Dilution is the whole problem
When two different parent materials are welded together the result is a heterogeneous weld metal — part filler, part parent from each side. How much parent ends up in the weld depends on the process and the parameters, and it is not a small effect: ESAB puts manual stick welding at 15 to 30 % pick-up and MIG/MAG at 25 to 40 %. A 316L consumable diluted by a third with plain carbon steel is no longer anywhere near 316L. It can fall short of the chromium and nickel needed to stay austenitic, and solidify partly martensitic — hard, brittle and liable to crack either on cooling or later under load.
The answer is to start richer
E309L is formulated at roughly 23 % chromium and 12 % nickel, substantially above the 304 or 316 it is usually joining. That excess is deliberate: after the joint has diluted it by 25 or 30 % with carbon steel, the deposit is still comfortably austenitic. E309L is not a better stainless electrode than E308L — on a matching joint it would be a wasteful choice — it is an electrode designed to survive being watered down. Where the stainless side is 316 or 316L, use E309LMo instead, so the molybdenum that side was specified for carries through the joint.
The 300 °C line
This is the fact that most often changes a procedure. ESAB is explicit: with austenitic consumables in a dissimilar joint, there should be no post-weld heat treatment and service use up to a maximum of 300 °C. Above that, chromium carbides precipitate at the grain boundaries and deplete the grains of chromium, reducing both strength and corrosion resistance. The handbook names the limit against a specific product in its own worked example — for OK 67.60, 300 °C should not be exceeded. If the joint will run hotter than that, 309L is the wrong answer however well it welds.
Above 300 °C, use nickel
Nickel-based consumables are suitable for operating temperatures above 300 °C and, unlike austenitic fillers in this application, they tolerate post-weld heat treatment. The Alloy 82 and 182 family is the usual answer — ESAB rates its Exaton Ni72HP grade for service in air to 1175 °C and in sulphur dioxide atmospheres to 800 °C, and names stainless-to-carbon-steel joints among its applications. They also resist the carbon migration that slowly weakens a hot stainless-to-carbon interface, where carbon moves out of the carbon steel into the higher-chromium weld metal and leaves a soft band on one side and a hard band in the weld. The cost is that dilution control matters even more — the handbook stresses paying particular attention to low dilution with nickel alloys.
The standard nickel procedure is a separate design
Where a nickel-based filler and post-weld heat treatment are both required, ESAB sets out the sequence: butter the fusion face of the low-alloy parent with the nickel-based alloy; carry out the necessary post-weld heat treatment, such as tempering or stress relieving, on that buttered component; then weld the joint between the nickel buffer layer and the high-alloy side with nickel-based filler. The PWHT happens while the joint is still two pieces, which is what makes it possible at all — the heat treatment the carbon steel needs would damage the finished austenitic joint.
Predicting the result: the Schaeffler diagram
The Schaeffler diagram plots chromium equivalent against nickel equivalent and shows what microstructure a given composition will produce. Plot both parent materials, draw a line between them, and the midpoint is the heterogeneous parent if both melt equally. Plot the filler, join it to that midpoint, and move along the line by the dilution percentage for the process to find the actual weld metal. ESAB works a full example for a P355GH to X15CrNiSi25-21 joint using OK 67.60 at approximately 25 % dilution. For multi-layer welds the previous layer counts as a third parent, but in practice it is enough to shift each subsequent layer's point along the line towards the filler.
Where the diagram stops being valid
Schaeffler has published validity limits and exceeding them makes the result unusable, not merely approximate: carbon below 0.2 %, silicon below 1 %, manganese below 4.0 %, molybdenum below 3 %, niobium below 1.0 %. ESAB gives the instructive failure case — an 18 8 Mn type consumable at roughly 6 % manganese is outside the range, and the diagram places it in the hot-cracking danger zone. In reality its high manganese gives it very good hot cracking resistance and the deposit does contain delta ferrite. The diagram is a tool, and a tool used outside its range gives a confident wrong answer.
Two zones to stay out of
The usable region on the diagram is narrower than it first looks, bounded on both sides. A fully austenitic deposit risks hot cracking during solidification, because there is no ferrite to interrupt the continuous austenite grain boundaries a crack runs along. Too far the other way and sigma phase forms, particularly during subsequent service at elevated temperature — sigma is very brittle and can threaten the integrity of the weld. What remains is a relatively small area in the middle, and hitting it is what consumable selection for a dissimilar joint is actually doing.
