Thermit Rail Welding: The Complete Guide to Aluminothermic Rail Joints, Process, Standards & Safety

What Is Thermit Rail Welding? A Portable Foundry on the Track

Thermit® rail welding is a fusion butt-welding process that joins two rail ends face to face using superheated molten metal produced by an aluminothermic reaction between a metal oxide and aluminum. If you have ever watched a crew light a small crucible of dark powder and step back from a blinding white plume, you have seen a foundry pour happen in the middle of a track. That is why track welders call thermit welding a portable foundry.

There is no arc, no electrode, and no power supply. The heat comes entirely from chemistry: aluminum strips oxygen away from iron oxide, and the reaction releases a pool of liquid steel hot enough to melt the rail ends and fuse them into one continuous length. The weld cools, and a train rolls over it roughly 45 minutes after work begins—sometimes more than an hour, depending on rail size and ambient temperature.

The naming trips up newcomers. “Thermite” is the generic spelling used in chemistry; “Thermit®” is the registered trade name tied to Goldschmidt. “Aluminothermic” is the technical adjective—reduced by aluminum. Exothermic welding is the broad category that also covers cadwelding of electrical connections. In railroading, all four point to the same process.

A classic arc weld melts metal with electricity or flame and adds filler. Thermit welding instead casts metal into a mold, much like a foundry pours an ingot. A sand mold wraps the rail ends, a crucible sits above it, and gravity does the rest. Because every component is consumable and self-contained, crews can carry the whole foundry to a remote mainline in a pickup or hi-rail truck.

The process dates to Hans Goldschmidt, who patented the Thermit reaction in 1895 after experimenting with aluminothermic reactions to produce high-purity chromium and manganese. The first rail weld was installed in Essen, Germany, in 1899, home of the Goldschmidt Corporation. American railroads adopted it in the early 1900s, and by 1980 continuous welded rail covered more than 80,000 miles of US main track. Eliminating bolted joints removed most of the clickety-clack and much of the maintenance burden of jointed track.

Thermit welding shines where portability matters. It needs no generator, no diesel flash-butt plant, and no road access for heavy equipment. A three-person crew can weld in a tunnel, on a bridge, at a remote siding, or between turnouts, then grind the joint and walk away. It also joins rails of slightly different sections and can repair broken or battered rail ends in place.

Because the rail ends melt and merge, engineers treat the finished joint as a butt weld in a continuous length of rail—not a mechanical joint, and not a lap. The filler metal is essentially rail steel that came from the crucible, so the weld shares the rail’s chemistry far more closely than a typical arc weld filler.

In practice, thermit welding serves both new continuous welded rail and day-to-day maintenance. Crews weld in new CWR strings, close out field welds between flash-butt plants, and repair service-damaged rail. Whatever the job, the principle is unchanged: molten steel, a sand mold, and a joint that outlasts the fasteners around it.

This guide walks through the science, the step-by-step pour, crucible and mold equipment, preheat and alignment, post-weld shearing and grinding, how thermit stacks up against flash butt welding, crew safety, common defects, and the AREMA standards that govern the work. It is written for railroad track welders, maintenance-of-way crews, welding students, and curious hobbyists who want to understand how a few pounds of powder becomes a permanent rail joint.

The Science: The Aluminothermic Reaction Behind Thermit Welding

Strip away the smoke and the sparklers, and a thermit weld comes down to one of chemistry’s oldest party tricks: a metal that loves oxygen stealing it from another metal, releasing enough heat to melt steel. Understanding that reaction explains why the process works, why it’s so portable, and why the metal chemistry must be tuned carefully before a spark is lit.

The Reaction: Aluminum Steals Oxygen

Thermit welding is an exothermic reduction. The “thermit” is a blend of iron oxide and fine aluminum powder. Aluminum has a stronger affinity for oxygen than iron does, so when the mixture ignites, aluminum rips oxygen away from the iron oxide, leaving molten iron and aluminum oxide (alumina). The two balanced reactions that matter are:

• Fe₂O₃ + 2Al → 2Fe + Al₂O₃ + 181.5 kcal
• 3Fe₃O₄ + 8Al → 9Fe + 4Al₂O₃ + 719.3 kcal

Those kcal figures are enthalpy—the heat released per reaction—and they come straight from the thermite rail welding literature used by the rail industry, including work cited by the Federal Railroad Administration. This is why aluminum is the reducing agent: it reduces the iron oxide, and the reaction releases energy rather than soaking it up.

The alumina doesn’t dissolve into the iron. It’s lighter, so it floats up as molten slag while the denser steel sinks—a natural separation that lets clean weld metal pour into the mold. That same principle of heat moving through molten and solid metal shows up across the trade, as we covered in our guide to metal rolling and welding.

