Print Page   |   Contact Us   |   Report Abuse   |   Sign In   |   Register
Community Search
Safety Guidelines for Handling Titanium
Share |

Table of Contents:


Antitrust & Liability


Titanium Sponge, Chip, And Scrap Handling

Handling, Storage and Disposal Of Titanium Fines, Powder and Grinding Swarf
    The five most common production techniques that generate either titanium fines or powder are:
    Flame Temperatures
    Common ignition sources
    General Housekeeping Tips
    Controlling Ignition Sources
    How to dispose titanium fines and powder

Basic Safety Principles to be Followed When Handling Titanium Powders From Aluminum Technology
    Why do titanium powders burn?
    Why do they explode?
    What are the primary causes of dust explosions?

Titanium Chip Material Handling System
    Material Storage
    In-Process Transporters
    Fire Detection System
    House Cleaning

Fire Safety

  1. A Thermodynamic Investigation into Reactive-Metal Melting-Furnace Explosions
  2. Safe Design of Melting Systems for Titanium
  3. Liquid Metal Cooling for Consutrode Melting
  4. Melting Safety: From a Directed Water Flow Perspective
  5. Safe Operating Precautions for Plasma Melting Systems
  6. Safety in the Skull-Melting Plant for Production of Titanium Slabs and Ingots
  7. Safety Philosophy in Titanium Melting Systems

Melt Crucible Issues

  1. A three-level safety strategy
  2. Crucible Wall Thickness and Melt Shop Safety And Crucible Inspection Tools, Process and Data Records
  3. "Are You Operating a Bomb?”

NaK Training

  1. The Potential Hazards of NaK
  2. A Brief Description of NaK Cooling on VAR Furnaces

NFPA Information Guidelines


This guide was developed by the Safety Committee of The International Titanium Association (ITA) and is intended to provide information on general guidelines for safe handling of titanium in processing/melting. The guidelines provided are based on the collective experience of members of the industry, but are not intended to be either exhaustive or inclusive of all pertinent requirements.

The information provided in this document is offered in good faith and believed to be reliable, but is made WITHOUT WARRANTY, EXPRESSED OR IMPLIED, AS TO THE MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR ANY OTHER MATTER.

The guidelines provided and the examples included are not intended to be directed to any particular product or business operation, nor are they claimed to satisfy all current legal requirements related to control of processing operations. Following this guide does not guarantee compliance with any regulation nor safe operation of processing facilities. Users are cautioned that the information upon which this guide is based is subject to change, which may invalidate any or all of the comments contained herein.

This guide also is not intended to provide specific advice, legal or otherwise, on particular products or processes. In designing and operating processing lines, users of this guide should consult with their own legal and technical advisers, their suppliers, Material Data Sheets and other appropriate sources (including but not limited to product or package labels, technical bulletins, or sales literature) which contain information about known and reasonably foreseeable health and safety risks of their proprietary products and processes.

ITA, its members and contributors DO NOT ASSUME ANY RESPONSIBILITY for the user’s compliance with any applicable laws and regulations, nor for any persons relying on the information contained in this guide.

Antitrust & Liability
The International Titanium Association (ITA) is committed to ensuring that you, your company, and ITA fully comply with the relevant antitrust laws as they pertain to the activities of ITA. To that end, the following Guidelines have been prepared by counsel for ITA to assist you and other members of the ITA in understanding the antitrust laws and ITA’s special resolution regarding prices, capacity, and sales. These Guidelines are intended to go beyond the minimal requirements of the antitrust laws, and set ITA a higher standard than merely "getting by."

It is your responsibility to familiarize yourself with these Guidelines and to comply with the antitrust laws. Antitrust compliance is important because the consequences of violations can be serious to ITA, your company, and to you. Violations of the Sherman Antitrust Act are felonies which can subject an individual to fines of up to $100,000 and to imprisonment for as long as three years, and subject ITA or your company to civil liability for treble damages and to injunctions that could impair your company’s ability to compete effectively.

The International Titanium Association Safety Committee has primarily been represented by Melters, Safety Managers, and Engineers.  The committee has evolved and now encompasses a Powder Safety Sub-committee.  The goal of the Safety Committee is to promote practices, procedures and systems that will protect people, equipment, the environment and the surrounding communities from the potential hazards associated with the melting and processing of titanium.

The Safety Committee has created this Safety Reference Site.  The Safety Reference Site will outline some of the hazards associated with manufacturing titanium.  The objective of this site is to establish a safety-oriented dialogue throughout the international titanium industry that will lead to a free flow of available information. The hope is that the use of this information should lead to the reduction in number and severity of accidents in melting of titanium. The available data on the cause and effect of the accidents, which have occurred in the melting of titanium over the years, is reviewed.   The topics discussed are the cause and effect of accidents in melting of titanium that result in an explosion. 

The primary goal is to improve the safety of the industry by developing the best management practices and provide the safest workplaces possible.  It is agreed that a problem for any one company is a problem for all the companies.

When it comes to issues regarding fire safety for the titanium industry, the key word to remember is: "housekeeping.”

Though it may seem like mundane advice, housekeeping—that is, maintaining a tidy, safe, well-though-out shop floor layout—is a common-sense concept that, unfortunately and all too often, gets lost in the shuffle of a company’s daily work flow.

Housekeeping, first and foremost, needs to be a priority when working with titanium.  Basic housekeeping guidelines typically are overlooked.  Refer checkpoints in the "5S+1” program, part of the Japanese "Kaizen” philosophy of continuous process improvement and lean manufacturing. 5S represents a management concept dedicated to organization and efficiency.  "Everything has its place; everything in its place,” he said. "These are life lessons.”

The five points of the 5S program are:

1.     Sort (having the facility cleaned and well organized);

2.     Set in Order (creating well-organized, individual work stations);

3.     Shine (regular maintenance and cleaning);

4.     Standardize (document maintenance procedures to make them consistent and easy to follow for all employees); and

5.     Sustain (perpetuating and building upon the improvements that have been achieved).

+1 Safety
The first rule of thumb for housekeeping in fire prevention is material segregation: safely storing titanium scrap, chips and "swarf” in metal containers. Swarf is a term used to collectively identify extremely fine unoxidized titanium scrap normally created from wet sanding operations. It represents a major fire hazard because small-particle scrap is more susceptible to ignition and burning. Small scrap always should be stored in quantities easily moved with the material handling equipment available. Any facility processing titanium scrap should have a "hot work” program that prevents torch cutting, grinding and welding in areas where titanium and other combustibles are stored.

Material segregation also involves separating small titanium scrap from other metal scrap, especially steel dust. The mixing of rusting steel and titanium scrap along with water could create an exothermic reaction or heat which leads to burning materials.  Titanium, a reactive metal, wants to return to its oxide state such as titanium dioxide when it burns.

Beyond material segregation, fire safety should include preventing the buildup of metal dust in a titanium facility.   Developing a plant-maintenance regime that involves regular inspection and cleaning of a facility’s equipment, air ducts and dust collectors. Titanium powder and unoxidized dust can flare and burn when suspended in the air.

Other housekeeping points involve the periodic inspection of a plant’s lighting and electrical systems. Static-charge sparks can be an ignition source for titanium. To avoid static buildup, use grounding and bonding wires on all equipment used to process titanium chip, dust and powder. Don’t forget to look at building support structures or suspended ceilings for dust accumulation.  Many facilities forget to look in areas that are not easily accessible.

Companies are urged to put together a comprehensive fire safety program to educate employees.  Companies should consult with an industry fire prevention expert or to contact state, county or municipal fire inspectors. (The ITA is a resource to obtain this type of information.) Of course, any common-sense fire safety program involves designating safe, restricted areas for those employees who smoke.

What happens if there is an accident and a fire breaks out at a titanium facility? A typical titanium fire burns hot and steady; rarely is there a sudden explosion or a violent burst of flame.  Users and processors should be aware of areas where titanium dust may build up and take action to prevent accumulations in tight areas.

If a fire does occur  DO NOT USE WATER to douse the blaze. The best way to extinguish a titanium fire is to smother it with salt (regular table salt), sand or dirt.   50-pound bags or 55-gallon drums of sand or salt can be purchased at industrial supply outlets. These bags and drums should be stored on the factory floor in specially designated areas where they can be easily reached.

Standard for Combustible Metals:

The National Fire Protection Association has published the NFPA 484-2012, "Standard for Combustible Metals”, available for sale at:

The Table of Contents for Chapter 12 Titanium is reproduced with permission from NFPA.  This reprinted material is not the complete and official position of the NFPA on the reference subject, which is represented only by the standard in its entirety. The NFPA is located in Quincy, MA ( The ITA is grateful to the NFPA for providing this information.

Click Here to View Table of Contents for Chapter 12 Titanium


Titanium Sponge, Chip and Scrap Handling

This chapter will discuss:

Handling, Storage and Disposal Of Titanium Fines, Powder and Grinding Swarf In Titanium Production and Finishing

    5 Common Production Techniques Generating Titanium Fines Or Powder
    Flame Temperatures
    General Housekeeping Tips
    Controlling Ignition Sources
    How to dispose titanium fines and powder

 Basic Safety Principles to be Followed When Handling Titanium Powders From Aluminum Technology

    Why Do Titanium Powders Burn?
    Why Do They Explode?
    What Are The Primary Causes Of Dust Explosions?

Titanium Chip Material Handling System
    Material Storage
    In-Process Transporters
    Fire Detection System
    House Cleaning

Handling, Storage and Disposal Of Titanium Fines, Powder and Grinding Swarf In Titanium Production and Finishing

The titanium industry produces thousands of pounds of titanium swarf and dust collector sludge per year. The handling and proper disposal of these materials has been a problem for the industry for the last 45 years.

This problem is common to all aspects of titanium processing where the titanium or alloys are generated in a finely divided state, having a large surface-to-weight ratio.

The problem involves titanium’s inherent propensity to return to the oxide state. The energy involved in the forming of Ti02 from titanium is adequate to break down water, absorbing the oxygen and thus freeing the hydrogen. This "mechanism” has been the problem in melt-furnace explosions, bag-house explosions and in the handling and disposal of grinding swarf. In addition this is why water is not recommended for fighting titanium fires.

There are several different ways that titanium fines and powder are produced. Each has its own unique set of problems. By common definition, titanium fines are material that will pass through a 20-mesh screen. The definition of a "powder” is any material that will pass through a 48-mesh screen.

The five most common production techniques that generate either titanium fines or powder are:

  1. Titanium sponge production and handling.
  2. Dust collectors and bag houses associated with titanium production.
  3. Turnings from machining or sawing operations. This is especially true of polishing operations.
  4. Grinding and polishing operations.
  5. Wash downs and crucible cleaning.

Titanium burns at a full-white heat. Even sparks coming off from a grinder are white hot. One test for the presence of titanium is to hold a piece of the metal against a grinding wheel. If the sparks are white hot, the material is titanium. With more iron present, the sparks are down in the red heat area. The following table shows the temperature of different color flames.

