The following article appears in the journal JOM,
52 (5) (2000), pp. 13-17

Titanium: Featured Overview

Safety-Related Problems in the Titanium Industry in the Last 50 Years

Eldon Poulsen


This article discusses fatal incidents in the titanium industry and explains the cause and effect. Safety recommendations are included for vacuum-arc-reducing furnace design and operation, handling and storage of titanium fines and sponge, and confined space entry.


The titanium industry dates back to the turn of the century, although commercial production of the metal actually started in about 1950. By the end of 1999, the industry was producing more than 100 million pounds per year. Early on, safety problems arose from a lack of knowledge regarding furnace design and related explosions. The knowledge at the time was based on steel technology, and the hydrogen explosions were a completely new problem. When molten titanium reacts with water, the titanium metal breaks down the water, absorbing the oxygen and liberating the hydrogen, which results in a major explosion. During the first five years of the industry, furnace explosions killed six employees.

The next problems that plagued the industry were fire and explosions from sponge and fines fire, which killed four employees. The third problem was confined space entry. Five fatalities have resulted from argon, nitrogen, and other inert gases.

To address these problems, safety committees were formed and safe operating equipment and procedures were developed. In this article, only operating areas with fatalities are discussed in the three problem areas. Due to constraints on information, only U.S. producer problems are discussed.


The very nature of melting titanium in a water-cooled furnace using copper crucibles establishes a risk. Water leaks can occur, and everything possible is done to prevent this from happening. The problem is that when water contacts molten titanium, the water turns to steam. Titanium has such an affinity for oxygen that it breaks down the water, absorbs the oxygen, and liberates the hydrogen. Under these circumstances, both steam and hydrogen explosions are possible. The goal is to design equipment, procedures, and facilities that will operate safely. If a problem occurs, the equipment and procedures must be designed to keep everyone as safe as possible, even under worst-case conditions.

Titanium melts at 1,635°C; if a 200°C super heat in the molten pool area is assumed, the operating temperature is around 1,835°C. This molten metal is contained in a water-cooled copper crucible. Copper melts at 1,100°C; hence, if the cooling water is lost, the copper will melt.

When water leaks into furnaces, it causes a two-stage explosion. The first is a steam explosion, which is then followed by a hydrogen explosion. In one of the early industry explosions, it was calculated that the combined explosion was the equivalent to a 500 pound bomb or 200 pounds of TNT—not a good thing to have in a melt shop.

There have been at least 50 documented VAR furnace explosions in the US titanium industry, the most recent being the one at Oremet on September 19, 1999. The last fatality from an explosion was in 1960, when Harvey Aluminum, Titanium Division, had a furnace blow in their California plant, killing one operator who was on top of the furnace at the time of the explosion. He was looking through a view port for a visual end of the melt when the furnace blew and decapitated him.

The first industry-wide safety committee was established in the late 1950s and was active until 1965. They developed guidelines for operation and equipment design that were adopted by the industry and are still largely used today. With these guidelines and safety training, VAR melting has become a relatively safe operation. The main results were improvements in furnace design, placement of the furnace melt zones in bunkers, and the moving of the operators away from the operating area. All auxiliary operations were moved out of the cone of the potential explosion. The industry has experienced a significant number of explosions since, but the tragic operator involvement has been minimized.

Following the 1965 downturn in the industry, the joint safety committee was disbanded. The industry worked well for 15 years, then, starting in about 1985, every major titanium company had a significant furnace explosion. The melt-shop managers from the major companies published a VAR safety paper and reestablished the VAR safety committee in 1992. The paper, “Safe Design of Melting Systems for Titanium,” was authored by Eldon Poulsen, Timet; Steven C. Stocks, Oremet; Steve Giangiordano, RMI: Eric Jarvis, IMI; and Jim Silvas, Titanium Melt Division, Teledyne Allvac.

The early furnaces were designed with the water jacket and melt zone above the operating floor. The actual melt zone was at eye level. The operator was stationed near the furnace for better control, since all control systems were visual. As an example, the water-flow meter was a weep hole on the backside of the water jacket. As long as a stream of water was weeping from the hole the operator knew there was cooling. There were no automatic shut-offs or electronic controls.

Figures 1a and 1b are photos of a state-of-the-art VAR furnace in 1955 before and after a catastrophic explosion that killed the operator. Following the explosion, the furnace design was changed, and the water jacket and melt zone were placed in reinforced concrete bunkers. All melting is now performed below the operating floor level. The bunkers were designed to direct the force of the explosion away from the working area, although there is still a hazard zone directly above the furnace. Operators are kept out of this area by procedure. Through the use of optics and video cameras, the control operator has been moved into a room that is remote from the furnace and is protected by bulk-heads.