Why the Theoretical Temperature Never Reaches the Weld

Textbook chemistry gives a theoretical reaction temperature of about 5,600°F—hot enough to vaporise almost anything nearby. In the real crucible, that number never survives. Some heat warms the slag, some radiates into the air, and some soaks into the crucible lining and the sand mold. Alloying additions also absorb energy as they dissolve. The result is a melt temperature around 3,500°F—still plenty hot enough to fuse rail steel, but far less dramatic than the theoretical peak.

The reaction itself can reach over 4,500°F (2,500°C), which is why the pour glows a blinding white and why the crew stands well back. The same obsession with controlled heat runs through other processes, from TIG on steel to robotic submerged arc welding.

Why Pure Reaction Iron Makes a Poor Rail

Pour pure thermite iron into a rail gap and you’d get a weld that’s soft, low in carbon, and nowhere near strong enough for a mainline joint. Iron obtained directly from the reaction lacks the carbon and alloy content that gives rail steel its wear resistance and yield strength. It behaves like near-pure soft iron—fine for filling a cavity, hopeless for carrying a loaded train. If you’ve seen what happens when soft metal meets a demanding joint, our breakdown of welding hardened steel tells a similar story.

Alloying Additions That Build a Rail-Grade Weld

So the powder isn’t pure thermite—it’s engineered. Ferro-manganese is added to control temperature and improve metal recovery, helping more of the iron report to the weld instead of the slag, while also contributing manganese for hardenability. Ferro-vanadium and ferro-molybdenum fine-tune strength and hardness so the finished weld matches the parent rail, and mild steel additions dilute the reaction and help moderate the heat. According to Goldschmidt’s Original Thermit documentation, manganese, ferro-vanadium, and ferro-molybdenum are all part of the welding powder—proof that the metallurgy is deliberately designed, not incidental.

Think of it as a recipe: the iron oxide and aluminum supply the heat and the base metal, while the ferro-alloys and steel supply the properties. Get the balance wrong and you cook the pour or soften the joint. Get it right—combined with disciplined practice, the kind covered in our guide to building real welding skills—and you finish with a joint that behaves like one continuous rail. The chemistry sets the ceiling; the crew’s control decides how close the finished weld gets to it.

Step by Step: How a Thermit Rail Weld Is Made

A thermit weld is a choreographed sequence—each step sets up the next. Miss one and you don’t just get an ugly joint; you risk one that fails under a train. Here’s how a trained crew builds a joint in the field, from job briefing to traffic release.

Site Prep and Job Briefing

The crew starts with a briefing: confirm the weld location, rail section, correct thermit portion size, and the traffic window. Tools, molds, crucible and PPE are staged on the ballast, clear of the four-foot. The track is protected with proper work limits, flags or lookouts—a thermit reaction cannot be stopped midway if a train appears, so a genuine window is non-negotiable. For teams new to the discipline, treating thermit like any high-stakes weld starts with fundamentals; see our guide to building real welding skills. A well-planned, safe work area matters as much trackside as in a shop.

Cropping the Rail Ends

Ends are cropped square with an abrasive rail saw or hydraulic cutter. Any damaged, cracked or battered steel is removed—the weld is only as good as the faces it joins. Ends must end up straight and dead square to the rail axis. Where the rail is under tension or the gap fights the pull, a rail puller adjusts it before final fit-up. The AWS D15.2 standard governs the process and quality requirements.

Gap, Cleaning, and Alignment

The correct welding gap is set with a gap gauge or feeler, matched to the rail section and mold. Both rail ends—and at least 4 inches either side of the weld on the top, sides and bottom—must be free of rust, dirt and grease. A wire brush, grinder and solvent do the job. Misalignment here shows up later as a kinked running surface, so align carefully before the mold goes on.

Clamping and Luting the Mold

Two-piece and three-piece refractory molds are clamped around the joint, then sealed—luted—with a sand/clay compound so molten steel can’t escape. Every gap gets packed, and the mold must seat fully on the web and foot. This is the step most welds leak from when rushed.

Charging the Crucible

Single-use or degradable crucibles are charged with the thermit portion before preheating begins; multiple-use crucibles are loaded and reused across joints. The portion is a precisely weighed blend of aluminium and iron oxide plus alloying additions matched to the rail steel, and crews double-check it against the section to avoid a mismatch.

Preheating the Rail and Mold

Torches or a preheat burner dry and heat the rail ends and mold to drive off moisture—critical, because any water flashes to steam and can blow molten metal back at the crew. Follow the maker’s preheat time for the section. Crews wear the full kit here, from face shield to eye protection; our guide to protective welding eyewear explains what that should be.

