alloy is deposited on another. If the chemistry changes across the fusion line, you are doing dissimilar metal welding, whether you meant to or not.

The definition is wider than most welders first assume. Work that Idaho National Laboratory prepared for ASME defines a dissimilar metal weld as a joint made with a filler metal different from the base metal (a heterogeneous weld), a cladding used to alter surface properties, or a weld between two different base metals (INL dissimilar metal weld report). The same report points out that even a joint between two austenitic stainless steels, such as 304L welded to 304H or 304H to 316H, technically qualifies, because the two base metals are not metallurgically identical.

Three ways a weld becomes dissimilar

• Two different base metals. Carbon steel to stainless, low-alloy chrome steel to a nickel alloy, duplex to 316L, cast iron to mild steel. This is the classic case, and the one that causes the most trouble.
• A filler that differs from the base metal. A heterogeneous weld. Running ER309L or a nickel filler into plain carbon steel pipe is dissimilar metal welding even though only one base metal is present.
• Cladding and overlay. A corrosion-resistant alloy welded onto a cheaper structural substrate so the part gets a strong body and a resistant surface at the same time.

Why you are probably reading this

Most people land on this topic with a job in front of them, not curiosity. Common triggers include:

• Tie-ins and retrofits where an existing run of pipe or plate in one alloy must be joined to new material in another.
• Cost control, by building the structure from inexpensive carbon steel and protecting only the wetted surface with an overlay.
• Corrosion resistance, where only the alloy that touches the product can survive the service environment.
• Code work, where a written procedure has to be qualified and traceable before an inspector will accept the weld.

Underneath all four sits the same logic: each side of the joint is optimized for its own job. One side carries the load, the other resists the corrosion, and the weld has to reconcile two materials that expand, contract and react to heat differently.

Which processes can be used

Fusion processes dominate DMW work: GTAW, GMAW, FCAW, SMAW, SAW and PAW. Only GTAW and PAW can be run autogenously, without filler metal, and autogenous welding is rarely appropriate for dissimilar joints because you cannot correct a composition mismatch with filler. GMAW, FCAW, SAW and SMAW are consumable-electrode processes, and all four are accepted for dissimilar metal welds in heavy fabrication practice and in ASME BPVC Section III, Division 5. Electrode choice matters more than usual here, which is why guides such as our alloy steel welding electrodes reference are worth reading before you order consumables.

When the mismatch becomes extreme, fusion welding stops being the right answer. Steel joined to aluminum, or copper to steel, tends to form brittle intermetallic compounds in the melt pool. For those cases, solid-state processes such as friction stir welding and explosion welding join the parts without ever melting them, and aluminum work generally follows its own dedicated practices (see our aluminum MIG welding guide for the fusion-side equivalent).

How this article is built

We start with where these joints are used, then move into the metallurgy of dilution and transition zones, filler metal selection combination by combination, buttering and joint design, the codes that govern qualification, the failure modes that show up in service, inspection methods, and fume safety. By the end you should be able to look at any two metals, name the risks they bring, and plan a joint that survives.

Where Dissimilar Metal Welds Are Used: Applications and Real Benefits

Not every weld is made between identical metals. In fact, some of the most critical joints in industry are between two materials that were never meant to be together. The reason they’re joined anyway comes down to one simple principle: use each metal exactly where it earns its keep. This is the logic behind dissimilar metal welding (DMW), and it’s what allows fabricators to optimize for strength, corrosion resistance, and cost all at the same time.

Power Generation and Nuclear Facilities

Nowhere is this balance more strict than the power industry. In light water reactors, INL reports that dissimilar metal welds have long been used to connect piping systems, with one of their primary motivations being to clad corrosion-resistant alloys onto higher-strength or lower-cost materials. The most well known of these are nuclear safe-end welds, which join the austenitic stainless steel nozzles of a reactor pressure vessel to the ferritic low-alloy steel piping, as well as buttered closure welds used during both fabrication and field repair.

Outside of the reactor core, power plants rely heavily on DMWs for transitions from ferritic to austenitic tubing within their boilers and steam lines. These transition joints allow for high-pressure containment on the hot, thick-walled ferritic side, whilst providing oxidation and creep resistance on the austenitic side that is exposed to superheated steam.

Oil, Gas, Chemical, and Subsea

Corrosion is the enemy in the oil, gas, and chemical processing industries, and it’s an enemy that is far too expensive to fight with solid, exotic alloys. Dissimilar metal welds allow for overlay and cladding of corrosion-resistant alloys, such as stainless or nickel-base metals, over cheaper structural or carbon steels for heat exchangers, pressure vessels, and piping.

During tie-ins and retrofits, it’s also incredibly common to join pipe runs of different schedules, grades, or materials to an existing line. For subsea applications, this extends to flange-to-pipe connections, where the flange may be manufactured from a high-strength low-alloy steel for structural integrity, whilst the wetted pipe material is selected to resist the harsh, chloride-rich environment.

In these environments, the selection of compatible filler metals is just as important as the base metals themselves. As noted within INL’s ASME BPVC review, common industrial DMW combinations include joining low-alloy steels to austenitic stainless steels or nickel-base alloys, all of which require careful procedure control to prevent premature failure. For shops working with tougher steel grades, utilizing the correct alloy steel welding electrodes can be an integral part of that compatibility.

Turbines, Marine, and the Logic of a Graded Joint

DMW isn’t only reserved for massive pressure boundaries. Turbine components, particularly in combined-cycle plants, will often utilize transition joints to connect dissimilar materials that are exposed to drastic temperature swings. Similarly, within the marine industry, hulls, piping systems, and fittings will see DMWs implemented where the corrosive splash zone calls for vastly different material properties than the load-bearing structure. In fact, joints found within projects such as aluminum boat welding showcase this exact ideology on a much smaller scale, trading a solid, expensive hull alloy for a smarter, joint-based approach.

