Types of Heat-Treating Furnaces

John W. Smith, Holcroft, A Division of Thermo Process Systems Inc.

Introduction

FURNACES commonly used in heat treating are classified in two broad categories, batch furnaces and continuous furnaces. In batch furnaces, workpieces normally are manually loaded and unloaded into and out of the furnace chamber. A continuous furnace has an automatic conveying system that provides a constant work load through the unit.

Batch Furnaces

The basic batch furnace normally consists of an insulated chamber with an external reinforced steel shell, a heating system for the chamber, and one or more access doors to the heated chamber. Standard batch furnaces such as box, bell, elevator, car-bottom, and pit types are most commonly used when a wide variety of heat-hold-cool temperature cycles are required. Other types of batch furnaces, discussed in separate articles of this Volume, are salt bath, vacuum, and fluidized-bed furnaces.

The use of batch equipment for heat treating usually requires considerable labor for loading, handling, and unloading of the work and work trays. High labor costs dictated by the process must always be considered in the selection of batch equipment.

Batch furnaces are normally used to heat treat low volumes of parts (in terms of weight per hour). Batch furnaces are also used to carburize parts that require heavy case depths and long cycle times. For example, integral-quench batch carburizers treat gears or rock bits, while pit-type carburizers process parts such as drill rods or bearing races. However, there is no real advantage in the use of batch methods for deep-case carburizing. Yet, very often batch furnaces are used for deep-case work because the volumes are so low that a pusher-type continuous furnace is not cost effective. Here again, the batch furnace has the advantage in terms of part volume, not case depth. Batch furnaces are also normally used:

• To handle special parts for which it would be difficult to adapt a conveying system for continuous handling (long drill rods processed in a pit furnace, for example)

• To process large parts in small numbers, for example, stress relief or annealing of large weldments or castings in a car-bottom-type furnace

• To process various parts requiring a wide range of heat-treat cycles that can readily be changed, either manually or automatically

Batch processing is especially appropriate when the work must be heated from room temperature to a maximum temperature at controlled rates, held at temperature, and cooled at controlled rates. For example, car-type furnaces are used for critical stress-relief work or carbon baking in saggers.

Box-Type Furnace. With the addition of powered work-handling systems--integral quench tanks, slow-cool chambers, and some automatic controls--the basic box-type batch furnace is upgraded to a semicontinuous batch furnace, which is a commonly used piece of heat-treating equipment. One type of semicontinuous batch furnace is shown in Fig. 1.

Fig. 1 Semicontinuous batch furnace with a controlled gas-heated box furnace. Courtesy of Seco/Warwick Corporation

The car furnace, also called a "bogie hearth," is normally considered an extremely large batch furnace. The bottom (or floor) of the furnace is constructed as an insulated movable car that is moved out of the furnace for loading and unloading, as shown in Fig. 2. When in position inside the furnace, the car is sealed to the furnace structure with granular-type "sand" sealing troughs or solid seals. Furnace cars can be self-propelled with the motor drive mounted on the car, or they can be moved in and out by a floor-mounted drive with a continuous chain or a rack-and-pinion drive. Most car furnaces are nonatmosphere type due mainly to the difficulty in sealing the car.

Fig. 2 Car-bottom batch furnace for homogenizing large cylindrical parts. Courtesy of Despatch Industries, Inc.

Heating systems normally are either direct fired or electrically heated with resistance elements. With direct-fired systems, it has proved advantageous to design a pressure-control system to control the flues. With the large difference in fuel burned during the heat-up portion of the cycle as compared to the soak portion of the cycle, it is extremely difficult to maintain a minimum acceptable furnace pressure with a fixed flue area. Most car furnaces are heated from room temperature with the load already in the furnace. A typical cycle would be to heat from room temperature to a control temperature at a specific rate, hold at the control temperature for a specified time, and then slow cool to discharge

temperature at a specified rate. Programmable temperature-control systems with stored menu programs are capable of performing a wide variety of heat-treat cycles, including process monitoring and recording of historical data.

The use of ceramic fiber insulation in a car furnace allows greater control of furnace temperature when following a programmed cycle. Because ceramic fiber has minimal heat storage capacity compared to hard refractories, it will heat and cool at faster rates. Also, less total heat is required to bring the furnace to the desired temperature, although the difference may be minimal because the total heat is governed more by the mass of the load. Further, continuous cyclic heating and cooling has little or minimal effect on ceramic fiber lining. Temperature limits of ceramic fiber blanketing are discussed in the article "Energy-Efficient Furnace Design and Operation" in this Volume.

Car furnaces are used from the lower stress-relieving ranges around 540 °C (1000 °F) to temperatures of over 1095 °C (2000 °F) for certain applications. Because many of the larger car furnaces are installed outdoors, increased allowances should be made for thermal holding losses caused by winds and other changes in ambient conditions.

Elevator-type furnaces are similar to car-bottom furnaces except that the car and hearth are rolled into position underneath the furnace and raised into the furnace by means of a motor-driven mechanism. Such furnaces are built to handle large, heavy loads and can be cooled rapidly by a high-velocity internal or external circulating gas system. For certain plant layouts, this type of furnace eliminates the need for crane facilities, transfer cars, and switching tracks and often saves floor space.

Elevator-type furnaces are suited for heavy work and for the precipitation-hardening nonferrous alloys, which must be quenched rapidly to retain a supersaturated solid solution. Either gas firing or electric heating is commonly used, with oil firing being employed less frequently. The temperature range for these furnaces is generally about 315 to 1200 °C (600 to 2200 °F).

Bell-type furnaces have removable retorts or covers called "bells," which are lowered over the load and hearth by crane. The inner retort is placed over the loaded hearth, sealed at the bottom, and provided with a constant supply of protective atmosphere; then the outer heating shell is lowered dover the assembly.

One bell furnace outer heating shell can take care of several retorts. For dense hearth loadings, a motor-driven fan for circulating the atmosphere inside the retort provides more rapid, uniform heating.

Pit furnaces, sometimes called pot furnaces, consist essentially of two parts: the furnace, which is placed in a pit and extends to floor level or slightly above, and a cover or lid, which extends upward from floor level. Large pit furnaces are generally installed with at least part of their heating chambers below floor level. Smaller furnaces are usually mounted on the floor. Workpieces are suspended from fixtures, held in baskets, or placed on bases in the furnaces. This type of furnace is particularly suitable for heating long parts, such as tubes, shafts, and rods suspended from a top supporting fixture or supported from the lower end and held in a vertical position. Loading in this manner gives minimal warpage.

Pit furnaces are available over a wide range of weight capacities and are well adapted to the use of devices for automatic carbon control. They are particularly suited to the processing of parts that must be cooled in the furnace. However, direct quenching is usually not feasible when large loads and large furnaces are involved.

An additional disadvantage of the pit-type furnace is that, if the work is to be direct quenched, the load must be moved from the atmosphere of the furnace into air before quenching. Although exposure in air is relatively brief, it results in the formation of an adherent black scale on the steel that for many applications must be removed by dilute mineral acids or by grit blasting. Thus, parts that must remain bright and scale free after furnace treatment, such as parts with internal threads, are processed in horizontal batch furnaces and quenched under a cover of protective atmosphere.

Continuous Furnaces

Continuous furnaces consist of the same basic components as batch furnaces: an insulated chamber, heating system, and access doors. In continuous furnaces, however, the furnaces operate in uninterrupted cycles as the workpieces move through them. Consequently, continuous furnaces are readily adaptable to automation and thus are generally (though not always) used for high-volume work. For example, continuous carburizing furnaces are generally preferred for production loads exceeding 180 kg/h (400 lb/h) and requiring the same case depth, or for loads of sufficient size that require 24-h continuous operation. Some types are equipped to provide cooling under a protective atmosphere.

Another advantage of continuous furnaces is the precise repetition of time-temperature cycles, which are a function of the rate of travel through the various furnace zones. In terms of atmosphere control, however, the frequent door openings can upset internal atmospheres during the charge and discharge cycles. Also, it is sometimes difficult to keep the atmospheres segregated by zone in single-chamber continuous furnaces, and often the zone controllers "fight" each other because of the interaction between adjacent zones. Therefore, in some applications (such as carburizing), high accuracy, temperature control, and atmosphere control may be easier to achieve in a multichamber pusher-type system where the heating, carburizing, and diffusion portions of the cycle are separated. A batch furnace could also be used.

In a general sense, continuous-type furnaces can be classified as either rotary-hearth furnaces or straight-chamber furnaces. In a rotary-hearth furnace (Fig. 3), the floor of the heating chamber rotates inside a stationary roof and inner and outer walls, with a sand or liquid seal between the floor and walls. Rotary-hearth furnaces have been in use for many years, mainly as single-chamber units for low-production manual operations with a variable-product mix. Rotary-hearth furnaces are useful in shops or operations where only one operator is employed, because the charging and discharging can be located next to one another. Rotary furnaces are also used in combination with straight-chamber furnaces (see the section "Other Continuous Furnace Systems" in this article).

Fig. 3 Schematic arrangement of a relatively small continuous rotary-hearth heating furnace. Larger furnaces

of this type have burners firing through both inside and outside walls above the hearth, while very large furnaces (up to 30 m, or 100 ft, in diameter) use multiple heating zones that can be fired either with roof-type burners or burners located in the vertical portion of saw-tooth roof construction. Source: Ref 1

Straight-chamber continuous furnaces include various types, which can be classified as:

• Pusher-type furnaces • Walking-beam furnaces • Conveyor-type furnaces that use rollers or belts • Continuous furnaces with tumbling or inertia action of the parts for movement

These four types of straight-chamber continuous furnaces are discussed in the following sections. Other types such as overhead monorail systems are also briefly mentioned. The common types of continuous furnaces are the pusher, rotary-hearth, roller-hearth, and continuous-belt furnaces. Walking-beam furnaces and the furnaces that impart tumbling or inertia of the parts (that is, the shaker-hearth and rotary-retort furnaces) are less common.

Pusher Furnaces

A pusher furnace uses the "tray-on-tray" concept to move work through the furnace, as shown schematically in Fig. 4. A pusher mechanism pushes a solid row of trays from the charge end until a tray is properly located and proven in position at the discharge end for removal. On a timed basis, the trays are successively moved through the furnace. Cycle time through the furnace is varied only by changing the push intervals.

Fig. 4 Schematic of tray movement in a pusher furnace

Pusher-type furnaces are quite versatile and, depending on the size and shape of parts and on permissible distortion, parts may be loaded randomly and free quenched in an elevator-type quench tank. Alternatively, parts may be removed individually from the furnace trays for plug or press quenching. Furnaces can be designed to provide a variety of special equipment and process requirements, such as number of trays, tray size, atmosphere control, atmosphere recirculation, temperature-control zones, and quenching facilities. They may also have more than one row of work in process at the same time.

Pusher-type furnaces are by far the most widely used continuous furnace for gaseous carburization. Construction usually consists of a gastight welded shell with radiant tubes for heating. The work is pushed through on trays with or without fixtures and after completion of the carburizing cycle may be quenched or cooled slowly. Circulating fans are almost always used for more uniform temperature and carburization. Most pusher-type furnaces are built with purging vestibules at the charge and discharge ends to reduce contamination of the atmosphere by air. In many instances, washing and tempering equipment is incorporated to provide a fully automated heat-treating line.

Skid-Rail Furnaces. In a skid-rail pusher furnace, the work is placed on flat, normally reversible cast-alloy grid trays. The trays, in turn, are supported through the furnace on skid rails. The total gross load on the tray determines the number of rails necessary to minimize wear by maintaining the bearing pressure between the tray and the rail within acceptable limits. In certain applications, particularly when an endothermic or an enriched endothermic atmosphere is used, the skid rails are lubricated by the atmosphere and the coefficient of friction is reduced, decreasing wear and increasing tray and skid-rail life. Skid rails are used normally for light-to-moderate tray loadings.

Skid rails that are cast or fabricated from nickel-chrome heat-resisting alloy have been commonly used in pusher furnaces. These alloy skid rails normally are supported by and anchored to a series of cast or fabricated alloy pier caps and supporting pedestals.

Because of the ever-increasing cost of nickel-chrome alloys, alloy skid rails are being replaced where possible by less expensive silicon carbide refractory rails. Silicon carbide skid rails are molded and prefired into various rectangular shapes and are then bricked in and supported on the lower piers or are supported directly on the furnace floor. With the

rails resting on edge, the alloy grid tray is thus supported on two or more rail faces, each normally 64 mm (2 12

in.) wide.

Because silicon carbide in contact with an alloy has a lower coefficient of friction than alloy on alloy, it makes excellent skid rails. However, in designing silicon carbide rails, precautions should be taken to eliminate severe temperature gradients or thermal shock in the rails, both of which could cause the rails to fracture.

Roller rails are used to both support and guide the trays as they are moved through the furnace. The rails are supported and anchored in a manner similar to that of alloy skid rails, but the mechanical advantage of the wheel and axle reduces the pushing force required to move the load when compared to skid-type support rails.

In certain instances, rails, wheels, and axles are all made of an alloy material. The use of a dissimilar material for the axle, however, can reduce the coefficient of friction between the wheel and axle, and the axle to rail, which can eliminate the natural "galling" effect of alloy on alloy.

The alloy roller rail tray usually has one or more runners and guides on its underside to keep the tray centered on the roller rails. Because the tray underside is normally not flat, it cannot be moved readily (skidded) at 90° to its normal travel, and a transfer carriage would be used. Also, the tray normally is not reversible. There are, however, designs of alloy trays employing recessed runner tracks, so that the tray will sit flat on a surface and is capable of being moved at 90° to its normal travel.

Because required pushing force is reduced, roller rails are used mostly for heavy tray loads and lengthy pushes. To keep wear in the wheel or axle to an acceptable minimum, the total load per wheel (line pressure), in pounds per inch (or kilograms per centimeter) must be kept within acceptable limits.

Buggy trays make use of the mechanical advantage of the wheel and axle in a manner similar to that of the roller rail. Wheels and axles attached to the underside of the tray are guided through the furnace in refractory or alloy troughs, supported on furnace piers or directly on the furnace floor.

This tray cannot be skidded at 90° to its normal travel, and a carriage would be required for such movement. Although buggy trays can be used for extremely heavy loads, the maximum recommended load per wheel should not be exceeded, to keep wear within acceptable limits.

A buggy tray with the wheels and axles attached normally results in a heavier tray weight per square foot. This could have an adverse effect in fuel economy as the weight of the tray to be heated and cooled is increased over other types of trays.

Walking-Beam Furnaces

A walking-beam furnace has movable rails that lift and advance parts along stationary rails inside the hearth. With this system, the moving rails lift the work from the stationary rails, move it forward, and then lower it back onto the stationary rails. The moving rails then return to the starting position and repeat the process to advance the parts again. Figure 5 shows a typical walking-beam furnace of this design used for moving steel slabs. In this type of system, the moving rail can be designed to move in an elliptical path or rectangular pattern. The frequency of lift and length of stroke determine the total processing time.

Fig. 5 Schematic of a walking-beam mechanism for advancing slabs through a furnace. In actual furnaces, water seals beneath the hearth prevent escape of furnace gases or air infiltration. Source: Ref 1

In another type of walking-beam furnace, both sets of rails move. One set of rails moves up and down, and the second set moves forward and backward. This system is known as a true rectilinear motion walking beam. The sequence normally is as follows: The lifting beam moves up and the traveling beam moves in reverse, then the lifting beam moves down and the traveling beam moves forward. The work is thus sequenced through the furnace.

Uses of Walking-Beam Furnace. Walking beams traditionally have been used in steel mills in reheat furnace hearth systems for slabs and billets. Walking-beam systems can be built ruggedly to move extremely heavy loads. In heat-treating operations, walking beams have been used successfully with flat-top beams to carry such work as flat plates or trays, or with pocketed-top beams to carry unstable parts such as rollers or shafts.

Advantages of Walking-Beam Furnaces. The typical advantages of walking-beam furnaces are:

• Only the work being processed has to be heated because normally trays or fixtures are not needed • Friction is reduced for heavy loads because the work is never skidded • The system can be loaded or unloaded automatically • A part can be picked from a specific spot and placed in a specific spot by using the walking-beam

mechanism • Equipment is self-emptying on shutdown

Disadvantages of Walking-Beam Furnaces. The typical disadvantages of walking-beam furnaces are:

• Mechanisms are usually more expensive than for pusher-type systems • On large high-temperature slab or billet reheat furnaces, there is a dramatic increase in thermal holding

losses and related fuel consumption due to the water-cooled insulated walking-beam rail system • Walking-beam mechanisms are not commonly used where protective atmospheres are required in the

furnace chamber due to the inherent problems in adequately "sealing" the moving walking beams and mechanisms

Conveyor-Type Furnaces

Roller-hearth furnaces move the work-piece through a heating zone with powered, shaft-mounted rollers that contact the work-pieces or trays. This type of furnace might be used to advantage in heating much longer slabs than would be practical in a pusher-type or walking-beam furnace. These furnaces are available as single units (Fig. 6) for zone heating or cooling in a line of furnaces.

Fig. 6 Section of a roller-hearth furnace. Courtesy of Seco/Warwick Corporation

Continuous-belt furnaces are similar to roller-hearth furnaces except that mesh or cast-link belts are used to move the parts. Such furnaces are preferred for small parts that cannot be moved satisfactorily directly on rollers. Conveyors used include woven belts of suitable material, and chains with projecting lugs, pans or trays connected to roller chains.

The parts are fed automatically onto a mesh belt at the front of the furnace, which can have a liquid or gas atmosphere seal to maintain atmosphere integrity in the furnace chamber. Cast belts return inside the furnace. Belt-type furnaces generally are furnished with fans for recirculating the atmosphere.

Reciprocating and Rotary Agitation Furnaces

Shaker-hearth furnaces (Fig. 7(a) and 7(b)) use a reciprocating shaker motion to impart inertia to the work; this motion may be regulated to control the time cycle. Heating is efficient and confined mainly to the work load. Parts may be fed into the furnace by hand or by means of automatic metering and are typically drop quenched individually. Use of this type of furnace is generally limited to lightweight parts that are to be carburized to case depths of 0.3 mm (0.012 in.) or less. The furnace hearth must be kept smooth, clean, and at the proper level. Either the time cycle of work going through the furnace or uniformity of case depth should be checked at frequent intervals to detect any unwanted change in the forward force exerted by the shaker mechanism. When these precautions are adequately observed and the necessary atmosphere and temperature controls are provided, the shaker-hearth furnace can produce satisfactorily uniform case depth.