Ignition and the Reaction

A sparkler or booster is lit and dropped onto the portion. The exothermic reaction erupts in seconds, superheating the charge to molten steel and slag.

Slag Separation and Tapping

After the reaction completes, about 20 to 25 seconds is required for slag to separate from the molten steel before tapping. Modern crucibles use a self-tapping thimble in the pouring nozzle. Slag, being lighter than steel, flows last and overflows into a steel catch basin as the metal fills the mold, where it is left to cool and then disposed of.

Filling, Waiting, and Mold Removal

Molten steel fills the mold cavity around the rail ends. Watch the risers to confirm a clean fill and that no metal escapes past the lute. The crew then waits the supplier’s mold waiting time and removes the mold and crucible, revealing the weld with its head riser and base risers still attached.

Shearing, Grinding, and Traffic Release

The head riser and base risers are sheared or trimmed off, then the head is rough-ground and finish-ground to the rail profile. A poor grind creates surface defects like those covered in our guide to weld defects. After cooling and inspection, the track is released to traffic.

Equipment and Consumables: Crucibles, Molds, Portions, and Torches

A thermit weld is a small, self-contained foundry, and everything it needs arrives in a kit: powder, crucible, mold, ignition, and heat. Get the kit wrong and no amount of field skill will save the joint. The controlling rule is simple — each manufacturer’s process is mutually exclusive, so only approved portions, molds, crucibles, and hardware may be used together. That is not bureaucratic caution; mixing brands risks a mold that will not seal or a portion that does not match the rail.

The Thermit Portion

The portion is the heart of the kit — a sealed canister of aluminum granules and iron oxide powder, plus steel and alloy additions. Ignited at the top, the aluminum strips oxygen from the iron oxide and the melt drops into the mold. Portions are matched to rail chemistry, running-surface hardness, and rail section. For rail steel with a running surface hardness of 370 HB and greater, the recommended process is the Thermit® 2.0 portion kit. Matching matters because the alloy additions decide the final weld metallurgy: a portion blended for standard carbon rail will not deliver the hardness a head-hardened rail demands. Each portion also ships with its own handling data — the Orgo-Thermit portion SDS spells out the reactive aluminum and iron-oxide hazards in plain terms.

Crucibles

Crucibles are single-use, degradable, or multiple-use. Multiple-use crucibles are lined with refractory material — magnesite or crushed alumina slag — and fitted with a bottom stone and thimble. They must be preheated before the first weld of the day to ensure freedom from moisture. Single-use crucibles are lighter but create waste every weld; multiple-use units lower consumable cost on high-production gangs but demand careful drying, inspection, and stone replacement. Moisture is the real enemy: a damp lining or wet mold can flash to steam and blow molten metal out of the cavity.

Molds and Consumables

Molds come as two-piece or three-piece hardened sand or high-silica designs that clamp around the rail ends to form the weld cavity. They are single-use and sized to the rail section. Luting sand or paste seals the mold to the rail and closes gaps so molten steel cannot escape, while a catch basin beneath contains overflow. Correct seating is critical — a poorly luted mold leaks molten steel, wasting the weld and endangering the crew.

Preheating Torches and the Weld Umbrella

Preheating drives out moisture and brings the rail ends to temperature. Air-petrol, compressed-air-petrol, and Oxy-LPG burners are all used, but Oxy-LPG is far faster — about 2.0 to 2.5 minutes versus roughly 10 to 12 minutes for air-petrol. Whatever the fuel, the goal is the same: a dry, uniformly heated rail end. A welding umbrella protects the weld area, welding materials, and finished weld from moisture.

Ignition and Finishing Tools

Ignition uses sparklers dropped into the crucible, though electric igniters improve consistency in wet or windy conditions. The rest of the kit handles alignment and finishing: hydraulic rail pullers bring the gap to the tension the weld needs, and shearing and grinding tools — plus hot chisels — knock off the still-red collar before it hardens. Sound workflow habits here mirror good welding practice: check every consumable, work clean, and inspect before the next joint.

• Portion canister matched to the rail grade and section
• Crucible (single-use or refractory-lined multiple-use) with bottom stone and thimble
• Two- or three-piece mold sized to the rail section
• Luting sand and a catch basin
• Preheat torch and fuel set (air-petrol, compressed-air-petrol, or Oxy-LPG)
• Sparklers or electric ignition
• Hydraulic rail puller, shear, grinder, and hot chisels

Preheating, Gap Control, and Alignment: Where Thermit Welds Succeed or Fail

A thermit weld is a controlled pour of liquid steel into a gap between two rails. Everything that happens before that pour—how hot the rail ends are, how wide the gap is, how straight the rail runs—decides whether the finished joint survives a million-tonne train axle or cracks within days. Get these three variables wrong and the hottest, most spectacular part of the process can’t save you.