Flame Temperatures:

Red Heat
525°C = 970°F
Dark Red Heat
700°C = 1292°F
Cherry Red Heat
850°C = 1562°F
Pale Red Heat           
950°C = 1742°F
Yellow Heat
1100°C = 1012°F
Gray-White Heat
1300°C = 2372°F
Full White Heat
1500°C = 2722°F

Titanium fines can be ignited with a common table match. The table above shows the ignition point to be 805°F, which is in the red heat zone.

In the handling and storage of finely decided titanium, the potential for fire in titanium is comparable to other combustible materials. Three things are required to start a fire: fuel; oxygen; and an ignition source. In most grinding operations, dust collecting or any fines-producing operation, there is abundant fuel (Ti fines) and more than adequate oxygen (from air).

The only controllable is the ignition source. Common ignition sources are:

  1. Open flame, torches, heaters, or welding operations.
  2. Electrical sparks, motors, broken light bulbs or shorts.
  3. Static charges from metal sliding across plastic such as buckets. Moving belts or even latex painted surfaces. (There are a number of incidents that have been reported where a technician dumped fine metal out of a plastic bucket and had it ignite.)
  4. Friction from belts rubbing together, metal sliding on metal, etc.
The size of the problem generated by a fire is in direct proportion to the amount of fuel present. It is recommended that wherever possible, the material be collected in small quantities and removed from the working area on a regular basis. The nature of the operation will determine the amount to be collected. A good rule of thumb is no more than a five-gallon bucket at any one time. The material should be moved to a safe storage area and collected in 55-gallon drums.

The Department of Transportation (DOT) requires that most fines be shipped in 10-gallon drums or less. Any collection system should have duplicate drop legs with Ducon-type valves that prevent backfiring into the collection system. Only one leg should receive material at a given time. All parts should be connected together with copper cables and alligator clips to bleed off any static charge.

The control of fires and explosions in the production, collection

Storage and handling of titanium fines and powder can be reduced by controlling sources of ignition. The severity of the problem is controlled by controlling the quantity that can be involved in a problem. In some extreme cases, inert gas can be used to control the availability of oxygen, but for the intent of this paper, we will look at controlling the ignition source.

General Housekeeping Tips

Keep all areas clean and clear of combustible materials.
Never allow materials to accumulate in work areas.
Never let excessive amounts of turnings accumulate on lathes or saws. A small titanium fire can destroy a lathe or saw.
Wash down the area on a regular basis at least quarterly.
Check areas to establish accumulation rates. Place cups on beams, and check them regularly to determine accumulation rate.


Controlling Ignition Sources

1.  Static Charge


No Plastic equipment where material is moving

No rubber separation in ducting or other handling equipment

All sections should be connected together and grounded to bleed off static charge

2.    Materials of Construction

All tools should be of the non-sparking type

Where possible, drop legs or collection units should be aluminum, stainless or brass

Never use plastic liners in pails or drums

Never use plastic pails or drums

3.    Dust Collectors and Bag Houses (see NFPA Standard No. 91)

Collect in small quantities and remove regularly.

Check airflow in ducts and at collection points.

Maintain 3500 ft. per min. in ducts and 200ft. per second at collection points.

Never store large quantities of material near collection points.

Dust collectors should be outside the buildings.

A firewall should be installed between collection points and building.

4.    Use explosion-proof motors and electrical gear.

5.    Electrical panels should be of the NEMA type that can be sealed against the collection of dust and fines. Where possible, the panel should have a positive pressure inside to keep out the dust. Filtered air or inert gas can be used. The compressed-air system in most plants can be filtered and used with a small flow.

6.    Area wash downs are a must. A complete wash down should be done quarterly. Wash all walls, beams and other collection areas. The slitter grinders and riffles should be washed down at the same time. The floor drain should have a Weir box to collect fines and powder so it is not carried into the sewer system.

7.    Once fines are water wetted, they should be maintained wet or they should be dried in a safe manner such as a vacuum dryer. Large quantities of wet fines should not be left to air dry without proper fire containment. Wet fines have been known to ignite spontaneously. This is especially true of sponge fines, where some Mg and MgCI2 are present.

8.    It is recommended that fines and powder be stored in steel containers. The containers should be purged with argon before sealing. Drums purged with argon have been checked and found to contain the argon even after five years and more in undisturbed storage.

9.    The Department of Transportation (DOT) regulates shipment of fines and powder. If planning to ship material, DOT regulations need to be checked. The regulations even apply to small quantities. A good technique for shipping small quantities, such as samples, is to use a 2-inch pipe nipple with caps on both ends to contain the sample container.

How to dispose titanium fines and powder:
  1. The ideal way is to sell or use the material. The aluminum companies use large quantities for grain refining.
  2. Mix with dry sand, using a 5-sand to 1-titanium ratio, then bury.
  3. Burn Ti02 in a controlled area, and then send clinkers to a landfill.
  4. Large quantities can be chlorinated to a usable product.
  5. The material can be digested in a HF-Nitric solution and sold to the chemical industry as a titanium fluoride.
  6. Field burn and landfill is probably the most common approach used. Many trucks and lugger boxes have been burned up using this technique.
  7. Some fines can be sold to the pyrotechnic industry for July 4 fireworks and displays.
  8. Materials can be compacted and handled safely or used for feed to melting or chemical applications.

In general, the best approach is to use the material for the designed application. That is, the sponge fines should be blended with the sponge and alloy, then pressed into compacts and melted into ingots. 


Basic Safety Principles to be Followed When Handling Titanium Powders From Aluminum Technology

Why do titanium powders burn?
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 and have much greater surface area per unit mass than atomized or granular particles. They will ignite more readily and burn faster than the coarser types of powder

Why do they explode?
Fine particles of titanium powder like some organic powders such as flour starch. and coat dust are easily dispersed in air where light mass allows them to remain in suspended or float in the air. 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 (rate of pressure in excess of 20,000 psi/sec) that a violent explosion results. Laboratory tests by the U.S. 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 as 25 milijoules may cause ignition.

What are the primary causes of dust explosions?
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. Anything producing a spark can set off an explosion—such as electric switches, broken light bulbs, electric motor commutators, loose electric power connections, and metal-to-metal impact. Even continued metal-to-metal rubbing (as in a dry-sleeve bearing) can generate enough heat to set off an explosion

Titanium Chip Material Handling System 

Material Storage
The chips are unloaded into what we call "transport” carts or transporters. Each cart can hold a total of 24,000 chips. This provides good segregation of materials for batch processing. The carts are wheel mounted for easy movement of the material. They are built with an oil trap and drain for easy clean out. The transporters are kept under covered storage and can be easily pulled to various processing areas in the plant.

In case of a fire, you can move transport carts into a remote area of the production facility to burn. The advantage of this system is to prevent the spread of fire into other storage carts and facility areas.

In-Process Transporters
The large transporters are emptied into the chip processor holding bend where the chips begin the cleaning process. The in-process transporters are small, portable and easy to move. While the system is a continuous process, the transporters are small and can maintain control of the grade of material

In case of a fire, the small portable concept allows one to move the transporter easily out of the building to a holding area and let burn.
The concept is to confine the fire to as small an area as possible.

Fire Detection System
Ultra-violet (UV) heat sensors are used in all areas of potential fires. Steel conveyers are used in high-risk areas of the system where chips are transferred to different stages of the process. The conveyers can be reversed and unloaded. If there is a fire detected the conveyer is reversed automatically, the load is dumped into a transporter and moved outside to burn off.

The (UV) detectors are programmed to automatically sound an alarm, stop the system, reveres the conveyor and start the unload process. The alarms and indicator lights are located at each transfer station so the plant operator can locate the problem quickly.

House Cleaning
House cleaning is the number-one concern for fires. The process generates a lot of finely divided titanium powder. The system should be kept clean to prevent the accumulation of this material.  Operator training is a must. They must understand the fire hazards of titanium. The supervisors must enforce safe-handling procedures continuously. Periodic fire drills should be required.

Fire Safety
This section will discuss:

A Thermodynamic Investigation into Reactive-Metal Melting-Furnace Explosions

Safe Design of Melting Systems for Titanium

Liquid Metal Cooling for Consutrode Melting

Melting Safety: From a Directed Water Flow Perspective

Safe Operating Precautions for Plasma Melting Systems

Safety in the Skull-Melting Plant for Production of Titanium Slabs and Ingots

Safety Philosophy in Titanium Melting Systems


A Thermodynamic Investigation into Reactive-Metal Melting-Furnace Explosions

Western Zirconium, a plant in the Nuclear Fuels Business Unit of Westinghouse Electric Company, embarked on a significant safety-centered improvement its vacuum arc remelting (VAR) process. The improvement involved the relocation at the control room and installation of a programmable logic controller control strategy for the VAR furnaces.

A critical consideration for the control strategy involved the implementation of the correct response to a breach in the water containment system that would allow water to contact the molten reactive metal in the furnace.

Western Zirconium performed a random sampling of reactive metal melters in the United States to ascertain and evaluate the range of industry responses. Two distinct, and differing, approaches were discovered. In order to develop a better understanding and evaluate the responses, Western Zirconium, in conjunction with the Westinghouse Science Technology Department, evaluated the thermodynamics involved in the reactions of molten reactive metals with water.

In the early 1950s, a process for making commercial quantities of reactive metals (titanium, zirconium and hafnium) was developed and since then, reactive metals have found uses in many unique applications. Titanium is used when a material with lightweight and high strength or good corrosion resistance is needed. Zirconium and hafnium are used predominantly in the nuclear industry for uranium fuel-rod cladding and reactor control rods, respectively, or for specific corrosion-resistant commercial applications.

Manufacturing these metals in large quantities required the development of a method for melting the metal, with appropriate alloying elements, into ingots that could be subsequently processed into wrought shapes.

Titanium, zirconium, and hafnium are labeled "reactive metals” because they very rapidly react with oxygen and/or nitrogen to form refractory compounds. Therefore, the most significant problem with the melting of reactive metals involved protecting the molten metal from sources of nitrogen and oxygen.

A key consideration in reactive metal melting is the correct response to a breach in the water containment system that would allow water to contact the molten reactive metal in the furnace, creating a potentially dangerous situation.

Early melting techniques favored melting the metal in enclosed furnaces, which were purged to remove most of the air, then controlled at atmospheric pressure under the cover of inert gases. When the use of inert gases proved less than satisfying, the furnaces were redesigned to operate under high vacuum (<10 Pal). The predominant melting technique in the last 50 years has been vacuum-arc remelting (VAR), a consumable-electrode melting method where an electrode consisting of reactive metal of the appropriate composition is drip melted—at low voltage and extremely high current—into a cylindrical copper crucible. The copper crucible is cooled by a high volume of cold water sweeping past the outside surface.