The Real Benefits: Strength Where Needed, Savings Where Possible

The biggest incentive for DMW will always be practicality. Building an entire chemical reactor out of solid Alloy 625 would be astronomically expensive, however, overlaying it onto a SA516 carbon steel shell grants near-identical corrosion protection at a fraction of the cost. This cladding method not only saves on material cost, but also reduces weight, lead times, and machining costs compared to its solid-alloy counterpart.

These economic gains are only made possible by the functionally-graded nature of the joint. Each side is tasked with its job: the base metal provides the tensile strength, fracture toughness, or rigidity required to contain pressure, whilst the weld metal or cladded layer provides the chemical resistance needed to withstand the process media. When executed correctly, the result isn’t two metals being forced together, but a single, purpose-built assembly that is stronger, more resilient, and far more cost-effective than either option could be alone.

The Science Behind the Joint: Dilution, Transition Zones and Microstructure

Here is the uncomfortable truth about dissimilar metal welding: you are not welding two known metals together. You are manufacturing a brand-new alloy in real time, in front of an arc, with no mill certificate and no property data sheet. The puddle pulls elements from both sides, mixes them with filler, and freezes into a gradient of compositions that changes character every fraction of a millimeter across the joint. Understanding what happens in that gradient is the difference between a joint that lasts thirty years and one that cracks on the night shift.

What Dilution Actually Means

The single most important number in any dissimilar weld is dilution. The Idaho National Laboratory defines it simply as the amount by which the filler metal is diluted by the base metal, and it can be measured directly from a macro-etched cross-section by comparing the cross-sectional area of the fusion zone against the areas of base plate consumed by the arc (INL, dissimilar metal weld guidance). Expressed as a percentage:

Dilution (%) = (area of base metal melted ÷ total fusion zone area) × 100

If a weld is 70% diluted, the final deposit is mostly base metal chemistry, not the filler you carefully selected. That is why process choice matters so much: high-penetration processes like submerged arc welding drive dilution up sharply, while a carefully controlled GTAW root pass keeps it low. Push the dilution too far and the weld metal lands in a composition range you never intended, with whatever cracking susceptibility comes with it.

Research published in the Journal of Materials Processing Technology on steel/nickel-alloy nozzle buttering welds showed just how dramatic this can get: dilution ratios of 40–60% produced distinct beach, peninsula and island structures in the fusion zone, and above 94.5% dilution the initial solidification phase switched entirely, from FCC austenite to BCC ferrite. Same two metals, same filler, completely different solidification behavior, driven only by how much base metal melted in.

The Five Regions of a Dissimilar Weld

Cut, polish and etch a dissimilar joint and you will see far more than “weld” and “base metal.” Each region behaves differently and fails differently:

Region	What it is
Fusion zone	Fully melted and mixed weld metal, including swirls of unrefined base-metal dilution carried by puddle convection.
Unmixed zone	Base metal melted at the fusion boundary but never swept into the mixed pool; composition close to the original base metal.
Transition zone	The steep chemical gradient where compositions shift from base metal to fully mixed weld metal.
Partially melted zone (PMZ)	Base metal heated above its solidus but not fully melted; grain boundaries liquate, making it prime territory for liquation cracking.
Heat-affected zone (HAZ)	Unmelted base metal whose microstructure was rewritten by heat, including the hard martensitic bands that invite hydrogen cracking.

The INL flags the weld metal swirls at the transition zone as areas of significant interest because of their high hardness and a documented history of hydrogen-induced cracking in subsea oil and gas service (INL). For a practical look at how these regions are characterized in real joints, the Ohio State metallurgical characterization work on F65-to-F22 dissimilar welds is a useful reference (MA2JIC/OSU).

A Functionally Graded Material You Built Yourself

Engineers use the term functionally graded material for components whose composition changes gradually so each side can do a different job: strength here, corrosion resistance there, cost control underneath. A dissimilar weld is exactly that, except nobody graded it on purpose. Because the composition varies continuously across the fusion boundary, there is no historical property data for the resulting alloy. You cannot look up its tensile strength, its elongation, or its corrosion rate. You can only measure it, which is why macro-sections and hardness traverses across the fusion boundary are treated as the honest answers, and why Alloy 20 welding procedures and other corrosion-alloy work lean so heavily on procedure qualification rather than assumption.

Predicting the Outcome Before You Strike an Arc

You do not have to guess. Constitutional diagrams give you a first-line prediction of what the puddle will become. The Schaeffler diagram maps chromium and nickel equivalents to a microstructure of austenite, ferrite and martensite; the DeLong diagram refined it by accounting for nitrogen and reporting ferrite number directly; and WRC-1992 is the modern standard for austenitic and austenitic-ferritic weld metals, giving a more accurate ferrite number in the range where cracking actually occurs. ASME BPVC Section III, Division 1, Subsection NB adopts the WRC-1992 diagram to predict primary ferrite, and delta ferrite in those specific A-Number 8 weld metals is limited to a minimum ferrite number of 5, because below 5 FN the risk of solidification cracking climbs (INL). That 5 FN floor is the practical takeaway: a little ferrite is insurance, too much invites embrittlement, and the diagram tells you where you will land before you commit to a WPS.

Choosing the Right Filler Metal for Dissimilar Metal Joints

On a dissimilar metal job, the base metals are usually set by the drawing, the service or the budget. The filler is the variable you get to design. Get it right and the joint outperforms both sides of it. Get it wrong and no technique can save you, because you are not repairing a weld problem, you are creating a new alloy with a chemistry nobody has tested.