Think of preheating and geometry as the welder’s version of precision setup. Just as polarity and current control decide heat input in a shop weld, preheat and gap control decide fusion in a field weld.

Why Preheat Matters So Much

Rail ends must be preheated to create the conditions for full fusion between the molten steel and the base metal. Per the BNSF thermite welding manual, rail ends are brought to roughly 1,100°–1,800°F. Indian Railways specifies a minimum preheat of around 650°C, with 650–850°C observed across RDSO-approved techniques.

The reason is physical. Pour molten steel onto cold rail ends and it chills and solidifies on contact before it can wash away the surface oxidation on the rail. The result is a joint with incomplete fusion (lack of fusion), a defect you can’t see from the track surface. Preheat does two jobs: it drives moisture out of the rail ends and mold, and it keeps the incoming steel fluid long enough to scrub oxides and knit into the base metal.

Flame Adjustment and Preheat Timing

A thermit preheat isn’t a single roaring torch. Most techniques use a propane or oxy-propane burner that must heat the head, web, and foot evenly. Aim the flame to sweep across all three sections—uneven heating leaves cold zones where fusion fails. Overheat the head while the foot stays cool and you get a weak, asymmetric joint.

Timing matters too. Preheat duration is set by the procedure for the specific portion and rail section, not by feel. Too short and you never reach temperature; too long and you waste gas and risk overheating the rail metallurgy. A clock, not instinct, drives this step.

Gap Control: The 25 mm Rule

Rail ends are typically spaced 25 mm (1 in) apart. That gap is the “mold cavity” the molten steel fills, so its width determines how much weld metal you get. Too narrow and you starve the weld; too wide and the portion can’t fill it.

If a weld fails, crews crop it back and set a 75 mm (3 in) gap, then rewelded with a special mold and a larger thermite charge. Getting the gap right is the same discipline as good welding practice: measure, don’t guess.

Alignment: Crown, Twist, and the No-Vibration Rule

Alignment tolerances cover vertical and horizontal straightness, crown (the slight rise over the joint), and twist. Get these wrong and you leave a rough-riding joint or a stress raiser that cracks.

For closure welds, a rail puller draws the opposing rails together and maintains pressure until the weld cools, keeping the gap stable as steel contracts.

Then comes the absolute rule: stop all vibration from alignment until the head riser is removed. No spiking, no anchoring, no concrete tie clips, no tamping. Movement after alignment or after the pour can change gap width, alter crown or alignment, and interrupt solidification—which creates a hot tear. This is closely related to how defects form in any fusion process; see our guide to the overlap welding defect for the same principle of letting molten metal freeze in peace.

Preheat, gap, alignment, stillness. Master those four and the pour takes care of itself.

After the Pour: Cooling, Shearing, Grinding, and Inspection

The blinding plume is gone, the crucible is tipped out, and the mold is full. Now the clock starts—and for the next half hour or so, the crew’s most important job is to leave the weld alone.

Let It Solidify: The Waiting Period

Molten thermit steel only a moment ago was over 4,000°F, and the rail ends around it are soaking up that heat. The weld must solidify and cool undisturbed. Any bump, vibration, or premature traffic can shift the rail ends, crack the still-hot weld, or trap a defect that shows up years later. On a single-alumino-thermic weld, the standard practice is to pull the mold shoes—the mold halves clamped around the rail—before the full waiting period ends. Removing the shoes early lets the weld cool in open air, but you do it carefully so nothing shifts the joint.

Once the shoes are off, the crew shears the head riser and base riser—the extra steel that fed the shrinking weld as it solidified. Hot chiseling follows, trimming the leftover riser stubs and flash close to the rail surface while the metal is still workable. Then the head is ground. Rough grinding removes the bulk of the surplus; finish grinding brings the running surface back to rail profile so wheels don’t trip over the joint. Every completed weld is then marked—usually with a painted or stamped identification that ties the joint to its welding record for later inspection.

Quality Inspection: Proving the Weld Is Sound

Grinding makes a weld look good. Inspection proves it is good. Thermit welds are checked several ways, and the crew should treat each one as non-negotiable.

• Ultrasonic testing (UT): The primary volume check. Per AREMA specification 3.13, thermite welds must show no reflective surface greater than 1/8 inch—a signal that flags porosity, inclusions, or lack of fusion.
• Visual inspection: Look for underfill, excess flash, burned or rough surfaces, and any visible cracks at the collar.
• Hardness checks: Weld metal hardness must fall within +/- 30 BHN of the manufacturer’s specified value. If it exceeds 410 BHN, the weld must be examined for a fully pearlitic microstructure—hard, brittle structures like martensite are a red flag.
• Dimensional tolerances: AREMA limits are 0.060 inch maximum vertical offset, 0.060 inch maximum horizontal offset, and 0.025 inch maximum horizontal kink. A bad kink or offset pounds the wheels and can flag a misaligned weld.
• Slow bend testing: On sample or qualification welds, the joint is bent slowly to reveal brittleness and internal defects that UT might miss.