Figure 1a
Figure 1b

Figure 1. A 1955 VAR furnace operating (a-left) before and (b-right) after an explosion that killed the operator and injured three other workers.

Figure 2a
Figure 2b

Figure 2. A 1999 VAR furnace (a-left) before and (bright) after a water leak and explosion.

Figures 2a and 2b show a 1999 state-of-the-art furnace before and after a water leak and explosion. There was equipment damage, but no injuries. With this design, operators are not to be within 15 ft. of the furnace when it is in operation.

Through the last 50 years and many safety-related committees, a set of guidelines was developed for the safe design and operation of VAR furnaces in the melting of titanium ingots; the sidebar lists these guidelines.


Design and Operation of VAR Furnaces

  Handling Titanium Powders*
1. Loss of water flow shuts off the furnace power. This feature is checked prior to the start of each melt. One rectifier is turned on, then the water is turned off; the rectifier must turn off, verifying the functionality of the safety circuit.
2. There must be an effluent water temperature monitor system. If the effluent water from the furnace has more than a 20°C temperature rise over the input, the system should be checked.
3. Each furnace should have a minimum of two redundant water supplies. If one fails, the system automatically switches to other. Ideally, one system is a gravity flow system with at least 60 ft of head.
4. There must be a vacuum system that continues to operate even with higher furnace pressure.
5. A pressure sensitive control system that terminates melt at one half an atmosphere (i.e., at 14 inches of mercury absolute or below) should be included.
6. Explosion ports should be placed on each furnace. The type recommended would vent, then reseal to prevent air in-rush.
7. The furnace operator should be in an isolated control room protected from the furnace area with barricades. The furnace may be housed in vaults to accomplish the operator's safety. In no case should the operator be located adjacent to the furnace without protection.
8. Where possible, operations, such as electrode preparation, should be remote from the operating furnace location.
9. With the current state of the art video cameras, operators should not be required to spend extended periods of time next to operating furnaces.
10. Inert gas flooding of the furnace prior to opening is recommended.
11. Anti-backup and short arc control systems should be incorporated into the control system for each furnace. Long arcs lead to arced crucibles.
12. A solenoid-type arc-focusing coil should be installed on each water jacket to prevent the arc from attaching to the crucible wall.
13. Every effort should be made to maintain straight electrodes that are concentric with the stub, ram, and crucible.
14. All parts and pieces of electrodes produced from bulk weldable material must be securely welded in place to avoid arcing to the crucible.
15. Where practical, furnaces should be designed with X-Y control and equipped with video cameras to maintain the electrode in the center of the crucible.
16. NaK cooling of crucibles is a desired alternative to water-cooling.

Positive-type bottom supports for crucibles, which are capable of holding the bottom in place if fasteners are arc damaged, should be included.

18. The operator and other personnel should not be allowed in the 30ft diameter at-risk zone around the top of the furnace when melting is in progress.
19. As a rule, manual shut-off valves should never be installed in the drain lines from the water jacket and furnace cooling system. If one is installed, it should be equipped with a safety monitor system that will not allow a melt to be started with the valve closed.
20. Only operators who have been trained and certified should be allowed to run a melt.
1. Avoid any condition that tends to suspend or float powder particles in the air.
2. Avoid every possible action that generates static electricity, creates a spark, or otherwise results in reaching the ignition temperature.
3. Where generation of static electricity may occur, utilize every means to minimize it and dissipate it (by grounding) to avoid a spark discharge.
4. Take steps to limit the size of a fire or explosion and to hold any resulting damage to the very minimum.
5. Do not allow powder particles to become suspended in the air in a room where titanium powder is being processed or handled. The less dust in the air, the better, as the lower limit is indefinite. It is presently considered to be 0.02-0.03 oz/ft3-an extremely small amount. Titanium dust will ignite with as little as 9% oxygen, with the balance helium or 3% oxygen with the remainder carbon dioxide. Titanium will ignite in carbon dioxide.
6. In transferring titanium powder, dust clouds should be kept at a minimum. Powder should be transferred from one container to another with as little agitation as possible.

In mixing titanium powder with other dry ingredients, frictional heat should be avoided. The best type of mixer for a dry-mixing operation is one that contains no moving parts, but affects a tumbling action such as a conical blender. The introduction of inert atmosphere in the blender is mandatory where dust clouds are generated. All equipment should be well grounded.