Misalignment of the electrode, volatile off-gassing of input materials, incorrect arc gap and many other conditions may allow the arc (approximately 1 MW) to impinge directly on the crucible wall. Copper has a very high thermal conductivity, allowing the crucible to disperse large amounts of heat to the water jacket without damage to the crucible itself. However, the copper crucible will melt readily if the arc impinges directly on the crucible surface. A high-power arc can melt through the crucible wall in seconds. During VAR melting, a large vacuum pumping system maintains a high vacuum inside the copper crucible. Meanwhile, the cooling water system moves more than 757 liters per minute of water with a large hydrostatic head pressure past the outside of the crucible. Once the containment has been breached, the differential pressure will force crucible cooling water through the opening, into the crucible and into direct contact with the molten reactive metal surface.

All molten reactive metals oxidize rapidly and exothermically in the presence of water, stripping the oxygen atoms from the water to form metal oxides and hydrogen gas. At the same time the water quickly flashes to steam, greatly increasing the gas pressure within the enclosed system. During the full-power portion of a VAR melt, an ingot can contain as much as 0.28 cubic meters of molten metal at the ingot top. The hydrostatic cooling water pressure outside the crucible coupled with the low operating pressure within the crucible can bring large amounts of water in contact with the molten metal very quickly. Extremely rapid internal pressurization may then result—a combination of steam and hydrogen generated by the high-temperature metal/water chemical reaction.

If the source of water is small enough, the volume of gas created will not overwhelm the vacuum pumping system and the steam-hydrogen mixture will be exhausted from the pumping system. The hydrogen discharged to the air at the exhaust port will quickly disperse and drop below the lower explosion limit of 4 percent in air. Should the volume of water be very large, however, the ingot top will be quickly submerged in water, quenching the ingot surface and slowing down the steam and hydrogen generation. If the amount of water entering the system falls between the two extremes, the volume of steam/hydrogen gas generated will overwhelm the vacuum pumping system, pressurizing the furnace. Once the pressure inside the melting system exceeds ambient pressure, the furnace seals will fail, exhausting steam and hot hydrogen to the atmosphere and allowing air to enter the enclosed furnace volume. The air containing free oxygen will react violently with the hot hydrogen, furthering the explosion. This scenario has the greatest potential for causing a large amount of damage and possibly the loss of life.

Many crucibles have burned through over the years; most have not resulted in an explosion. Over the years, several spectacular melting furnace explosions have occurred, however, resulting in the deaths of operating personnel and millions of dollars in equipment and building damage. A great deal of effort has been expended in the VAR industry to reduce the opportunity for crucible burn-through and to limit other sources of water from reaching the molten metal. Melt shops are now laid out with the furnaces located in pits below floor level, some with steel bunkers built surrounding the furnace tops. Even so, water incursion (and potential explosions) will continue to happen in reactive metal melting.

The Burning of Reactive Metal Exposed to Water

If a VAR melt should breach the containment of the copper crucible, what is the appropriate emergency response? Observations from reactive metal fires can add insight to the reaction of water with molten metal at the top of a VAR-melted ingot in a crucible burn-through.

At Western Zirconium, a plant in the nuclear fuels business unit of Westinghouse Electric Co., the burning reactive metal fines has been observed for many years. These fines are typically band saw chips, grinding "swarf,” or sponge fines. The plant emergency response team has trained multiple times with burning reactive metal fines in an effort to find the most appropriate methods to extinguish these fires.

Adding water to reactive metal fines works only if the volume of the burning fines is small enough for the water to remove the heat faster than the exothermic reaction can create heat. Typically these fires are small and the ratio of surface area to volume is large. If the fines are constrained in a tight volume (such as a barrel), the heat generated cannot be rapidly removed by the water. As the water drops through the mass of fines, fresh hot surface is presented to the water. The hot reactive metal reacts with the oxygen in the water, providing the oxidizer to the burning metal and releasing hydrogen gas.

Western Zirconium has tried many types of water-based foams and found that the reactive metal pulls the oxygen from the water in the foam, as well. Fires continue as long as fresh, hot surface can be exposed to water. The burning stops once all the available surface has oxidized.

Response to Exposure of Molten Reactive Metal to Water

What is the proper response to a metal/water event to reduce the risk to personnel, equipment and buildings? A random sampling of reactive metal melters in the United States indicates that there are two entirely different response methods to water contacting molten reactive metals in a VAR furnace.

The most common response (given by four out of five melters in this limited survey) is to immediately terminate power while continuing to operate the vacuum pump system in an attempt to maintain as low a pressure as possible—for as long as the steam/hydrogen generation rate will allow. The intent of this response is to prevent the gases in the furnace from exceeding atmospheric pressure. If the furnace does not exceed atmospheric pressure, the furnace clamping system and seals will retain their integrity, preventing hydrogen from escaping the system and also preventing air from entering the furnace in an uncontrolled manner. If the pressure can be maintained below atmospheric pressure until the metal pool cools below the melting point and the solid surface is passivated, or made less reactive, by an oxide layer, the hydrogen evolved will be safely exhausted to the atmosphere from the furnace through the vacuum pump system.

The downside risk to this response is a pressure rise greater than the pumping system can handle. If the pressure inside the furnace significantly exceeds atmospheric pressure, the clamps will be forced open, breaching the furnace seals. The result will be a high-pressure exhaust of steam and hydrogen from the furnace accompanied by a possible influx of cold air into the enclosed volume. The oxygen in the air may rapidly react with the hot hydrogen in both spaces, resulting in a furnace explosion.

The alternate response (indicated by one melter out of five) is to immediately isolate the furnace from the pumping system and to rapidly backfill and purge the furnace with the heavier-than-air inert gas (argon). This response essentially guarantees that the furnace system will exceed atmospheric pressure as the steam and hydrogen are generated. The vacuum pumping system and pressure relief valve(s) attempt to maintain the furnace at near-atmospheric pressure.

Venting of the steam and hydrogen gases in this manner will result in any hydrogen burning taking place outside of the furnace containment. In addition, dilution of the steam-hydrogen with large-volumes of argon reduces the flammability of the mixture or duration of the hydrogen burn. The high partial pressure of argon in the enclosed volume may also prevent outside air from entering the furnace system and may prevent furnace explosions. The high internal pressure and temperature of the argon/steam mixture may allow cold outside air into the enclosed volume through a "chimney effect,” causing an explosion.

An additional difficulty with this approach is the volume of argon required. Cooling the ingot to the point where metal/water reaction stops may take several minutes. Therefore, sufficient argon feed must be available to maintain a blanket until H2 generation ceases, while being continuously exhausted by the pumping system. The vacuum system must also be designed to pump this large volume of gas over an extended period of time without overheating or damage.

Each of the two responses, described above, presents a number of positive and negative potential outcomes. While this article is not an attempt to justify either response, it should be a call to companies involved in reactive-metal melting to evaluate the appropriate responses to the inadvertent mixing of water and molten reactive metals. Each of the current responses may be completely appropriate, but each carries a certain amount of risk.

Perhaps an alternate approach, not yet developed, may be more appropriate. In an effort to better understand the mechanisms involved in molten reactive metal/water explosions, and to help develop the correct response for Western Zirconium, a detailed evaluation of the probable reaction thermodynamics was undertaken by the Westinghouse Science and Technology Department.

Thermodynamics of Steam/Hydrogen Explosions

A review of the past furnace explosion near-miss events in the reactive-metal melting industry finds no evidence of explosions involving zirconium or hafnium. Explosions appear to be limited exclusively to the melting of titanium. Significantly, much more titanium is and has been melted since the advent of VAR melting of reactive metals.

Due to the higher production rate, there has been more opportunity for furnace explosions to occur during the melting of titanium ingots. Zirconium and hafnium have been melted for many years, so while there has been more titanium melted, one would expect zirconium or hafnium melting to have experienced at least a few furnace events in the past.

The melting point of titanium is 1,671°C. Therefore, at the VAR furnace operating temperatures the preferred oxide that will form is Ti40. Tetratitanium heptoxide has a melting point of 1,677°C, only six degrees above the melting point of titanium. Reaction of molten titanium will therefore form a liquid phase oxide over the entire VAR operating temperature range from 1,677°C to nearly 2,200°C.

Since the melting points of zirconium oxide and hafnium oxide are much higher (2,677°C and 2,900°C, respectively), and no other oxides are formed at equilibrium, the reactions of molten zirconium and hafnium with steam may both form a solid oxide crust at the reaction surface. The solid reaction crust may inhibit further reaction by preventing or significantly reducing the contact of the steam with the underlying molten metal. By contrast, the reaction of the titanium with the steam will create an oxide, which is an easily disrupted liquid film at the reaction temperature. Therefore, little or no protective solid oxide layer forms until the surface temperature has been reduced below the titanium freezing point. The titanium/steam reaction will continuously break the protective layer of the liquid oxide exposing the molten metal to the steam. The heat of reaction will also continue to impart additional temperature to the molten surface.

Although the details of the actual interaction during a crucible breach are complex and not amenable to modeling, formation of a liquid oxide at the molten titanium temperature may significantly increase the extent of the metal-water reaction, the amount of hydrogen generated, and therefore the probability and consequences of a furnace explosion.

Water leaks into VAR furnaces have taken place since the advent of vacuum arc melting and will continue to occur. A thermodynamic investigation into the reactive metal/water reaction indicates that titanium is more likely to react explosively, due to the liquid-phase Ti0 oxide that forms at and above the titanium melting point.

While zirconium and hafnium are more likely to passivate during a crucible breach, steam and hydrogen will still be generated and therefore the threat of explosion cannot be ignored. Reactive-metal melters have developed their own individual responses to potential explosion threats. However, no industry wide response has been embraced. The appropriate response to molten reactive metal/water contact should be undertaken as an industry initiative.

Safe Design of Melting Systems for Titanium

Webster Dictionary defines an accident as an unfortunate event resulting from carelessness, unawareness, ignorance or unavoidable cause. It defines the term explosion as, to burn suddenly so that there is violent expansion of hot gases with great disruptive force and a loud noise, or a large scale rapid expansion, outbreak or upheaval.

These definitions are accurate in discussing any of the furnace explosions that have occurred in the melting of titanium. When conditions are exactly right, an accidental water leak into molten titanium will result in an explosion, which manifests itself as a rapid expansion of gases resulting in great disruptive force with loud noise and considerable outbreak and upheaval. As a side note, in the mid-1950, the U.S. Bureau of Mines was commissioned by the titanium/zirconium melting safety committee to set up a VAR furnace and duplicate the factors, which resulted in explosions. After several months of experimentation, not one explosion was experienced. They used a VAR furnace and introduced controlled amounts of water into the melt zone. They introduced the water at different points in the melt cycle—for example: early with a limited amount of molten titanium; midway through the melt; and at the end of the melt with the maximum amount of molten metal. We are certain that they had some steam problems, but not one explosion was experienced.