The Cardinal Rule: Never Use 308 to Join Stainless to Carbon Steel

Start with the mistake that cracks the most welds in this trade. Type 308 filler (ER308, E308) is balanced for joining one austenitic stainless to another, at around 19% chromium and 9% nickel. When carbon or low-alloy steel melts into that pool, both elements dilute downward. Hobart Brothers’ technical guide on dissimilar metal welding states the rule plainly: never use 308 to join austenitic stainless to plain carbon or low alloy steel, because dilution can push the deposit below roughly 17Cr/7Ni. The weld then solidifies fully austenitic, with no ferrite to absorb shrinkage strain, and cracks along the fusion line while it is still cooling.

309L and 312: The Two Austenitic Workhorses

For stainless-to-carbon or stainless-to-low-alloy joints, the same Hobart guide identifies Type 309(L) and Type 312 as the two leading austenitic fillers. They work for opposite reasons:

• Type 312 carries a large alloy reserve near 29Cr/9Ni and deposits with ferrite well above 25 FN. That ferrite is the crack-resistance insurance, absorbing solidification shrinkage strain, which makes 312 the forgiving choice for unpredictable dilution, field fit-up and high-penetration processes.
• Type 309(L) deposits with roughly 7 to 12 FN: enough ferrite for crack resistance without the heavy alloy loading of 312, and the preferred choice for stainless-to-carbon joints operating above about 800°F (425°C).

That temperature split matters for two long-term reasons. Above 800°F the ferrite in 312 welds can transform into sigma phase, a hard, brittle intermetallic that embrittles the joint over months or years. And 310 filler, reflexively grabbed for stainless-to-stainless work, has cracked over phosphorus-bearing weathering steels, where phosphorus segregates into low-melting films along the fusion boundary. Check the base metal certificate before defaulting to 310.

If you need a ready-made qualification route, AWS publishes Standard Welding Procedure Specification B2.1-1/8-230, a GTAW procedure with a consumable insert root for joining carbon steel (M-1/P-1) to austenitic stainless (M-8/P-8) using IN309 and ER309(L). It is a useful reference even when you write your own WPS.

Match Thermal Expansion, Not Just Chemistry

Above about 800°F, or once you move from stainless into nickel alloys, austenitic fillers stop being the first choice, because the driver becomes coefficient of thermal expansion rather than dilution alone. A stainless filler on ferritic low-alloy steel leaves the two sides expanding and contracting at different rates through every heat cycle, concentrating strain at the fusion boundary.

A nickel-base filler sits between ferritic steel and austenitic stainless in expansion behaviour, so Hobart recommends nickel-base ERNiCr-3 (Alloy 82) for 2.25Cr-1Mo ferritic steel joined to austenitic stainless in high-temperature service. For low-alloy steel to nickel alloy, the usual answers are ERNiCrMo-3 (Alloy 625) for corrosion and strength-critical service, or ERNiCr-3 for simpler high-temperature transitions. Nickel fillers also tolerate dilution far better than stainless types and avoid forming hard martensite when a little carbon steel melts in. Our alloy steel welding electrodes guide covers the low-alloy side of this pairing, and the Alloy 20 welding procedures guide shows how nickel-alloy fillers are handled in aggressive chemical service.

Quick Reference: Combinations and Recommended Fillers

Base Metal Combination	Recommended Fillers	Key Reason
Austenitic stainless to carbon or low-alloy steel	E/ER309L, E/ER312	Ferrite resists dilution cracking
Same joint, service above 800°F	E/ER309L (7–12 FN), or ERNiCr-3	Avoids sigma-phase embrittlement
2.25Cr-1Mo ferritic to austenitic, above 800°F	ERNiCr-3 (Alloy 82)	Expansion sits between the two
Low-alloy steel to nickel alloy	ERNiCrMo-3 (625), ERNiCr-3	Dilution tolerance, corrosion, strength
Duplex stainless to carbon or low-alloy steel	E309L, E309LMo, E2209	Controlled ferrite, matched corrosion
Cast iron to steel	ENiFe-CL nickel-iron electrode	Absorbs carbon, stays machinable
Precipitation-hardening stainless	Matching PH filler or 309L/312 buffer	Preserves ageing response
Titanium to steel or copper	V or Nb buffer layers, or solid-state joining	Avoids brittle intermetallics

Duplex, Cast Iron and the Exotic End of the Scale

Duplex-to-carbon joints have their own consumable family. IMOA’s shop sheet on dissimilar welds in duplex lists E309L, E309LMo and E2209 for joining duplex to carbon and low alloy steels, with E2209 preferred where corrosion resistance must be retained on the duplex side. IMOA adds an easy-to-miss caution: nickel-base fillers used on duplex must be free of niobium, since niobium combined with duplex chemistry promotes harmful intermetallic precipitation.

Cast iron to steel puts a brittle, high-carbon member on one side. Nickel-iron electrodes such as ENiFe-CL give a deposit that absorbs the carbon, tolerates the mismatch in ductility and can still be machined afterwards. For precipitation-hardening grades, the filler depends on your post-weld heat treatment plan: matching PH filler if you will solution treat and age, an austenitic buffer such as 309L where you cannot.

Titanium to steel or copper is the case where fusion filler selection runs out of answers. Direct fusion welding produces brittle iron-titanium and copper-titanium intermetallics that crack on cooling, so the practical routes are buffer layers of vanadium or niobium built up gradually, or explosion bonding and friction stir welding. Anyone doing exotic-metal TIG work should be comfortable with aerospace TIG welding standards and techniques before attempting these joints.

Whichever combination you face, verify the filler against a constitution diagram such as Schaeffler, DeLong or WRC-1992 before ordering. Filler classification is only half the answer; dilution, heat input and pass sequence are the other half, and that is where joint design and buttering take over.

Buttering, Joint Design and Welding Procedure Techniques

Choosing the right filler metal is only half the battle. Two welders can run the same ERNiCrMo-3 bead on the same joint and get different results, because joint shape, dilution, temperature and pass sequencing control the weld composition. Technique is how you design the new alloy you create.