The point of all this is simple: a thermit weld is only as good as the evidence behind it. If hardness drifts high or a UT reflection shows up, the weld gets flagged, records updated, and the joint may need to come out. Good crews log everything—temperatures, wait times, inspector initials—so a weld can be traced years down the line. That same discipline of documenting, testing, and dialing in technique is what separates a shop that welds by guesswork from one that builds real welding skill.

It’s worth noting that the tools around a thermit crew—irons, grinders, even the small torches used for cleanup—reward the same fundamentals as any other process, whether that’s understanding TIG polarity or keeping contact tips in good shape on a wire feeder. A thermit joint is finished work, but it still lives or dies on the basics.

Once UT passes, hardness is in range, dimensions are within tolerance, and the surface is at profile, the joint is signed off—marked, recorded, and ready for the tonnage to come.

Thermit vs. Flash Butt vs. Other Rail Welding Methods

Thermit welding never works alone. On any railway it competes with electric flash butt welding (FBW), gas pressure welding, and arc-based repair methods — and each process owns a different niche. The standard that governs all of them, AWS D15.2/D15.2M, covers thermite welding (TW), flash welding (FW), SMAW, GMAW, and FCAW for rails and related components — a reminder that no single technique wins everywhere.

Flash Butt Welding: The Plant Champion

Flash butt welding clamps two rail ends, drives current across the tiny contact area, and forges them together once the metal softens. The result is a clean, high-integrity weld with a narrow heat-affected zone and strong fatigue performance. The trade-off is hardware: a flash butt machine is heavy, expensive, and usually confined to a welding plant or a containerized mobile unit. That makes FBW ideal for new continuous welded rail (CWR) construction and long strings welded in a yard, yet awkward for a single broken rail miles from anywhere.

Gas Pressure and Arc Welding

Gas pressure welding heats rail ends with oxy-acetylene torch banks and squeezes them together under pressure. It needs no electricity and makes respectable joints, but it is slow, fussy about alignment, and largely superseded by flash butt and thermite. Arc processes — SMAW, GMAW, and flux-cored — still appear for railhead build-up, repair, and small fabrication rather than full structural joints, and the gasless MIG and polarity rules for those wires apply here just like any shop job.

The Heat-Affected Zone Problem

Every rail weld leaves a band of altered metallurgy. Rail welds have a wide heat-affected zone — approximately 45 mm for flash butt welds and even more for aluminothermic and gas pressure welding — which can cause uneven wear, cupping, and batter at the joint, as AREMA guidance notes. A wider or mismatched HAZ concentrates fatigue and residual stress, and that is where many field failures begin. Historically, thermite welds have shown higher breakage rates than flash butt welds, though modern practice has narrowed the gap.

Raising Thermit Quality: Forging and Heat Treatment

Additional forging — the ALFONS/WRIST module — marginally reduces the fusion-zone and heat-affected zone width but appears to have little negative effect on the residual stress field. Post-weld heat treatment does more for toughness: in a 2,000-pound weight-drop test, heat-treated (STTW) thermite welds broke from 9 to 11 feet, versus only a few feet for traditional thermite welds, while flash butt welds endured drops of about 14 feet. The gap narrows, but FBW still leads on raw impact toughness.

Which Process Fits Which Job?

Field versus plant is the real dividing line. Flash butt welding lives where rails and power are staged; thermite lives wherever a truck can stop. Insulated joints — which must carry a glued, non-conductive gap — are almost always made with thermite molds or special arc techniques because they cannot be flash welded.

Continuous welded rails have been in use since the 1930s, and aluminothermic welding remains the most common continuous welding method performed in track. That staying power comes from portability: a three-person crew, a crucible, and a mold can join rail where no flash butt machine can reach. Where new track is being laid in a plant, flash butt wins on quality; where the track is already in service, thermite usually wins on logistics — and the same visual weld-judgment instincts from any good and bad weld guide still apply.

Safety and Health: Burns, Fume, Fire, Moisture, and PPE for Thermit Crews

Thermit welding packs the hazards of a foundry into a two-foot square of track. There’s no switch to turn off once the reaction starts—molten steel at over 4,000°F, blinding light, toxic fume, and a violent sensitivity to water all arrive within seconds. Good crews don’t treat safety as paperwork; they treat it as the reason everyone goes home.