At some point, every titanium melter has experienced a water leak into an operating VAR or non-consumable furnace where no explosion has occurred. At other times, with what appear to be the same operating conditions, we have a significant explosion. In all of the reported accidents where significant damage occurred, there have been not one but two or more distinct and separate explosions.

Years ago, during an explosion at a TIMET facility, two operators were looking directly at the furnace when two explosions occurred. They reported that the first explosion produced a red/orange colored flame. The second explosion was a blue/white flame with a louder noise.  TIMET’s E.D. Dilling obtained the heats of reaction for the following equations from Dr. K. K. Kelley of the U.S. Bureau of Mines in 1957.

The theory is that the first explosion is from superheated steam and the second is a hydrogen explosion resulting from the molten titanium absorbing the oxygen from the water and releasing hydrogen. When the steam explosion occurs it expels all of the gas from the furnace cavity and creates a vacuum in the furnace chamber, which is replaced with a hydrogen-oxygen mixture. This mixture of hydrogen and oxygen combines exothermically resulting in the second more violent explosion. Operators and engineers alike have reported accidents where the steam explosion has lifted the furnace straight up and released the steam pressure, then the furnace dropped back into place and resealed the system and no secondary explosion occurred.

The following set of circumstances is common to each of the significant furnace explosions in the titanium industry:

  1. All reports indicate two or more distinct explosions.
  2. It is interesting to note that not all water leaks into molten titanium result in explosion. Conditions must be exactly right for an explosion to occur. The researchers at the U.S. Bureau of Mines spent months trying to produce an explosion in a VAR furnace and did not succeed. Explosion can occur with a small amount of molten titanium or a large amount; that is at the start of the melt or any time during the melt.
  3. The severity of the resultant explosions may not be proportional to the quantity of molten titanium.
  4. The conditions required for an explosion are difficult to reproduce or define.

If we accept as a given that, in melting of titanium in water-cooled crucibles, violent explosions are possible and unpredictable, then we must design for safe operation even with an explosion.


The maturing of titanium from a laboratory curiosity to a production operation occurred in about 1950. The pioneers in the industry were joint ventures between titanium pigment manufacturers and stainless steel melters. The pigment companies had developed the required technology to produce pure TiCl4 Dr. Kroll and others plus the U.S. Bureau of Mines had developed the technology required to reduce the TiCl4 to titanium metal.

The steel producers had developed the vacuum arc melting technology required to convert the raw titanium sponge into solid titanium ingot and the process required to convert the ingot metal to usable form such as plate, sheet and bar.

Early Years of the Titanium Industry

In 1955  four melters and processors of titanium metal  dominated the market—Titanium Metals Corp. of America, Rem-Cru Titanium, Inc., Mallory-Sharon Titanium Corp. and Republic Steel Corp. Two newcomers started operations in 1956 : Oregon Metallurgical Corp. at Albany, OR, initiated an ingot melting facility, and Harvey Machine Co., Inc., Torrance, CA, installed VAR furnaces to melt extrusion rounds.

For the sponge metal, the major producer was Titanium Metals Corp. of America; others arranged in probable order of output were E.I. DuPont, Electro Metallurgical Co., Cramet Inc., and Dow Chemical Co. Sizeable expansion plans were announced in 1956 by Titanium Metals Corp. of America and DuPont.

The available technology from the steel industry was inert gas melting or vacuum arc melting. During the early 1950’s it was catastrophically discovered that the steel melting practice had to be modified for titanium. The positive pressure argon melting produced an extremely rough melt. On June 11, 1954 a Mallory-Sharon furnace, melting titanium in an argon atmosphere, had an arc-through with the resultant steam explosion followed by a hydrogen explosion. The final result was major equipment and structural damage with four people killed and five severely injured, all from burns.

Following this experience, all titanium melting was converted to vacuum arc type. The titanium-zirconium melting-safety committee was formed. The committee was to share safety related information on the melting of titanium and/or zirconium. The committee members were Mallory-Sharon, Republic, Titanium Metals Corp. of America and Rem-Cru.

Here is a list of recommendations from 1957:

  1. Melting should be done at less than 1000 microns.
  2. Melt with as short an arc as possible.
  3. Electrode to crucible wall clearance should be greater than 2 inches.
  4. High frequency arc starters should be used.
  5. The interior walls of crucibles should be buffed and polished.
  6. Melting Practice: Reverse polarity• (i.e., electrode positive) was practiced by Mallory-Sharon. All others used the straight polarity or electrode negative approach.
  7. Cooling mediums other than water were tested.
  • Heavy walled copper crucibles, i.e., 2 ft thick, were tested.
  • Water spray instead of submersion was tested.
  • Non-hydrogen cooling medians were tested.
  • NaK cooling was evaluated.

The problems experienced in titanium melting were unique and different than those in steel. The major difference is the secondary hydrogen explosion. With steel the steam explosion occurred as a result of thermal expansion with some damage, but nothing to compare with the damage experienced in the titanium industry where the secondary explosion was a detonation type caused by oxygen-hydrogen combination. The difference, of course, is that the steel does not break down the water to free the hydrogen. Without the hydrogen the detonation does not occur.

The major explosions in the titanium industry occurred prior to 1958. By 1958 the melting controls had been developed for melting using short arc conditions, this minimized crucible arcing. In addition, all furnaces were equipped with magnetic focusing coils, which prevent the arc from attaching to the crucible wall and melting through to the water.

By 1958 the titanium and/or zirconium melters realized that even with the best of controls, explosions are possible. When they do occur the forces developed are very large. In general, the philosophy of the titanium industry has been that we do everything we can to prevent explosions but at the same time we design our systems to protect people and equipment if one should occur. The major melt shops are constructed with the furnaces in bunkers. The bunkers are designed to withstand the forces of the explosion and to direct the force in a controlled direction. Video cameras and optics have been developed which allow the furnace operators to be isolated in control rooms, away from the furnaces and out of harms way.

Even though we have discussed a number of melting related explosions, the titanium industry has a very good overall record. Only three explosions have resulted in fatalities and all of them were during the development stage of the industry, prior to 1957.  There are significant factors that must be taken into account when designing a titanium melting facility. It must be recognized that explosions are possible and people and equipment must be protected from the forces involved.

Everyone must be aware of the potential hazards involved and work to eliminate problem areas in their system. There are major concerns with the installation of the new Plasma or cold-hearth systems where potential hazards exist. Across the board, systems are being installed and the operators sitting close to the furnace at the melt level and significant amounts of molten metal with a potential for water leaks. The designers and owners of these systems should evaluate their system based on the idea that they could explode.

List of Recommended VAR Furnace Controls and Equipment Design

  1. Loss of water flow shuts off the furnace power. This feature is checked prior to start on each melt. One rectifier is turned on, the water is turned off which turns off rectifier.
  2. Effluent water temperature monitor system. If the effluent water from the furnace has more than a 20° temperature rise over the input, the-system should be checked.
  3. Minimum of two redundant water supplies to each furnace. If one fails, system automatically switches to other. Ideally, one system is a gravity flow system with at least 60 feet of head.
  4. 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. That is, at 14 inches of mercury absolute or below.
  6. Explosion ports 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 into the control system should be incorporated 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 equipment with video cameras to maintain the electrode in the center of the crucible.
  16. NaK coaling of crucibles is a desired alternate to water-cooling.
  17. Positive type bottom supports for crucibles, which are capable of holding the bottom in place if fasteners are arc damaged.

Pros and Cons of NaK Cooling

NaK was originally developed by the U.S. Navy for use as a heat-transfer medium on nuclear submarines. The system performed as designed and satisfied the needs of the Navy. The Navy has since quit using NaK on submarines, not because it did not perform as designed but because it reacts violently with water; not a good thing to have on a submarine under attack.

NaK was first used as a cooling medium for titanium melting at Titanium Metals Corp. of America in 1957. A test system was installed and operated for several months to develop experience and operating data. When the new research and development melt shop was constructed in 1958 all of the experimental furnaces were installed with NaK cooling. This system has operated continuously for 33 years with a minimum of problems. In 1985 IMI in England converted a portion of their production melt shop to NaK cooling.

NaK is a silver gray liquid metal similar to mercury in appearance. It is an alloy of approximately 50 percent sodium and 50 percent potassium. The melting point of the typical alloy is below 25°F. It has a heat capacity of approximately one fourth that of water. When a NaK leak does occur and NaK leaks into or onto molten titanium there is no violent. reaction. In addition, there is no hydrogen release; consequently, there is no secondary explosion.

NaK cooling is accomplished using a totally self-contained sealed system. NaK reacts with the moisture in the atmosphere and forms oxides of both sodium and potassium. The ideal system uses NaK to air heat exchanger to remove the heat from the melting system, NaK to water systems have been used but they are not recommended due to the violent reaction of NaK and water. NaK is pumped through the system using totally enclosed pumps.

Furnace Design
Since the NaK in the system cannot be exposed to the atmosphere, each and every component must be self-contained in the NaK circuit. Typically, the cooling circuits for a given furnace are connected in series. The NaK is fed in through the bottom, then up through the annulus between the copper crucible and the stainless steel containment jacket. The ram and furnace parts are on a separate circuit. The NaK to air heat exchanger is normally in a remote location.

Drawbacks to NaK
The major drawback to a NaK system is the violent reaction between NaK and water. Properly designed systems keep NaK and water separated. Handling of NaK is hazardous because it will react with the moisture in the skin and result in serious burns. The main point is that with properly designed systems and well-trained people NaK is a safe cooling medium for titanium melting, If you are looking at a greenfield site for a melt shop, NaK cooling should at least be evaluated.   One major rule of safety is, never design a system or operate a system where NaK and water can combine to produce an explosion. The suppliers of NaK will provide all of the necessary measures for a safe operation.

Liquid Metal Cooling for Consutrode Melting
Experiments by Titanium Metals Corp. of America over a two-year period indicate that NaK cooling is both practical and safe for consumable-electrode arc furnaces.

Explosion hazard has always been a problem to melters of reactive metals.  Development of a liquid metal-cooling system, which utilizes a sodium-potassium (NaK) alloy indicates that this may become a thing of the past. The NaK alloy, which has fluidity similar to water at room temperature, has proven safe and reliable in almost two years of operation at the Henderson, NV, research melt shop of Titanium Metals Corp. of America, which developed the process.

The principal advantage of the NaK system is the elimination of the explosion hazard, thus avoiding the high cost and relative inconvenience of explosion-resistant vaults. System evaluation in routine melting of several hundred ingots up to 18 inches in diameter shows maintenance cost is minor; melted material is salvageable on crucible failure (any wetting by water of a hot ingot of reactive metal in conventional systems results in total loss of the ingot); repair and alteration can be done by normal plant maintenance groups; crucibles can be interchanged by means of specially designed valves; potential safety problems engendered by NaK’s own reactivity are overcome through proper system design; and operating procedures are similar to those for conventional systems.

