Chapter 6 of 10 · 10 min read
Which metals forge weld
Compare the welding ranges of wrought iron and carbon steels, see why carbon narrows the window and alloy oxides resist flux, which other metals weld in the solid state, and how steel joins iron.
Not every metal forge welds, and those that do differ greatly in how easily. Wrought iron, the metal smiths worked for centuries, is the most forgiving: it welds over a wide range of temperature, brings its own slag to help, and tolerates a slow transfer. Mild steel welds well with care. As the carbon rises, the window narrows and the risk of burning grows, until high-carbon tool steels need the most skilful work. Steels alloyed with chromium form oxides that ordinary fluxes barely touch. Cast iron cannot be forge welded at all. Hart surveyed the metals one by one in 1910 and found reasons in each case, and those reasons still hold.
This chapter compares the welding ranges of iron and steels, explains why carbon and alloying elements make welding harder, surveys the other metals that can be joined in the solid state, and describes the old and still useful practice of welding a steel edge to an iron body.
The welding range
A metal forge welds between a lower temperature, below which its surfaces will not bond in the time available, and an upper one, at which it begins to burn or melt. The width of that range is the smith’s margin.
Hart described nearly pure wrought iron as weldable from an imperfect weld at a cherry red to good welds at a dazzling white, the widest range of any common metal. Mild steel’s range is still generous. Higher-carbon steels must be welded at lower temperatures, because their melting and burning points fall as the carbon rises, while the temperature needed to bond them falls more slowly, so their window closes from the top.
Why carbon narrows the window
Carbon lowers the melting point of steel considerably, so a high-carbon steel begins to melt along its grain boundaries, and to burn, at a temperature where mild steel is still comfortably solid. At the same time, carbon at the surface reacts with oxygen in the fire, and a high-carbon steel held too long at heat loses carbon from its surface, decarburising it, so that the welded edge is softer than the rest. The smith welding tool steel must take a lower heat, a bright yellow rather than a white, use a flux that melts low enough to work at that heat, such as borax, and work quickly.
Cast iron, with several percent of carbon, sits at the far end of that trend. Hart explained that, being high in carbon and silicon, it passes suddenly from a crystalline solid to a liquid, with no range in which it is soft but solid. It cannot be forge welded. It is joined by fusion welding or brazing, as Weldability by the Numbers and Brazing and Soldering by the Numbers describe.
Impurities
Hart also noted that small amounts of some impurities spoil welding even at fractions of a percent. Sulphur causes red-shortness, cracking at red heat; Chapter 2 explained how it enters from the fuel as well as from the steel. Elements that oxidise to form a tough skin, such as aluminium in iron, must be avoided, because their oxide resists the flux. Modern steels are cleaner than those Hart knew, but free-machining steels, which contain added sulphur or lead to help machining, forge weld poorly and should be kept out of the welding stock.
Alloy oxides
Steels alloyed with chromium, and to a lesser extent with silicon, manganese and aluminium, form surface oxides that are much more stable and harder to dissolve than iron oxide. Borax barely touches chromium oxide at forge welding temperatures, so a weld in a chromium steel is spoiled by an oxide film that the flux has not removed. That is why stainless steels cannot be forge welded in an ordinary fire. Smiths who make laminated blades with stainless layers weld them in sealed containers or under a protective atmosphere, keeping oxygen away from the faces entirely, rather than relying on flux. Modern alloy tool steels, with their chromium, vanadium and molybdenum, sit between plain carbon steel and stainless: some forge weld with care and suitable flux, others reliably only in a sealed stack.
Wrought iron’s advantage
Wrought iron’s advantages for forge welding are its low carbon, which gives it a high melting point and a wide range, and its slag. As Chapter 3 described, fine threads of iron silicate slag run all through it, and at welding heat those at the surface melt and help flux the joint. That is why wrought iron was the forge welder’s metal for so long, and why heritage repairs in wrought iron are usually best made in wrought iron, as Chapter 10 discusses. New wrought iron has not been made commercially for decades, so heritage repairers rely on reclaimed stock, sorted and tested.
Welding steel to iron
A traditional use of forge welding was to put a hard steel edge on a softer iron body: the steel bit of an axe, the edge of a chisel or a plane iron, the face of an anvil, the point of a pick. The tough iron gave the tool its strength and resistance to breaking; the hardened steel gave it an edge. The weld had to be made at a heat the steel could bear, below the iron’s best welding heat, with a flux that worked at that heat. Smiths often used a cleft weld, setting the steel into a split in the iron, or laid a thin steel plate onto the iron body and welded it on, then shaped and hardened the edge.
The same principle, a hard layer bonded to a tough body, lives on in modern clad and laminated metals, as Chapter 9 shows, and in hardfacing by fusion welding, which Hardfacing and Surfacing by the Numbers covers.