Thermal Burns: The 8–10 Foot Rule

Molten steel and slag spatter unpredictably during the pour, and the crucible itself can spit if anything goes wrong. The single most important field rule: stay at least 8–10 feet away from the crucible from the moment of ignition until the metal has fully poured into the mold. Only the igniter should be near the crucible, and only long enough to light the portion. Once the plume erupts, everyone else is behind that line. Sparks can carry well beyond it when wind is up, so keep the safety line upwind and never stand in line with the crucible mouth.

Moisture: The Silent Killer

Water is thermit’s worst enemy. Any moisture in the portion or crucible flashes to steam almost instantly: 1 part water becomes roughly 1,000 parts steam by volume, and that expansion can eject molten metal straight out of the crucible mouth. Wet thermit welding powder that ignites will react violently. Manufacturer guidelines prohibit welding in heavy rain, sleet, or blowing snow, and every mold, crucible, and rail end must be dry before you start. Inspect portions for damp packaging and never use one that has been wet—and a well-organized welding facility makes drying and staging part of the routine.

Fire Prevention and Weather Limits

Clear all combustibles—ties, dry vegetation, spilled fuel—within a generous radius, and keep standing water, ice, and snow out of the work zone. On bridges and other spots where a spill could ignite the structure below, pre-wet the surrounding area and post a dedicated fire watch during and after the pour. Wind direction matters too: a downwind spark shower can carry ignition sources far from the weld. Pay attention to rail temperature limits and weather restrictions in the weld procedure; cold, wet, or windy conditions change how the reaction behaves.

Fume and Respiratory Exposure

Thermit welding produces a heavy plume of metal-oxide fume. NIOSH identifies welding fumes as a potential occupational carcinogen linked to metal fume fever, interstitial pneumonitis, and nervous-system effects. Manganese exposure is a particular concern with thermit portions (see the SDS), so work upwind where possible, ventilate enclosed spaces, and follow OSHA 29 CFR 1910.134 for respirator selection, fit testing, and use. Early signs of overexposure—flu-like symptoms, fever, chills, a metallic taste—mean removal from the plume and medical evaluation.

Eye and Radiation Hazards

The preheat flame and the reaction throw intense light and infrared radiation that can cause painful “welder’s flash” the next day. Use shade 5 filtered lenses for preheating and shade 2.5 for checking rail ends—the same logic behind any protective eyewear choice. Side shields, impact goggles, and a face shield round out the eye protection.

PPE and Firefighting Rules

Recommended minimum PPE:

• Hard hat
• Face shield
• Safety glasses with side shields and impact-resistant goggles
• Shade 5 lenses for preheating, shade 2.5 for checking rail ends
• Leather gloves
• Leggings
• Protective footwear
• Welder’s jacket

Leather gloves and a welder’s jacket shed spatter; cotton or leather beats synthetics, which can melt into skin. Good welding practice means this gear goes on before ignition, not after.

Never use water, carbon dioxide, or foam on a thermit fire. Smother metal-dust fires with dry silica sand or a Class D extinguisher—staged before ignition, not after—and remember the reaction itself cannot be stopped once ignited. OSHA’s welding hazards guidance is worth reviewing before your next weld.

Common Defects, Mistakes, and How to Troubleshoot Thermit Welds

Even a textbook thermit pour can turn into a rejected joint in the seconds it takes for something to go wrong. The good news: almost every defect traces back to a handful of preventable causes—moisture, cold steel, movement, or a sloppy mold. Read a finished weld and it will tell you exactly where the process slipped.

Porosity and Gas Defects

Porosity is the most common thermit weld defect, and moisture is almost always the culprit. Damp molds, wet crucibles, sweaty portions, or a water-soaked rail end flash to steam at pour temperature and trap gas bubbles in the solidifying steel. Remedy: store consumables sealed and dry, bake or pre-dry molds per the manufacturer, and never pour into a wet joint. On high-risk pours, a light preheat drives off surface moisture before the reaction starts.

Lack of Fusion

Insufficient preheat, a contaminated rail end, or a cold mold leaves the weld metal sitting on top of the rail instead of fusing into it. That is a crack waiting to open under load. Proper preheating brings the rail ends and mold up to the temperature specified on the portion tag, while clean, bright rail ends free of rust, scale, oil, or grease give the molten steel something to bond to. Flame control matters—an oxidizing flame or shortcuts on preheat time show up later as incomplete fusion. For a deeper look at how flame and current discipline build sound joints, see our guide to welding practice.

