Thermit in Practice

Early Applications of Aluminothermic Welding in America before June 1905

Aluminothermic reactions generate heat by converting Fe2 O+ 2Al to  Al2 O+ 2 Fe. This heat can be used for Thermit Welding and similar processes.


Thermit powder was invented and initially only made in Germany, but after it could be produced in the United States with effect from June 1904, the benefits of the Thermit process were quickly recognised and it was quickly applied to locomotive repairs, track laying and shipbuilding as well as to the production and casting of steel and iron. A paper by Ernest Stütz describes the early applications as shown below: 


The Thermit Process in American Practice

By Ernest Stütz

Read at the June, 1905, meeting of the American Society for Testing Materials[1]

The Thermit Process

Just a year ago (i.e. in June 1904) the first Thermit was manufactured in this country (i.e. the U.S.A.) and the applications developed in Europe by Dr. Hans Goldschmidt, at the works of Th. Goldschmidt, Essen-Ruhr (founded 1847), were transplanted to American soil and have since blossomed forth under the fostering care of American ingenuity.



Thermit welding in the USA: Thermit Welding of Rudder Stock on Tugboat 'Schenck' on Marine Railway, Sault Ste. Marie
Thermit Welding of Rudder Stock on Tugboat 'Schenck' on Marine Railway, Sault Ste. Marie
The principle of the Thermit process can now be said to be known to the technical world, and it will be sufficient to state that through the ignition of finely divided aluminum and metallic oxide a reaction is started which produces heat at about 5400° F (3000° C) and at the same time reduces the iron oxide to a metallic iron almost free from carbon, in a highly superheated liquid state. Thermit steel has practically twice the temperature of open-hearth steel, and a correspondingly greater fluidity. By suitable additions of carbon, in the form of steel punchings, chilled iron shot or ferro-silicon, its hardness, and, by addition of manganese, its toughness, can be increased to any suitable degree.
The following analyses will confirm this:

Analysis of Thermit Steel: Illinois Steel Company, the Rookery, Chicago, Illinois and Pennsylvania Railroad, Altoona

Analysis of Thermit Steel: Illinois Steel Company, the Rookery, Chicago, Ill. and Pennsylvania Railroad, Altoona


The first is one of pure Thermit steel; the other of the steel in the riser of a welded-steel locomotive frame, drawn out under the hammer into a bar some three feet long and turned down and broken.


The simplicity of outfit and manipulation and the speed with which the reaction does its work are its chief recommendations for industrial purposes.


In a crucible some 20 inches (500 mm) high and therefore easily transportable, in half a minute can be produced 30 pounds (15 kg) of liquid steel, so hot that it will melt a steel bar of 4 inches square (2500 mm²) section and fuse with it to one homogeneous mass.


The essential characteristic of Thermit is that it welds by fusion, and, by reason of this fact, calls for the foundry man's experience more than the blacksmith's. Its success depends on the proper material, shape and condition of the mold.


The mold into which the contents of the crucible are run must be of refractory material.


The general instructions must, of course, be broad and cannot go beyond stating that a mixture of equal parts of sharp sand and ordinary brickmaker's clay has given satisfaction. The formula has been varied sometimes, according to local conditions, in some cases flour, in the proportion of 6 to 100, being used as binder for the sand. Some shops have already evolved their own particular formulas, which they treat as secret.


The mold always must be dry — burnt dry. In some cases, for instance, at the Elkhart shops of the Lake Shore & Michigan Southern, the difficulty has been overcome by using firebrick cut down to size. This certainly overcomes the question of drying molds.


The shape of the mold must next be considered. It must be so constructed that the steel flowing down through the gate will not strike direct on to the casting or forging, but will flow underneath the lowest part and rise around and through it. What is required is good circulation for the Thermit steel. It must flow around all the welding surfaces, and as it gets chilled in contact with these it must be driven up into a riser and be followed by a sufficient supply of fully heated Thermit steel to effect the actual weld, which takes the shape of a collar or reinforcement, cast on or over the fracture.


The mold must, therefore, allow fo

  1. a gate;
  2. a collar, shoe or other reinforcement on the surface of the welded piece and overlapping the edges of the break or joint;
  3. a riser;
  4. a skim gate, to prevent the slag from getting mixed with the steel.

The formula for calculating the amount of Thermit must also allow not only for the cubic space of this reinforcement, but further, for again as much Thermit, to supply the contents of gate and riser. These are the general instructions for welding, for instance, locomotive frames — a problem which some thirty railroads in this countrv (i.e. in the U.S.A.) have investigated with more or less success. These frames are of wrought iron or cast steel and vary from 3½ x 3 ½ to 5 x 6 inches (90 x 90 mm to 125 x 150 mm) in section. They are very liable to break and their repair without dismantling the engine means a very large saving per engine. It has been stated that an engine the frame of which is repaired in the forge remains a fortnight out of commission and the actual weld costs $250 to $300. The work by Thermit can be done comfortably in three or four days, at a cost of about $50.