What Buttering Actually Does

Buttering means depositing one or more beads onto one member of the joint before it is assembled and closed. Those beads become a transition layer, so the final weld never bridges two incompatible metals in one pass. Butter a carbon steel flange face with nickel filler, machine it smooth, then weld the stainless spool with a stainless-compatible consumable: the carbon steel never dilutes the stainless weld, and the composition gradient spreads across controlled layers instead of one fusion line.

The payoffs: fewer hard martensitic zones, lower cracking risk, and a buffer for thermal expansion mismatch. It also gives you a second chance. If buttering dilution runs high, grind it back and add another layer before anything structural depends on it.

Qualifying Buttering Under ASME Section IX QW-283

Buttering is a code-defined operation, not a shop improvisation. ASME BPVC Section IX covers it at QW-283: the qualification coupon must replicate the production sequence — buttering first, heat treatment second, the joining weld last. A coupon that tests only the closing weld proves nothing about the buttered interface in service. This procedure qualification guidance for buttering shows how those coupons are built and tested.

Section IX also draws a thickness line that catches people out:

• If buttering thickness on the test coupon is less than 3/16 in (5 mm), heat input becomes an essential variable — deviate from the qualified range and the WPS no longer covers you.
• If it is 3/16 in (5 mm) or more, buttering thickness becomes a non-essential variable, but the WPS must still state it as the minimum thickness remaining after machining.

Joint Design: Keep Undiluted Base Metal Out

Good dissimilar joint geometry guarantees full penetration and keeps unmolten, undiluted base metal out of the solidifying weld. The IMOA shop sheet on dissimilar joints says designs must facilitate full penetration and avoid undiluted base metal in the weld, and bevels should be machined rather than ground so land thickness stays uniform around the full circumference.

Symmetrical (double-V) prep halves pass depth, so each bead dilutes less and shrinkage stays balanced. One-sided prep is faster but puts root passes where dilution is highest and mixing is poorest. Also watch the unmixed zone along the fusion boundary: bands of base metal that melted but never mixed with the filler. Sharp included angles, excessive root gaps and wandering arcs all make it worse.

Heat Input, Pass Sequencing and Preheat

Heat input sets dilution, and dilution sets composition. Recent work buttering AISI 4130 steel with ERNiCrMo-3 (Alloy 625) and ERNiCrMo-15 fillers found that pulsed GMAW produced noticeably lower dilution than short-circuit GMAW — same filler, same steel, different waveform, different weld chemistry. Pulsed modes, smaller wire and stringer beads with modest weave keep base-metal pickup down.

• Temper-bead / cold-pass control: overlap each bead so it tempers the one beneath, avoiding PWHT on repair work.
• Preheat and interpass limits: keep them low on the ferritic side, because excess preheat grows coarse grain and widens the hard transition zone. Low-alloy consumables and their preheat needs are covered in our alloy steel welding electrodes guide.
• Pass sequence: put the most dilution-tolerant passes first against the carbon or low-alloy side, finishing with passes the corrosion-resistant side will accept.

The PWHT Dilemma

Postweld heat treatment relieves stress and removes hydrogen, but on buttered dissimilar joints it can backfire. The same 4130/Alloy 625 research found PWHT made the buttered interface more brittle regardless of the welding procedure used, as diffusion at temperature forms brittle intermetallics and carbon-rich carbide bands. So decide PWHT during planning and check hardness traverses on the qualified coupon rather than assuming. If PWHT is mandatory, qualify the actual cycle; if it is discretionary, weigh hydrogen benefit against embrittlement risk. The same planning discipline applies to other corrosion-resistant alloys — see our Alloy 20 welding procedures piece for how service environment drives the heat-treat decision.

Field Buttering: A Practical Sequence

1. Identify both base metals by certificate or positive material identification.
2. Machine bevels and lands to drawing; clean at least 1 in (25 mm) from the edge.
3. Preheat only as the WPS requires, checked with a calibrated thermometer, not color.
4. Deposit the buttering layer with the qualified process; keep heat input and interpass in range.
5. Cool as specified, then PT for cracks before machining.
6. Face back to the minimum remaining thickness stated in the WPS.
7. Join with a filler compatible with the buttered layer, not the base metal.
8. Record it on the traveler; QW-283 expects production to follow qualification, bead for bead.

Get this sequence right and the joint stops being a gamble. Every variable that matters — dilution, hardness, stress, composition — is decided deliberately, before the arc is struck.

Codes, Standards and Qualification Requirements for DMW

Here is the awkward truth about dissimilar metal welding: the codes were written mostly for joining like to like. A welding procedure qualification record assumes you know what the base metal is, what the filler is, and roughly what the finished weld metal will look like. In a dissimilar joint, the finished weld metal is a brand new alloy that nobody has ever spec’d, tested or published data for. Qualifying one therefore takes more thought than a routine carbon steel hookup, and it is where a lot of otherwise good weld procedures fall apart in review.

ASME BPVC Section IX: The Qualification Backbone

For pressure-retaining work in the United States, ASME Boiler and Pressure Vessel Code Section IX is the governing rule for procedure qualification (PQR/WPS) and performance qualification of the welder. Section IX classifies base metals by P-numbers and weld metals by A-numbers, and consumables by F-numbers, so a reviewer can quickly see whether a proposed combination has been qualified before. A change in filler classification, heat input, preheat/interpass temperature, or material thickness can move you outside your qualified range and force a new test.

For dissimilar work, the qualification coupon must be built from the actual base metal combination, not a convenient stand-in, because dilution and the resulting transition zone are the whole point. The standard test set is a transverse tension test plus guided bend tests. Side bends are the default choice for dissimilar joints, and deliberately so: they load the fusion boundary across the face of the specimen, which is exactly where a dissimilar weld wants to fail.