Fig. 7(a) Shaker-hearth furnace with conveyor removal of quenched parts. Courtesy of Lindberg Heat Treating Company

Fig. 7(b) Schematic of shaker-hearth furnace for continuous carburizing

Unless a specially designed hearth, such as a corrugated hearth, is provided, heavy, flat parts are not suited to processing in shaker-hearth furnaces, because of the difficulty of obtaining adequate and uniform case on the part surface making contact with the hearth. Delicate parts and parts with fine threads may be mutilated by the action of the conventional shaker hearth. Without a special hearth, balls and cylinders will not move and progress uniformly in this type of furnace.

Two aspects of the design of shaker-hearth furnaces require special attention:

• Hearth plates should be of suitable weight to respond to the reciprocating motion of the shaker mechanism. This action slides the parts forward randomly and ensures exposure and equal treatment to that portion of the part that rests on the hearth

• Adequate exhaust facilities should be provided to handle quenching-oil fumes. These fumes are highly carburizing and, unless properly disposed of, will interfere with control of the carbon potential of the atmosphere. They will also soil the surfaces of parts being processed

Rotary-retort continuous furnaces (Fig. 8) are used to carburize the same types of small parts that can be handled in a rotary-retort batch furnace. The advantage of the continuous rotary furnace is that it can be loaded and unloaded automatically, thus eliminating the need for removing and replacing the head.

Fig. 8 Rotary-retort furnace for continuous carburizing

The inside of the retort is provided with a mechanized spiral rib throughout its length. The spiral rib can move the work load in either the forward or reverse direction. The frequency and duration of cycles of forward and reverse motion can be varied over any desired range. By this means, furnace length can be minimized, and a reasonable agitation or tumbling action of the parts is obtained. The tumbling action provides for a better uniformity of case depth. However, it can also serve to damage certain parts by nicking.

Continuous retorts cannot be tilted at will. Parts are fed at the front end and automatically issue from slots in the rear of the retort directly into the quenching medium. Because the front end of the furnace must be open to allow continuous charging, sufficient carburizing gas must be fed into the furnace to prevent the admission of outside air. These furnaces are suitable for carburizing to case depths of 0.4 to 2.5 mm (0.015 to 0.100 in.).

Other Continuous Furnace Systems

Combination Pusher-Type and Rotary-Hearth Furnace Heat-Treat System. The material above basically describes the traditional continuous-type furnace that operates on the principle of work being processed through the furnace on a "first in, first out" (FIFO) sequence. With recent industry trends toward the physical downsizing of heat-treatable components, lowering of manufacturer stockpiles and inventories to reduce costs, and moving into Just In Time (JIT) manufacturing, a need was created for a completely flexible continuous heat-treat system capable of simultaneously running a wide variety of parts to variable cycles. To accomplish this type of heat-treat processing using conventional FIFO furnaces would be extremely cumbersome and time-consuming. When changing the product mix or cycle in the FIFO-type furnaces, it would be necessary to either empty the chambers partially or completely, using empty trays, thus greatly affecting the operating efficiency of the equipment.

A multichamber, combination pusher and rotary-hearth heat-treat system as shown in Fig. 9 has only recently been conceived and built to address many of the shortcomings encountered when adapting existing continuous FIFO furnace systems to the new JIT manufacturing demands of industry. With pusher-type, continuous-type furnaces, it is not possible to simultaneously process side by side, a random mix of work parts, each requiring a different cycle time. When changing the process variable of "time" in a pusher-type continuous FIFO furnace, it is necessary to run a number of empty trays between the different types of tray loads in the furnace chamber. This has a negative effect on furnace efficiency, particularly if process cycle time changes are frequent.

Fig. 9 Pusher-type and rotary-hearth heat-treat system

When shutting down a continuous pusher-type furnace, it is necessary to run empty trays into the charge end of the system to allow the work parts in the system to empty. The reverse is true on start-up, in that the system must first push out all the empty trays in the system before the product emerges. This arrangement wastes fuel and manpower in the heating and handling of empty trays.

The unique rotary hearth and pusher chamber concept allows tray loads with variable cycle times to be run simultaneously, side by side, in the same furnace chamber. This system also is built to be completely self-emptying without the use of empty trays. A complex system of this type is shown in Fig. 9, comprising three donut-shaped rotary furnaces: a carburize chamber, a diffusion chamber, and an equalize chamber, each having a circular rotatable hearth for supporting and moving trays of parts within an annular furnace chamber. Each rotary furnace is connected to the adjacent rotary furnace or pusher chamber through a tunnel with a special sealing door. The rotary hearths permit movement of any tray in any position within a furnace, at any time, by rotation of the selected position on the hearth to the charge or discharge door of that furnace, thus providing a high degree of flexibility in operation of the system. One or more pusher mechanisms are included within the circular space or "hole" of each donut-shaped furnace for moving the trays of parts between chambers.

The third rotary furnace of the system (equalize chamber) serves as a cooling chamber to 840 °C (1550 °F), a mechanism for transporting trays of parts to a selected quench system or to an atmosphere cooling chamber, and as a reheat chamber for trays of parts returned from the atmosphere cooling chamber. Trays that have been pushed into the atmosphere cooling chamber from the equalizing chamber can, after cooling, be reintroduced into the equalizing chamber for reheating and quenching or can be removed directly from the atmosphere cooling chamber to a tray-return transfer line.

Strand-Type Furnaces. Continuous strand-type furnaces for heat treating uncoiled strip reduce the handling and cycle times required with batch-type furnaces for sheet in coil form. Furnaces of this type also permit combining other operations, such as cleaning and/or coating. These furnaces usually have high production rates.

Overhead-Monorail Furnaces. In these furnaces, workpieces to be heat treated are suspended from rods attached to carriers on a monorail. Most continuous enameling is performed in overhead monorail furnaces.

Reference cited in this section

1. H.E. McGannon, Ed., The Making, Shaping and Treating of Steel, 9th ed., United States Steel Corporation, 1971

Direct-Fired Furnace Equipment

With direct-fired furnace equipment, work being processed is directly exposed to the products of combustion, normally referred to as flue products. To minimize the scaling (oxide) effect on the work, the flue products can be controlled or varied by adjusting the fuel-air ratio of the combustion system. Although fuel-air adjustments can be made manually, more precise control can be achieved automatically by a wide variety of fuel-air ratio control systems on the market today. When direct-fired burner equipment is used in a heat-treating furnace, the parts being processed often are in some primary or intermediate stage of manufacture. The oxide formed is not detrimental to the part because it will be removed later in the manufacturing process. For example, forged parts are sometimes hardened in the as-forged condition, and rough castings are annealed prior to machining.

Gas-Fired Equipment. Gaseous fuel used in heat-treating furnaces can be natural gas, straight propane, a propane-air mix, or a relatively low-energy manufactured gas.

With the proper selection of burners, controls, orifices, and pipe sizes, a combustion system can be designed to operate on 2500 Btu (2635 kJ) propane gas, 1000 Btu (1055 kJ) natural gas, or 160 Btu (170 kJ) producer gas. The number refers to the energy contained in a cubic foot (0.028 m3) of the gas. Manual adjustments are required for conversion from one gas to another.

Oil-Fired Equipment. Almost any grade of oil that can be satisfactorily atomized can be burned in direct-fired equipment. Lower-viscosity oils such as diesel fuel and No. 2 fuel oil can be easily atomized with pressurized (room-temperature) air. These are probably the fuel oils most commonly used for heat treating. Even with easily atomized oils, caution should be employed in using them on flame-supervised furnaces operating below 760 °C (1400 °F) with interrupted pilots. At low oil flows and excess air conditions, nuisance shutdowns can occur from the flame supervision devices. In certain instances, as dictated by the National Fire Prevention Association, "constant pilots" may be used to eliminate the shutdowns. Insurance carriers must approve the use of constant pilots for the particular application, however. Heavier grades of oil must be atomized by a method other than low-pressure air. Normally, high-pressure air and steam are used.

Burners that can be fired by either gas or oil are available. In most instances, oil is used as the standby fuel to be used in peak periods when natural gas supplies are curtailed. Oil is considered desirable by some in the forging industry because it creates a "softer" scale on the billet, which is more easily removed in forging. High vanadium, potassium, and sulfur contents in fuel oil burned in the direct-fired process can reduce the useful life of various furnace components, especially the nickel-chrome heat-resisting alloys.

Advantages of Fuel-Fired Furnaces. The following advantages are common to fuel-fired furnaces:

• Lower energy cost • Easy to adjust or alter connected input • Recuperator heat-saving devices can be added easily, and controlled cooling can be initiated easily with

proper design of combustion systems • Faster heat-up times because inexpensive control factors can be added to accommodate the difference in

fuel burned during heat-up

Disadvantages of Fuel-Fired Systems. The following disadvantages are common to fuel-fired furnaces:

• Requires extensive ventilation systems • Potential explosion or fire hazard • Requires more manpower for start-up and shutdown • Adjustment more difficult to maintain, resulting in excessive fuel use • Only certain materials or types of products can be run in direct-fired furnaces due to the effect of high

dew point and oxidizing flue gases on the part surface

Electrically Heated Furnace Equipment

Electrically heated furnaces are commonly found in all temperature ranges: from low-temperature tempering furnaces, through the heat-treating range, and up to forging temperatures. The basic consideration in selecting the type of heating element is to determine whether the elements are to be the open type, which are exposed to the furnace environment, or the indirect type, which are protected from the furnace environment by some means such as a radiant tube, muffle, or retort. Factors affecting this decision are furnace atmosphere, need to protect the element from mechanical damage, and space required for placement of the element.

Material Selection. Almost all furnace atmospheres other than air will in some way affect the overall performance and subsequent life of each type of heating-element material. Manufacturers of heating-element materials provide charts that allow designers to predict the material performance with any of the given atmospheres. Each heating-element material can be exposed to the different furnace atmospheres with varying degrees of success.

The notable exception is with a carburizing-type atmosphere. The conventional nickel-chrome strip heating element does not perform well in a carburizing atmosphere because the element itself carburizes, affecting element performance. Generally, in a carburizing atmosphere, heating elements are placed inside radiant tubes or outside work-protecting muffles or retorts. However, some alternative elements are designed specifically to operate when exposed in a carburizing atmosphere.

The selection of open or indirect elements is a choice also determined by the need to protect the element against mechanical damage from parts being heated, from accumulations of metallic scale, or from broken refractories. In furnaces where bottom heat is mandatory and scale can be formed readily on the parts, or where parts may fall from a tray or conveyor, electric elements should be protected in radiant tubes below the hearth. Open elements could still, however, be used throughout the upper portion of the furnace.

In some furnace designs, the physical space available determines the design of the element.

A further consideration is whether to use an element material other than nickel-chrome strip or rod. Silicon carbide elements (globar) or molybdenum disilicide rod elements have been used with success when directly exposed to various atmospheres, although the former normally is not recommended for use in a carburizing atmosphere. Silicon carbide elements have been used on occasion inside radiant tubes for protection against carburizing atmospheres.

Metallic Resistance Heating Elements. The following are general types of furnaces, with a description of the kinds of heating elements used in each. More detailed information on resistance heating elements is given in Volume 2 of ASM Handbook, formerly 10th Edition Metals Handbook.

Low-Temperature Furnaces with Open Elements. The temperature range of this type of furnace varies from approximately 150 to 675 °C (300 to 1250 °F), and the furnace is normally a recirculated-wind convection heating type. The simplest type of heating element is a commercially available duct heater, usually full-voltage, 440 V or 220 V heaters. These are quite useful when the designer can stay within the manufacturer limitations. Maximum use temperature of the commercial duct heater element is normally limited to 400 °C (750 °F).

The heater should be large enough to cover the entire recirculated wind-duct cross section, but designs are limited to the heater sizes available. The design watt density for this type of duct heater is normally 34 000 W/m2 (22 W/in.2). Watt density is the expression commonly used for connected power of each element in watts divided by its total surface area. Watts per unit area is an important design consideration, and the allowances vary greatly with temperature, type of element, and furnace.

As an alternative to the commercial unit, a custom-built duct heater can be used. A steel or alloy frame can then be designed to completely fill the air-duct cross section, and such a unit can be removed easily through a sidewall bulkhead.

The nickel-chrome ribbon material is supported in ceramic insulators mounted in tiers. A common element material is 35Ni-18Cr-44Fe, and it is normally selected in the lighter gage thicknesses and narrower widths. A typical cross section for the ribbon material would be 13 by 0.8 mm (0.50 by 0.030 in.).

Variable with temperature and wind flow, the design watt density would be in the 23 250 to 46 500 W/m2 (15 to 30 W/in.2) range. This heater is also normally designed to operate at full-line voltage of 440 or 220 V.

High-Temperature Furnaces with Open Elements. The temperature range of this type of furnace varies from approximately 675 to 955 °C (1250 to 1750 °F), and the furnace is normally a radiant-heating type.

Where large wall areas are available inside a furnace, a common method of mounting the nickel-chrome strip element is to attach it in a serpentine pattern from insulated alloy or ceramic anchors normally on the vertical walls only. With this design, especially at the higher temperatures, the structural strength of the element material and configuration for expansion must be considered. The element must support itself at operating temperatures without excessive droop or warping. Sufficient warping to cause the element to touch at various points could shorten the effective length of the element, decrease the resistance, and cause premature failure due to excessive currents and watt densities.

On larger furnaces with accessible wall areas, maintenance on this design element, although done from inside the "cold" furnace, is relatively easy. On smaller furnaces, accessibility for replacement of wall elements becomes a problem.

Other types of modular or drawer-type elements are available, which makes element removal and maintenance much easier.

The element strip material used with open elements is generally one of the following types: 80Ni-20Cr, 68Ni-20Cr, or 35Ni-18Cr-44Fe. The nickel-chrome element is generally selected in the heavier gage thicknesses and wider widths. A

typical cross-section range would be from 1.3 to 2.3 mm (0.050 to 0.090 in.) thick and 19 to 38 mm ( 34

to 1 12

in.) wide.

Variable with temperature and location of elements, the design watt density would be 12 400 to 23 250 W/m2 (8 to 15 W/in.2).

An alternative to the nickel-chrome strip element is the cast nickel-chrome heating element. This element has good structural strength, stability, and resistance to atmospheric attack. The quality control necessary in the manufacture of this element has made it slightly less flexible and popular than nickel-chrome strip. The casting must have uniform density and cross section to ensure a guaranteed resistance without danger of hot spots. Castings made with the investment casting (or lost-wax) process generally meet the quality requirements. Low voltages at the element and high currents tend to make the control hardware and wiring quite expensive. A common cast element material would be 35Ni-15Cr, and the watt-density range again would be 12 400 to 15 500 W/m2 (8 to 10 W/in.2).

Nonmetallic Resistance Heating Elements. In general, the nonmetallic heating elements are used in furnaces operating above 1010 °C (1850 °F). Silicon carbide elements are generally used in temperature ranges of 1010 °C (1850 °F) and above. They tend to be very fragile, so care should be taken in the design to allow for proper support and freedom for the element to expand and contract as the furnace is heated and cooled.

Silicon carbide elements undergo resistance increase with age; thus, to maintain constant power over the life of the elements, it is necessary to have a voltage adjustment available, usually with a step transformer. The useful life of a silicon carbide element is usually established at the point when its resistance has increased four times. To maintain constant power would mean that a total voltage demand of twice the initial voltage would be required because power, P, is equal to E2/R.

Silicon carbide elements are provided in various diameters and lengths, with published "hot" resistances. Design watt densities vary with such factors as temperature and atmosphere, but with silicon carbide elements, conservatively designed watt densities result in better element life.

In a sintering furnace operating at 1150 °C (2100 °F) with an endothermic atmosphere, a design watt density of 31 000 to 46 500 W/m2 (20 to 30 W/in.2) would be considered appropriate. Molybdenum disilicide elements are commonly formed in U-shaped rod configurations and normally are mounted vertically. The published maximum temperature-use range in air is above 1650 °C (3000 °F), which covers all furnace temperatures up through the forging range.

Element location for molybdenum disilicide elements is an important consideration because these elements are designed to operate at very high watt densities and related high thermal heads. These elements undergo a high resistance change from cold to hot, with resistance increasing with temperature. The control hardware and wiring must be properly designed to handle the high initial currents. A typical selection for a 955 °C (1750 °F)furnace with an endothermic atmosphere

would be a 9 mm (0.35 in.) rod element, rated at 244 900 W/m2 (158 W/in.2), with an element temperature of 1430 °C (2610 °F).

Advantages of Electrically Heated Furnaces. The following are advantages associated with electrically heated furnaces:

• Systems are clean and free of the pollution normally found with fuel-fired systems • Cooler plant environment without exhaust stacks and hoods for some furnaces (atmosphere furnaces

may still require stacks and hoods) • Quieter because of the absence of blowers and combustion noise • More uniform heat pattern from grid elements or from side-to-side of chamber temperature uniformity

on electric radiant tubes • No exhaust system required to affect building air pressures; no make-up air system required • Does not generally (though not always) require purge or flame safety systems • Electrical power available almost everywhere

Disadvantages of Electrically Heated Furnaces. The following are disadvantages of electrically heated furnaces:

• Inflexible system makes changing connected heating capacities or varying individual element capacities in the same zone difficult

• High initial equipment costs • With numerous electric furnaces in a plant and no form of peak-demand control, the user pays at a high

demand rate for all power • Higher operating costs • Cool-down times are longer because no combustion air is available • Nonmetallic elements tend to become brittle as they age and are subject to breakage from handling,

vibration, or shock

Radiant-Tube-Heated Furnace Equipment

With fuel-fired, radiant-tube-heated furnaces, the work chamber is protected from the products of combustion. With a radiant-tube furnace, the work chamber normally contains a controlled atmosphere as dictated by the process. There are, however, cases where the chamber remains filled with air, and the only purpose of the radiant tubes is to protect the work from the high dew point flue gas products. Electrically heated radiant tubes normally are used to protect the heating-element material from attack by the furnace atmosphere.

Gas-fired radiant tubes, as shown in Fig. 10, are by far the most common type of fuel-fired indirect method of heating. This is mainly due to the wide availability of natural gas. The proper selection of combustion system components such as burners, controls, orifices, and piping, will allow the same radiant tube to be fired with a wide variety of gases: natural gas, propane-air mix, straight propane, and certain low-Btu manufactured gases.

Fig. 10 Radiant tubes for indirect gas-fired furnace heating. Courtesy of Despatch Industries Inc.

Radiant-tube burners are of two basic types, sealed-head and open-type burners. A particular advantage of the sealed-head burner is that it is readily recuperable by using the products of combustion and an air-to-flue-gas heat exchanger to preheat the combustion air prior to its entering the burner. This advantage can result in considerable fuel savings (see the article "Energy-Efficient Furnace Design and Operation" in this Volume).