Cooling System Research
Consumable electrode melting of titanium has been the subject of intensive research at Titanium Metals Corp. with concentrated study directed toward methods for cooling the crucible. Water, of course. became the first candidate for this cooling job. A water-cooling system is deceptively simple in appearance. Actually, in operation, many details of the system turn out to be extremely critical because: the electric arc represents an extremely concentrated form of energy, which operates at extremely high temperatures, making the possibility of a perforated crucible ever present; and many molten metals  such as titanium may react violently on contact with water.

To minimize the chance of crucible failure, adequate water flow must be maintained to every square inch of the crucible during melting. This means that the entrance of water, obstructing flanges, and spaces for steam pockets to develop are carefully considered. Failure-save instrumentation is provided to assure a steady flow of water for cooling in the event the main system fails, since a large quantity of sensible heat remains in the ingot upon termination of melting power. And finally, since all means of protecting against the possibility of a crucible arc-through are subject to human and mechanical error, protection is provided against crucible failure and the potentially hazardous explosion itself by use of protective vaults.

The use of a liquid-metal coolant, replacing water, would eliminate  the potential hazard. A preliminary apparatus was set up for testing NaK cooling. It consisted of a 12-inch crucible especially designed for NaK; an electromagnetic pump; an electromagnetic flow meter; a water-cooled heat exchanger; and other necessary accessories. These were arranged in a closed loop system. On the basis of four successful 350-lb heats melted at a maximum of 7000 amp, NaK cooling was considered to be adaptable to the furnace cooling problem.

The decision was then made to enlarge and perfect the system to include several furnaces in the Titanium Metals Corp.’s Melting Research division. This was based on the objective of piloting a larger-scale system and of providing a safe system for melting research, which would eliminate the need for vault protection, since vaults would limit the value of these furnaces for research purposes. Both of these objectives have been accomplished.

All controls for operating the NaK system and furnaces were placed on a control panel near the furnaces to facilitate study of furnace operation. To date, several hundred melts, ingots ranging in size from 5 to 18-inches in diameter, have been made using NaK cooling.

The air-cooled heat exchanger was found satisfactory for its design limit. But when shop facilities  were expanded to utilize more melting power and make larger ingots, it was necessary to increase the cooling capacity. This was done by the arbitrary choice of the water-cooled heat exchanger in addition to the air-cooled one. This system works satisfactorily, but water-cooled heat exchangers would not be recommended for plant installations. The air-cooled heat exchanger alone is used for melts up to 12 inches diameter. Air-cooled and water-cooled heat exchangers are used in series for larger melts,

Safety Considerations
The NaK piping system was constructed with all welded joins as far as possible to avoid leakage, which might develop in gasketed joints. Welds that may appear sound might develop leaks in time because of the attack of NaK on slag inclusions.

The main object of using NaK as the heat-exchange medium in titanium melting is to avoid furnace explosions upon crucible arc-through. A crucible arch-through is defined as a perforation in the crucible wall caused by the electric arc. A perforation would allow the coolant fluid to enter the melt zone, where a sizable quantity of liquid titanium is present along with the highly concentrated arc heat source. Several explosions throughout the industry, resulting from such crucible arc-throughs using a water cooling, demonstrate the potential hazard of that cooling method.

Liquid metal should not cause such an explosion upon a crucible arc-through, since no chemical reaction is believed possible between the liquid metal of the coolant and the liquid titanium. However, a pressure rise might be expected since the vapor pressure of NaK at the melting point of titanium is quite high, approximately 20 times atmospheric pressure, It must be remembered, on the other hand, that the pressure in a vessel containing vapors is controlled by the vapor pressure at the coldest point in the vessel. The vapor pressure of NaK at 441°F is of the order of 1.5 XXX. Therefore, under normal operating conditions, when the cooled portions of the furnace are 400°F or less, there can be no appreciable pressure rise.

Cost of a NaK System

Cost comparisons with water systems are difficult because of the multitude of ways either system can be designed. An examination of water usage on a water-cooled system, which does not recirculate the water shows that, where water costs are high, cost of the NaK system can be repaid in a few years by water savings alone. Where a recirculation system is used, the NaK system may appear favorable from the standpoint of the original investment and maintenance or operating costs.

Installation cost of a NaK system is comparable to a water system when the cost of vaults is included. The danger of loss of equipment and valuable melting stock is always present with water cooling. The NaK system eliminates such danger, and should prove to be more economical in the long run. As an example, the loss of a 2000-lb ingot of columbium—even if it was just wetted without an explosion—would pay for the NaK system. Therefore, when a new melting facility is planned, it may be wise to consider the NaK system from the standpoint of economics. in addition to the improved safety features.

Melting Safety: From a Directed Water Flow Perspective
Contributors:  Ross Snyder, Project Engineer and Mike Dagle, Sales Manager Titanium Safety Committee Meeting: Feb. 21, 2002 (RMI and Titanium)

Zak Inc., Green Island, NY, has nearly 50 years of experience in designing, fabricating and repairing the water-cooled copper components for the metal melting industry. Our experience includes the manufacture and repair of crucibles, melting hearths, transfer hearths, molds, pullers, pour lips and a host of accessory components. Our experiences in the techniques of electron beam (EB), plasma arc melting (PAM) and electroslag remelting (ESR) have proven beyond a doubt that directed "high-velocity” water flow, from both a safety and economic standpoint, is by far the most advantageous method of cooling copper.

In 1990, we talked about "Proper Copper,” various alloys and their applications. We also indicated that 10 feet per second water velocity was providing excellent service life in EB applications. In 1993, we talked about our experiences and observations associated with VAR crucible repair, suggesting that high velocity water flow held promise for both VAR base-plate and stool applications. In 1995, we talked about heat transfer in VAR applications and provided updates on the continued success experienced by EB/plasma processors utilizing high velocity water flow. In the fall of 1998 we talked about directed water flow in ESR and went one step further, suggesting that the time had come for directed water flow designs for VAR ingot melting crucibles. In 1999, we said that it was "Time to Take Controlled Water Cooling of VAR Crucibles Seriously.” In 2000, we analyzed the "Safety Features of a Drilled Water Passage Cooled VAR Crucible Design” and presented a paper entitled "Drilled Water Passage Cooling of Copper—The Right Stuff and a Step in the Right Direction for Safety.”

In 2001, we presented a paper entitled "Proper Cooling of Copper” at the International Symposium on Liquid Metal Processing and Casting. We showed the progression of copper cooling methods across the range of EB, ESR, PAM and VAR melting. We demonstrated that the industries that have adopted cooling methods that ensure the 10 feet per second cooling water velocity and the principles of directed water flow cooling have realized significant economic benefits.

Throughout the past decade we have discussed the safety and economic advantages of directed water flow designs. Unfortunately, we’ve also heard others speak at several of these meetings about recent VAR events that have included explosions, facility damage, equipment downtime, lost time injuries, permanent disability and financial considerations associated with productivity losses. If memory serves us correctly, many if not all of the explosions experienced with VAR melting have been directly associated with the use of conventional water wall cooled crucibles.

Within the context of this paper and over a time span now approaching two decades, however, we are not aware of one EB, ESR or plasma processor, utilizing drilled water passage cooling, who has experienced a catastrophic processing event (explosion). In fact, many if not all EB, ESR and plasma processors have taken advantage of the beneficial derivatives associated with directed water flow cooling.

Directed water flow, when properly implemented, has proven itself from both an economic and a safety perspective. Water wall cooling in the conventional VAR application does not insure 10 ft per second cooling water velocity and it does not insure directed and controlled water flow. If the liner fails, water-wall cooling has the potential to introduce large volumes of water into the melt with the expected results. The solution is to adopt a directed water flow philosophy. Based on studies performed by us and based on our observations of new and repaired equipment we can say that there are three main considerations to insure that copper is properly cooled.

The first requirement is the minimum velocity of the cooling water. Independent engineering studies and our experience shows us that 10 ft per second is necessary to insure the optimal conditions for copper survival. Anything less allows vapor bubbles to coalesce while anything more will nullify heat of vaporization efficiency. The second requirement is directed and controlled flow of the water. This is generally accomplished through three methods: water-wall cooling, drilled-passage cooling and finned cooling. The third requirement is the use of conditioned water at a uniform inlet temperature.

Studies and experience have shown that deposits formed on a copper surface can raise the operating temperature of the copper by 50°F or more. Anything that raises the operating temperature of the copper shortens its service life (distortion, creep, recrystallization and grain growth).

  • Why is directed water flow better than water wall cooling? From an equipment longevity standpoint it is superior for the following reasons:
  • In order to accommodate the cooling holes, the wall thickness of the copper is greater than that in water wall design. The end result is a stiffer wall with reduced creep deformation characteristics.
  • The cold face to hot face ratio is also greater than that associated with a water wall design. The end result is more cold face constraint, which further reduces creep deformation.
  • The outer periphery of the copper can be reinforced with a stronger metal such as stainless steel. This drastically reduces creep deformation. We use stainless steel welded to the copper for non-magnetic purposes and also because of their similar coefficients of thermal expansion.
  • Being an all welded structure, where the crucible or mold is the jacket, there are no mechanical seals to create leak problems. The only seals are on the pipe fittings of the water supply lines and the vacuum seals of the crucible.
  • By using a combination of parallel and series water-flow circuits that are incorporated within the vessel, water volumes and therefore pumping requirements are significantly less than in a water wall system.
  • Water pressure in the cooling holes, that should not have to exceed 60 PSI, is easily handled within the hole. This negates the unit becoming a pressure vessel as it does in water wall cooling. Therefore, hydrostatic stresses on the vessel are not a concern.
Many of these same factors relate to operational equipment safety advantages, as well. Several years ago Zak Inc. performed a study where a hypothetical 030 x 150-inch crucible was subjected to two methods of failure. The first failure mode studied was what would happen if the melt were initiated with no water flow. The second failure mode studied the effects of penetrating the copper liner with an arc during the melt, assuming nominal water flow conditions.

Two different crucible designs were compared. The first design was a conventional liner/jacket water wall cooled crucible; the second design was a drilled water passage crucible. This analysis found the drilled passage crucible to be far superior and safer than that associated with a conventional design.

For the water wall cooled VAR crucible, a melt started with no water flowing would cause a massive failure of the liner around the area of the melt level. Up to 350 gallons of water would be rapidly introduced into the melt. The resultant steam and hydrogen explosion would be massive, an unfortunate occurrence that has been experienced too many times, by various melt shops. Under the same no-flow conditions, the drilled passage design performs much better. The greater wall thickness of a drilled passage design acts as a heat sink, therefore, the time between melt initiation and liner failure in increased. While the copper is heating up, the residual water contained in a drilled passage will be percolated out and/or vaporized to steam so that when the liner does fail there is little or no water left to be introduced into the melt. Therefore, without the presence of water, no explosion will result.