Section III Division 5, Section VIII and High-Temperature Service

High-temperature nuclear components fall under ASME Section III, Division 5, which Idaho National Laboratory’s research on dissimilar metal welds notes permits fusion welding by GTAW, PAW, GMAW, FCAW, SAW and SMAW for service above the creep range. The same body of INL work makes the central qualification problem explicit: DMWs effectively create alloys that lack the historical property data used to qualify materials in the first place. That gap is why code bodies keep expanding their allowances for dissimilar joints rather than closing them off.

The practical evidence of how demanding this is comes from INL’s own experimental qualification of Grade 91 ferritic steel to Alloy 800H using FM82 nickel filler: two of four transverse side bends came back with rejectable defects running along the fusion boundary. Two passing specimens out of four would not buy you a WPS. Section VIII, Division 1 pressure vessels draw on Section IX for the same qualification mechanics, with additional examination requirements where service conditions are severe.

AWS Routes: B2.1, D1.1, D1.6 and D10.4

AWS offers parallel paths. AWS B2.1 is the standard welding procedure specification series, and AWS B2.1-1/8-230:2023 is an approved ANSI/AWS SWPS covering carbon steel to austenitic stainless steel from 1/16 in (1.5 mm) up to 1-1/2 in (38 mm) in the as-welded condition. Using an approved SWPS can let you skip writing your own PQR, provided your joint, thickness and process fit its limits exactly. Structural contexts bring in AWS D1.1 for carbon and low-alloy steels (see Alloy Steel Welding Electrodes: Selection, Use & Safety) and AWS D1.6 for stainless structural work.

For piping, AWS D10.4M/D10.4:2023 is the guide to welding austenitic stainless steel piping and tubing. Originally issued in 1955, the current edition has been expanded with a delta ferrite discussion and annexes covering high-carbon castings, both of which matter directly when you are judging cracking resistance in a dissimilar root pass.

What a DMW Qualification Record Must Contain

• A PQR citing both base metals by specification and grade, the filler classification, F-number and A-number, and the process used.
• Transverse tensile results with the required minimum tensile strength for the weaker side of the joint.
• Guided bend specimens, with side bends called out where the fusion boundary is the concern.
• Toughness or impact testing where service temperature demands it.
• Hardness traverses or micro-Vickers maps across the fusion boundary when required by the service code.
• Essential variables recorded as actually run: heat input, preheat, interpass temperature, thickness and welding position.

When a dissimilar procedure is genuinely new, ask for a macro-etched cross-section in addition to the standard tests. It is the cheapest way to see unmixed base metal, excessive dilution or a buttering layer that did not bond, long before a bend specimen tells you in the hard way. Qualification for dissimilar metal welding is slower than for similar materials, but it is the only place where the new alloy you are creating gets proved.

Common Failure Modes, Cracking Mechanisms and Mistakes to Avoid

Dissimilar metal welds rarely fail in the middle of the bead. They fail at the boundary: the narrow band where two metals that were never designed to meet are suddenly forced into a single crystal structure. Knowing where that band breaks, and which human decisions cause it, is what separates a joint that lasts three decades from one that cracks on the night shift.

Hydrogen-Induced Cracking in Hard Transition Zones

When carbon steel dilutes into a stainless or nickel deposit, the dilute zone right at the fusion line can transform into hard, brittle martensite. Three things then have to coincide: a susceptible microstructure, hydrogen, and tensile stress. Hydrogen comes from damp electrodes, oily fit-ups, and ambient humidity; stress comes from joint restraint. The result is a crack that often appears hours after the weld has cooled and looks clean on the surface. Corrective habits are unglamorous: dry low-hydrogen consumables, controlled preheat and interpass temperature, and clean grinding-free surfaces. The electrode choice itself matters more than most welders expect, which is why alloy steel welding electrode selection deserves real study before anyone strikes an arc on a dissimilar joint.

Solidification Cracking From Too Little Ferrite

Austenitic weld metal that solidifies fully as austenite traps low-melting contaminants (sulfur, phosphorus, trace elements) at the final grain boundaries and tears open while still hot. The classic defense is a controlled amount of delta ferrite; ASME-referenced practice targets a minimum of 5 FN in the deposited weld metal. The trap is dilution. Hobart warns that submerged arc welding’s high penetration raises dilution and can push a 309 deposit below the required ferrite number, which makes parameter selection and joint design critical rather than cosmetic. A 2025 study on iron dilution confirmed the pattern: where base-metal dilution was not controlled, Fe content rose sharply while Ni, Cr and Mo fell, directly degrading corrosion performance of the finished weld (research summary).

Type II Boundary and Liquation Cracking

Type II grain boundaries run parallel to the fusion line inside the weld metal and become preferential crack paths under stress. Idaho National Laboratory’s review of dissimilar weld research lists in-service cracking along Type II boundaries in a 309L-on-A508 steel DMW, and identifies mismatched yield strength and coefficient of thermal expansion as the drivers of large residual stress, with thermally induced strains capable of reaching yield-strength magnitude. Nearby, in the partially melted zone, liquation cracking occurs when grain-boundary films remelt during the thermal cycle of the adjacent pass. Both mechanisms are amplified by excessive heat input and by poorly sequenced passes.

Slow Degradation: Migration, Fatigue, Galvanic and Service Cracking

Not every failure announces itself. Carbon migrates from carbon steel into stainless under sustained heat, forming a soft decarburized band on one side and a hard carbide-rich band on the other, a recipe for both cracking and crevice attack. Cyclic service adds thermal fatigue: the two sides expand and contract at different rates until the interface accumulates plastic strain. Dissimilar joints also create a galvanic cell wherever the corrosion potential differs, which is exactly why aluminum boat welding guides hammer on isolating aluminum from steel hardware. In nuclear service, primary water stress corrosion cracking has repeatedly found DMW interfaces, and long-duration operation keeps exposing new degradation modes.