Open-type radiant-tube burners use room air for combustion, which is entrained into the radiant tube by an eductor on the exit end of the radiant tube. Open-type radiant-tube burners cannot be recuperated, although in most cases, they can be replaced with a sealed-head burner. Radiant tubes are normally constructed from centrifugally cast tubing of various

diameters, with wall thicknesses of 3 to 8 mm ( 18

to 516

in.). In many cases, tubes fabricated from 3 and 5 mm ( 18

and

316

in.) wrought alloy are used.

With the high metal temperatures attained in radiant tubes in most heat-treating furnaces, the alloys commonly used are the higher-grade cast alloys HT, HK, and NA22H, or the wrought alloys 330, 601, and Incoloy 800. Additional information on the materials for radiant tubes is given in the article "Heat-Resistant Materials for Furnace Parts, Trays, and Fixtures" in this Volume.

Oil-Fired Radiant Tubes. Straight oil-fired radiant tubes are somewhat uncommon in heat-treating furnaces and are used mainly where an adequate supply of gaseous fuel is not available. Radiant tubes equipped to burn both oil and gas are more common, with the oil, usually No. 2 fuel oil, used as the standby fuel.

High vanadium, potassium, and sulfur contents in oil have a great effect on the useful life of nickel-chrome heat-resisting alloys. With the high temperatures attained inside the radiant tubes acting as a catalyst, attack on the radiant-tube alloy is accentuated.

When oil-fired radiant tubes are used, the construction and types of materials used are similar to those described for the gas-fired radiant tubes. Recuperation is also possible with certain sealed-head oil burners. When recuperation is employed, the burner manufacturer should be consulted regarding possible damage to or problems with the atomizing system in the burner, from the preheated combustion air, normally supplied at 370 to 540 °C (700 to 1000 °F).

Electrically Heated Radiant Tubes. With this design, the radiant tube protects the resistance heating element from the furnace atmosphere. A common design uses nickel-chrome alloy rod inside the radiant tube. These rods are formed into continuous hairpin shapes supported and contained by ceramic spacer discs.

With the heating element contained in a tube, it is very important to conservatively select the heating element watt density. The design watt density of the internal element is a direct function of furnace temperature and varies from 18 600 to 46 500 W/m2 (12 to 30 W/in.2). Some designs use nickel-chrome strip material rather than rod, formed into similar hairpin patterns.

If a gas-fired "U"-tube furnace is to be converted to electrically heated radiant tubes, it is desirable to directly replace the "U" tube with two straight tubes, rather than a single straight tube, for improved uniformity and reduced watt density. Thus, the conversion from natural gas to electricity can be accomplished without loss in production capacity.

Other types of electric radiant tubes are available, where the radiant tube itself becomes the resistive heating element. As with cast elements, however, the quality and condition of the radiant tube will determine whether it functions properly as the resistance element. Silicon carbide elements have also been used inside radiant tubes to protect them from carburizing atmospheres.

General Furnace Maintenance

Maintenance on a furnace should be performed on a regular basis to prevent unscheduled shutdowns. In most plants, major maintenance is performed once each year, normally during the plant vacation period.

Because unscheduled maintenance is very disruptive to production, especially in plants without backup heat-treating capabilities, some plants prefer a number of smaller furnaces rather than a single large furnace capable of handling all production.

Many components of furnaces must be considered as consumable items, although lifespan normally can be predicted from accurate maintenance records. Many furnace owners regularly inspect and change internal furnace components such as radiant tubes, thermocouples and wells, retorts, and electric heating elements. Trying to get a few more months of life from a certain internal component could result in an unscheduled shutdown, and an extended loss of production can occur if a replacement part is not readily available. Most furnace equipment manufacturers provide a recommended list of spare parts that the furnace owner should maintain in stock to ensure reasonably uninterrupted production.

In addition to the consumable items that have to be regularly replaced, many furnace components must be adjusted and/or calibrated at regular intervals to maintain the efficiency and accuracy of the heat-treating operation--once each shift, daily, weekly, monthly, or annually. Components that require regular monitoring for adjustment and calibration are mainly those that control the quality of the heat-treating process, such as thermocouples, temperature- and carbon-potential-control instruments, and gas analyzers. For example, some furnace operators regularly change all thermocouples at specified intervals to avoid the gradual deterioration that occurs prior to indiscriminate failure.

In addition, maintenance required to minimize wear and thus prolong component life must also be considered. This form of maintenance usually consists of a well-planned and documented adjustment and lubrication schedule. Lubrication can be accomplished manually or with an automatic system.

Care should be taken in the proper selection of the various greases and oils to ensure compatibility with the various furnace components and furnace atmospheres. Excessive greasing should be avoided because it will limit the life of the bearing seals. Excessively long grease supply lines should also be avoided because grease may harden before it reaches the point of use. In addition, supply lines are not normally routed through high-temperature areas to the lubrication points.

On manual systems, lubrication points should be carefully coded. Many special greases are incompatible with each other, and lubricating components with the wrong grease can have disastrous results.

For major maintenance, such as complete refractory replacement or complete rebuilding of mechanisms and systems, it is normally best to consult the original equipment manufacturer. Because this type of maintenance usually occurs when the equipment is quite old, the manufacturer will normally make recommendations that will improve the equipment and upgrade the system to the present state of the art and to comply with current recognized industry safety standards.

Salt Bath Heat-Treating Equipment

W. James Laird, Jr., The Metal Works Industrial Furnaces, Inc.

Introduction

HEAT TREATERS can no longer use heat-treating methods that are only uniform, quick, efficient, and economical. They must also be environmentally responsible. Salt bath heat-treating methods are uniform, quick, efficient, economical, and environmentally responsible.

Salt baths are used in a wide variety of commercial heat-treating operations including neutral hardening, liquid carburizing, liquid nitriding, austempering, martempering, and tempering applications. Salt bath equipment is well adapted to heat treatment of ferrous and nonferrous alloys.

Parts that are heated in molten salt baths are heated by conduction; the molten salt bath provides a ready source of heat as required. Although materials being heated come in contact with heat through their surfaces, the core of a part rises in temperature at approximately the same rate as its surface. Heat is quickly drawn to the core from the surface, and salt baths provide heat at an equal rate over the total part.

Neither convection nor radiation heating methods are able to maintain the rate of heating required to reach equilibrium with the rate of heat absorption. The ability of a molten salt bath to supply heat at a rapid rate accounts for the uniform, high quality of parts heat treated in salt baths. Heat-treating times are also shortened; for example, a 25 mm (1 in.) diam bar can be heated to temperature equilibrium in 4 min in a salt bath, whereas 20 to 30 min would be required to obtain the same properties in either convection or radiation furnaces.

Salt baths are very efficient methods of heat treating; about 93 to 97% of the electric power consumed with a covered salt bath operation goes directly into heating of the parts. In atmosphere furnaces, 60% of the energy goes for heating, and the remaining 40% is released up the furnace stack as waste. Steels that are heat treated in molten salts typically are processed in ceramic-lined furnaces with submerged or immersed electrodes containing chloride-based salts.

Applications

Applications of the various furnace designs and auxiliary equipment to specific heat-treating processes, such as austempering and martempering, are described in this and other articles in this Volume. Basic advantages of salt bath treatment include surface protection and control of distortion.

Surface Protection. Parts immersed in a molten salt bath develop a thin cocoon of solidified salt, which can be easily washed from the surface after treatment. This surface protection afforded by salt baths can eliminate the formation of damaging oxide scales. Moreover, because salt baths do not contain the oxygen, carbon dioxide, and water vapor levels found in most non-vacuum (or atmosphere) furnaces, immersed parts are protected further from scale formation. Decarburization of steel parts from contact with oxygen and carbon dioxide are also eliminated by the use of molten salts. Vacuum furnaces provide similar advantages in surface protection.

Control of Distortion. Salt baths offer a way to minimize the bad effects of nonuniform heating, lack of support, and poor quenching that may cause size and shape distortion. Unlike parts in an atmosphere or vacuum furnace, parts immersed in molten salts are supported by the density of the medium. Due to its buoyancy, sagging or bending of the parts is minimized in a molten salt bath.

Heating in molten salts is also very uniform. The temperature uniformity in a molten salt bath averages ±3 °C (±5 °F) throughout the bath, depending on furnace design. The layer of solidified salt around a part can also protect the part from rapid initial heating and the resulting thermal shock. As the cocoon of salt melts, the part is gradually and uniformly heated, minimizing distortion and preventing cracking.

Selecting a Salt for a Given Application. Information concerning the various salts suitable for heat-treating furnaces is available from many sources, such as the many competent salt companies. Also, military specification MIL-10699 describes the salts in detail. When selecting a salt for a given application, the following must be considered:

• The salt must have the proper working range to meet the operating temperature requirements • The salt must have the proper melting point to avoid prolonged heat-up times for heavy loads • The salt must be compatible with other salts and oil used in the same heat-treating line • The versatility of the salts application • The ease with which the salt is washed from the work after heat treatment and affinity of the salt for

moisture

By balancing these factors, a salt best suited for a particular application can be chosen. Naturally, if a single salt must perform several functions, it will be necessary to make compromises and sacrifice some advantages to obtain the required versatility. Salts used in heat treating tool steels are described in the following example.

Example 1: Molten Salt Bath Treatment for Hardening High-Speed Tool Steel.

The use of molten salt baths for high-speed steel hardening has progressed during recent years. The greater degree of control and versatility, coupled with a simplicity of process, has been the chief reason for this progress. In addition to these factors, salt bath hardening ensures greater uniformity and rapid heating as well as freedom from scaling of the tools being heated.

The greater rate of heat transfer obtained through the use of molten salts permits the use of hardening temperatures approximately 15 °C (25 °F) below those in muffle-type furnaces, thereby reducing the possibility of sweating (melting of the surface) as well as undesirable grain growth. For best results, high-speed steel should be hardened in four steps. One procedure is described below.

1. Preheating. To safeguard against cracking and distortion, preheating tool steels before austenitizing is recommended. Small pieces of simple shape, which are not as susceptible to the damaging effect of thermal shock as are larger, more complex pieces, may not require preheating. Other more intricate shapes may require one or more preheating steps. Preheating is usually done in a ternary eutectic chloride mixture, which melts at approximately 541 °C (1006 °F) and is usable from 600 to 1010 °C (1100 to 1850 °F). These salts are formulated to preheat or harden steel without decarburization and with a minimum amount of sludge formation. Usually, the replenishment offered by mechanical dragout is sufficient to control bath chemistry. After a period of idling, however, it is usually necessary to rectify the bath as well as remove any sludge to be certain that the bath is neutral. The heating rate in this type of salt is rapid. It thoroughly heats most tools in 10 min or less. Prolonged heating times are not detrimental to steel and will help achieve thermal equilibrium negating the need for long austenitizing cycles.

2. Austenitizing. The most critical treatment in tool steel hardening is that of austenitizing. Austenitizing of high-speed tool steel is done at temperatures close to its melting point. Long heating times, or excessively high temperatures, will cause increased grain growth, distortion, loss of strength, and loss of ductility. Low hardnesses and low wear resistance will result from inadequate heating of the steel. Achievement of thermal equilibrium is a must before quenching to

eliminate damaging the workpieces, Heating for 1 min for each 6 mm ( 14

in.) of total steel section thickness being heated

is a good rule of thumb.

Austenitizing baths generally consist of anhydrous barium chloride, which begins to melt at approximately 960 °C (1760 °F) and has a working range of 980 to 1315 °C (1800 to 2400 °F). Neutrality of barium chloride operating above 1090 °C (2000 °F) is maintained by using a 50 mm (2 in.) diam carbon rod with a length of 300 or 600 mm (12 or 24 in.). The carbon rod reacts chemically to reduce metal oxides that are then deposited in droplets on the carbon rod. Frequency of immersion is usually 1 h, divided into 15 to 20 min intervals, for every 5 h of heat-treating operation. Too long an immersion will allow droplets of metal to flow together and run off the rod. The user should take caution not to allow this to happen, as it will recontaminate the bath.

Methyl chloride rectification is highly recommended in neutral salt baths operating between 1040 and 1315 °C (1900 and 2400 °F). The bath temperature should be kept below 1040 °C (1900 °F) while methyl chloride is being introduced. Methyl chloride is introduced under the surface of the bath to avoid a chemical breakdown above the surface of the bath, which would result in gas loss. The broken down methyl chloride supplies chloride ions, which combine with the harmful metallic oxide to form neutral metallic chlorides. Self-rectifying neutral salts are also available from most salt suppliers. It

is recommended that the users thoroughly familiarize themselves with the disadvantages and detrimental effects rectifying salts can have on equipment before deciding to use self-rectified salts.

3. Quenching. For the quenching bath, the use of water is recommended for most cold heading tools, while the use of triple eutectic chloride salts is suggested for high-speed steels. For quenching air hardening steels, it is recommended that a 540 °C (1000 °F) triple eutectic chloride salt be used to reduce the temperature of the steel to a point where the heat color has been taken out of the workpiece so that it will not rapidly oxidize when air cooled. These are typical salts mentioned previously for use as a preheating medium.

As a majority of high-speed steels are quenched to temperature between 540 and 700 °C (1000 and 1300 °F), the triple eutectic chloride salts will have sufficient fluidity at these temperatures to allow effective quenching without the serious corrosion resulting from the use of salts containing highly hygroscopic constituents. Lower melting point salts usually contain calcium chloride and, if allowed to remain on the workpiece until it is cooled to room temperature, may severely corrode the surface.

The exact quenching temperature to use is dependent chiefly upon the composition of the steel. In some instances, use of low quenching temperatures cause the salt to freeze on the surface. This condition is caused by contamination of quench salt by barium chloride carried from the high heat bath, which gradually raises the freezing point of the quench bath. Proper salt handling techniques will eliminate this problem.

4. Drawing or tempering is done to obtain the desired strength, hardness, and toughness by modifying the microstructure of the quench hardened tool steel. More than one tempering cycle may be required to alter the ausquenched and heterogeneous mixture of retained austenite, untempered martensite, and carbides. All retained austenite and untempered martensite will be transformed when tempered, using a few short cycles rather than one long one. Some steels may require three or four cycles before the optimum structure is obtained. Each cycle will require not less than one hour at tempering temperature. Tempering is accomplished in nitrate/nitrite mixtures, which melt at approximately 140 to 600 °C (290 to 1100 °F). Neutral chloride salts can be used to temper small parts and light loads, but they are not recommended for production unless the temper temperatures are above 600 °C (1100 °F).

Externally Heated Furnaces

Externally heated salt bath furnaces may be fired by gas or oil, or heated by means of electrical resistance elements. Figure 1 shows typical externally fired furnaces used in liquid carburizing applications. Pots may be press formed from a single piece of low-carbon steel or iron-nickel-chromium alloy; a composition of Fe-35Ni-15Cr is usually preferred for the latter. Less expensive welded pots may be fabricated from either of these materials.

Fig. 1 Externally heated salt bath furnaces for liquid carburizing

A flange usually supports salt pots; consequently, pot size is limited by the strength of the flange material. Round pots for gas- and oil-fired furnaces range from 250 to 900 mm (10 to 35 in.) in diameter and from 200 to 750 mm (8 to 30 in.) in depth; they are about 10 mm (0.4 in.) thick. Larger sizes have been built for special applications and have operated

successfully. Pots larger than about 350 mm (14 in.) in diameter and 450 mm (18 in.) deep are rarely used for electrical resistance furnaces. Although it is physically possible to support the bottom of a large pot on a refractory pier, excessive temperature gradients may result.

Gas- or oil-fired salt bath furnaces (Fig. 1 a) are generally lower in initial cost than electrode- or resistance-heated furnaces and are simple to install and operate. As described above, gas- and oil-fired salt bath furnaces also have larger salt pots than resistance-heated furnaces.

To contain the molten salt, fuel-fired furnaces employ a round or rectangular pot made of either steel or alloy. Heat is applied by two or more self-cooling burners that fire tangentially between the outer wall of the pot and the inner surface of the furnace lining. The hot gases are vented through a flue located near the top for atmosphere-type type burners, or near the bottom for pressure-type burners and atmosphere-type burners for which the flue is connected to a stack about 1 to 2 m (3.3 to 6.6 ft) high. The height and placement of the flue allows a negative pressure to be maintained within the firing chamber. Firebrick and additional required insulation lines the combustion chamber. A steel casing completely surrounds all sides of the furnace housing and provides adequate safety in the event of pot failure.

Electrical resistance furnaces (Fig. 1b) for neutral heating of liquid baths are less widely used than furnaces fired by gas or oil. A series of resistance heaters surrounding the salt pot heat these furnaces. For this reason, pot failure may result in the total destruction of the electrical heating elements. Operating temperatures below 900 °C (1650 °F) are used to reduce pot failure.

Pot Service Life. In a well-designed furnace, the life of a round alloy pot will vary with the maximum operating temperature approximately as follows:

Temperature

°C °F

Service life, mo

840 1550 9-12

870 1600 6-9

900 1650 3-6

In one installation, the placement of an additional control thermocouple in the combustion chamber to prevent the temperature of the chamber from exceeding 1095 °C (2000 °F) served to extend the life of high-temperature (HT) alloy pots to 2 years (previous life had been 6 months). Pot temperature was maintained at 900 °C (1650 °F) during a work week of 120 h (24 h/day, 5 days/week). Other factors affecting pot life are considered in the section "Design and Operating Factors."

Temperature of the salt is measured by a thermocouple and suitable pyrometer. Operating within the range from 790 to 920 °C (1455 to 1690 °F), the externally fired furnaces may vary as much as 10 °C (18 °F) above and below the set temperature when using on-off or high-low control systems. This is considered acceptable for many applications. Where closer control of the temperature is required, a proportional control system, which will hold temperature variations to less than ±5 °C (±9 °F), should be used.

Design and Operating Factors. In the design of fuel-fired furnaces, ample space must be provided for combustion so that the flame will not impinge on the pot. If flame impingement is unavoidable, the pot should be rotated slightly at least once a week. Rotating the pot and/or using a sleeve reduces local deterioration in the region of flame impingement and prolongs the service life of the pot. The combustion-chamber atmosphere also has important effects on pot life. A system with a control range from high-fire to low-fire is preferable to an on-off system because the latter allows air to enter the combustion chamber during the "off" portion of the cycle, thereby increasing the rate of sealing of the outer surfaces of the pot.