Hot Tears and Cracking

Thermit steel shrinks as it freezes. If the rails move—from vibration, traffic on adjacent track, a passing train, or a rail that springs—the joint tears while it is still weak. This is the cardinal rule of thermit welding: if the rail pulls apart during initial cooling, even slightly, the weld is defective and must be replaced. Stop every vibration source, brace the rail, hold the gap, and let the pour freeze undisturbed.

Misalignment, Crown, Twist, and Offset

Gap and alignment errors show up as step, crown, or twist outside tolerance. Verify the gap with a feeler gauge and wedge, check for vertical and lateral offset, and confirm the rail ends are square and level before lighting up.

Hardness and Martensite

Hydrogen, moisture, and too-fast cooling can form martensite—brittle, crack-prone steel. In an AREMA-compliant thermit weldment, the presence of martensite is not acceptable at any location. High-hardness or alloy rail demands correct portion selection and controlled cooling.

Slag Inclusion and Self-Tapping

Poor mold sealing lets slag and turbulence enter the weld. A vigorous or boiling reaction, or a loose closing pin, can cause self-tapping—weld metal entering the mold prematurely. Self-tapping means the joint must be rejected, cut out, and rewelded. Out-tapping (metal escaping a leaking mold) requires letting the joint cool to ambient temperature before restarting. If a mold starts to leak, do not try to save it—let it go. Safety guidance from OSHA and the field procedures used on Class I railroads stress exactly this point.

When to Crop and Reweld

Crop and reweld whenever the weld is rejected, the rail pulled apart, or inspection finds unacceptable defects. Cut back beyond the affected zone, re-square the ends, set the 75 mm gap, and repeat the full process with fresh, dry consumables. Precision troubleshooting here echoes the fundamentals covered in our guide to telling good welds from bad and our breakdown of the overlap defect—the same discipline of reading a joint applies. Ultimately, dry materials, correct portion selection, verified gaps, and zero vibration prevent most thermit failures before they ever reach the mold.

Standards, Quality Assurance, Training, and the Future of Thermit Rail Welding

A thermit weld either holds a mainline together for decades or becomes a defect report nobody wants to write. The difference almost always comes down to standards, documentation, and training—not luck. Get the process approved, the people qualified, and the records straight, and the weld becomes boring in the best possible way.

The Standards Ecosystem That Governs Thermit Welding

In North America, AREMA Chapter 4 is the anchor. Specification 3.13, “Quality Assurance of Thermite Welding of Rail,” sets mechanical properties, dimensional tolerances, and test procedures. It doesn’t micro-manage your pour—it defines what a passing weld must prove afterward, from tensile strength to internal soundness, so every joint can be judged against the same yardstick.

AWS D15.2/D15.2M covers procedure qualification, welder qualification, and general welding safety for rails and related components. In Europe, the EN 14730 series defines the approval procedure for aluminothermic welding processes through laboratory tests on welds made under workshop conditions—slow bend, hardness, fatigue, and ultrasonic testing. Australia leans on AS 1085.20, and Indian Railways follows IRS T-19-2021. Different documents, same philosophy: prove the process, then prove the people running it.

Qualification, Sampling, and Inspection

Process qualification comes first. A manufacturer’s portion, mold, and crucible combination is welded under controlled conditions and destroyed in the lab to confirm strength and soundness. Once approved, production sampling keeps it honest. Under IRS T-19-2021, Indian Railways randomly selects one in every 100 joints for hardness, transverse load, and deflection testing.

Ultrasonic testing is the routine non-destructive check. Operators scan the weld zone for inclusions, porosity, and lack of fusion—the internal flaws a visual pass can’t reveal. The same instinct applies to any fusion process; learning to read a bead critically, as covered in our guide to good and bad MIG welds, builds the eye that rail inspectors rely on. A weld can look flawless on the surface and still hide a cold lap or trapped slag underneath.

Training and Documentation

Workers performing thermite welding must be trained by a qualified instructor. Welders are then qualified in their railroad’s procedures, the specific manufacturer’s processes, preheating equipment, and rail end gap preparation. Job briefings capture the rail profile, gap measurement, preheat time, portion type, and weather. That paperwork isn’t bureaucracy—it’s the trail you follow when a joint fails three years later, and often the only evidence that the crew followed the procedure correctly.

Where the Technology Is Heading

Automation is reshaping the consumables more than the reaction. Single-shot crucibles with automatic tapping reduce the human timing error that causes cold pours. Three-piece molds speed setup, and WRIST/ALFONS forging plus post-weld heat treatment refine the weld’s microstructure for longer service life. These align with the broader automation trend seen in robotic submerged arc welding, where consistency beats raw speed.

Still, flash butt welding dominates new construction because it’s fast and repeatable in a plant. Thermit keeps its niche where flash butt can’t reach: isolated repairs, insulated joints, remote track, and emergency breaks. The exothermic reaction remains unbeatable for portability—a crew with a few cases of portions can weld rail where no welding machine could ever be trucked in. That field flexibility is why the process has survived more than a century of mechanization.