Troubleshooting: Symptom to Likely Cause

Symptom	Likely cause
Crack along the fusion line after cooling	308 filler on carbon steel, or hydrogen with no preheat
Centerline tear in the final bead	Ferrite number pushed below 5 FN by high dilution
Under-bead cracking in the carbon steel HAZ	No preheat, wet consumable, or excessive restraint
Porosity and intermittent lack of fusion	Grinder-scaled, uneven fit-up instead of machined bevels
Rust staining at the seam in service	Galvanic attack plus uncontrolled iron dilution
Leak or crack after thermal cycling	CTE mismatch fatigue; PWHT never planned

The recurring theme is that each failure traces back to a decision made before welding: wrong filler, skipped preheat, uncontrolled heat input, ignoring dilution in high-penetration processes like SAW, grinding instead of machining fit-up, or never deciding whether postweld heat treatment will be applied. Plan those six, and most dissimilar weld failures never start.

Inspection and Testing: How You Prove a Dissimilar Weld Is Sound

A dissimilar metal weld (DMW) can look flawless and still be failing internally. The bead is convex, the color is right, the tie-in is wetted — and along the fusion boundary there is a hard martensitic stringer, a cluster of micro-voids, or a crack that only opens days later. That gap between appearance and reality is why DMW inspection is a discipline of its own. You are checking an interface between two materials that respond to every inspection method differently.

Visual and Liquid Penetrant: The First and Best Line

Visual inspection (VT) still starts every exam, but on a DMW it has a specific job: confirm profile, check for arc strikes and wrong filler, and spot the faint discolored line that often tracks the fusion boundary. Liquid penetrant (PT) is the workhorse for dissimilar interfaces because most cracking there is surface-breaking and the corner geometry is awkward for other methods. In Idaho National Laboratory’s dissimilar weld qualification work, PT was performed 24 hours after welding to confirm no surface-breaking cracks at the interface. That waiting period matters: hydrogen-assisted and delayed cracking may not be open yet when the welder arcs out.

Magnetic particle (MT) is the companion method wherever one side is ferritic, since it picks up slightly subsurface indications that penetrant misses. The limitation is obvious: it loses sensitivity in austenitic and nickel-base regions where the material is non-magnetic.

Where Radiography and Ultrasonics Struggle

Volumetric NDE has an uncomfortable truth on dissimilar joints. Radiography is weak at detecting tight, planar, vertically oriented fusion-line cracks, and the two materials produce very different film densities in one exposure. Ultrasonics is worse in practice: the acoustic impedance mismatch at the boundary reflects and refracts the beam, and the coarse, directional, anisotropic grain structure of the weld metal scatters it further. The result is noise and false echoes exactly where you want a clean look. Examiners compensate with side-view probes, timed gates and experienced operators — but they rarely trust UT alone on a critical DMW.

Hardness Mapping: The Most Informative Single Check

If you can run only one meaningful evaluation across a dissimilar fusion boundary, make it a micro-Vickers hardness traverse. INL’s qualification program used micro-Vickers traverses at a 10-gram load stepping across the fusion line of Grade 91 to Alloy 800H joints, and that single dataset tells you more than three radiographs: where the hard zone formed, how wide it is, whether preheat and heat input controlled it, and whether post-weld heat treatment softened it. A hardness spike above roughly 250–300 HV on a carbon- or low-alloy-steel side is a hydrogen-cracking and service-cracking red flag that no amount of penetrant will reveal.

Macro-Sections, Composition and Ferrite Number

A macro-etch cross-section remains the definitive proof that a procedure works. It exposes fusion-boundary geometry, unmixed zones, lack of fusion, dilution and penetration profiles in one image — and for a new DMW procedure it should be a required deliverable, not a nice-to-have. Paired with chemical analysis by optical emission spectroscopy or energy-dispersive X-ray spectroscopy, it confirms that dilution stayed inside the composition window the procedure assumed.

Modern characterization goes further. A Journal of Materials Processing Technology study identified beach, peninsula and island structures at steel/nickel-alloy fusion boundaries using optical microscopy, SEM, EDS and EBSD — that technique set is what serious DMW research now uses to describe the unmixed zone. On the shop floor the practical proxy is a ferrite meter: verify the weld metal landed near the target ferrite number and comfortably above the 5 FN minimum referenced in ASME practice, because insufficient ferrite is the most common root cause of solidification cracking.

Destructive Tests and Interpreting Results

Qualification ultimately rests on destruction. ASME Section IX and AWS B2.1 qualification procedures call for transverse tensile specimens and guided bend tests, and side bends are the telltale: they open the fusion boundary face-on and reveal lack of fusion, lack of sidewall penetration and incipient cracks that a face bend simply closes up. OSU’s Ma2JIC research on F65-to-F22 welds overlaid with Alloy 625 documents this full package — macrographs, hardness traverses, tensiles and metallurgical investigation of whether low-alloy-steel butter layers could replace Alloy 625 — and it is a useful template for anyone building a qualification record from scratch.

Read results as a system: PT says the surface is intact, hardness says the transition zone is safe, macro-etch says geometry and dilution match the design, and tensiles and bends say the joint meets minimum strength and ductility. When those four agree you have evidence. When they disagree, the macro-section and hardness traverse almost always explain why — and tell you whether to fix heat input, filler or joint prep before the next coupon. For corrosion-resistant alloy fabrication the same inspection mindset carries across; our Alloy 20 welding procedures guide covers the equivalent verification approach.

Health and Safety Hazards of Dissimilar Metal Welding

Every welding hazard guide warns you about arc flash, electric shock and fire. Dissimilar metal welding (DMW) adds a quieter hazard most welders underestimate: what is actually in the fume when you run a nickel-alloy or stainless pass inside a hood. DMW jobs almost never involve plain carbon steel; they pull in 309L, 312, Inconel 625, Monel, Hastelloy and duplex stainless, and each one changes the chemistry of the smoke you breathe.