Electrical-resistance-heated furnaces should be equipped with a second pyrometer controller whose thermocouple is located within the heating chamber. This will prevent overheating of the resistance elements, particularly during meltdown, when the thermocouple that controls the temperature of the main bath is insulated by unmelted salt. Because heating elements and refractories are severely attacked by salt, all salt must be kept out of the combustion chamber. For this purpose, a high-temperature refractory fiber rope may be used to seal joints where the pot flange rests on the retaining ring at the top of the furnace.

Externally heated pots should be started on low fire (low heat input) regardless of the method of heating. Once the salt appears to melt around the top, heat can be gradually increased to high fire to complete meltdown. Caution: Excessive heating of the sidewalls or pot bottom during startup may create pressures sufficient to expel salt violently from the pot. For added safety, the pot should be covered during meltdown with either a cover or an unfastened steel plate.

The waste heat of flue gases may be fed to an adjacent chamber and used to preheat work. Flue gases should always be visible to the operator. The appearance of bluish-white or white fumes at the vent indicates the presence of salts within the combustion chamber; prompt action is required.

Advantages and Disadvantages. Because of the ease with which they can be restarted, externally heated furnaces are well suited to intermittent operations. Another advantage of furnaces of this type is that a single furnace can be used for a variety of applications by simply changing the pot for one containing the proper salt composition.

Externally heated furnaces do have several characteristics, however, that limit their usefulness in certain operations. They are less adaptable to close and uniform temperature control because the furnace dissipates heat by convection, creating temperature gradients in the bath. Also, the temperature lag of the thermocouple and the recovery time of the furnace may result in overshooting or undershooting the desired temperature by 15 °C (25 °F). In addition to requiring an exhaust system for generated flue gases, externally heated furnaces may overheat at the bottom and sidewalls in restarting, which creates a pressure buildup in the thermally expanding molten salt and may cause an eruption. Finally, externally heated furnaces are seldom practical for continuous high-volume production because of the limitations of pots with respect to size and maximum operating temperature. High maintenance cost is also a factor.

Immersed-Electrode Furnaces

Ceramic-lined furnaces with immersed (over-the-side) electrodes (Fig. 2), when compared to externally heated pot furnaces, have greatly extended the useful range and capacity of molten salt equipment. The most important of these technical advances are:

• The electrodes can be replaced without bailing out the furnace • Immersed electrodes allow more power capacity to be put into the furnace, thus increasing production • Immersed electrodes permit easy startup when the bath is solid. A simple gas torch is used to melt a

liquid path between the two electrodes, thus allowing the electrodes to pass current through the salt to obtain operating temperatures

Fig. 2 Internally heated salt bath furnace with immersed electrodes and ceramic tiles

Immersed-electrode furnaces, however, are not as energy efficient as submerged electrode furnaces. The area in which the immersed electrodes enter the salt bath allows additional heat loss through increased surface area. As exhibited in Table 1, the surface area of the salt bath (A) in the submerged-electrode furnace is smaller than the surface area plus the immersed electrodes (A + B) in the immersed-electrode furnace. However, a good cast ceramic and fiber-insulated cover placed over the bath and electrodes will reduce surface radiation losses up to 60%.

Table 1 Service life of electrodes and refractories

Operating temperature Service life, years

°C °F Electrodes Refractories

Submerged-electrode furnaces

Furnace A

535-735 1000-1350 15-25 15-25

735-955 1350-1750 6-12 6-12

955-1175 1750-2150 5-7 5-7

1010-1285 1850-2350 2-4 2-4

Furnace B

535-735 1000-1350 10-20 10-20

735-955 1350-1750 4-8 4-8

955-1175 1750-2150 3-4 3-4

1010-1285 1850-2350 1-3 1-3

Immersed-electrode furnaces

Furnace C

535-735 1000-1350 2-4(a) 4-5

735-955 1350-1750 1-2(a) 2-3

955-1175 1750-2150 12

-1(a) 1-2

1010-1285 1850-2350 14

-12

(a) 112

Furnace D

535-735 1000-1350 2-4(a) 4-5

735-955 1350-1750 1-2(a) 2-3

955-1175 1750-2150 12

-1(a) 1-2

1010-1285 1850-2350 14

-12

(a) 112

Furnace E

535-735 1000-1350 2-4(a) 4-5

735-955 1350-1750 1-2(a) 2-3

955-1175 1750-2150 12

-1(a) 1-2

1010-1285 1850-2350 14

-12

(a) 112

Note: Service life estimates are based on the assumption that proper rectification of chloride salts is being done, as well as routine unit maintenance and care.

(a) Hot leg only

Super-duty fireclay brick lines the immersed-electrode furnace. Approximately 130 mm (5 in.) of castable and insulating brick then surrounds the fireclay brick on five sides. Figure 2 is a schematic drawing of an immersed-electrode furnace with interlocking tiles and removable electrodes. The removable electrodes enter the furnace from the top, and a seal tile is located in the front of the electrodes to protect them from exposure to air at the air-bath interface. This protection helps prolong electrode life. Table 1 compares service life of electrodes and refractories for some basic furnace designs.

Over-the-top (or over-the-side) electrodes are usually built with laminated cold legs, and water cooling is always required. A typical life expectancy for electrodes operating in such a furnace at 840 °C (1550 °F) is approximately 6 mo to 2 y for over-the-top electrodes, compared to 4 to 8 y for submerged electrodes.

The salt is heated by passing alternating current through it with immersed electrodes. As a result of the resistance built up to passage of current through salt, heat is generated within the salt itself. This heat is quickly dissipated by a downward stirring action created by the electrodes. The electrodes are attached by copper connectors to a transformer that converts the line voltage of the plant to a much lower secondary voltage (approximately 4 to 30 V) across the electrodes. Temperature is measured and automatically controlled by a system containing a thermocouple, pyrometer, relay, and magnetic contactor.

The energy required by an immersed-electrode furnace is a function of:

• Furnace size necessary to hold the load and electrode well • The energy (Qw) needed to heat the load to the desired temperature. (The value of Qw is a function of

load mass, the specific heat of the load, and bath temperature) • Energy losses and safety factors

Once energy requirements are determined, then electrode number, size, and spacing can be determined.

Microcomputers are used to calculate the rate of heat generation per unit length of the electrode to ensure that the current is uniform from the top and bottom of the electrodes, taking into account the complexity of the current paths between the electrodes, the electromagnetic forces, and the circulation (influenced by the viscosity of the salt).

The electrode spacing is usually selected between 25 and 100 mm (1 and 4 in.); the height of the electrode should be smaller than the depth of the pot, the difference depending on electrode spacing. The electrode width is usually 50 to 75 mm (2 to 3 in.) and rarely exceeds 125 mm (5 in.). Transformer voltages usually range from 4 to 30 V, with the ratio of maximum to minimum voltage of a given transformer approximately 4.5 (Ref 1).

Steel-Pot Furnaces. Some metal-treating processes are performed in salt compounds that cannot be contained in a ceramic liner. For these applications, furnace manufacturers make use of a welded steel pot with immersed electrodes. This type of furnace is suitable for special applications such as case hardening in straight cyanide baths, tempering, and marquenching.

The steel pot often has a sloped back wall, which produces a bottom heating effect resulting in better circulation and uniform temperature. This is accomplished by sloping the electrodes shown in Fig. 3 and 4. As the current passes through the salt between the electrodes, the salt is heated, decreasing its density and causing it to rise toward the bath surface. Control of the rate of rise of the salt is effectively gained by decreasing the distance from the electrodes to the steel pot. At the lower extremity of the electrode, the current enters the metal pot upon leaving the electrode to follow a shorter path to the other electrode. This arrangement ensures current flow through the salt along the entire electrode length. Due to the close proximity of the lower portion of the electrode to the pot, most of the heating is done in the lower part of the bath. This is the desired method of heating any liquid.

Typical standard sizes

Working dimensions Temperature range

(A) Length (B) Width (C) Depth

Heating capacity

°C °F mm in. mm in. mm in.

Input, kW

kg/h lb/h

540-150 1000-300 457 18 457 18 610 24 25 45 100

540-150 1000-300 457 18 686 27 610 24 25 68 150

Fig. 3 Metal pot, immersed-electrode salt bath furnace for ferrous tempering and isothermal annealing

Typical standard sizes

Working dimensions Temperature range

(A) Length (B) Width (C) Depth

Heating capacity

°C °F mm in. mm in. mm in.

Input, kW

kg/h lb/h

955-650 1750-1200 305 12 305 12 455 18 25 34 75

955-650 1750-1200 305 12 455 18 610 24 40 68 150

955-650 1750-1200 455 18 610 24 610 24 75 159 350

Fig. 4 Metal pot, immersed-electrode salt bath furnace for liquid carburizing, cyaniding, and carbonate baths

The metal pots are made of either plain steel or hot-dipped aluminized steel, depending on the application. Thicknesses

range from 12 to 38 mm ( 12

to 1 12

in.). Reinforcing members for light plate, usually angular in shape, are welded from

the top. Where depth of the pot so requires, additional members are used at the midsection.

The pot is housed in an insulated 230 mm (9 in.) thick wall furnace either with a brick outside wall contained in a rigid welded steel frame or in a steel-clad frame, depending on personal preference. In either type of construction, the frame is self-supporting on a lattice formed by welding channels or beams to the underside of a steel base plate. The pot is supported on an insulated refractory pedestal.

Electrode Arrangement. Immersed electrodes are made of either mild steel or an alloy "hot" leg welded to a mild steel "cold" leg. As previously mentioned, these are shaped to follow approximately the slope of the pot wall. The portion of the electrode that crosses over the top of the salt bath and is connected to the plant power source is referred to as the cold leg. This is welded to the hot leg, the portion of the electrode that is immersed in the bath, with sufficient weld cross section to provide necessary current conductor capacity. The shanks are drilled and tapped at the tinned terminal connection end for water cooling when necessary. If the latter is not required, the electrical connection is water cooled. Suitable clamping devices are used to facilitate electrode replacement.

Electrode arrangements can vary as follows:

• Single-phase operation with metal or ceramic pots: Several electrode arrangements can be used, depending on the size of the bath. If only two electrodes are required, they are normally positioned on the sloped-wall side and at least 125 mm (5 in.) apart. Three electrodes are usually placed so that the center electrode, equal in size to two of the other electrodes, is used as a common conductor with equal current paths to each of the outer electrodes. More than three electrodes would be arranged in multiple groups

• Three-phase operation with metal pots: Three electrodes are used and spaced in a manner similar to the spacing described above. They are connected to three single-phase transformers that have Y-connected secondaries and delta-connected primaries. The current flows from the electrodes to the metal pot, which is the neutral point. Several variations of the three-phase connections are used, depending on the type of furnace and load requirements

All accessories, such as starting units, sludging tools, and secondary connectors, are the same for steel-pot immersed-electrode furnaces as for ceramic furnaces.

Advantages and Disadvantages. Immersed-electrode furnaces do not require the use of iron-chromium-nickel alloy pots.

These furnaces require minimum floor space and maintenance and can be used for all types of neutral salts. Electrodes made of alloy steel should have an average service life equivalent to that indicated for steel pots in the section "Pot Service Life." Worn electrodes can be replaced while the furnace is in operation.

Depending on the positioning of electrodes, control to within ±3 °C (±5 °F) is easily obtained with immersed-electrode furnaces. Heat is generated within the bath, and overshooting is readily avoided. These furnaces lend themselves to mechanization and are suitable for high-volume production in the range of 815 to 1300 °C (1500 to 2370 °F).

The depth of salt pots for immersed-electrode furnaces is not restricted for ceramic or ceramic-lined pots. Metal pots may be restricted to depths of about 0.6 m (2 ft). Pots may vary in length and width to suit requirements, and multiple pairs of electrodes can be installed to furnish the necessary heating capacity.

The immersed-electrode furnace is not recommended for intermittent operation. Depending on furnace size, reheating the salt charge may require a day or more. Pots are not intended to be interchangeable. Removal of the pot usually involves replacement of the surrounding insulation.

Reference cited in this section

1. V. Paschkis and J. Persson, Industrial Electric Furnaces and Appliances, Interscience, 1960 Submerged-Electrode Furnaces

Submerged-electrode furnaces (Fig. 5 and furnaces A and B in the figure to Table 1) have the electrodes placed beneath the working depth for bottom heating. Many submerged-electrode furnaces are designed for specific production requirements and are equipped with patented features, which offer certain economical and technical advantages. General characteristics of submerged-electrode furnaces include:

• Maximum work space with minimum bath area: The electrodes do not occupy any portion of the bath surface, so that they only come in contact with the salt. Bath size is consequently smaller, and electrode life increases many times over by incorporating unidirectional wear and eliminating excessive deterioration at the air-bath interface

• Circulation-convection currents: Bottom heating provides more uniform bath temperatures and bath movement through the use of natural convection currents

• Triple-layer ceramic wall construction: The temperature gradients through the wall cause any salt penetrating the wall to solidify before it can penetrate the cast refractory material that forms the center portion of the wall construction. The design requires from 5 to 8% of the initial salt charge to fill the ceramic pot. By comparison, in some designs 140 to 150% of the initial charge is needed to seal the ceramic walls of furnaces built with two layers of ceramic brick, backed up and supported by a steel plate. Salt penetrates the ceramic walls of any furnace and distorts the geometry of the walls. Reducing the amount of salt allowed to penetrate the ceramic walls aids in maintaining dimensions and in promoting a longer furnace life

• Electrode placement: Enclosing the electrode in a clear rectangular box, free of any protruding obstructions, eliminates any potential hazards to operating personnel during cleaning. Any sludge formed in the furnace is removed easily by operating personnel

Fig. 5 Internally heated salt bath furnace with submerged electrodes. This furnace has a modified brick lining for use with carburizing salts.

Frame Construction. A typical submerged-electrode furnace is made of brick and ceramic material reassembled, regardless of size, in a rigid, self-supporting welded steel frame (see, for example, Table 1). This frame consists of supporting channels or beams welded to the underside of a heavy steel plate that forms the frame base. To this base are welded lengths of heavy angle iron around the outside and on top of the plate. These pieces are notched to permit welding of the heavy angle-iron posts to the plate and vertical sides of the base-plate angle iron. Lengths of heavy angle iron are welded similarly to the top of the posts. When required, additional vertical reinforcing members are welded between the bottom and top pieces of angle iron, and prestressed horizontal members also are used to ensure that the refractory material cannot move after the furnace has been brought to operating temperature.

Brick Construction. Three types of refractory materials are commonly used in submerged-electrode furnaces. A typical design is shown by furnace A in Table 1.

Submerged-electrode furnace liners are constructed with 230 mm (9 in.) thick high-temperature burned bricks. Consisting of approximately 42% alumina and 52% silica, the brick material is used in standard brick sizes such as 60 by 115 by 230

mm (2 12

by 4 12

by 9 in.) and in various brick shapes, such as straights, flat backs, and splits. The bricks are laid with a

high-quality air-setting mortar that resists abrasion, erosion, and chemical attack by chloride, fluoride, and nitrate-nitrite salts. The mortar offers sufficient wear and corrosion resistance to be economically used with some salts containing cyanide. For straight cyanide or carbonate salts, a welded steel pot or a furnace with a modified brick lining (Fig. 5) is used.

The outer wall of the salt bath furnaces is made of a second-quality firebrick with the same dimensions as brick used for the liner. The important qualities of this brick are the strength of the material and uniformity in size and shape.

The inner castable wall is constructed with a maximum of refractory cement and aggregate that is poured between the liner and outer wall to form a 240 mm (9.5 in.) thick monolithic wall structure. This dimension provides a temperature

gradient sufficient to cause the salt to freeze in the wall, thus making the wall self-sealing. With this design, salt penetration into the wall amounts to less than 8% of the bath volume. The maximum temperature of the outside wall during furnace operation is 60 °C (140 °F).

Electrode Construction. The electrodes used in submerged-electrode salt bath furnaces vary widely in size and shape, depending on the geometry of the furnace and the power requirements. All of the electrodes are located near the bottom of the bath and are built into the wall (furnace A in Table 1) so only one face of the electrode is in contact with the salt. This placement leaves the bath area free of obstruction for ease of cleaning and eliminates the possibility of touching the electrodes to the work.

Alloy electrodes are made by welding a 1610 mm2 (2.50 in.2) alloy material to a mild steel backing, or by welding a 125 by 125 mm (5 by 5 in.) alloy material directly to the mild steel tank. The spacing between electrode pairs is usually 65 mm (2.5 in.), or 190 mm (7.5 in.). The spacing is fixed and nonadjustable. For this reason, computation of the secondary tap voltages is critical to the successful operation of the furnace throughout its lifetime.

The durability of typical electrode and ceramic components of submerged-electrode furnaces is described in Table 1. Alloy electrodes can be replaced with graphite electrodes, which are renewed as they become consumed without disconnecting them (Fig. 5) or shutting off the power.

Startup and Shutdown. The submerged-electrode furnace can be started by adding molten salt from another furnace or by using a gas-fired torch or electric starting coil to melt a pool of salt that will wet both electrodes and provide molten salt for the current path. After the current path has been established in the molten salt between the electrodes, salt may be added to bring the bath up to working level. Additional salt will be required to maintain this level because a small amount, approximately 5%, will seep into the brickwork and freeze.

If the furnace must be shut down, the molten salt should be bailed from the furnace before it freezes. However, if the salt is allowed to remain in the furnace, a resistance-heated starting coil should be submerged in the bottom of the furnace while the salt is still molten. This coil remains in the frozen salt and it is connected to the transformer leads to start up the furnace.

Newer designs have one pair of electrodes close to the surface of the bath. When the furnace cools, the surfaced electrode pair is exposed, thus simplifying startup.

Advantages and Disadvantages. In common with the immersed-electrode furnaces, submerged-electrode furnaces require minimum floor space and maintenance and are highly adaptable to mechanization.

Because the submerged-electrode furnace employs water to cool the electrodes and transformer, it may be operated at 50% overload without overheating the transformer, whereas the immersed-electrode furnace, being air-cooled, should not be operated at an overload above 10%.

Because a ceramic pot is used, unexpected pot failure is rare with submerged-electrode furnaces, and the furnaces can be rebuilt on a planned schedule during annual shutdowns. In common with other electrical equipment, submerged-electrode furnaces are at a disadvantage where electric power rates are high, but this can be overcome to some extent by working the furnace in nonpeak periods when lower power rates are applicable.