  * Taken from Aluminum Technology Edited for Titanium, by Eldon Poulsen.


The most prevalent accident occurring in the industry during the past 50 years has been metal fires. The major cause is poor housekeeping and inexperienced operators. Using the fire triangle as a base, there is always fuel for the fire when working with finely divided metals. When handling the material in the atmosphere, there is always oxygen available. Thus, the controllable leg of the triangle is the ignition source, even though the extent of the loss can be controlled by design and housekeeping.

In the past year, the titanium industry has experienced five major fires at a cost of well over $1 million. All were preventable. We can do better.

In 1974, an employee was working as a press and weld operator. He was running titanium sponge through a 24 to 1 splitter to produce containers of sponge for feed to the press. The material was a blend of Kroll sponge and Hunter process sponge. (The latter material was very fine due to the sodium-reduction manufacturing route.) With 30% of the material split, the material began catching in the tote bin. The operator went to the top of the splitter, which was 30 ft. above ground. He used a metal bar to rap the tote bin to get the metal to flow out of the bin. Evidently, he created a spark that ignited a dust cloud of titanium fines in the area. In an instant, the fire flashed to the ground level and back up. He was completely engulfed in flames and died from his injuries. The ensuing fire is depicted on the cover of this issue of JOM.

The investigation of this accident revealed several shortcomings of the system. First, the dust-collecting system was not adequate to handle the fines content of this material mix. Second, the use of the bar was an ignition source. Third, employees should not be allowed to work in dangerous areas.

Following this accident, the dust-collector system was replaced. The bar was taken out of the area, and procedures were established to keep operators from going to the top of the splitter when it was in operation.

Three lessons were learned from this accident.

There have been other significant metal fires, related fires, and explosions experienced in the industry. When it is realized that even wheat flower in a dust cloud can cause a catastrophic fire, all efforts must be taken to eliminate the dust cloud and the ignition source.

Chemically, titanium has an enormous affinity for oxygen. This results in a thin film of titanium oxide being produced almost instantaneously on the surface of the titanium when exposed to the atmosphere. The titanium-oxide film is inert and protects the underlying metal from further attack.

When a titanium-powder particle is heated to a certain temperature (known as the ignition point), the mass of the particle is so small that the entire particle may oxidize almost instantly. Thus, a pile of such particles will burn. Since sponge particles are much smaller in mass than atomized or granular particles, they will ignite more readily and burn faster than the coarser types of powder.

Fine particles of titanium powder, like some organic powders, such as flour, starch, and coal dust, are easily dispersed in air, where their light mass allows them to remain suspended. Like particles in a pile, they will burn when the ignition temperature is reached; but when dispersed in the air (mixed with oxygen) in a certain proportion, the burning extends from one particle to another with such rapidity (pressure rise in excess of 20,000 psi/s) that a violent explosion results.

Laboratory tests by the US Bureau of Mines and others have established the proportions required for an explosion. They extend throughout a wide range, and very little titanium powder is needed. Very small amounts of energy are required to ignite certain mixtures of titanium powder and air. In some cases, energy as low 25 millijoules may cause ignition.

The discharge of static electricity will produce an electric spark that raises the powder particles in its vicinity above the ignition point, resulting in an explosion. Electric switches, broken light bulbs, electric motor commutators, loose electric power connections (even a metal-to-metal impact), anything producing a spark can set off an explosion. Even continued metal-to-metal rubbing (as in a dry-sleeve bearing) can generate enough heat to set off an explosion. Safety principles for handling titanium powders are shown in the sidebar.


The most recent problem in the industry has been the loss of life due to errors in judgement. This occurs when a person is entering a confined space that does not have sufficient oxygen to support life. A confined space is defined as any area that has limited or restricted means for entry or exit, is large enough and so configured that an employee can bodily enter and perform work, and is not designed for continuous employee occupancy.1 The major gases of concern in the production of titanium are argon, helium, nitrogen, chlorine, and titanium tetrachloride. The fatalities that have occurred in the past five years have been caused by argon and nitrogen.

In 1996, a crew was working on a reduction vessel to return it to service. The vessel had been placed back in service and had been pumped to a vacuum and backfilled with argon. The vessel was approximately 6 ft. in diameter and 22 ft. deep. It had an open top and a closed dished bottom. A piece of equipment had dropped to the bottom of the vessel and had to be removed before the unit could be returned to service. The metal was lying on the bottom of the vessel, 22 ft. down from the top flange.