Why DMW Fume Is Not Ordinary Welding Fume

Ordinary mild-steel fume is mostly iron oxide with a little manganese. DMW fume is a chemically unstable cocktail that shifts as the arc alternately melts base metal and filler. OSHA’s welding fact sheet puts it plainly: prolonged welding fume exposure may cause lung damage and cancers of the lung, larynx and urinary tract, while manganese exposure can produce Parkinson’s-like symptoms including tremor, slowness and gait problems.

The Canadian Centre for Occupational Health and Safety notes that fumes from nickel alloys such as Inconel, Monel and Hastelloy carry much more nickel and very little iron, while stainless steel fume carries more chromium, including hexavalent chromium (Cr VI). That distinction matters: a joint alternating between a carbon steel root and an Alloy 625 fill exposes the welder to two toxicant profiles in one shift.

Hexavalent Chromium: The Number That Governs the Job

Hexavalent chromium is what turns a routine stainless tie-in into a compliance event. OSHA sets the permissible exposure limit at 5 µg/m³ as an 8-hour time-weighted average under 29 CFR 1910.1026 for general industry and 1926.1126 for construction. A pinch of salt in a warehouse-sized room, across a full workday. Formed during welding, cutting and gouging of stainless and chromium-bearing alloys, Cr VI is a confirmed human carcinogen, and the standards require written exposure control plans, respiratory protection, protective clothing and hygiene areas, plus medical surveillance once triggers are met.

NIOSH’s criteria document 89-110 puts welders at roughly a 40% increase in relative risk of lung cancer attributable to their work experience, with risk among stainless steel welders tied specifically to nickel- and chromium-containing fume. On years of DMW work, that is no abstract number.

Controls That Actually Work

The control hierarchy is unchanged, only applied harder:

• Ventilation first. 29 CFR 1910.252(c)(2) requires mechanical ventilation where the shop holds less than 10,000 cubic feet per welder or overheads under 16 feet; above that, natural ventilation is allowed only in limited conditions.
• Local exhaust at the arc. A fume extractor hood a foot or two from the joint removes fume at the source and beats diluting the whole shop.
• Respiratory protection. Where engineering controls cannot hold exposure under the Cr VI PEL, an OSHA-compliant respiratory program with fit testing and medical evaluation is required.
• Confined space discipline. A DMW inside a vessel or pit multiplies exposure and adds atmospheric hazards. Test, ventilate, monitor continuously and follow a formal entry program.
• Clean the base metal. Solvents, coatings and galvanizing make their own toxic fume. Strip paint, primer and plating before the arc strikes. Good prep is a fume control as well as a weld-quality control, which is why prep matters on aluminum work covered in aluminum cleaner for welding.
• Positioning. Work upwind or crosswind from the plume, never downwind, and keep your head out of the fume column.

Electrode choice helps too. Low-hydrogen consumables with controlled alloy additions generate less fume than older high-manganese stick electrodes, a practical reason to review alloy steel welding electrode selection before committing to a procedure.

Process Hazards and the Unknown-Alloy Problem

Beyond fume, DMW carries the standard arc hazards unchanged: ultraviolet and infrared radiation burning eyes and skin, electric shock, hot work and fire risk, and inert-gas asphyxiation in poorly ventilated spaces. Nickel-alloy fillers often run with GTAW, which looks cleaner but produces no less UV.

Repair work adds a trap unique to dissimilar joints: the base metal may be unidentifiable. A decades-old fitting marked “316” might be a ferritic casting, a chrome-moly grade or an unknown proprietary alloy. OSHA’s shipbuilding and repair guidance emphasizes identifying materials before hot work, and positive material identification with a handheld XRF analyzer or OES spark testing is the only honest way to know what you are mixing into the weld pool. Welding an unknown alloy against stainless can put lead, beryllium or copper in the fume with no visible warning.

Bottom line: joint metallurgy gets the engineering attention, but fume chemistry deserves equal respect. Stainless and nickel-alloy passes in an enclosed joint mean carcinogenic hexavalent chromium, nickel and manganese well above mild-steel levels. Ventilate, extract, monitor, identify your metals, and treat every dissimilar pass as if the fume is as engineered as the weld.

Best Practices, Quick Tips and Conclusion

You have now walked the full path of dissimilar metal welding: what it is, where it is used, what happens metallurgically inside the joint, how to pick a filler, how to butter and qualify, and how it fails. What is left is compressing all of that into something you can carry to the bench, the tie-in, or the procedure-writing desk. Here is the consolidated playbook.

The Pre-Job Checklist: Run This Before You Strike an Arc

1. Identify both base metals precisely. Pull the material test reports if they exist, and confirm them with positive material identification using an XRF gun or OES analyzer. Never trust a tag, a paint mark, or a customer’s memory. In repair work the alloy is often genuinely unknown, and welding an unidentified casting is guesswork dressed up as work.
2. Predict dilution and check a constitution diagram first. Run your expected dilution through a Schaeffler, DeLong, or WRC-1992 chart and confirm the predicted ferrite number before anyone strikes an arc. The WRC-1992 diagram is published in ASME BPVC Section III, Division 1, Subsection NB precisely because ferrite prediction matters so much in arc welding of these joints.
3. Choose filler by property compatibility, not corrosion alone. Thermal expansion behavior, hardenability, and crack resistance outrank matching the corrosion rating of one side. For the low-alloy side of the joint, our alloy steel welding electrodes guide covers the consumable fundamentals you will build on.
4. Plan buttering and postweld heat treatment before fabrication starts. Decide which member gets buttered, how thick the layer will be, and whether PWHT helps or hurts that specific combination. Retrofitting a buttering decision after fit-up is how projects lose three days.
5. Control heat input and pass sequence deliberately. Stringer beads, modest current, controlled interpass temperature, and a sequence that keeps the heavy, heat-sinking member from starving the lighter one.
6. Qualify a real PQR. Do not improvise, and do not borrow a WPS written for a similar but not identical combination.
7. Inspect with liquid penetrant plus a hardness traverse. PT finds surface-breaking fusion-boundary defects; a micro-Vickers hardness map across the fusion line is the single most revealing check of whether a hard zone formed.
8. Manage fume exposure from the first pass. Stainless and nickel-alloy fillers put chromium, nickel, and manganese into the breathing zone, and OSHA’s hazard communication standard obliges employers to provide information and training on the hazardous materials involved, including the alloys and consumables used in dissimilar metal welding work.