In the second scenario where the melt is started and an arc penetrates the liner, the results are similar. A conventional liner/jacket design can have up to 800 GPM (gallons per minute??) of water available. As an arc penetrates the liner, the size of the arc hole will determine the amount of water that enters the melt. For a 1-inch hole it could be 200 GPM for a 1.5-inch hole it could be up to 500 GPM if the pumps are capable. The results are predictable, a large steam explosion followed by a large hydrogen explosion. A drilled water passage design, however, will perform much better in the same situation.

Based on our experiences with repairing arc damage on conventional VAR crucibles, we believe that an arc would only expose one drilled passage. Assuming it exposed the drilled passage in such a way that it were fully open then 14 GPM of water would be introduced into the melt. The resultant explosion would therefore be about 1/25th the intensity of a conventional unit. It is also possible that the arc damage could occur between holes. Such a failure would introduce no water into the melt and would result in no explosion. Based on our experience with drilled passage cooling in EB, Plasma and non-consumable arc melting we believe that only 1\4-inch holes might be opened into a drilled passage. This would introduce only about 1.5 GPM into the melt, resulting in an explosion of even smaller intensity. The vacuum pumps would draw off the pressure and the pressure rise sensors would immediately terminate the melt.

EB, Plasma and ESR processors have used water wall cooling in the past but in many cases found it to be an ineffective cooling method. Many have moved towards directed water flow designs and have realized not only the economic benefits of increased equipment service life, but also the inherent safety factors associated with those designs. We firmly believe that VAR processors who would adopt directed water flow technology would see long term economic benefits as well as significant increases in operating safety.

Recently a large drilled water passage ESR slab mold was returned to Zak Inc. for repair. This slab mold had been in service for several years and had seen several hundred heats. However, it had also undergone a very large amount of deformation.

As mentioned before, our experience has indicated that drilled passage equipment when properly designed resists large amounts of deformation similar to what we were seeing in this mold. A brief investigation revealed that the mold had been piped incorrectly ever since it was installed. Instead of cooling water flowing at 10 ft per second, one of the wide side plates had simply tilled with water and remained stagnant. The melt shop even reported boiling of the cooling water. The conditions being experienced by this mold were well outside any of the design considerations, but one important fact was clear: it did not leak.

After years of service and hundred’s of heats, this mold continued to safely produce acceptable product. Although we noted a gross amount of deformation at time of receipt, it remained watertight. This is due to the superior design of the drilled passage concept. In the past, we have built similar ESR slab molds that were cooled with the water-wall concept. Water wall designs are prone to large amounts of deformation and leakage even under normal operating conditions. The thicker wall of the drilled passage design, with all welded construction, combined with the integral stainless steel support structure all worked together to keep this mold from leaking and failing catastrophically. If this mold had utilized the less effective water-wall construction it would not have lasted as long as it did. Severe deformation would have taken place, which in turn would have broken the mechanical water seals causing water leakage. Despite the severe deformation of the mold, we were able to repair it successfully, a true testament to the durability and safety of a directed water flow design.

In summary, Zak Inc. has conducted studies that indicate that drilled passage cooling not only offers significant economic benefits it the form of equipment durability and longevity but also offers significant safety benefits. Now is the time to embrace the economic and safety benefits of directed water flow cooling in VAR applications.

Safe Operating Precautions for Plasma Melting Systems
Contributors:  David Warren, Retech, a division of Lockheed Martin, 1979; San Francisco, CA

While operating a plasma system, good safety precautions for foundry processes should be followed. Safety concerns are trip hazards; lifting and rigging; crane operations; steam and hot surfaces; falls; confined spaces; eye protection; arc light; dust grinding; high-pressure hyd and water; heavy moving equipment; fork truck operations, burning and grinding. Safety programs should be implemented to address these concerns.

Although asphyxiation is a concern in titanium processes where inert gas is used, it is more so for plasma systems. Helium and argon are used for plasma system operations. Do to the design of the chambers, feed systems, ingot cans and the volume of these gases, being used, extra precautions should be practiced. Employees are required to enter chambers for routine maintenance on a frequent basis.

Where argon is used, pits should be equipped with oxygen monitors. A confined space program should be established to OSHA guidelines. Where helium is used, personnel should be aware that the gas rises to the top of chambers. It may appear to be safe at the bottom but you wouldn’t want to stick you head in the top.

Finely divided metal dust can be highly pyrophoric. It has always been a safety concern in the titanium industry. Metal dust is much more of a concern in a plasma system do to the type of feed materials and the way the material is fed into the system. Dust is more apt to accumulate in the system than in VAR furnaces. It will tend to collect in feed chutes vacuum pipes, chamber view ports ledges and other penetrating ports.

A routine cleaning program should be established to clean these areas to minimize the accumulation of metal dust. Employees should be trained in how to prevent flare ups and wear fire-retardant clothing when cleaning. They should not enter a chamber without a burn out of the system. They should follow good confined space practices.

Unlike the VAR furnaces that operate at 40 to 50 volts, where the chances of a shock is very low, the plasma furnace operates at 300 to 400 volts with open circuit voltage in the 1000 volt range. The operators should not have access to the top of the furnace while the power is on. All bus bars should be insulated, and the operators should be aware of the hazard. Plasma is much safer that EB in this area as EB operates in the 25k range and higher.

What is needed for an explosion to occur?

In order to have an explosion. three things are required: fuel (in this case hydrogen); oxygen (air); and an ignition source. This maybe from molten metal, chlorides or iron oxides that have formed as a finely divided dust on the chamber walls. This dust maybe ignited when air has entered into the chamber.

Water and molten titanium react to form hydrogen. For every mole of titanium you will combine two oxygen. Two water 2H20 in molten titanium will form four hydrogen 2H2.  Water enters a molten pool of Ti. The worst condition is when this happens under a molten pool (a leak in the crucible). There is a less chance of having an explosion when the water is dumped on top of the pool (leak in ram). However, the potential for explosion is still there.

Hydrogen stratifies at the top of the furnace. Steam is generated and over pressures the relief port, the port blows off and air enters the chamber. The air mixes with the hydrogen at the top of the chamber. An ignition is generated either by hot metal or with a slow burn of chlorides or ion oxide dust that has formed on the chamber walls. When the ignition source reaches the hydrogen, an explosion occurs.

Safety in the Skull-Melting Plant for Production of Titanium Slabs and Ingots
Contributors:  By Filippo Oreglia, plant manager of a skull-melting furnace, Titania SpA May 10, 2006 Houston, TX

Titania is the owner of a titanium skull melting plant (started in June 2001) dedicated to the production of titanium, titanium alloy, and zirconium slabs and ingots. To date the plant has performed about 1,100 melts. The plant can melt 100 percent of titanium scrap in different shapes (crops, plates, tubes, electrode ends, billet portions, chips), which is charged into the copper crucible and melted by means of a titanium consumable electrode called "skull,” which is generated in a previous melt.

The skull is extracted from the crucible to be used in the next melting run as an electrode. In our skull-melting plant it is possible to produce "as-cast” material (ingots and slabs) with a weight of 1,600 Kg (about 3,500 lbs) or slightly more. The scrap charged into the furnace mostly results from our domestic production (plate, sheet, tube, forged pieces), and, for the rest, from our sister company Deutsche Titan GmbH in Essen, Germany. If necessary, it is purchased from the market.

Scrap is processed in our plant (spectro-controlled for the sorting of materials, cut, sand blasted, pickled) and is melted by an electric arc under dynamic vacuum atmosphere.  When the electrode is completely consumed and the scrap is melted, it is possible to cast the liquid metal inside the uncooled mold.  During the melting process, vacuum is attained by two vacuum skids, roots and mechanical pumps, to permit extraction of about 18,000 cubic meter per hour of gas (hydrogen). Before pumping, a pre-vacuum up to 60 mbar is created by means of a liquid ring pump to allow the process pump to activate.

Some of the plant parts such as the crucible, the ram, the body of the chamber, and the cables are water-cooled. The total quantity of water is about 150 cubic meter per hour. The cooling system is divided into two branches, one being closed-circuit and the other open-circuit for crucible, chamber, and casting chamber. Each branch of the cooling system is controlled by flow meters and thermometers connected to the programmable logic controller (PLC).

During melting, the operators in the control room can supervise from a monitor the changes of the process parameters in the plant. Valves, for both the water and the vacuum circuits, are motorized by air and by an electrical device during cooling, pumping of vacuum, or when stopping the plant. However, if, for any reason, a cut-off of the air occurs, immediately a switch activates an argon flow inside the air line to drive the valves and prevent them from closing.

The furnace is positioned in a bunker with walls 500 mm (20 inches) in thickness, built in reinforced-concrete. The door of the bunker is built in carbon steel (thickness about 250 mm (10 inches) and is closed during the melt. Over the roof of the bunker (where the furnace is located) a tank with 10,000 liters (2,650 gallons) of water is installed. This quantity can ensure the cooling of the crucible in case of any problem in the cooling system or if the power energy is cut off. Moreover, a safety power electrical station by gas/oil-engine is immediately activated (after 8 seconds ca.) to keep the vacuum pump and water pump working. This emergency power system is not able to continue the melt.

On the top of the bunker there is a "vent” to permit evacuation of heavy materials  in case of explosion. The copper crucible is monitored by twenty-one thermocouples that are positioned within its walls. The operators (the melter and his assistant) can control all temperatures at any time during the melting, so that the operators can adjust the power of the system to obtain a proper quantity of liquid metal and at the same time to prevent over heating of the crucible.

One of the aims of this process is to generate a solid skull inside the crucible to protect the side wall and bottom. Actually, the melting takes place inside the solid titanium skull that is generated during the melt when the liquid metal comes into contact with the scrap and immediately freezes due to the cool crucible walls. Thus "another crucible” containing molten titanium is created.

At the end of the melting process an amount of 1600-1700 kg (3500-3700 lbs) of liquid metal remains in the crucible and is ready for casting. The control-room, located not far from the bunker of the furnace, hosts all the devices necessary to drive the melt and to control process parameters. The melting process itself is controlled by PLC and the operator can adjust the values of the power during the melt. Voltage is automatically obtained by the values set in the PLC and depends on the metal melted.

In any case, if the sequence automatically set is not correct or if necessary to reduce the power of the arc (to prevent the overheating of parts of the crucible or other parts), it is possible to change to manual mode during the melt. It is possible to check the consumption of the electrode and the production of liquid metal by means of two video cameras positioned on the roof of the furnace. Another camera is installed over the casting chamber in order to enable the operator control the casting operation after the tilting of the crucible when filling the mold.

The duration of the cooling of the product (ingot or slab) and of the skull is about three hours. During this time an argon/helium atmosphere is maintained at 40 mbar to prevent the oxidation of the material inside the furnace. Before opening the furnace, it is necessary to wash the furnace three times with an air flow (up to 0.350 mbar) and the air must be extracted by liquid ring pump. This operation allows to prevent firing of the powder before the final opening.