The foreman on duty decided to use the overhead crane and lower himself to the bottom of the vessel to retrieve the piece of metal. He entered the vessel without the required safety harness secured to the crane and without any atmosphere testing of the vessel—factors that contributed to the ultimate fatality. When he was halfway down into the vessel, he lost consciousness and fell to the bottom. A further operator, seeing the problem, slid down the crane cables to try to rescue the foreman. When he got halfway down, he too lost consciousness and also fell to the bottom. A third operator saw the problem, called for help, then rigged two three-quarter inch diameter compressed air hoses, dropped them to the bottom of the vessel, and turned them on to plant air. The fire department was called and arrived on the scene.

The firemen put on breathing apparatus, lowered ladders into the vessel, and retrieved the two men. All the time the two air hoses were blowing compressed air into the bottom of the vessel. The foreman died and the other man was released from the hospital the next day without problems.

Later that day, the atmosphere in the vessel was tested. At the top of the vessel it was 98% air. At 5 ft. down it was 70% argon. At 8 ft., it was 98% argon. With all of the activity in the vessel and the two air hoses blowing air into the bottom of the unit, it might be expected that a lot more of the argon would have been blown out.

It is evident that the air just blew through the argon as if it were water and very little of it was removed. Other studies have shown that the best way to remove argon from a vessel is to use suction and pull the argon from the bottom and pump it away. Another is to be able to open a port at the bottom and let the argon drain out. Another is to turn the vessel over and pour the argon out in a similar manner to pouring water.

Another factor that came out of the investigation is that a person loses consciousness after eight seconds of breathing argon. Evidently, the argon quickly displaced the oxygen in the lungs and, thus, the bloodstream. This is the reason the two men lost consciousness so rapidly and fell to the bottom of the vessel.

At another plant, three operators were cleaning out a rail car. The car had been washed out and backfilled with nitrogen. They had the worker door open on the top of the car, which was designed as a TiCl4 shipping vessel.

When the 30 in. diameter door on the top of the car was opened, material from the door fell down 12 ft from the top flange to the bottom of the car. One of the operators dropped down into the car to retrieve the material. There were placards on the car stating that it had been backfilled with nitrogen, but the operator did not follow procedures and did not have a retrieval harness on. By the time he could be pulled from the car, he had lost consciousness and died.

Several lessons can be learned from these examples.

Entry Procedures and Guidelines

Training should be provided to each affected employee before an employee is assigned duties that require confined spaces duties. Training is also a necessity whenever there is a change in confined space operations that presents a hazard for which an employee has not been trained, and whenever there is reason to believe there are deviations from the confined space entry procedures. Inadequacies in the employee’s knowledge or use of these procedures also warrants safety training.

Before entering a confined space, the entrant(s), attendant(s), and supervisor(s) should meet and discuss the job in detail to ensure that everyone understands what is to take place. Entry into a confined space should not be made unless the following procedures have been completed.

After the previous procedures have been carried out, the following guidelines should be adhered to upon entry.

Confined Space Rescue

Confined space rescue should be initiated immediately only if the attendants or other rescue personnel present meet the requirements of OSHA’s 29 CFR 1910.146 Section (k). The individuals must be properly trained in personal protective equipment and rescue equipment usage, trained to perform assigned rescue duties and have such training at least once every 12 months, by means of simulated rescue operations,2 and be trained in basic first-aid and cardiopulmonary rescuscitation.3


The help and assistance of Matt Mede of Retech is gratefully acknowledged. The sponsorship and assistance provided by Retech in the preparation of this material is greatly appreciated. The help and support of all the members of the International Titanium Association’s Safety Committee is gratefully acknowledged and appreciated. The help of Julia Garland of Retech in the typing, editing, and preparation of this paper is gratefully acknowledged. The author thanks the various companies that comprise the industry for the opportunity to present a perspective on these incidents.


1. OSHA 29 CFR 1910.146 Section (k).
2. Confined Space Entry, Item No. 1910-146 (OSHA, 1995).
3. NFPA 481 Standard for the Production, Processing and Storage of Titanium (Quincy, MA: Nat. Fire Prot. Assoc.,1995).

Eldon Poulsen is a retired from Timet.

For more information, contact E. Poulsen, 4360 Malaga Drive, Las Vegas, Nevada 89121; (702) 451-1809.

Copyright held by The Minerals, Metals & Materials Society, 2000

Direct questions about this or any other JOM page to jom@tms.org.

If you would like to comment on the May 2000 issue of JOM, simply complete the
on-line critique form
Search TMS Document Center Subscriptions Other Hypertext Articles JOM TMS OnLine