Other metals
Forge welding is not confined to iron. Hart’s survey recorded several other metals that bond in the solid state when hot and clean.
| Behaviour | Practice today | |
|---|---|---|
| Platinum | does not oxidise in air; welds at white or even red heat without flux | historically hammered together; now rarely needed |
| Nickel | pure nickel welds at bright red to white heat; impure nickel does not | nickel cladding on steel by rolling in protective conditions |
| Copper | can be smith welded with difficulty; oxidises and scales | joined by brazing, fusion or solid-state industrial processes |
| Gold and silver | clean surfaces bond readily | jewellery and laminated decorative metals |
| Aluminium | its oxide film prevents bonding in an open fire | friction, ultrasonic, roll and explosion bonding |
| Cast iron | no plastic range; passes from solid to liquid | cannot be forge welded |
Hart noted that platinum, which does not oxidise, needs no flux, only polished surfaces, and that pure nickel welds readily at a bright red or white heat while ordinary commercial nickel, brittle from its impurities, will not. Hart recorded that copper had been forge welded to full strength with skill, though most smiths and editors of the day thought it impossible, and that brazing or soldering was generally preferred. Aluminium’s tough oxide film defeated the forge, though the oxide is broken up and the metal bonded in modern processes that scrub or deform the surfaces, such as friction stir and ultrasonic welding.
Testing a new stock
Because stock varies, and because a weld’s quality cannot be seen once it is made, the shop tests any new stock before using it for work that matters. A pair of test pieces is forge welded from the stock, allowed to cool, and bent in the vice, nicked and broken, or both, as Chapter 7 describes. A weld that bends with the bar and, when broken, shows the bar’s own grain across the joint passes; one that opens at the joint, showing a flat, grey interface, fails. The test takes a few minutes and has saved the shop from making whole batches of work in steel that would not weld reliably.
Identifying steel without a certificate
Much of the steel a small shop welds comes without a certificate: offcuts, scrap, old springs and files reused for tools. The smith identifies it as best they can. Spark testing on a grinding wheel, comparing the sparks with samples of known steels, distinguishes low-carbon from high-carbon steel and suggests some alloys; a magnet distinguishes most stainless and manganese steels from ordinary steels; a trial hardening in water or oil shows whether the steel hardens. Unidentified steel is kept for non-critical work, and a test weld is made before it is used for anything that will be loaded.
Mixing metals in one weld
Two different steels welded together must both be within their welding ranges at the moment of welding, which means bringing them to heat together at the lower of the two ranges’ upper limits. A mild steel body and a high-carbon steel edge are welded at the high-carbon steel’s heat, a little low for the mild steel but within its range. The smith places the high-carbon piece higher in the fire, or puts it in later, so that it does not overheat while the larger mild steel piece comes up. Two steels with very different ranges may not overlap at all, and cannot be forge welded together reliably.
Grain size and overheating
Even below burning, steel held too long at a very high temperature grows coarse grains, which make it weaker and more brittle when cold. Forge welded joints, heated to near the top of the range, can be coarse-grained around the weld for that reason. The smith refines the grain afterwards by working the weld at progressively lower heats, finishing at a red, and, for tool steels, by normalising: heating to a moderate red and cooling in still air, which refines the structure before hardening. Weldability by the Numbers explains how heat and cooling affect steel’s structure.
Working the weld to size
Once a weld has taken, it is thicker than the bar and rough, and it must be worked back to size and shape while it is still hot. The smith takes a second, lower heat if needed, and hammers the joint all round, turning the piece, until it matches the bar’s section and the line of the joint can no longer be seen. Cravens advised shaping the scarf weld all round by turning it while hammering. Working the weld also consolidates it, closing any fine gaps at the edges of the joint, and on wrought iron it helps carry the fibre through the weld. A weld left thick and unworked is weaker at its edges, where the faces may not have bonded fully, and looks clumsy, so the time to dress it is part of the weld’s cost.
The running case: the shop’s stock rules
This is an illustration. The shop’s owner set rules for welding stock after a run of failed tool welds. Mild steel bar from the regular supplier, with a known grade, could be used for any welding. Reclaimed wrought iron was used only for heritage work, after a test weld from each batch. High-carbon steel for tool edges was bought new, of a known grade, rather than salvaged from old files and springs. No stainless or unidentified alloy steel was used for forge welding at all, and free-machining bar was kept on a separate, marked rack. The failures stopped.
A failure: the salvaged spring
This is an illustration. A student made a small hatchet by welding an edge cut from an old car leaf spring into a mild steel body. The weld looked sound, but when the edge was hardened and the hatchet used, the edge split away from the body along most of the weld line.
The spring steel turned out to contain silicon and chromium, as many spring steels do, whose oxides the borax had not fully removed; the weld had bonded only in patches. Remade with a plain high-carbon steel edge of known grade, welded at the right heat with fresh flux, the next hatchet held. The owner keeps the split hatchet as a lesson in knowing the steel before welding it.
Exercise: welding two steels together
A smith wants to weld a 1% carbon steel edge into a mild steel body. Using the illustrative ranges above, mild steel 1,200–1,450 °C and 1% carbon steel 1,100–1,250 °C, what is the overlap, and at what temperature should the weld be made? If the mild steel body comes to heat more slowly because it is larger, how should the smith manage the fire?
The overlap is between about 1,200 °C, the mild steel’s lower limit, and about 1,250 °C, the carbon steel’s upper limit, a window of only about 50 °C. The weld should be made at about 1,225 °C, the middle of the overlap, a bright yellow rather than a white.
The smith should put the larger mild steel body into the fire first, or deeper in the reducing zone, and add the smaller carbon steel edge later, or hold it higher in the fire, so that both reach the overlap together. Watching the edge closely for the first sparks, and taking the heat as soon as the body is ready, keeps the edge from burning, and a low-melting flux such as borax lets the weld take at that lower heat.