Key Takeaways

• AREMA 3.13, AWS D15.2, EN 14730, AS 1085.20, and IRS T-19-2021 form the quality backbone.
• Random sampling—like one in 100 joints—plus ultrasonic testing catches flaws before trains do.
• Qualified instructors and documented job briefings turn a chemistry trick into a reliable process.
• Automation and forging extend service life, but portability keeps thermite relevant.

Thermit rail welding rewards crews who respect the paperwork as much as the pour. Keep learning—whether that’s standards updates or hands-on skill building through deliberate welding practice—because the joint you make today carries tomorrow’s traffic.

Frequently Asked Questions

What is the difference between thermit welding and thermite welding?

They describe the same aluminothermic process. ‘Thermit®’ is the trademarked name owned by Goldschmidt, while ‘thermite’ is the generic spelling commonly used in North America. Both refer to joining metals using superheated molten steel produced by an exothermic reaction between aluminum powder and a metal oxide.

How hot does a thermit rail welding reaction get?

The reaction can reach over 4,500°F (2,500°C), and the theoretical temperature of the reaction itself is about 5,600°F. After heat losses, the molten steel poured into the mold is typically around 3,500°F, which is still hot enough to fuse the rail ends.

How long does a thermit rail weld take from start to train traffic?

Most field welds take roughly 45 minutes to more than an hour from the start of work until a train can pass. The weld must cool below 700°F (about 370°C) before traffic, heavy equipment, or a rail puller is released, which typically happens 30 to 35 minutes after the head riser is removed.

Why is moisture such a serious problem in thermit welding?

One part water expands into roughly 1,000 parts steam by volume when the reaction ignites, which can violently eject molten metal from the crucible. Even small amounts of moisture cause porosity and defective welds, so portions, crucibles, and molds must be kept completely dry.

Can you put out a thermite fire with water or a regular extinguisher?

No. Water, carbon dioxide, and foam can create explosive hydrogen or flammable methane gas and will not stop the reaction once it is ignited. Metal dust fires should be smothered with dry silica sand or a Class D extinguisher, and if the reaction cannot be extinguished, crews should withdraw and let it burn out.

What is the typical gap between rail ends for a thermit weld?

The standard gap is about 25 mm (1 inch), which ensures consistent results when the molten steel is poured. If a weld fails and must be replaced, the damaged ends are cropped to roughly a 75 mm (3 inch) gap and a special mold with a larger thermite charge is used.

How are thermit rail welds inspected and tested?

Welds undergo ultrasonic testing, visual inspection, hardness testing, and dimensional checks for offset, kink, and crown. Under AREMA Specification 3.13, ultrasonic acceptance requires no reflective surface greater than 1/8 inch, hardness must be within +/- 30 BHN of the manufacturer’s specification, and martensite is not acceptable anywhere in the weldment.

Is thermit welding better than flash butt welding for rails?

They serve different purposes. Flash butt welding generally produces higher-quality welds with a narrower heat-affected zone but requires heavy plant equipment. Thermit welding is highly portable, needs only a small crew and kit, and is the go-to method for in-track repairs, broken rail replacement, and closure welds.

What temperatures are too cold for thermit rail welding?

Manufacturer and railroad guidelines recommend welding at rail temperatures of 40°F and above. Welding between 0°F and 40°F is permitted only with additional preheating and slow-cooling procedures, and welding is prohibited at 0°F and below.

Authoritative Sources & Further Reading

• https://www.osha.gov/welding-cutting-brazing/hazards-solutions
• https://www.cdc.gov/niosh/npg/npgd0666.html
• https://pubs.aws.org/Download_PDFS/D15.2-D15.2M-2013PV.pdf
• https://publications.arema.org/publications_toc/MRE_004/AREMA%20MRE%202023%20Chapter%204,%20TOC.pdf
• https://www.goldschmidt.com/wp-content/uploads/2019/09/Goldschmidt-Original-Thermit_english.pdf
• https://en.wikipedia.org/wiki/Exothermic_welding
• https://irfca.org/docs/thermit-welding.html
• https://orgothermit.com/downloads/SDS%20Thermit%20Portion.pdf
• https://data.ntsb.gov/Docket/Document/docBLOB?FileExtension=pdf&FileName=BNSF+thermite-welding-manual-rules-and-procedures-Rel.pdf&ID=17174551
• https://www.tenderfiles.com/tenderdocuments/52026/1/30b7201a-10f8-47d1-928d-6c8ae0afb0d2?FileName=ATWeld_Manual-2022.pdf