Because of the erosive effects on ceramic pots of water-soluble salts with high sodium carbonate or high sodium cyanide contents, submerged-electrode furnaces can be used with only low-cyanide, low-carbonate salts. Baths with high cyanide or carbonate salt require a modified basic brick. The furnace with modified brick and submerged alloy electrodes provides many years of service in noncyanide and cyanide operations. To increase furnace life, the furnace shown in Fig. 5 is recommended. This furnace has a modified basic brick lining for use with basic carburizing salts. The alloy electrodes are replaced with continuing graphite electrodes. The electrodes are renewed as they become consumed without disconnecting them or even shutting off the power.

Air-Quality Assurance

Salt bath furnaces that operate at temperatures above 650 °C (1200 °F) will fume. An open furnace containing a 50-50% NaCl/KCl mix, operating at 870 °C (1600 °F) at sea level, will fume at a rate of 0.2 kg/m2 per h (0.04 lb/ft2 per h).

Sodium chloride and potassium chloride are both edible; however, in large quantities they can be a nuisance. The best way to overcome this nuisance is to capture it at the source.

Figure 6 illustrates two ways of capturing fumes from a salt bath furnace. The 380 mm (15 in.) location of a capture hood (Fig. 6a) requires treatment of 200 m3/min (7120 ft3/min) of air and fumes, whereas a canopy hood (Fig. 6b) at 305 mm (120 in.) requires treatment of over 900 m3/min (32 000 ft3/min) of fumes and air. When the basket and parts are lifted from the salt bath, fumes are greatly increased, probably in proportion to the total surface area of the basket and parts exposed to air (plus the bath surface fumes). It is important to remember that the fumes coming off a salt bath are hotter and have more energy than fumes at standard temperature and pressure. To calculate the type and amount of ventilation required, consult Ref 2.

Fig. 6 Ventilation of a salt bath furnace with (a) a capture hood and (b) a canopy hood. The capture hood in (a) requires a ventilation rate of 200 m3/min (7120 ft3/min), whereas the canopy hood in (b) requires a larger ventilation rate of 905 m3/min (32 000 ft3/min). All dimensions given in inches

Reference cited in this section

2. Industrial Ventilation, 20th ed., American Conference of Governmental Industrial Hygienists, 1988 Isothermal Quenching Furnaces for Austempering or Martempering

Isothermal quenching furnaces are pot-type furnaces with salt agitation, cooling, and chloride-elimination features. As little as 10% chloride salt will cause the quench rate of a salt quench to be reduced by 50%. Isothermal quenching furnace systems were designed to eliminate the occurrence of chloride carryover from the austenitizing bath to the quench bath, through salt separation and uniform vertical lamellar flow agitation. The three most common approaches to alleviating the salt concentration are chemical, temperature, and gravity separation.

Chemical Precipitation. Chemical agents have been used to attempt to lower the solubility of the chloride salts so that they will precipitate in the quenching salt. When the salts settle to the bottom of the quench tank, they are removed as sludge. This method offers little success because the precipitate that forms is fine textured and buoyant and therefore tends to remain in suspension rather than to precipitate out.

Temperature Precipitation. The elimination of carryover salts has also been attempted by continuously pumping salt through a small auxiliary chamber whose temperature is maintained at a lower level than the main chamber. As the salt is processed through the auxiliary chamber, chlorides are continuously precipitated out.

Although this method appears practical, a fundamental error exists in its application. The salt is cooled by air blown through a space between the pot and the outer shell of the precipitation chamber. Air is blown through this space to maintain the temperature levels of the main chamber and precipitation chambers. The moving air cools the pot walls below the salt-precipitation point so that the salt freezes and cakes to the sides. Salt buildup continues until the bath is

unusable. Consequently, depending on the level of salt concentration, the bath would have to be shut down, possibly after only a few weeks of operation, to remove the remaining molten salt and chip away the caked salt.

Gravity Separation. This system of carryover salt removal also uses a two-chamber design. The caking problem is eliminated by heavily insulating the pot walls at all points and using an internal air-water heat exchanger. Because the pot walls and the salt are at the same temperature, there is no caking action. The chloride salts settle into an easily removable shallow pan at the bottom of the precipitation chamber, or, if they are fine textured and buoyant, the salts float to the top of the tanks and are easily skimmed off.

The main advantages of two-chamber gravity-separation equipment include:

• Easily removable variable-speed propeller-type agitator with suitable baffling to provide vertical lamellar flow within the quench area, therefore ensuring maximum quench power and minimum distortion

• Separate chloride precipitation chamber with adjustable weirs to maintain a low chloride level and subsequently high quenching power

• Easily removable internal heat exchanger to maintain quench temperature and precipitate chlorides • Easily removable settling pan to ensure maximum efficiency in removal of chlorides • Heavily insulated pot and precipitation chamber to eliminate salt caking on walls

Furnace Heating. Generally either gas or electricity may be used to heat isothermal quenching furnaces. When gas heating is desired, immersion tubes are recommended because they are usually made of mild steel and provide long service life.

Further, if the pot should develop a leak, the insulation and outer shell will contain the salt. Caution: If a furnace with an externally heated pot were to develop a leak, the nitrate-nitrite salt would drip on the hot refractory and may cause a fire hazard. One or more immersion tubes normally are used, depending on bath size. Generally, they will have nozzle-mix sealed-in burners and will be available to Factory Mutual or Factory Insurance Association specification.

Electric heating may be by one of the following methods, depending on the maximum operating temperature:

• Sheathed resistance strip heaters are mounted externally to the side walls near the bottom. Maximum operating temperature is 425 °C (800 °F). They are easily removable through the insulated plug-type door. Protection against overshooting is achieved by locating a sensing device close to the heaters. The sensors operate directly on line voltage

• Sheathed resistance immersion heaters have a maximum operating temperature of 425 °C (800 °F). They can operate without a transformer but are susceptible to premature burnout due to the sludge accumulation or operator tampering and abuse

• Immersed-electrode heaters operate in the same manner as electrode pot furnaces for carburizing and tempering

Furnace Construction. The pot is fabricated from firebox-quality steel plate, double welded inside and out and properly supported to maintain its shape. Steel plate offers adequate resistance to chemical attack by the standard alkaline nitrate-nitrite salts at normal austempering and martempering temperatures. The pot is insulated with 100 to 150 mm (4 to 6 in.) of slab-type mineral insulation to prevent the chloride-saturated nitrate salt from freezing to the side walls or the bottom. The insulation is externally contained by a continuously welded outer steel shell. The shell is reinforced to ensure retention of the original shape and dimensions throughout its designed operating temperature range.

Automatic and Semiautomatic Lines

The use of automated hoists makes possible the combination of austempering, martempering, and tempering or carburizing in one line. One or more hoists travel back and forth, automatically advancing the fixture carriers of work through the required stations.

The hoist movement is controlled by a solid-state programmable control with functions that would normally require hundreds of relays, counters, switches, and extensive wiring. Once programmed, the controller performs the desired commands and functions. Time cycles, sequences, drills, and skips are easily entered or changed to meet metallurgical requirements. For instance, parts can be programmed to be carburized, air cooled, washed, rinsed, and returned for unloading. A push-button command then returns the program to standard processing.

Parts suitable for fully automatic or semi-automatic installations are those that can be fixtured by wiring, racking, or placing in baskets and that do not present problems in either buoyancy or drainage.

Fluidized-Bed Heat-Treating Equipment

Revised by Robert F. Sagon-King, Can-Eng Ltd

Introduction

FLUIDIZED-BED TECHNIQUES are not new to the metalworking industry. A 19th century American patent describes the roasting of minerals under fluidized-bed conditions. Other established applications include potter's clay and miner's hydraulic slurries. Systems of fluidized solid particles, such as quicksand, occur in nature.

Early attempts to use fluidized beds in the heat treatment of metals were limited in the temperatures that could be employed. Electrically heated furnaces capable of maintaining fluidized beds at temperatures up to 500 °C (930 °F) could be produced commercially, but difficulties were encountered when attempts were made to attain higher temperatures. A principal problem was the high rate at which refractory distributors, which distribute the hot fluidizing gases, were consumed.

In early gas-fired fluidized-bed furnace design, gas entered the base of the container after being mixed with air to make it ignitable at the point of entry. With newer designs, the mixtures are introduced separately and thus cannot be ignited accidentally. This design eliminates the danger of explosion at the point of entry. The surface of the bed is heated first, and the heating of surface particles causes progressive ignition downward through the container until the entire contents of the bed achieves uniform heat-treating temperature. Newer furnace designs extend fluidized-bed technology into the higher temperature ranges (540 to 1040 °C, or 1000 to 1900 °F) required for most common heat treatments.

Principles of Fluidized-Bed Heat Treating

In fluidization, a bed of dry, finely divided particles, typically aluminum oxide in the heat-treating context, is made to behave like a liquid by a moving gas fed upward through a diffusor or distributor into the bed. A gas-fluidized bed is considered a dense-phase fluidized bed when it exhibits a clearly defined upper limit or surface. At a sufficiently high fluid-flow rate, however, the terminal velocity of the solids is exceeded, the bed goes into motion, and the upper surface of the bed disappears. This state constitutes a disperse, dilute, or lean-phase fluidized bed with pneumatic transport of solids. The general phases or stages of fluidization are shown in Fig. 1. Usually the aggregative or bubbling-type stage is used for heat-treatment processes.

Fig. 1 Various types of contacting in fluidized beds

Although the properties of solid and fluid alone determine the quality of fluidization (that is, whether smooth or bubbling fluidization occurs), many factors influence the rate of solid mixing, the side of the bubbles, and the extent of heterogeneity in the bed. These factors include bed geometry, gas-flow rate, type of gas distributor, and internal-vessel features such as screens, baffles, and heat exchangers.

Determination of Fluidization Velocity. In determining the quality of fluidization, a diagram of pressure drop (∆p) versus velocity (μ0) is useful as a rough indication when visual observation is not possible. A well-fluidized bed will behave as shown in the diagram in Fig. 2, which has two distinct zones. In the first, at relatively low flow rates in a packed bed, the pressure drop is approximately proportional to the gas velocity and usually reaches a maximum value (∆pmax) slightly higher than the static pressure of the bed. With an increase in gas velocity, the packed bed suddenly "unlocks" and becomes fluidlike.

Fig. 2 Pressure drop versus gas velocity for a bed of uniform-sized particles. Mmf, minimum fluidization velocity. Source: Ref 1

When gas velocity increases beyond minimum fluidization (μmf), the bed expands and gas bubbles rise, resulting in a heterogeneous bed. This is the second zone, in which, despite a rise in gas flow, the pressure drop remains practically unchanged. The dense gas-solid phase is well aerated and can deform easily without appreciable resistance. In its hydrodynamic behavior, the dense phase can be likened to a liquid. If a gas is introduced into the bottom of a tank containing a liquid of low viscosity, the pressure required for injection is roughly the static pressure of the liquid and is independent of the flow rate of the gas. The constancy in pressure drop in both the bubbling liquid and the bubbling fluidized bed may be taken intuitively to be analogous.

The diagrams in Fig. 3 show poorly fluidized beds. The large pressure fluctuations in Fig. 3(a) suggest a slugging bed. In Fig. 3(b), the absence of the characteristic sharp change in slope at minimum fluidization and the abnormally low pressure drop suggest incomplete contacting, with particles only partly fluidized.

Fig. 3 Pressure drop diagrams for poorly fluidized beds. Source: Ref 1

One of the most important factors influencing the quality of fluidization is the uniformity of gas flow across a constant pressure drop. Figure 4 illustrates this schematically.

Fig. 4 Quality of fluidization as influenced by type of gas distributor. Source: Ref 1

Temperature Effect on Minimum Fluidization Velocity. One of the most important parameters of a fluidized bed is the minimum fluidization velocity. In simplified terms, minimum fluidization velocity (μmf) approximates to a function of the square of the particle diameter (d) and a linear function of particle mass (p) as:

μmf ≅ d2p (Eq 1)

In the design of heat-treating furnaces, the effect of temperature must be considered. Figure 5 shows that the flow of gas required for fluidization decreases rapidly with increases in temperature.

Fig. 5 Effect of temperature on the flow corresponding to minimum fluidization for particles 0.1 mm (0.004 in.) in diameter having an apparent density of 2

Defluidization. One of the common concerns about fluidized beds is that, because of their principle of operation, they are not well suited for large, solid parts with horizontal surfaces that remain stationary in the bed. This a result of the incorrect belief that fluidization occurs only in a vertical direction. With parts of this type, a cap of nonfluidized particles collects on the horizontal surfaces, forming a thermal screen. The higher the temperature of operation, however, the greater the energy and agitation of the bed and the smaller the likelihood that the bed will collapse. Moreover, various methods can be used to overcome this apparent disadvantage, and these are designed into most fluidized beds. These methods are:

• Movement of the part being treated • Introduction of additional agitation in the zone of fluidization around the part, either by localized

injection of fluidizing gas or by careful design of the outline of the basket that holds the parts • Increased fluidizing velocity • A more favorable orientation of the part

Selective Heat Treatment. Bed collapse can be turned to advantage for special heat treatments in which one area of the path must be hard and tough and the remainder must be soft and more ductile, as in the case of the engineered parts of the shape described above. In this case, after uniform heating, the part is removed from a hot fluidized bed and partially submerged in a fluidized quenching bed, with the part to be hardened facing down. The top horizontal surface becomes covered with a cap of particles that form a thermal screen, which retards the vigorous cooling caused by the fluidized bed.

Reference cited in this section

1. R.W. Reynoldson, Controlled Atmosphere Fluidized Beds for the Heat Treatment of Metals, Heat Treatment of Metals, University of Aston in Birmingham, 1976

Heat Transfer in Fluidized Beds

An important characteristic of fluidized beds is high-efficiency heat transfer. The turbulent motion and rapid circulation of the particles in the fluid furnace provide a heat-transfer efficiency comparable to that of conventional salt bath or lead bath equipment.

The heat transfer coefficient of a fluidized bed is typically between 120 and 1200 W/m2 · °C (21 and 210 Btu/ft2 · h · °F). The turbulent motion and rapid circulation rate of the particles and the extremely high solid-gas interfacial area account for this feature. The following factors are important in heat transfer.

Particle Diameter. Of all the parameters that affect the heat transfer coefficient in fluidized beds, particle diameter exerts the greatest influence. Particle diameter is generally a compromise between conserving fluidized gas flows and avoiding entrainment or carry-out. Normally a sieve size of 80 to 100 grit is used.

Bed Material. The governing physical property of any bed material is its density. There appears to be an optimum density for bed materials: about 1280 to 1600 kg/m3 (80 to 100 lb/ft3). High-density materials tend to produce lower heat transfer coefficients and in addition require more power for fluidization. Carry-out problems occur with low-density materials. Other properties, such as thermal conductivity and specific heat, are less important.

Fluidization Velocity of Gas. It is essential to use the optimum flow rate, that is, one that provides the maximum heat transfer rate for a particular particle density and diameter. Generally, this flow rate is considered to be between two and three times the minimum fluidization velocity. Too high a velocity leads to particle entrainment, high consumption of fluidizing gas, and poor heat transfer; too low a velocity leads to poor heat transfer and lack of uniformity in processing.

Heating Rates. Relative heating rates of a 16 mm (0.6 in.) steel bar in salt, in lead, in a fluidized bed, and in a conventional furnace are illustrated in Fig. 6(a); relative cooling rates for air, oil, water, and a fluidized bed are shown in Fig. 6(b). Figure 7 presents heating and recovery rates for a fluidized bed. Results of both hardening and isothermal quenching of type D3 tool steel with salt baths and with fluidized beds are given in Table 1. The difference between the two installations in total time for final heating and holding resulted from a difference in preheating conditions.

Table 1 Comparison of the effects of hardening and isothermal quenching of type D3 tool steel in salt baths and in fluidized beds

Diameter of test pieces

Preheating temperature

Hardness, HRC Heating or cooling medium

mm in. °C °F

Total time for final heating and holding at 960 °C (1760 °F), min

At surface

At center

Salt bath 80 3.2 500 930 44 65.5 65

Fluidized bed(a) 80 3.2 490 915 51 65 65

Salt bath 40 1.6 540 1000 36 64.5 64

Fluidized bed 40 1.6 500 930 41 64.5 64

(a) Small parts of the same steel but with a diameter of 8 mm (0.3 in.) were treated at the same time; hardness of these parts was 66 HRC.

Fig. 6 Relative heat transfer rates. (a) Heating rates for 16 mm (0.6 in.) diam steel bars in lead, in salt, in a fluidized-bed furnace, and in a conventional furnace. (b) Quenching rates for 16 mm (0.6 in.) diam steel bars in air, in oil, in water, and in a fluidized-bed furnace. Source: Ref 1

Fig. 7 Recovery rates for 25 mm (1 in.) diam steel parts in a 0.3 m3 (10 ft3) fluidized bed furnace

Reference cited in this section

1. R.W. Reynoldson, Controlled Atmosphere Fluidized Beds for the Heat Treatment of Metals, Heat Treatment of Metals, University of Aston in Birmingham, 1976

Control of Atmospheres

A full range of atmospheres can be used within the work zones of fluidized beds. The volume of gas used is clearly dictated by particle size, temperature of operation, and optimum fluidization velocity. However, it can be shown that, with careful design and the use of low-cost carrier gases such as nitrogen, even low-temperature surface treatments can be both effective and economical. In addition, one of the major advantages of a fluidized bed is that expensive gas need not be consumed while there is no work in the bed. Atmosphere conditioning is rapid: within about 30 to 60 s after an inert gas is introduced into the bed, the purity of the atmosphere is equivalent to that of the gas supply. In fluidized beds, various types of atmospheres can be obtained, as discussed below.

Reducing or Oxidizing Atmosphere. Adjustment of a gas-air mixture to the bed so that it is either gas-rich or oxidizing causes some decarburization or oxidation reactions in the materials being processed (the gas-rich mixture produces somewhat less severe reactions). However, these are time-dependent reactions, and, because of the rapid heating rates of parts being processed and the subsequent short immersion times needed to obtain the correct structure and through hardness, little surface effect other than discoloration and slight scaling is exhibited in section sizes up to 25 mm (1 in.). For larger sizes, the user must be aware of surface reactions that can occur, particularly as the processing temperature increases. Figure 8 shows the relative decarburization bands for steels held in a fluidized bed.

Fig. 8 Representative decarburization bands for steel held in a fluidized bed. Steels used: type O1 and type D3 tool steels and 0.75% C plain carbon steel. Source: Ref 1

Neutral Hardening and Carburizing. Atmospheres for the neutral hardening of tool steels or the carburizing of low-carbon steels can be used for bed flotation. This practice allows oxygen-free heating of tool steels. However, care must be taken during the transport of workpieces to the quench tank to prevent decarburization or oxidation.