During extraction of the skull, each operation inside the crucible is carried out with special copper tools to prevent the risk of fire by generation of sparks. Generally the temperature of the skull extracted from the crucible is about 350° to 400° C.

Safety Philosophy in Titanium Melting Systems
Contributor:  Matt Mede, Retech, A Division of LESAT Oct. 16, 1995

When melting reactive metals we must consider the safety hazards, which can potentially arise. A cooling water leak can result in water coming in contact with molten metal. This creates two separate phenomena. First, steam is generated which will result in over pressure in the chamber. The Retech designed relief valve will open and reseal. Second, hydrogen can be generated when the reactive metal getters the oxygen, leaving explosive hydrogen.

The conditions that lead to an explosion include the combination of hydrogen and air in the presence of hot metal. In order to avoid this, we design our furnaces to relieve the steam pressure then reseal the chamber. We also recommend opening the vacuum valve fully and pumping on the chamber to remove evolved hydrogen outside of the room.

Steps Leading to Dangerous Conditions in a Titanium Melting System

First Step—

Water leak occurs, caused by and incorrect assembly or maintenance of water-cooled portions of the system; an incorrect application of the electric current, side-arcing with VAR; or plasma melting through a water-cooled hearth.

Water hits hot metal and steam is generated, creating a pressure rise in the melting chamber. Some of this steam will condense on the chamber walls. Some of this steam will be pumped away by the vacuum system. The rest of the steam raises the chamber pressure. If the pressure rise exceeds the working limit of the melt chamber, then the gas should be relieved in a controlled manner to prevent damage to the vessel.

After this pressure is permitted to escape, the vessel should be resealed. Retech uses a spring-loaded relief port. This device performs the very important function of relieving pressure and resealing to prevent air from entering the vessel.

Second Step—

Water hits hot metal and steam is generated, creating a pressure rise in the melting chamber. Some of this steam will condense on the chamber walls; some of this steam will be pumped away by the vacuum system. The rest raises chamber pressure.

If the pressure rise exceeds the working limit of the melt chamber, then the gas should be relieved in a controlled manner to prevent damage to the vessel. After this pressure is permitted to escape, the vessel should be resealed. Retech uses a spring-loaded relief port. This device performs the very important function of relieving pressure and resealing to prevent air from entering the vessel.

Third Step—

Water and molten titanium react to form hydrogen Ti + 2H20 --7 Ti02 + 2H2. This is the point when hydrogen can come in contact with air and hot metal or other ignition sources to create a large explosion. To prevent the catastrophic third step from occurring, Retech recommends the following:

As soon as the operator confirms that there is a water leak, all vacuum valves are to be fully opened and all vacuum pumps should be used to the maximum extent possible. Depending on the process, the vacuum valves and vacuum pumps may already be on. If the process involves an inert gas back fill, this should be defeated and the valves should be opened. Pumping should be continued until the metal is fully cooled before the chamber door is opened. This can take hours to days.

Careful location of relief ports can reduce explosion potential. Hydrogen being the lightest of gases will stratify to the top of a chamber; therefore, we locate relief ports at the top of vessels to vent hydrogen away from ignition sources.

Retech’s philosophy is not to back-fill with inert gas when a confirmed water leak occurs. This will add gas volume to a vessel, which is already in danger of being over pressurized. If the vessel is over-pressurized beyond its working limit; an uncontrolled condition results, possibly causing doors, lids, and feeders to fly through the facility, perhaps hitting personnel.

Alternate Second Step
The course of events may be different under some conditions. When water is injected below the surface of molten metal, steam is formed. This cools the surrounding metal and may form a containing skin. As the steam continues to expand, this creates high pressure and great stored pressure until the containment yields in a powerful release of the stored energy, which is almost certain to do major damage.

Titanium melting systems  (listed in order of potential danger):
  1. VAR
  2. Consumable casting furnace
  3. Rototrode melting
  4. Cold-wall induction
  5. Plasma hearth furnace
  6. EB hearth furnaces
We consider VAR to have the highest potential danger since there arc large volumes of molten metal and large volumes of cooling water involved. In addition VAR has the greatest potential for mechanical damage (i.e., heavy electrodes falling, large hardware and scaling surfaces involved).

Rototrode melting and consumable casting arc probably equally hazardous since they involve similar amounts of molten metal. Rototrode has less potential for mechanical damage but has a high flow of cooling water in close proximity to the molten pool. This water can be injected into the molten bath if the rotating electrode tip is lost. Consumable casting furnaces have electrode damage potential and skull shrinking, which can damage the crucible.

Cold-wall induction (also known as cold crucible, induction skull melting) and hearth melting are inherently safer, but are not inherently safe. It’s possible to have a titanium explosion any time you have molten titanium and water.

Cold-wall induction and hearth furnaces have a protective skull layer between the molten metal and cooling water, but it is still possible to flood the molten metal with water. The amounts of molten metal in these systems are relatively small, although, cold-wall induction systems are growing and new potential danger exists. Previously, cold-wall induction systems in the United States were a maximum capacity of 100 to 150 lbs. of titanium, with most in the 50 lb. range. New systems are moving to 250 lbs.

Cold-wall induction, consumable casting and hearth melting systems have the additional benefit of large volume chambers relative to the molten metal volume. The chamber volume absorbs pressure rises reducing the risk of uncontrolled explosions.

EB hearth melting systems represent the lowest risk due to the high vacuum operating level and large vacuum systems, which will absorb pressure rises and remove hydrogen at a high rate.

For reference purposes, the volume of steam generated from a gallon of water is 662 ft. at 1500°F and 1 atmosphere. The amount of hydrogen generated from a gallon of water being combined with 11 lbs of molten titanium is 180 standard ft. After 180 ft. of hydrogen is mixed with 430 ft. of air, the explosive energy potential is equal to that of six pounds of TNT.

Melt Crucible Issues

A Three-Level Safety Strategy
Contributor:  Graham Keough, Vice President of Technology, Consarc Corp., Rancocas, NJ, Feb. 21, 2002

Ever since the tragic events of Sept. 11, 2001 there appears to be a new dynamic for safety in the United States. For example, yesterday I arrived at the airport two hours before take-off. I waited patiently in line while my identification was examined and re-examined and we all went through security checkpoints before boarding the plane. It occurred to me it was remarkable that people accepted this activity in the interest of safety.

However, aside from this example, I suspect the concept of safety is not always given the priority it deserves. In a recent article I read about the urgent need to "increase productivity and efficiency while maintaining” The difference in emphasis was subtle, but this is indicative of an attitude that, when it comes to industry and business, safety may not always be our first priority.

The good news is that, through advances in modern manufacturing and process control technology safety done right can be merged with improvements in productivity and reliability of equipment.  A three-level safety strategy was developed for two Consarc reactive vacuum arc-remelting (RVAR) furnaces, which were installed at Timet’s  melt shop?? in Morgantown, PA. A similar strategy was implemented on RVAR furnaces.

The first level was to limit by design all foreseen risk of operating in a mode which could lead to a hydrogen explosion. This included:

  • A full set of devices and interlocks continuously monitored by the programmable logic controller (PLC)
  • Dual on-line PLC processors in hot-backup mode
  • Reliable Ram/stub interface
  • Critical alarms with automatic power shut-off
  • Remote TV monitoring of the melt
  • Electrode weight monitoring and limit alarms
  • Arc-gap monitoring and alarms
  • Precision stirring current control and interlock
  • Optimized crucible cooling with emergency cooling
  • Redundant pumps and sensors where appropriate

Level 2 assumes that, despite the best intentions of Level 1, we cannot guarantee to eliminate  a risk of water entering the crucible. As a result, sensors and controls are added to detect the onset of an unsafe condition and automatically initiate action to mitigate the effects of a steam explosion and try to minimize the risk of turning it into a hydrogen explosion. This includes:

  • Pressure rise detection
  • Vacuum system isolation
  • Rapid argon back-fill
  • Over-pressure relief
  • Argon flood the over pressure relief device (to minimize the risk of air ingress to the furnace)
  • Head hold down (mechanically bolted) to minimize the risk of air ingress to the furnace


Level 3 accepts that, if the above levels fail to contain the event and a hydrogen explosion does occur, then additional measures are required to reduce the danger to people and equipment. This includes:

  • Each melt station is protected by thick concrete walls above ground and a vented pit below ground
  • No one is allowed inside the protective enclosure during RVAR operation (melting)

Consarc is rightfully proud of the controls developed in implementing the above three-level safety strategy. But, when it comes to overall safety concerns, modern control technology is only part of the story.

Two particular examples of detailed interlocks illustrate this fact. First, one of the most critical interlocks when melting titanium in a water-cooled crucible is the cooling water flow interlock. A simple interlock would simply check the status of the flow switch and such an interlock would be relatively easy to "jumper out.” For the above controls, before starting each melt the operator manually opens in inlet valve and observes full flow detection. Then the operator throttles back the flow until loss of flow is alarmed. Then the valve is returned to full open and flow detection re-confined. This "full function check” is certainly more powerful than a simple interlock, but even this could be defeated so part of the strength of the function check is to emphasize to the operator the importance of the cooling water flow detect function.

The second example of the importance of the user in implementing any safety strategy is the pressure rise detection of Level 2. As long as Level 1 and good melt practices maintain  optimum process conditions and we do not breach the crucible we are unlikely to experience a pressure rise. But, if water does enter the crucible during melting it is critically important that the safety circuits function correctly.

To assure proper function the user must set up a routine maintenance and calibration plan so that a melt start can be simulated and then the pressure rise manually initiated to fully test the safety circuit response.

Modern control technology allows us to improve safety while also improving product quality and yield and improving productivity and equipment reliability. But to achieve the benefits requires close cooperation between vendor and user and a program to test and maintain the controls.

Crucible Wall Thickness and Melt Shop Safety And Crucible Inspection Tools, Process and Data Records
Contributor:  Mike Dagle, Vice President, Zak, Incorporated Crucible Wall Thickness and Melt Shop Safety

Although there is no uniform rule or titanium industry standard, each melt shop should have minimum wall-thickness criteria for their operation. Some people are of the opinion that 50 percent of the original wall thickness is a safe cut-off point, which might be good if the starting wall thickness is greater than 1 inch. Under that scenario, would the same rule apply if the original wall thickness were somewhere between 0.500  and 0.750 inch?  The depth of your average arc damage needs to factor heavily into what you establish for a minimum-acceptable wall thickness for your operation. It might be worthwhile to keep a closer eye on those tubes that have a 25-percent loss of wall thickness.

You also might consider the frequency of crucible usage, the tubes age and its repair history when establishing your minimum-acceptable wall thickness criteria. Do you specify or require a 100-percent wall thickness check to be recorded over a 12 x 12 inch grid pattern over the whole crucible body? Do you require spot wall thickness checks in the active melt areas on your crucible?