Repair of Locomotives

In reply to a circular letter of inquiry, about twenty railroads have supplied data, which, however, cannot be considered complete, as some of the most regular and extensive users of Thermit did not care to supply the information asked for.


The first successful weld it has been possible to get a record of was made by Mr. Sanderson, superintendent motive power, Seaboard Air Line, on October 19, 1904. This engine has continued in service ever since. It is one of eight engines welded on that road which has given satisfaction, which speaks highly for the care used at the Portsmouth shops in handling a new and therefore difficult problem.    


Locomotive Frame: Ready for Ignition & Albany Line
Locomotive Frame: Ready for Ignition
Another series of successful welds is reported by the Boston & Albany Line, where Mr. Fries welded five engines quite successfully, one being in continuous service since the end of November (i.e. November 1904). One, welded in the jaw, broke again, but four inches away from the weld.
Of late the Lake Shore & Michigan Southern has shown great interest, and its perseverance has been crowned by success in some very good welds at their Elkhart shops, about which Mr. Webb read a very interesting paper at the last annual meeting of the American Foundrymen's Association, giving a full account of each step in the operation. On a preliminary test, a welded bar 2½ x 2¾ (64 x 70 mm), stood a pressure of 50 tons on supports 20 inches (508 mm) apart, before breaking, and that after two sides of the reinforcing collar had been machined off.
Aluminothermic welding: Break of Welded Bar, 2½ x 2¾  (64 x 70 mm), after Pressure of 50 Tons
 Break of Welded Bar, 2½ x 2¾  (64 x 70 mm), after Pressure of 50 Tons
In all, there are records of thirty engines with welded frames that have been in service for three months or longer. Failures are recorded only in isolated instances and are assignable to two different reasons:
  • First, wrong construction of mold.
  • Second, insufficient Thermit;
in other words, insufficient circulation — therefore, insufficient fusion. 
Thermite welding: Locomotive Frame: Welded in the Jaw
Locomotive Frame: Welded in the Jaw
For those familiar with the process, a weld that breaks on account of lack of cohesion at the welding surface is attributable under all circumstances to lack of experience or care, except in one particular case. It is possible for Thermit welded frames to break in spite of proper execution of the work. The original break is due, in the first place, to a structural defect. With the break in such a position as to necessitate the entire removal of the reinforcing collar, it is too much to expect the mere bridging of the broken ends by Thermit steel to overcome this innate weakness.
An important factor in success in welding locomotive frames is to allow for equal shrinkage of parallel parts; also, wherever possible, to spread the ends apart in order to let them come back when the iron begins to set.
Another operation of interest to railroad men is the welding of spokes of drivers.
In making tests of the metal of such welds, the Chicago, Milwaukee & St. Paul R. R. found a tensile strength of 93,900 pounds per square inch. The analysis agreed with that of the Pennsylvania R. R., with the exception of manganese, which in this case was only 0.74. 
Aluminothermic Welding Spoke of Locomotive Driving Wheel

Welding Spoke of Locomotive Driving Wheel



Next came repairs in marine engineering, which are mostly successes obtained by Mr. Des Anges, superintendent floating equipment of the Long Island Railroad. A 12-inch (304 mm) crank shaft, 13⅝ inches (346 mm) at point of fracture, of the ferry-boat Manhattan Beach was welded with 400 pounds (200 kg) of Thermit. The break was in the "wheel center," necessitating the shifting of the center to a new position and shortening the paddle boxes. The shaft was pre-heated, by a charcoal fire and hand-blower, to black heat. To protect the woodwork of the ferrv-boat an asbestos curtain was hung around the crucible, which served its purpose admirably. The ferry-boat has been in uninterrupted service for nearly three months, and continues so now.


Thermite Weld of Crank Shaft, 'Manhattan Beach'
Weld of Crank Shaft, 'Manhattan Beach'
A rudder stock 5 inches (127 mm) in diameter was welded with 50 pounds (25 kg) of Thermit and 10 pounds (5 kg) of punchings. The collar in this case had to be entirely removed, but the welded rudder-stock has now been m service for eight months.
On the Great Lakes, through the enterprise of Captain Johnson, at that time with the Dunham Towing & Wrecking Company, the rudder shoe of the tugboat Schenck was welded, 125 pounds (62.5 kg) of Thermit being used. The weld was sound — in replacing the propeller a chain broke and the propeller dropped on the welded shoe without injuring it.
Thermit Welded Rudder Shoe, Tugboat 'Schenck'

Welded Rudder Shoe, Tugboat 'Schenck'