Reference cited in this section

1. R.W. Reynoldson, Controlled Atmosphere Fluidized Beds for the Heat Treatment of Metals, Heat Treatment of Metals, University of Aston in Birmingham, 1976

Surface Treatments

Fluidized beds, using atmospheres composed of ammonia, natural gas, nitrogen, and air, or similar combinations, are capable of performing low-temperature nitrocarburizing treatments equivalent to conventional salt bath processes or other atmosphere processes. High-speed steel tools oxynitrided in a fluidized bed are comparable to similar tools treated by the more conventional gaseous process. Carburizing and carbonitriding in a fluidized bed can yield results similar to those achieved in conventional atmosphere furnaces.

Mixtures of propane and air produced the results shown in Fig. 9, which compares the case depths obtained on SAE 8620 steel bearing rings carburized in a fluidized bed and by the conventional atmosphere process. An effective case depth of 1 mm (0.04 in.) was achieved in 1.5 h using the fluidized-bed technique. Developmental work on this process is still being performed, but sufficient knowledge exists to compare the mechanisms of conventional gas carburizing and the fluidized-bed process.

Fig. 9 Comparison of hardness profiles obtained by fluidized-bed and conventional gas carburizing. SAE 8620 steel, rehardened from 820 °C (1510 °F). Source: Ref 1

Conventional Gas Carburizing. Carburizing occurs through the catalytic decomposition of CO according to:

CO + H2 → CFe + H2O (Eq 2)

Propane enrichment aids this reaction according to:

C3H8 + 3CO2 → 6CO + 4H2 and C3H8 + 3H2O → 3CO + 7H2

(Eq 3)

Fluidized-Bed Carburizing. The relatively large volumes of propane consumed during fluidized-bed carburizing, together with high gas velocities, favor carburization by the thermal decomposition of propane to precipitate carbon in accordance with:

C3H8 → C ↓ + 2CH4 (Eq 4)

The amount of carbon precipitated is proportional to the number of carbon atoms in the hydrocarbon fuel gas; that is, propane forms more carbon than does methane. In addition, the purity of propane is important, especially with respect to unsaturated hydrocarbon content, which increases its carbon-forming capability.

The precipitated carbon reacts instantaneously with the oxidizing products of combustion:

C3H8 + 5O2 É 3CO2 + 4H2O (Eq 5)

to form carbon monoxide and hydrogen:

C + H2O → CO+ H2 and C + CO2 → 2CO

(Eq 6)

Carburization then proceeds by the catalytic decomposition of CO by H2 as in conventional carburizing. It is possible that carburization is further complemented by thermal dissociation of the methane formed during carbon precipitation:

CH4 → CFe + 2H2 (Eq 7)

The carbon potential of the atmosphere varies with the air-to-gas ratio. For each type of hydrocarbon gas (typically propane, methane, or vaporized methanol), a relationship can be established among air-to-gas ratio, temperature, and carbon potential. Control of the reaction and carbon potential of the atmosphere by conventional gas analysis is possible, and fluidized-bed furnaces are equipped with sample ports and probes so that suitable measurements can be taken.

Reference cited in this section

1. R.W. Reynoldson, Controlled Atmosphere Fluidized Beds for the Heat Treatment of Metals, Heat Treatment of Metals, University of Aston in Birmingham, 1976

Types of Furnaces for Heat Treating with Fluidized Beds

The type of fluidized bed most widely used for heat treatment is the dense-phase type, although units based on the dispersed-phase bed have been constructed, with particle circulation for the heat treatment of long, thin metal parts such as shafts and plates. In a typical dense-phase fluidized bed, the parts to be treated are submerged in a bed of fine, solid particles held in suspension, without any particle entrainment, by a flow of gas.

Liberation of adequate quantities of heat within fluidized beds is a prime consideration in adapting them for metal processing. Because transfer of heat from the bed to the workpiece is usually much more efficient than transfer of heat from the heat source to the fluidizing medium, the greatest difficulty is encountered in transferring suitable quantities of heat to the fluidizing medium. In addition, the major part of the heat loss from any practical fluidized system is the heat content of the spent fluidizing gas. In instances in which thermal efficiency is unduly influenced by this factor, recirculation of the fluidizing gas or installation of a recuperative system may be justified. Each has been used in practical applications. Heat input to a fluidized bed can be achieved by several different methods; the most accepted, however, are described in the paragraphs below.

External-Resistance-Heated Fluidized Beds. A fluidized bed contained in a heat-resisting pot can be heated by external resistance elements (Fig. 10). Waste heat recovery can be used to increase thermal efficiency, and the fluidizing gas can be maintained at any desired composition. Heat-up time from ambient to operating temperatures of 815 to 870 °C (1500 to 1600 °F) typically takes 3 to 4 h.

Fig. 10 Fluidized-bed furnace with external heating by electrical resistance elements

External-Combustion-Heated Fluidized Beds. A fluidized bed contained in a heat-resisting pot can be heated by external gas firing (Fig. 11). In this arrangement, a fuel-air mixture is introduced through a standard commercial burner. The burner can be controlled very accurately down to low temperatures for low-temperature tempering. The products of combustion are then removed by flue in the normal fashion.

Fig. 11 Externally gas-fired fluidized-bed furnace

Submerged-Combustion Fluidized Beds. The technique of submerged combustion consists of passing the combustion products directly through the mass to be heated. This method provides an excellent rate of heat transfer and is now well established for a wide range of liquid-heating applications, from the heating of swimming pools to the concentration of acid solutions. The application of this method to the heating of a fluidized bed requires that the burner be used such that it provides strong agitation of the suspended particles, thereby achieving the desired properties of excellent heat transfer and uniformity of bed temperature.

Equipment developed for this purpose consists essentially of a burner, two concentric tubes, and a particle separator. A suitable gas mixture is fed through the burner into the central tube, where it is ignited. The flame develops in the tube, and the combustion products escape at its lower end, where they impart heat to the suspended particles before moving up through the annular space between the two tubes. As they rise, a quantity of particles is entrained. These are separated from the gas stream by the deflector plate and fall back into the bed by virtue of gravity. Figure 12 shows a system that incorporates submerged combustion with a controlled atmosphere for the low-temperature treatment of metals.

Fig. 12 Controlled-atmosphere fluidized-bed furnace heated by submerged combustion. 1, burner; 2, combustion tube; 3, tube through which combustion gases and particles rise; 4, particle separators; 5, heat exchanger; 6, gas recycle compressor for fluidization; 7, distributor plate; 8, parts to be treated

Internal-Combustion Gas-Fired Fluidized Beds. A major development in the heating of fluidized beds occurred when an air-gas mixture was used for fluidization and was ignited in the bed, generating heat by internal combustion. Prior to this breakthrough, many technical difficulties prevented the use of this mode of fluidized-bed heating. A typical furnace design incorporating this technique is shown in Fig. 13.

Fig. 13 Gas-fired fluidized-bed furnace with internal combustion. 1, insulating lagging; 2, refractory material; 3, air and gas distribution box; 4, fluidized bed; 5, parts to be treated

The advantage of this system is that the bed is fluidized by burning gases, and thus the heat is generated within the bed. In gas-fired fluidized beds, the supporting gas or fluidizing medium is a near-stoichiometric mixture of gas and air. This combustible mixture is ignited above the bed and quickly imparts its heat to the particles, which in turn heat the incoming gas further down the bed. After a period, combustion takes place spontaneously within the bed and is complete within the first 25 mm (1 in.) of the diffuser once the spontaneous combustion temperature for the gas being used is reached. This temperature commonly varies between 600 and 800 °C (1110 and 1470 °F). If the vessel is well insulated, the bed temperature can rise to a theoretical combustion temperature, and heat-up times from cold to 800 °C (1470 °F) are

typically between 1 and 1 12

h. However, problems inherent to the basic technique are:

• The bed is fluidized by burning gases. To obtain good temperature control and optimum fluidizing conditions, however, it is desirable that the fuel input rate and fluidizing velocity be independently variable

• Combustion is somewhat unstable below the spontaneous combustion temperature • Very high temperatures can occur in the immediate vicinity of the distributor/diffuser tile. When the bed

is incorrectly fluidized so that this heat cannot be removed from the top of the distributor, theoretical flame temperatures are achieved with consequent deterioration of the distributor. The thermal stresses of expansion and contraction on the distributor tile at these high temperatures tend, even with the best fixing techniques available, to cause failure of joints, which enhances the problem

Two-Stage, Internal-Combustion, Gas-Fired Fluidized Beds. The basic problem of separating the control of heat input from the control of fluidizing velocity has been overcome in two alternative designs (Fig. 14). In both designs, the initial heat-up from cold to operating temperatures is carried out by two-stage internal combustion. A noncombustible mixture of gas and air is introduced beneath the distributor tile. Secondary air is added to make up a stoichiometric or slightly gas-rich mixture immediately above the tile by means of jet holes drilled into heat-resisting tubes. This is done to reduce the possibility of explosion and to avoid high flame temperatures at the surface of the tile. The technique has an

adverse effect on good fluidization, but this is unimportant during initial heat-up, in which the prime objective is to raise the temperature of the bed to operating temperature as quickly as possible. Once this has been accomplished, the remaining objective is to isolate the heat-up control from the control of the fluidizing velocity. This is achieved in two ways:

• Three-chamber design: In this design (Fig. 14a), the heat control outer chambers are separated from the treatment zone by a muffle. The fluidizing velocity and atmosphere are independently controlled in the inner chamber, while the outer two zones are still supplying heat by internal combustion. To achieve adequate heat input, fluidization levels in these outer chambers are above the optimum for heat transfer and surface reactions, but this is relatively unimportant

• Back-radiation design: When fuel-rich gases are permitted to burn by the injection of secondary air immediately above the control chamber of the fluidized bed, a back-radiation effect causes a rise in bed temperature. This design (shown operating in the heating/controlling and cooling modes in Fig. 14b and c) makes use of this effect and at the same time utilizes heat that is normally dissipated when gases are burned outside the furnace. It therefore uses fuel more economically. In principle, the gas-rich mixture is supplied to the central chamber, and extra air is added to produce stoichiometric conditions during initial heating of the bed. When cold work is loaded for treatment, the extra air is injected above the bed to produce a radiating flame and recover bed temperature. If bed temperature exceeds set temperature, the extra air is switched to the outside of the furnace wall to provide cooling and finally is mixed with the rich gas/air to produce combustion at the top of the specially constructed hood

Fig. 14 Two-stage, gas-fired, internal-combustion fluidized beds. (a) Three-chamber design. (b) Back-radiation design in heating mode. (c) Back-radiation design in cooling mode

Internal-resistance-heated fluidized beds are not accepted by users. The elements and work load will make contact if insufficient care is taken.

Applications of Fluidized-Bed Furnaces

The potential applications of fluidized-bed technology to heat treating are many. Figure 15 specifies those applications in which fluidized beds can compete with conventional furnaces.

Fig. 15 Fluidized-bed applications; decision model. Source: Ref 1

Applications of fluidized-bed furnaces to the heat treatment of metals include continuous units for all types of wire and strip processing (patenting, austenitizing, annealing, tempering, quenching, and so on) and all configurations of batch-type units for general heat-treating applications. A typical batch-type unit with an output of approximately 150 kg/h (330 lb/h) is available as a standard furnace. Using mechanical handling equipment, it can be automated into a continuous heat-treatment line. The following example describes one firm's decision to install fluidized-bed furnaces for heat treatment.

Example 1: Improved Turnaround Time with Fluidized-Bed Treatment.

A company specializing in the design and production of aluminum extrusion dies had relied on sub-contract heat-treatment facilities for the hardening of dies. The decision to install in-house facilities came as a result of difficulties in meeting the 7- to 14-day turnaround of dies required by customers. Previously, hardening, case hardening, and tempering had been done by salt bath immersion. After studying alternatives, the firm decided to employ the latest fluidized-bed technology. Approximately one year later, the firm installed a second fluidized-bed furnace and made available its surplus capacity to other firms on a subcontract basis.

Carburizing, Nitriding, and Carbonitriding. In recent years, design innovation has led to the use of fluidized-bed furnaces as a practical tool for carburizing, carbonitriding, nitriding, and nitrocarburizing processes. In this technique, 80 mesh or 180 μm aluminum oxide particles produce a fluidizing effect so that the bed behaves like a liquid. When gas or electricity is used as the heat source, the bed provides a faster heat transfer medium. This is provided with quench and tempering furnaces.

Previously, gas-fired internal-combustion units or submerged combustion units were used successfully to provide both heat source and fluidizing/carburizing medium. Recently, more attention has been directed toward the use of externally heated fluidized beds, which is claimed to allow greater control over the carburizing process as a result of separate heating and fluidizing functions (Ref 2, 3). The advantages of the fluidized-bed process include:

• High rates of heating and flow cause the utilization of higher treatment temperatures, which, in turn, provide rapid carburizing

• Temperature uniformity with low capital cost and flexibility is ensured • A fluid bed furnace is very tight; the upward pressure of the gases minimizes air leakage • The process produces parts with very uniform finish

References cited in this section

1. R.W. Reynoldson, Controlled Atmosphere Fluidized Beds for the Heat Treatment of Metals, Heat Treatment of Metals, University of Aston in Birmingham, 1976

2. A.J. Hicks, Met. Mater. Technol., Vol 15 (No. 7), 1983, p 325-330 3. K. Boiko, Heat Treat., Vol 18 (No. 4), 1986, p 65, 66 Operational Safety

As with all forms of gas heating, normally accepted safety devices are incorporated on the majority of beds presently manufactured. The "flexible-tile" concept ensures that any failure of joints does not influence the performance of the bed.

Parts carrying surface oil or moisture do not create an explosion risk because the contaminants simply vaporize and are removed with the waste gas, as in conventional furnaces. The heat transfer medium (aluminum oxide) is nonhazardous and as such is not subject to disposal restrictions.

Cleaning Operations

Fluidized solids are nonabrasive and non-corrosive and do not wet immersed objects. There is some drag-out loss of the aluminum oxide, however, because some particles accumulate on flat surfaces as work loads are removed from the fluidized bed. These particles can be removed in part by agitation, bouncing, or blowing with an air pipe. Particles can be reused by being dried, sieved, and returned to the bed. When parts already scaled or preoxidized are placed in a fluidized

bed, particles tend to adhere to the scale to a greater degree than if the workpieces were clean. These particles can be removed by water spraying.

Heat Treating in Vacuum Furnaces and Auxiliary Equipment

Revised by the ASM Committee on Vacuum Heat Treating*

Introduction

VACUUM HEAT TREATING consists of thermally treating metals in heated enclosures that are evacuated to partial pressures compatible with the specific metals and processes. Vacuum is substituted for the more commonly used protective gas atmospheres during part or all of the heat treatment. Furnace equipment used in vacuum heat treatment differs widely in size, shape, construction, and method of loading.

Although originally developed for the processing of electron tube materials and refractory metals for aerospace applications, vacuum furnaces are now employed in brazing, sintering, heat treating, and the diffusion bonding of metals. Vacuum furnaces also are used for annealing, nitriding, carburizing, ion carburizing, heating and quenching, tempering, and stress relieving. Furnaces for vacuum heat treating are equipped for workloads ranging from several pounds to 90 Mg (100 tons), and heated working chambers range in size from 0.03 m3 (1 ft3) to hundreds of cubic feet. Although most vacuum furnaces are batch-type installations, continuous vacuum furnaces with multiple zones for purging, preheating, high-temperature processing, and cooling by gas or liquid quenching also are used. Vacuum heat-treating furnaces also:

• Prevent surface reactions, such as oxidation or decarburization, on workpieces, thus retaining a clean surface intact

• Remove surface contaminants such as oxide films and residual traces of lubricants resulting from fabricating operations

• Add a substance to the surface layers of the work (through carburization, for example) • Remove dissolved contaminating substances from metals by means of the degassing effect of a vacuum

(removal of H2 from titanium, for example) • Remove O2 diffused on metal surfaces by means of vacuum erosion techniques • Join metals by brazing or diffusion bonding

Note

* Roger C. Anderson, Abar Ipsen Industries; David Scarrott, Scarrott Metallurgical; Roger Keeran, Metal-Lab Inc.; Walter Prest, Seco-Warwick Corporation; and Roger F. Carlson, Lindberg, A Unit of General Signal

Vacuum Measurements

A theoretical or ideal vacuum is an empty space that does not contain either vapors, particles, gases, or other matter and consequently has no absolute pressure. Because this condition does not exist, even in outerspace, an ideal vacuum cannot be achieved. Normally, when the term vacuum is used, it refers to an absolute pressure below that of the normal atmosphere.

The standard absolute pressure of the atmosphere at sea level, 45° latitude, and 0 °C (32 °F) can be expressed in various values and units:

1 atm = 760 torr = 760 mm Hg = 760,000 μm Hg = 29.921 in. Hg = 14.696 psia

The standard absolute pressure of the atmosphere is the reference or 0 gage pressure for a normal pressure gage. Hence, gage pressure is negative for a vacuum condition. For some technologies other than vacuum furnaces, a degree of vacuum is measured by pressure below gage pressure. It is important to know how degrees of vacuum are expressed in the various technologies.

Most vacuum furnace pressure levels are expressed in terms of absolute pressure rather than gage pressure. Normally the units of measure used are torr, mm Hg, or μm Hg. When vacuum furnaces are pressurized above atmospheric pressure, such as for gas quenching, the pressure is expressed in terms of bars. One bar is slightly less than one standard atmosphere of absolute pressure. A bar is equal to 14.50 psia, 29.53 in. Hg, 750 torr or mm Hg, or 750,000 μm Hg.

The vacuum or pressure value of Hg refers to the height of a mercury column sustained by the differential between standard atmospheric pressure and an attained level of vacuum (or, more accurately, partial pressure) or pressure level (above standard atmospheric pressure) being measured.

Table 1 compares vacuum and pressure to standard atmospheric pressure. The normal pressure range of vacuum heat treating should be noted.

Table 1 Pressure ranges required for selected vacuum furnace operations relative to standard atmospheric (0 gage) pressure

Equivalent pressures Gage pressure classification

Furnace application

Vacuum classification

Pa torr mm Hg(a)

μm Hg in. Hg psia(b) psig atm bar

Pressure quenching

. . . . . . . . . . . . 177.17 87.02 72.32 5.92 6

. . . . . . . . . . . . 147.65 72.52 57.82 4.93 5

. . . . . . . . . . . . 118.12 58.02 43.32 3.95 4

High gas

. . . . . . . . . . . . 88.59 43.51 28.81 2.96 3

Pressure

Gas

. . .