Five Quick Tips That Save Repairs

• Never put 308 filler on a stainless-to-carbon joint. It is the single most common and most avoidable mistake in the trade.
• Machined fit-up beats ground fit-up. Grinding leaves geometry you cannot predict, which makes dilution unpredictable too.
• Side-bend your qualification coupon. Transverse tensiles rarely fail at the fusion line; side bends often do.
• Watch 312 in hot service. Sigma-phase embrittlement makes it a poor choice above roughly 800°F for long service.
• Check the alloy family before assuming nickel filler works. If you regularly weld corrosion-resistant nickel-ferrous alloys, our Alloy 20 welding procedures article walks through filler and heat-input decisions for exactly those joints.

When Solid-State Joining Is the Smarter Answer

Sometimes the honest answer is that fusion welding is the wrong tool. As INL’s review of dissimilar metal joining notes, solid-state processes are routinely used to mitigate the metallurgical challenges of joining drastically different alloys such as steel to aluminum. Friction stir welding, explosion welding, and roll bonding all avoid melting, which means no dilution, no solidification cracking, and no carbon migration. If your mismatch is ferrous-to-nonferrous, or if service temperatures sit where creep and thermal-fatigue mismatch dominate, bring a solid-state specialist in early rather than debugging fusion cracks for a month. And if your work leans toward aluminum joints on the fusion side, our aluminum MIG welding guide covers the practical fundamentals.

Common Questions Before the Final Pass

Can I use the same WPS as a same-material job? No. A dissimilar joint creates a new alloy with no historical property data, so the procedure must be qualified for that specific combination and sequence. Is PWHT always required? No, and sometimes it is actively harmful, particularly at buttered nickel-alloy interfaces. What if the base metal cannot be identified? Stop and test it. Guessing is how end-users inherit cracking six months into service.

Conclusion: Design the Alloy You Are Creating

That is the whole principle of dissimilar metal welding in one sentence. Every dissimilar joint is a functionally graded material you are fabricating in real time, and every decision, from filler chemistry to pass sequence to PWHT, writes a line in that alloy’s recipe. Identify both metals precisely, predict your dilution, choose filler for property compatibility, plan buttering and heat treatment, qualify honestly, inspect seriously, and control fume exposure from the first pass. Do that, and a dissimilar joint stops being the most failure-prone corner of the trade and becomes what it should be: a controlled, deliberate piece of metallurgy that lasts as long as the equipment around it.

Frequently Asked Questions

What exactly counts as a dissimilar metal weld?

A dissimilar metal weld is any joint where the two base metals differ chemically or metallurgically, where the filler metal differs from the base metal (a heterogeneous weld), or where a corrosion-resistant alloy is overlaid on another material. Even joining 304L to 304H technically qualifies, because the compositions and their responses to heat are not identical.

Why is dissimilar metal welding harder than welding matching metals?

The two sides expand, contract, conduct heat and transform at different rates, which generates large residual stresses and unpredictable microstructures across the fusion line. On top of that, base-metal mixing creates a composition gradient that can form hard martensite, brittle intermetallics or ferrite-starved zones prone to cracking.

Which filler metal should I use for carbon steel to stainless steel?

Type 309L or 312 austenitic filler is the standard choice, because its higher chromium and nickel content absorbs carbon steel dilution while still leaving enough ferrite to resist solidification cracking. Never use 308 for this joint — dilution can drop the deposit below roughly 17Cr/7Ni and cause hot cracking.

What is buttering and when is it required?

Buttering is depositing one or more layers of filler onto one member before the final joint is welded, creating a transition layer that the second member then joins to. It is used in highly restrained joints, in low-alloy-to-nickel-alloy connections, and whenever heat treatment or dilution control is easier to manage as a separate step.

Why does dilution matter so much in dissimilar welding?

Dilution determines the final chemistry of the weld, and that chemistry determines which phases form and whether the joint cracks or corrodes in service. A high-penetration process like submerged arc welding can dilute a 309 deposit enough to lose the ferrite needed for crack resistance, so parameters and joint design must be controlled deliberately.

Do dissimilar metal welds need post-weld heat treatment?

It depends on the combination and the code. PWHT can relieve residual stress and temper hard zones, but studies on nickel-alloy buttered joints have found it can embrittle the interface, and expansion mismatch means one side always sees a different effective tempering temperature. The WPS and the applicable code must decide this case by case.

How do you inspect a dissimilar metal weld?

Visual and liquid penetrant testing catch surface-breaking cracks, while macro-etch cross-sections, microhardness traverses and chemical analysis confirm the fusion boundary is sound and dilution stayed within limits. Conventional ultrasonic testing is less reliable at dissimilar interfaces because the abrupt change in grain structure scatters the beam.

What are the main health risks when welding dissimilar metals?

Welding stainless and nickel alloys generates fume containing hexavalent chromium, nickel and manganese, which OSHA and NIOSH link to lung cancer, respiratory disease and nervous-system effects. Controls include local exhaust ventilation, respiratory protection, working upwind, and never welding coated or contaminated surfaces without cleaning first.