Repair of Grey Iron Castings

Some important repairs of gray iron castings are also reported. At the Renovo shops of the Pennsylvania Railroad a hydraulic wheel press was repaired, the part welded having to stand a pressure of 60 tons per square inch (927 N/mm²). The original "strong back" holding the wheel against which the axle was pressed was not strong enough for the purpose until repaired by Thermit.
Cylinder covers are also repaired by Thermit and have been made as good as new.
Work with gray iron castings requires more experience, in regard to pre-heating and cooling down gradually — more Thermit is necessary to effect the weld, on account of a hard, glassy scale on such castings, which resists fusion, and an addition oi fero-silicon (about 2%) is advisable to prevent hard spots at the lines of junction between Thermit steel and cast iron. 
Thermite Weld of 5-in. (127 mm) Rudder Stock
Weld of 5-in. (127 mm) Rudder Stock

Continuous rail for electric railroads

The most important application of the Thermit process is for making a continuous rail.
The process having been brought to a high state of perfection in Europe before coming here, there was little room for changes in practice. About 30 different cities are investigating the process in actual operation and about 5,000 joints have been put in up to date. All these roads recognize in the Thermit process the best and simplest means of joining rails for electric traction, as long as care is taken to do small and simple things right. Competitors in the field of rail welding may send out fanciful blue-prints about broken joints, to create unfavorable impressions, but such manoeuvers prove nothing beyond the fact that they admit the success of the Thermit process in this field.
Some tests may be of interest. A heavy double trolley car was taken over a welded joint with supports 13 feet (4 m) away, without breaking it.

To decide whether the head of the rail got softer, micrometer caliper measurements were taken of depressions made under equal blows of a steam hammer, by a blunt tool hardened at the head, ¼ inch (6,35 mm) in diameter.
One-half inch away from the joint the depression was 0.1432 inches (3,6373 mm).
Three feet away from the joint the depression was 0.1596 inches (4,0538 mm).
The electric conductivity of the Thermit joint is recognized to be higher than that of the rail, due to increase of area, and is permanent.
Thermit Welding Trolley Rail at Holyoke, Mass.

Steel Foundries

That steel foundries should have been the first to recognize the possibilities of liquid steel that can be produced anywhere in half a minute goes without saying. There are already several of the largest with whom Thermit is as much a necessity as foundry sand. Some prefer — for no apparent reason — not to disclose the fact that they repair faults in castings by Thermit, but all can openly admit that they use it to reduce the size of their risers, an application which, through its simplicity, recommends itself to all foundries — gray iron as well as steel. 


Thermit thrown loosely or in a paper parcel on steel will ignite and keep the contents of the riser fluid even after the metal has become plastic in the casting. Liquid cast iron will only ignite Thermit in the presence of the ignition powder.

The application of Thermit to reduce the piping in ingots, although very simple in itself, necessitates some liquid steel being held in readiness to fill up the piping after the solidification has been interrupted by a thermit reaction. This should not be impossible to arrange.


Improvement of Grey Iron Castings

Another branch of alumino-thermics which will be of interest is the improvement of gray iron castings, by the introduction of titanium Thermit in the ladles, by immersing it in a cartridge below the surface of the metal. 


Riehle Bros., Testing Machine Company: Tests on Malleable Iron Bars Cast at Pennsylvania Malleable Company's Works, McKees Rocks, PA.

Riehle Bros., Testing Machine Company Tests on Malleable Iron Bars Cast at Pennsylvania Malleable Company's Works, McKees Rocks, PA.


Some experiments, thanks to our fellow-member's, Dr. Moldenke's, kind intercession, were made at the Pennsylvania Malleable Works, with the foregoing results, the bars having been poured out of the same ladles, one before, the other after, the titanium Thermit reaction.


Experiments with lower grades of iron showed the same favorable results. At the Featherstonc Foundry, Chicago, titan Thermit treated test bars showed a tensile strength of 3,550 Pounds (15.8 kN), against average untreated, 3,250 Pounds (14.5 kN). The metal, after treating, is much denser, but can be easily machined.

Incidentally it may be mentioncd that by the introduction of a 1½-pound (750 g) cartridge of ordinary black Thermit into an 800 pound (400 kg) ladle 40 pounds (20 kg) of steel borings can be melted without difficulty.

This necessarily very short account of what is doing in Thermit cannot, of course, cover the entire field of the applications, but will perhaps tend to convince those who had rather be guided by results obtained elsewhere than spend time and money for what they think experiments, and encourage others who are doubtful from lack of experience, by showing them what has been accomplished in actual practice.


  1. Ernest Stütz: The Thermit Process in American Practice. Read at the June, 1905, meeting of the American Society for Testing Materials. In: The Iron and Steel Magazine, September 1905, p. 212-221.