. . . . . . . . . . . . 59.06 29.01 14.31 1.97 2

Zero . . . . . . 1.01×105 760 760 7.6×105 29.92 14.696 0 1 1.01

Vacuum treatment

. . .

1.00×105 750 750 7.5×105 29.53 14.50 . . . 0.99 1

1.3×104 100 100 105 . . . . . . . . . . . . . . .

Negative

Normal backfill

Rough

1.3×103 10 10 104 . . . . . . . . . . . . . . .

130 1 1 103 . . . . . . . . . . . . . . .

13 0.1 0.1 100 . . . . . . . . . . . . . . .

1.3 0.01 0.01 10 . . . . . . . . . . . . . . .

Normal range

Soft

0.13 10-3 10-3 1 . . . . . . . . . . . . . . .

0.013 10-4 10-4 0.1 . . . . . . . . . . . . . . .

1.3×10-3 10-5 10-5 0.01 . . . . . . . . . . . . . . .

1.3×10-4 10-6 10-6 10-3 . . . . . . . . . . . . . . .

1.3×10-5 10-7 10-7 10-4 . . . . . . . . . . . . . . .

Maximum Hard

1.3×10-6 10-8 10-8 10-5 . . . . . . . . . . . . . . .

(a) Equal to 133.322387415 Pa, it differs from torr by one part in 7 × 106.

(b) psia = psig + 14.7 psi.

Comparison of Vacuum and Atmosphere Furnace Processing

In most heat-treating processes, when materials are heated, they react with normal atmospheric gases, which consist of approximately (by volume) 21% O2, 77% N2, 1% H2O vapor, and 1% other gases. If this reaction is undesirable, the work must be heated in the presence of some gas or gas mixture other than normal air. This is done in normal atmosphere furnace processing.

The gas or gas mixture may be varied to cause desirable reactions with the material being processed or it may be adjusted so that no reactions occur. At different temperatures, different reactions may occur with the work and furnace atmosphere. In most atmosphere furnaces it is not possible to change the atmosphere composition rapidly enough for optimum reactions or to control the atmosphere composition with the degree of precision required for some heat-treating processes. Vacuum furnaces allow gas changes to be made quite rapidly because they contain gases of low weight.

Vacuum furnace technology removes most of the components associated with normal atmospheric air before and during the heating of the work. An analysis of the residual atmosphere in a leakproof vacuum furnace at a vacuum of about 0.1 Pa (10-3 torr) indicates that less than 0.1% of the original air remains. The residual gases primarily consist of water vapor, with the remainder largely comprised of organic vapors from the seals, vacuum greases, and vacuum oils. The oxygen content at 0.1 Pa (10-3 torr) is less than 1 ppm. If all of the residual gas in the vacuum furnace were converted to water vapor, the water vapor content would be approximately 1.5 ppm, or equal to that of a gas with a dew point of about -80 °C (-110 °F). At a vacuum level of 10 Pa (10-4 torr), the equivalent dew point of gas is estimated to be approximately -90 °C (-130 °F) or less.

These low dew point equivalents compare favorably with the driest inert gases available from highly efficient gas dehydration equipment. With suitable vacuum pumping systems, the concentration of oxygen and water vapor can be reduced to lower levels than those achieved in inert or reducing-gas atmospheres.

After a vacuum heat-treating furnace has been evacuated, gaseous reactions such as those encountered with atmosphere heat treatment are virtually eliminated. Moreover, the vacuum extracts many gases, surface contaminants, and processing lubricants that would be difficult and costly to remove by any other method. Gases drawn from the metal surface into the vacuum surrounding the charge are trapped by the vacuum pumps and exhausted from the system as the work is being processed. This advantage of a vacuum system is of greater significance when parts with complex shapes, blind holes, or deep recesses are heat treated. A complete purging of such parts in a protective atmosphere requires an extended purging period. Even long-time purging, however, may not ensure the complete removal of entrapped air, other contaminants, or contaminants generated by reactions with the atmosphere.

When more thorough purging is required, the furnace can be evacuated with a simple vacuum system, and the enclosure or retort can be backfilled with the desired protective or reactive atmosphere (see the article "Furnace Atmospheres" in this Volume). This method markedly reduces the amount of protective atmosphere and time required to produce satisfactory results.

Volatilization and Dissociation

In a vacuum furnace, materials can be pressed at temperatures and pressures at which the vapor pressure of the materials becomes an important consideration. Vapor pressure, which is the gas pressure exerted when a substance is in equilibrium with its own vapor, increases rapidly with temperature because the amplitude of molecular vibration increases with temperature. Some molecules in the outer surface of the solid material have higher energies than others, and they escape as free molecules or vapor. If a solid substance is contained in an enclosure devoid of any other material, molecules will continue to escape from the solid surface until their rate of escape is exactly balanced by the rate of condensation or recapture of the gaseous molecules. The equilibrium pressure developed is the vapor pressure of the substance at that temperature. The vapor pressure of a metal is dependent on temperature and pressure only but the effect is time dependent.

It is normally desirable to use a vacuum-temperature combination that accelerates the desorption of gases without producing the vaporization of more volatile alloy constituents. Alloys with high concentrations of volatile elements, such as brass, are not heat treated in vacuum furnaces.

If brass is heated in a vacuum at a temperature of 540 °C (1000 °F) and a vacuum level on the order of 13 mPa (0.1 μm Hg), the zinc component will vaporize (volatilize) and the brass will eventually be converted to copper sponge. The zinc will deposit in the cold section of the furnace and can revolatilize on subsequent runs at higher temperatures, causing unwanted pitting or other surface reactions on the work load.

Metals such as lead, zinc, and magnesium have relatively high vapor pressures; if heated above a temperature at which the vapor pressure of the element exceeds the pressure in the furnace, they will evaporate or sublime rapidly. Thus, high-vacuum heat treatment is not applicable to some metals and alloys. To handle certain metals and alloys properly, either the pressure must be limited to the soft (fine) vacuum range (Table 1) or a backfill to a higher vacuum pressure level must be employed.

Alloys with lower concentrations of volatile elements can be processed in vacuum by using the backfill pressure of an inert gas such as nitrogen or argon that exceeds the sublimation pressure of the element at the temperature involved. A backfill pressure of a few hundred μm Hg at temperatures of about 980 °C (1800 °F) precludes the vaporization of elements such as chromium, copper, or manganese from steels processed at these temperatures.

For example, if pure manganese were heated to approximately 790 °C (1455 °F) at a pressure of 13 mPa (10-4 torr), it would vaporize. If the material were held at a higher temperature or lower pressure for an adequate period of time, the metal would become depleted and would eventually disappear, and the vapors would condense on the colder areas of the furnace and/or pumping system. Backfill or higher pressures greatly slow the rate of evaporation or volatilization. The vapor pressures of carbon and selected pure metals, as related to temperature, are shown in Fig. 1 and Table 2.

Table 2 Vapor pressures of various elements

Vapor pressure at Element

0.013 Pa 0.13 Pa 1.3 Pa 13 Pa 1.0 × 105 Pa

10-4 mm Hg 0.1 μm

10-3 mm Hg 1.0 μm

10-2 mm Hg 10 μm

10-1 mm Hg 100 μm

760 mm Hg 760,000 μm

°C °F °C °F °C °F °C °F °C °F

Aluminum 808 1486 889 1632 996 1825 1123 2053 2056 3733

Antimony 525 977 595 1103 677 1251 779 1434 1440 2624

Arsenic . . . . . . 220 428 . . . . . . 310 590 610 1130

Barium 544 1011 625 1157 716 1321 829 1524 1403 2557

Beryllium 1029 1884 1130 2066 1246 2275 1395 2543 . . . . . .

Bismuth 536 997 609 1128 699 1290 720 1328 1420 2588

Boron 1140 2084 1239 2262 1355 2471 1489 2712 . . . . . .

Cadmium 180 356 220 428 264 507 321 610 765 1409

Calcium 463 865 528 982 605 1121 700 1292 1487 2709

Carbon 2290 4150 2471 4480 2681 4858 2926 5299 4827 8721

Cerium 1091 1996 1190 2174 1305 2381 1439 2622 . . . . . .

Caesium 74 165 110 230 153 307 207 405 690 1274

Chromium 992 1818 1090 1994 1205 2201 1342 2448 2482 4500

Cobalt 1362 2484 1494 2721 1650 3000 1833 3331 . . . . . .

Copper 1035 1895 1141 2086 1273 2323 1432 2610 2762 5003

Gallium 859 1578 965 1769 1093 1999 1248 2278 . . . . . .

Germanium 996 1825 1112 2034 1251 2284 1420 2590 . . . . . .

Gold 1190 2174 1316 2401 1465 2669 1646 2995 2996 5425

Indium 746 1375 840 1544 952 1746 1090 1990 . . . . . .

Iridium 2154 3909 2340 4244 2556 4633 2811 5092 . . . . . .

Iron 1195 2183 1310 2390 1447 2637 1602 2916 2735 4955

Lanthanum 1125 2057 1242 2268 1381 2518 1549 2820 . . . . . .

Lead 548 1018 620 1148 718 1324 820 1508 1744 3171

Lithium 377 711 439 822 514 957 607 1125 1372 2502

Magnesium 331 628 380 716 443 829 515 959 1107 2025

Manganese 791 1456 878 1612 980 1796 1020 1868 2151 3904

Molybdenum 2095 3803 2295 4163 2533 4591 3009 5448 5569 10056

Nickel 1257 2295 1371 2500 1510 2750 1679 3054 2732 4950

Niobium 2355 4271 2539 4602 . . . . . . . . . . . . . . . . . .

Osmium 2264 4107 2451 4444 2667 4833 2920 5288 . . . . . .

Palladium 1271 2320 1405 2561 1566 2851 1759 3198 . . . . . .

Platinum 1744 3171 1904 3459 2090 3794 2293 4159 4407 7965

Potassium 123 253 161 322 207 405 265 509 643 1189

Rhodium 1815 3299 1971 3580 2150 3900 2357 4274 . . . . . .

Rubidium 88 190 123 253 165 329 217 423 679 1254

Ruthenium 2058 3736 2230 4046 2431 4408 2666 4831 . . . . . .

Scandium 1161 2122 1282 2340 1422 2593 1595 2903 . . . . . .

Silicon 1116 2041 1223 2233 1343 2449 1485 2705 2287 4149

Silver 848 1558 920 1688 1047 1917 1160 2120 2212 4014

Sodium 195 383 238 460 291 556 356 673 892 1638

Strontium 413 775 475 887 549 1020 639 1182 1384 2523

Tantalum 2599 4710 2820 5108 . . . . . . . . . . . . . . . . . .

Thallium 461 862 500 932 606 1123 660 1220 1457 2655

Thorium 1831 3328 2000 3630 2196 3985 2431 4408 . . . . . .

Tin 922 1692 1010 1850 1189 2172 1270 2318 2270 4118

Titanium 1250 2280 1384 2523 1546 2815 1742 3168 . . . . . .

Tungsten 2767 5013 3016 5461 3309 5988 . . . . . . 5927 10701

Uranium 1585 2885 1730 3146 1898 3448 2098 3808 . . . . . .

Vanadium 1586 2887 1725 3137 1888 3430 2079 3774 . . . . . .

Yttrium 1362 2484 1494 2721 1650 3000 1833 3331 . . . . . .

Zinc 248 478 290 554 343 649 405 761 907 1665

Zirconium 1660 3020 1816 3301 2001 3634 2212 4014 . . . . . .

Note: The vapor pressure of metals is fixed with probable values at a given temperature, and the temperature at which the solid is in equilibrium with its own vapor descends as the pressure to which it is exposed descends. For example: iron must be heated to 2735 °C (4955 °F) at atmosphere before its vapor pressure is greater than atmosphere (760 mm Hg); this point is reached at 1311 °C (2390 °F) at a pressure of 130 mPa (10-3 mm Hg).

Fig. 1 Vapor pressure versus temperature for carbon and various pure metals

Alloy Vapor Pressures. The vapor pressures of pure metals are constant, well-established values. The vapor pressure of a given alloy varies according to conditions. The vapor pressure of an alloy is governed in part by a law analogous to Dalton's law of partial pressures: The total vapor pressure of an alloy, under ideal conditions, is equal to the sum of the partial vapor pressures of its constituents. However, the partial pressure of each element in the alloy is lower than its normal vapor pressure and is proportional to its concentration.

In processing at temperatures where the vapor pressures of more volatile minor constituents are still in the micron range, alloys behave in accordance with Dalton's law. For example, if pure manganese is heated to 790 °C (1455 °F), its vapor pressure will reach 13 mPa (0.1 μm Hg), making it impossible to evacuate to lower pressures without evaporating all of the manganese. However, when manganese is alloyed with other elements, as a solid solution in iron, for example, its effective vapor pressure is lowered. The total vapor pressure for the alloy is the sum of vapor pressures of the individual elements multiplied by their concentrations in the alloy. The vapor pressure of manganese in a 1% Mn alloy at 790 °C (1455 °F) is about 0.13 mPa (10-6 mm Hg). When alloys such as stainless steel are processed at high vacuum levels (theoretically exceeding the vapor pressures of some of its pure metal components), the volatilization is only a few molecules thick. This volatilization tends to draw the stable elements with it in a complex molecular destabilization that results in a surface chemistry similar to that of the core material. It is this molecular surface activity that can remove thin film oxides even though their theoretical combined vapor pressure has not been exceeded.

Many metals form compounds by reaction with oxygen, hydrogen, and nitrogen. These reactions are usually exothermic, and the possibility for dissociation of the resulting compound increases with higher temperatures. Some oxides, such as water, vaporize at temperatures so low that dissociation occurs only in the vapor phase. For an oxide, nitride, or hydride that remains a solid over a wide range of temperatures, a dissociation pressure exists at any temperature that represents an equilibrium between the compound, gas, and the metal.

All metallic compounds decompose into constituent elements when heated to sufficiently high temperatures. However, many of the metal oxides are quite stable, requiring low pressures at high temperatures to effect dissociation. It is impractical to dissociate many of these compounds because of the combination of vacuum level and temperature required. When a metal oxide dissociates, the metal remains and the oxygen is evacuated.

For example, chromium oxide will dissociate in a 1.3 mPa (10-5 torr) vacuum at 1300 °C (2370 °F). The dissociation of a metal oxide usually depends more on temperature than on pressure. Most oxides can be dissociated under normal operating vacuum levels at approximately their reduction temperature in a highly reducing hydrogen atmosphere.

The nitrides and hydrides often have higher dissociation pressures, making many of them unstable when heated in a vacuum. For this reason, vacuum heat treating can be used both to dissociate these compounds and to remove the evolved gas without disturbing the base metal.

It is believed that when oxidized surfaces brighten during vacuum heat treating, the mechanism involved is not simply thermal dissociation of the oxide. Bright surfaces do not discolor, or become brighter, when they are exposed to a vacuum atmosphere that is theoretically oxidizing. A metal surface can be maintained almost free of visible oxidation at a partial pressure several decades higher than that suggested by theoretical calculations. The following theories have been proposed to explain this apparent anomaly:

• The solution and diffusion rate for oxygen exceeds its surface absorption rate • Oxide nucleation occurs at discrete sites rather than as a continuous film • The effective concentration of oxygen is reduced by carbon and hydrogen in the solid metal and by the

vacuum atmosphere

Heat Treating in Vacuum Furnaces and Auxiliary Equipment

Revised by the ASM Committee on Vacuum Heat Treating*

Vacuum Furnace Design

Although conventional atmosphere furnaces can be adapted for vacuum heat treating by adding a vacuum-tight retort connected to a suitable pumping system, furnace equipment developed especially for vacuum heat treating is generally used. There are two distinctly different types of vacuum furnaces: hot wall (no water cooling of the exterior walls) and cold wall (water-cooled walls).

Vacuum furnaces can be grouped into one of three basic designs:

• Top-loading, or pit, furnaces • Bottom-loading, or bell, furnaces • Horizontal-loading, or box, furnaces

Furnace designs can be varied to fit a wide variety of processing requirements by changing the chamber length or by adding internal doors, circulating fans, recirculating gas systems, and/or internal quenching systems.

Every vacuum furnace, regardless of its end use and basic hot- or cold-wall design, requires:

• Heating elements controlled to generate proper processing temperatures and cooling rates • Suitable vacuum enclosures with access openings • Vacuum pumping system • Instrumentation to monitor and display critical processing data

Production furnaces may be single-chamber units, batch-type units, or multichamber, semicontinuous units.

Hot-Wall Vacuum Furnaces

Vacuum furnaces are classified according to the location of the heating and insulating components. Hot-wall furnaces were the first type to be designed. Because of the demand of the heat-treating industry for higher temperatures, lower pressures, rapid heating and cooling capabilities, and higher production rates, hot-wall vacuum furnaces have become essentially obsolete--with the exception of low-pressure chemical vapor deposition (LPCVD) and ion-nitriding processes--and have largely been replaced by cold-wall vacuum furnaces.

The entire vacuum vessel is heated by external heating elements in the hot-wall construction. The heat is contained by insulation materials similar to the materials used in electrically heated heat-treating furnaces. Hot-wall furnaces have limited use because of slow heating and cooling capabilities. They are also limited in temperature because the strength of materials is reduced at elevated temperature. However, hot-wall equipment is readily adaptable to low-temperature operations not exceeding 980 °C (1800 °F), with moderate-sized chambers.

The double-pump modification of the hot-wall furnace permits the construction of larger vessels and the use of operating temperatures approaching 1150 °C (2100 °F). This system incorporates a second vacuum vessel outside the vacuum retort to maintain a roughing vacuum during the heating cycle. This removes the stress of the atmospheric pressure on the heated retort or vacuum vessel.

Bell-Type Furnace. A bell-type hot-wall furnace is shown in Fig. 2. The workload is placed on an elevated refractory metal hearth that rests on an insulated base clad with an alloy plate material. A water-cooled circumferential flange and vacuum gasket are located on the vacuum-tight base cover adjacent to the heated zone but in an unheated area. A retort made with a heavy-walled heat-resisting alloy covers the work load. A flange at the bottom of the retort fits on top of the base gasket to provide a vacuum-tight enclosure. The bell-shaped furnace equipped with internal electrical heating elements is lowered into position over the retort by a vertical hoist. The vacuum pumping system is connected through the insulated base.

Fig. 2 Bell-type hot-wall vacuum furnace

Because this furnace cannot be heated or cooled rapidly, even when the bell-shaped vessel is removed, production rates and the number of thermal cycles within a given time period are limited. Moreover, because the hot retort must support the entire pressure of the external atmosphere, its wall must be quite heavy. Practical operating temperatures for a furnace of this type are generally limited to approximately 925 °C (1700 °F).