Considering the importance of wall thickness on the safety of your operation, do you need to have internal procedures and processes in place to monitor crucible wall thickness? What are your options in this regard?

Mechanical thickness gages are available for measuring the wall thickness at either end of a crucible. Hand-held ultrasonic thickness gages are also available, which allow you to monitor wall thickness anywhere on the body of the crucible.

"Are You Operating a Bomb?”
Contributor:  By David Warren and Matt Mede, Retech, a division of M4

What is needed for an explosion?

In order to have an explosion three things are required: first, fuel, in this case hydrogen; second, oxygen (air); and third, an ignition source. This ignition source may be from molten metal, mag. chlorides or iron oxides that have formed as a finely divided dust on the chamber walls. This dust maybe ignited when air has entered into the chamber.

Water and molten titanium react.  Water and molten titanium react to form hydrogen. For every mole of titanium you will combine two oxygen. Two water 2H20 in molten titanium will form four hydrogen 2HZ.

Dangerous condition, what happens?  Water enters a molten pool of titanium. The worst condition is when this happens under a molten pool (such as leak in the crucible). There is a less chance of having an explosion when the water is dumped on top of the pool (leak in ram). However the potential for an explosion is still there.

Hydrogen stratifies at the top of the furnace. Steam is generated and over pressures the relief port, the port blows off and air enters the chamber. The air mixes with the hydrogen at the top of the chamber. An ignition is generated ether by hot metal or with a slow burn of mag. chlorides or iron oxide dust that has formed on the chamber walls. When the ignition source reaches the hydrogen, an explosion occurs.

Vacuum arc remelting (VAR) furnace basic bunker design
The operator should be located in a remote location or a bunker type constructed room. The design goal of the furnace and bunker is to retain the blast below the operator floor level in the pit. Direct the blast out of the building and up. The pit floor and walls should be constructed of steel reinforced concrete.

VAR casting furnace basic bunker design
The operator should be located in a remote location or a bunker-type constructed room. Casting furnaces are assembled at operator floor levels. A steel-reinforced concrete blast wall should be built around the furnace. All access doors should have steel constructed sliding doors. The chamber door can act as a blast port. The blast should be directed in an upward direction.

Control room bunker
If possible the control room for the operators should be located remotely from the furnace. If this is not possible a steel reinforced concrete bunker should be constructed. The roof should have a steel sub-floor with a steel reinforced concrete slab on top. A viewing window of the furnace area can be added if required. The window should be made of 1-inch thick acrylic plate with a steel support frame. Supports should not be more than 12 inches apart.

NaK Training
Contributor:  Kevin Berry, Creative Engineers Inc., Phoenix, MD

The Potential Hazards of NaK
NaK is a silver/gray liquid metal similar to mercury in appearance, used as a cooling medium for titanium melting. It is an alloy of approximately 50 percent sodium and 50 percent potassium. When it comes in contact with water it creates a violent to explosive reaction because hydrogen is generated and usually ignited. NaK can easily bum when exposed to air. The reaction with water or moisture from the air is caustic, which is corrosive to many materials and to human skin. It reacts with many common materials, such as Teflon, or anything with moisture, any materials that are halogenated.

Why is NaK used for this heat-transfer application in titanium melting?
NaK is a tremendous conductor of heat. It stays liquid over a wide temperature range, 9°F to 1445°F. NaK’s elementary materials (potassium and sodium) do not decompose, regardless of temperature. It has very low vapor pressure, even at high temperature (NaK boils at 1,445°F). This means that the NaK system can operate at relatively low pressure for high-temperature applications. NaK has a low density (lighter than water). It’s easy to pump. It’s one of the few thermal fluids that can operate normally (without phase change, without significant pressure and without decomposition) as a liquid above 600-800° F.

  • NaK firefighting procedures
    • NO water
    • NO dry chemical
    • A Class D fire extinguisher is recommended
    • Dry soda ash is also frequently used to smother a NaK fire
  • A Nak fire creates a significant amount of dense white smoke, consisting of sodium and potassium hydroxide particles. The smoke is an irritant to skin, eyes, respiratory system and is a visual obstruction. Respiratory protection is highly recommended.
  • Personal Protective Equipment (PPE)
    • Fire retardant clothing such as Nomex with multiple layers. No cuffs
    • Leather gloves
    • Glasses with side-shields
    • Face shield with hard hat
  • Spill cleanup recommendations
    • Contain the spill to keep it from flowing
    • Extinguish the fire if burning
    • The residues should be swept up and stored in a covered container. The residues must be deactivated using extreme care by qualified individuals.
  • Building and Storage Issues
    • No sprinkler systems
    • No wood or combustible materials of construction. Cement and steel are best suited for working areas where there is a chance of spilled or splashed NaK
    • No floor drains because of the water trap in the drain
    • Have the capability to ventilate the upper elevations of the building where hydrogen could pocket and present an explosion hazard
    • Secondary containment for vessels or systems is recommended, consisting of a steel or stainless steel pan, which is maintained in a dry condition
  • Special Considerations
    • Contrary to procedures for other alkali metals, such as sodium, do NOT use mineral oil to cover NaK to shield it from the atmosphere
    • Organic materials present the potential for an extraordinary explosion hazard, which is shock sensitive
    • Use an inert gas such as nitrogen or argon to shield the NaK from the atmosphere
    • Avoid the use of organic fluids with NaK where possible
  • Materials of Construction for Handling NaK
    • Normally, steel or stainless steel are ideal based on temperature; nickel and cobalt alloys also are suitable
    • Elastomer selection should be carefully evaluated
    • NO aluminum
    • NO Teflon
  • Piping Issues
    • Minimize threaded connections for high temperatures
  • Valve selection
    • No Teflon-seated ball valves
    • Consider using "bellows-sealed” valves
    • Ensure that any receiving tanks (or tank jackets) and piping components are dry and purged with inert gas before introducing NaK to the system
    • Piping system should be designed with proper slope for easy draining; no "U’s” or dead pockets in piping system that cannot be fully drained
    • Special consideration should be applied to pump selection; systems should operate at low pressures (less than 15 psi), when possible
    • Tank or containment vessels should have no bottom drains.

A Brief Description of NaK Cooling on VAR Furnaces
Contributor:  Dr. Dave Collins, IMI Titanium Ltd., Birmingham (West Midlands), UK October 1995

Following the explosion of one of our water-cooled VAR furnaces in 1966, IMI decided to look for a safer alternative. In 1968 IMI installed its first NaK-cooled VAR furnace. Since then we have added four additional NaK-cooled furnaces and in 27 years have not experienced another major incident.

The basic NaK-cooled VAR furnace differs very little either in design or operation from a conventional VAR furnace. The main difference, particularly in the way we choose to operate them, is that once the crucibles and retractable crucible bases have been assembled into the NaK loop they are not usually removed again until they fail or suffer damage. This clearly limits  the flexibility of the furnace, since it is not practical to keep changing crucible to accommodate different size of ingots and could therefore be considered one of the main drawbacks of this type of furnace.

In a typical layout of a NaK-cooled VAR furnace, cool air is drawn by the blower through a filter from outside the building and passed over the heat exchanger in the "NaK Pack” before exhausting via the stack. The NaK Pack is a caged self-contained structure, standing above a deep, dry drip tray. It houses the NaK pumps, control valves, flow meters, heat exchanger, cold trap, expansion tank and the cover gas (argon) control system.

Should any of these components start to leak, it is the purpose of the drip tray to contain the leakage so it is essential it be kept dry. A smoke detector positioned above the pack will sound an alarm should a leak occur. This has proved to be an extremely sensitive alarm system since even the smallest of NaK leaks tends to give off copious amounts of smoke.

Housed in the control room under the constant observation of the melting controller are the NaK flow and temperature gauges along with a mimic of the NaK system displaying the current status of all of the components in the system and array alarm conditions.

As previously stated the VAR furnace and its vacuum pumping system are exactly the same as for a conventional water-cooled system. However, the electrode ram has its own self -contained NaK cooling system and the crucibles housed in the furnace pit, for practical purposes, can be considered "fixed” to the furnace since they are changed only for repair or refurbishment, which is generally every three to five years.

Since the crucibles are not removed from tile furnace, they have removable NaK-cooled bases, which are retracted to allow the hydraulic ingot ejector to push the cooled ingot out of the crucible. Situated at the lowest point in the pit is the NaK dump tank, which is sized to safely contain all of the NaK in the system.

In an emergency, should a NaK leak or fire develop in the bottom of the pit, (detected by the smoke detector) the NaK in the furnace system would be dumped into the dump tank. Then, after checking to ensure the pit is clear, argon gas would be manually discharged to fill the bottom of the pit to a level of approximately eight feet. The ventilation to the bottom of the pit is shut off leaving the ventilation to the intermediate  level to remove any smoke or fumes. IMI has had to use this procedure only once, when an oil fire developed on the ingot ejector on one of the furnaces, and it worked very well.

Another safety feature of this system is the cold-trap NaK pump. In normal operation this pump circulates a small quantity of NaK through the cold trap, which acts as a filter to remove any system debris or oxides. However, should the power to the main NaK pump fail or its flow be reduced, then the power to the cold-trap pump is automatically increased to compensate.

Important Features to Consider When Installing a NaK System

  • Safety must be the prime consideration in the system’s design, operation and maintenance
  • Ensure the system is correctly sized, has separate emergency power feeds and is installed with good accessibility in a clean, dry, well-ventilated area, which preferably is separate from the main working area
  • Make sure the local county or municipal fire department is aware of the facility, knows how to fight NaK fires and has the appropriate equipment to do so
  • A safe, environmentally friendly method of disposing of used NaK and cleaning components, including crucibles, will be required
  • Have clear written procedures and trained personnel to implement them, especially in the event of an emergency
  • Install effective smoke detectors and have and sufficient plant fire-fighting equipment with trained personnel in the facility, in order to contain any accidents before they get out of hand.

References:1.1. Barin, Thermochemical Data of Pure Substances (Weilhem, Federal Republic of Germany: VCH Verlagsgeselschaft, 1989).2. 8.J. McBride and S. Gordon, "Computer Program for calculation of Complex Chemical Equilibrium Compositions and Applications,” NASA Reference Publication 1311 (Cleveland, OH: Lew Research Center, 1996).Note about the authors:Steven C. Evans is principal engineer with Westinghouse Electric Company. Western Zirconium plant located in Ogden, UT. David F. McLaughlin is a fellow engineer with Westinghouse Electric Company, Science and Technology Department, Pittsburgh.

Titanium Fire Demonstrations:

GSL Inc. Titanium Extinguishment Demonstration

Automated Fire Extinguishing System by TLI Group Ltd.

Titanium Fire Extinguished with FEM-12 SC  


LAFD / Titanium Scrap Warehouse / Part One of Three

Sign In

Forgot your password?

Haven't registered yet?