Pit-Type Furnace. Figure 3 shows a pit-type hot-wall furnace. The work load is placed in a top-loading muffle or retort made from a heat-resisting alloy. The upper end of the retort is provided with a water-cooled flange and vacuum gasket that interlock with a flange on the upper part of the furnace above the heated zone. The muffle is lowered into the furnace by an overhead hoist, providing vacuum connections for the furnace and retort.

Fig. 3 Pit-type hot-wall vacuum furnace

With this construction, the space between the muffle and heating furnace can be evacuated by a roughing pump so that the pressure at the exterior surface of the muffle is essentially 0. This evacuation permits the use of a muffle with a much thinner wall and raises the maximum operating temperature of the system to approximately 1175 °C (2150 °F). When the heating cycle is completed, inert gas is bled into the retort, and air is bled simultaneously into the heating furnace so that the pressures remain balanced on both sides of the retort. The retort then can be removed from the furnace to a cooling stand, and another retort can be inserted in the hot furnace. This construction increases heating and cooling flexibility, which in turn increases cycle frequency and production.

Horizontal and vertical two-zone hot-wall vacuum furnaces are shown in Fig. 4 and 5. In both types, the heat-resisting alloy muffle is extended much further beyond the heating section of the furnace. This extended section has a water-cooled jacket to provide accelerated cooling. In the horizontal furnace, the charge is carried on an alloy hearth that can be moved in and out of the heated zone by a push rod extending through a seal in the outer fixed end of the muffle. This hearth has vertical heat baffles or multiple radiation shields at each end to confine the heat to the heated portion of the muffle. An equivalent means of transferring the work load is necessary in the vertical furnace as well, although this is not shown in Fig. 5. By using this technique, the charge can be cooled much faster because it is necessary to remove the heat from the hearth and work load only, not from the hot end of the muffle. This increased cooling rate permits the hardening of air-hardening steels and is adaptable to certain solution treatments not possible with other hot-wall furnaces.

Fig. 4 Horizontal, two-zone, hot-wall vacuum furnace

Fig. 5 Vertical, two-zone, hot-wall vacuum furnace

Cold-Wall Vacuum Furnaces

By far the most widely used, practical cold-wall furnace units consist of a water-cooled vacuum vessel maintained near ambient temperature during high-temperature operations. Consequently, because the operating temperature does not affect the strength of the vessel material, large units can be constructed for use at high operating temperatures.

In the cold-wall design, the water-cooled vacuum vessel contains and supports the internal insulation, the electrical heating elements, and the hearth upon which the work load rests. The vacuum acts as:

• A substitute for the normal heat-treating atmosphere to protect the work load • An insulating medium in the furnace because the thermal conductivity of a vacuum is essentially 0 • An effective protective coating around the heating elements, heat shields, and supporting hearth

There are three forms of heat transfer in a furnace--radiation, conduction, and convection--but the only effective method of heat transfer in a vacuum is radiation. Heat transfer by conduction or convection is negligible because little or no gas is generated.

The use of a vacuum as the insulating medium has permitted the use of multiple radiation shields of very low mass or special lightweight ceramic, graphite laminate, or felt insulations that facilitate rapid heating and cooling. Rapid rates of heating and cooling are important because usually each treatment cycle is started at ambient temperature and must be cooled to near ambient temperature at completion. As the protective medium, vacuum has permitted the use of materials such as graphite, tungsten, molybdenum, and tantalum for heating elements and hot furnace structures. Such materials normally cannot be used in other furnace constructions without more elaborate, expensive, and sometimes hazardous protective-atmosphere environments.

Cold-wall vacuum furnaces can be classified as either batch vacuum furnaces or semicontinuous vacuum furnaces. In a batch operation, the work load remains stationary inside the furnace during heating. On the other hand, in semicontinuous

vacuum furnaces with multiple chambers, the work load is moved within the vacuum usually after completion of a processing step or segment.

Batch Vacuum Furnaces. By using high melting-point materials in furnace structures, extremely rapid rates of heating and high temperatures can be attained in batch furnaces. Heat is transferred to the work load almost entirely by radiation.

Radiation cooling in a hot batch vacuum furnace is extremely slow, however. To reduce furnace time and shorten quenching time, pressure in the vacuum chamber is usually increased to either just below atmospheric pressure or up to six times the atmospheric pressure by introducing a pure inert gas such as nitrogen or argon. This gas is rapidly recirculated within the furnace and then through cooling coils or through an external heat exchanger and back into the furnace, with high-powered large-capacity gas pumps.

The advantages of batch cold-wall vacuum furnaces over atmosphere furnaces include:

• Reliability • Repeatability • Cleanliness • Bright, oxide-free treatment of most metals and alloys • Outgassing and purging of entrapped volumes of gas • Retained surface finish • Removal of surface volatiles • No heat added to local environment • No chemical effect on furnace or work during treatment • Easy control of furnace environment by controlling backfill gases • Highest-quality workpiece produced • No furnace conditioning • Instant pushbutton start from cold • Low pollution • Wide operating temperature range in one unit • Easy maintenance • Safe operation • Fully automatic processing • Wide range of programmable heating and cooling rates • Minimal distortion of treated work • High heating rates and temperatures resulting from use of high-melting-point materials • Complete shutdown when not in use; no need to maintain heat to maintain low dew point • Blanketing gas usually not required during heating

Disadvantages of batch cold-wall vacuum furnaces over atmosphere furnaces include:

• Large amount of floor space required in relation to load size • Lower productivity • Longer cycle times • Furnace must withstand atmospheric pressure and be free of leakage • High capital cost • High maintenance cost

Semicontinuous furnaces are constructed of modular units of three or more vacuum chambers. The heating chambers can be maintained at heat as required during loading and unloading. Each unit has a work carrier transfer system, an internal assembly of heating elements and shield package, pumping system, and temperature-controlled power system. Isolation locks or doors at each end of the vacuum heating or brazing environment separate these modular units from the entry and exit vacuum vestibules. These vestibules, in turn, have doors for access to and from atmospheric

pressure and other assembly operations. Usually included with the furnace are fans for fast cooling at the exit end of the furnace and an overhead conveyor system to transfer work carriers from the exit to the furnace entrance. Work carrier loading and unloading stations are incorporated in the external overhead conveyor system. A high-volume-production semicontinuous vacuum furnace is shown in Fig. 6. These furnaces are used for the fluxless brazing of aluminum heat exchangers at production rates of 100 to 250 parts per hour.

Fig. 6 Automatic vacuum furnace for fluxless aluminum brazing. Dimensions given in millimeters

For high-volume production and easy flow of work loads, semicontinuous vacuum furnaces are equipped with electrical controls that can also be computer programmed for automatic operation. The internal and external work carrier transfer systems, door operations, pumping systems, heating systems, gas backfill, and external cooling, loading, and unloading stations are all electrically interlocked and controlled for completely automatic operation. Automatically controlled mechanical loading and unloading of the work carriers can also be incorporated into the complete system.

Types of Cold-Wall Vacuum Furnaces. Cold-wall vacuum furnaces can be divided into bottom-loading, top-loading, and horizontally loading types.

Bottom-Loading Furnaces. As shown in Fig. 7, the furnace is stationary and elevated well above floor level. The bottom descends to floor level for ease of loading. The work is loaded on trays that are placed on the hearth by a fork lift when the bottom is in the lowered position. Such furnaces are built to handle large, heavy loads and are cooled rapidly by a high-velocity internal or external circulating-gas system.

Fig. 7 Bottom-loading cold-wall vacuum furnace

Top-loading furnaces, as shown in Fig. 8, are not as widely used as bottom-loading furnaces. However, they are useful in processing long and relatively thin workpieces such as slender shafts. The workpieces are suspended vertically from hangers attached to the removable top of the furnace or placed on the hearth in the base. Adequate head room and a vertical hoist are required. These furnaces are cooled in the same way as bottom-loading units.

Fig. 8 Top-loading cold-wall vacuum furnace

Horizontally Loading Furnaces. A box-shaped or rounded design horizontally loaded furnace consists of a cylindrical shell free of gas leaks, with circular convex end plates or doors (Fig. 9). In some designs, both of the end plates are hinged to permit easy access to the furnace interior. Furnaces are also constructed with a stationary rear plate or with a second hinged access door at the front that is smaller than the main front plate. The cylindrical shell and the end plates are water cooled by copper coils soldered to the exterior surface or through the use of double-walled construction. The gastight shell is made of stainless or carbon steel, depending on the intended use. The movable end plates are sealed by O-rings at the end faces of the cylindrical section. The pressure of the outside atmosphere on the convex ends supplies the pressure for vacuum sealing. Usually, auxiliary clamps are provided to supply sealing pressure and to prevent the door from becoming unsealed when positive pressure inside the furnace is used during inert-gas quenching.

Fig. 9 Horizontal vacuum furnace configurations. (a) Single-chamber vacuum furnace with gas/fan quenching. (b) Two-chamber vacuum furnace; one chamber for heating, with integral second chamber for gas/fan quenching only and mechanism for internal in-and-out transverse movement of work load to and from heat chamber. Unit can be loaded and unloaded only from cooling or quenching chamber. (c) Same as (b) but gas/fan cooling is also included in heat chamber and unit can be loaded or unloaded from either chamber. (d)

Two-chamber vacuum furnace; one chamber for heating, with integral second chamber for oil quenching only and mechanism for internal movement of work load to and from heat zone. Unit can be loaded/unloaded from oil-quenching zone only. (e) Same as (d) but gas/fan cooling is also included only in quenching or cooling chamber. (f) Same as (d) but gas/fan cooling is also included only in the heating chamber and unit can be loaded or unloaded from either chamber. (g) Same as (d) but gas/fan cooling is included in both chambers and unit can be loaded or unloaded from either chamber. (h) Three-chamber vacuum furnace; middle chamber is for heating only. One end chamber contains gas/fan quenching only and internal mechanism for work movement to and from heat zone. Other end chamber contains oil quenching only and an internal mechanism for work movement to and from heat zone. Unit can be loaded or unloaded from either end chamber. (i) Three-chamber vacuum furnace with middle chamber for heating only. End chambers both contain gas/fan quenching only and internal mechanisms for workload movement to and from heat zone. Unit can be unloaded or loaded from either end.

Many horizontally loading furnaces are equipped with a special lifting and transfer truck that is stationed in front of the furnace. Frequently, this truck rolls on a track so that it is permanently aligned with the heating chamber. A hydraulic fork lift raises the work basket, and the truck moves forward to transfer the basket into the furnace, where it is lowered onto the hearth or work pedestal. This mechanism avoids damage to the interior of the furnace, which could occur if the transfer were attempted without controlled movement.

Horizontally loading furnaces may have several chambers, depending on the heat-treating operation to be performed (Fig. 9). Special systems have been designed to transfer work loads inside these furnaces. The conveyor, walking beam, roller-hearth, and pusher-type furnace designs can be adapted for vacuum furnaces.

The hearth is supported on wheels that roll on rails that are installed below and outside the heated zone and are protected by movable heat baffles. The longitudinal motion can be supplied by a sealed push rod extending through the furnace wall to an air or hydraulic cylinder or by an internal chain-driven conveyor in the cool area.

Another method of work transfer within a horizontally loaded furnace uses an internal rack-and-pinion drive. An overhead chain-driven conveyor can also be used. The work trays may also be transferred longitudinally to the second chamber hearth by a lifting mechanism installed in the unheated chamber that is exposed to the furnace heat only during the short transfer time. Rack-and-pinion drives and pneumatic cylinders are often used to execute vertical elevator movements into and out of liquid quench tanks, as well as to open and close internal heat shields and vertical doors.

A horizontally loaded vacuum furnace equipped with radiation shields is shown, in a vertical section, in Fig. 10. The work load is exposed directly to radiation from the electrical heating elements. The multiple radiation shields are made of thin sheets of heat-resisting material, such as molybdenum, in parallel layers between the heating elements and the chamber shell. An alternative construction using thermal insulating material instead of radiation shields is shown in Fig. 11. This insulation may be a fiber fill, graphite felt, or special low-density fiber ceramic material.

Fig. 10 Radiation shield cold-wall vacuum furnace

Fig. 11 Insulated cold-wall vacuum furnace

Cross-sectional views of a three-chamber oil quench furnace are shown in Fig. 12. The front chamber is equipped with internal cooling coils and a circulating fan for accelerated gas cooling. The center chamber is the heating chamber, which can be sealed at both ends during the heating cycle by internal moving heat shields and doors equipped with O-rings. The third chamber contains the oil quench and the vertical transport system required to immerse the work in the circulated quenching oil.

Fig. 12 Three chamber vacuum oil quench furnace

Two-chamber oil quench furnaces are also common. One chamber is used for heating, and the other for loading and oil quenching. A double high elevator in the quench chamber allows loading from either end of the unit.

Heating Elements

Resistance heating and induction heating were formerly the two most commonly used methods of heating within the cold-wall furnace; induction heating is now rarely used. When vacuum furnaces are heated inductively, a graphite cylinder is used as a susceptor; the graphite is heated by induction and radiates the heat to the work inside the cylinder. When heating is provided by the more common resistance elements, the heat transfer is also completed by radiation; therefore, the active heating surface should be large enough to effect a rapid and uniform transfer of heat.

Essentially all vacuum furnaces use three-phase 60 Hz power supplies. Three types of power supplies and controls are used:

• Controllable variable reactance transformers • Saturable core reactors • Silicon-controlled rectifiers

Low voltage (generally, <70 V or a maximum of 100 V at a pressure of 13 Pa, or 100 μm) should be used in the vacuum chamber because a high electrical potential can produce short circuiting of the elements by ionizing the residual gases within the chamber.

Resistance heating elements operating in a vacuum do not require oxidation-resistant properties equal to those required in oxidizing atmospheres. To improve operating efficiency, resistance heating elements are heated to higher temperatures than are the elements used in conventional furnaces because the transfer of radiant energy is proportional to the fourth power of the absolute temperature. Higher temperatures require heating elements with low vapor pressures to ensure long life. Materials meeting these requirements are:

• Refractory metals, such as tungsten, molybdenum, and tantalum • Pure solid graphite in the form of bar, rod, or tube • Pure graphite cloth woven from fine filaments of pyrolyzed graphite

• Chromium-nickel elements for operating temperatures of less than 980 °C (1800 °F)

Properties of these materials are compared with those of iron in Table 3.

Table 3 Characteristics of heating elements used in vacuum furnaces

Vapor pressure at Melting point Upper operating temperature limit

1600 °C (2910 °F) 1800 °C (3270 °F)

Material

°C °F °C °F Pa torr Pa torr

Molybdenum 2617 4743 1705 3100 1.3×10-6 10-8 1.3×10-4 10-6

Tantalum 2996 5425 2500 4530 1.3×10-9 10-11 1.3×10-7 10-9

Tungsten 3410 6170 2800 5070 1.3×10-11 10-13 1.3×10-9 10-11

Graphite 3700 6700 2500 4530 1.3×10-11 10-13 1.3×10-8 10-10

Refractory Metals. The high melting points of tungsten, molybdenum, and tantalum make these metals ideal for use as heating elements in vacuum furnaces.

Tungsten is capable of withstanding higher operating temperatures than the other refractory metals (see Table 3). As a heating element material, it is used as wire or rod, thin sheet, or sections of woven wire screen. Wire screen is less likely to be damaged from thermal stresses that occur during heating or cooling.

Molybdenum in the form of solid rod, strip, or thin sheet material is the most widely used metallic element. Material in sheet form is normally preferred because the electrical power density (watts per square inch of radiating surface) is low compared to that of cylindrical rod, resulting in lower operating temperatures and thus longer service life for the elements. Also, thermal expansion and contraction and the resulting stresses are handled more easily from a design standpoint. However, thin sheets are subject to mechanical damage. Hangers and supports for metallic heating elements must have good insulating properties and must be chemically stable at the temperatures and pressures encountered in service. Heating elements must not be restrained by the support system from free movement during the thermal cycle.

Molybdenum exhibits extremely brittle characteristics at low temperatures after being in service. If deformed at low temperatures, molybdenum heating elements can fail because of the brittle nature of the metal.

Molybdenum undergoes a large change (500% increase) in electrical resistance between room temperature and the normal operating temperatures; consequently, the power supply must control the current during the early stages of heating to avoid damaging the elements, as well as the furnace heating systems.

Solid Graphite Heaters. All metals lose some strength when heated, whereas crystalline carbon in the form of graphite increases in strength as the temperature increases. Pure graphite in the form of flat bar or rod is less expensive than other high-temperature metallic resistors. Graphite also has a much lower heat expansion coefficient and is more resistant to thermal shock than most metallic materials.

As shown in Table 3, graphite also has a high melting point and a low vapor pressure; thus, it is an excellent choice for a vacuum furnace heating element material. In processes in which possible minute concentrations of carbonaceous material in the vacuum atmosphere will not have an adverse effect on workpieces, such as in the processing of the refractory alloys, where the use of graphite is not recommended (forbidden by most specifications), graphite resistors are commonly chosen. Moreover, the presence of incandescent carbon may provide additional cleansing or gettering action with respect to oxygen or water molecules in the residual vacuum atmosphere. This subtle action is not provided by any of the metallic resistors. The replacement of graphite resistors may be less expensive than the replacement of metallic ones because often graphite elements can be replaced without disturbing the surrounding insulation and because electrical connections are mechanical and require no welding.

Graphite has up to a 20% decrease in resistance as it heats, requiring wiring and power sources to be sized to handle the increased amperage should full power be applied to the cold elements.

Graphite Cloth Heaters. A third type of material used for vacuum heating elements is a cloth composed of fine graphite fibers. This material is made from rayon cloth pyrolyzed at high temperature to convert the carbon in the rayon to crystalline graphite. The cloth is strong and very flexible. It can be cut with ordinary scissors to the desired size and shape. Because the cloth is flexible, the supporting system can be considerably simplified. The ends are usually clamped in graphite electrodes. This graphite cloth is also available as a hollow cylinder with solid graphite ends of high electrical conductivity, forming a self-supporting electrode with a large radiating surface area.

The direction of the weave of cloth used in making elements is a critical factor in determining the resistance and voltage values of the graphite cloth. In most weave patterns, the fibers in one direction are straight, and those in the other direction are woven over and under the first. The actual cloth resistance obtained will exceed the calculated resistance value because the over-and-under cut length will exceed the straight cut length of the cloth.

Heat Insulation

Part of the insulation in a cold-wall vacuum furnace is provided by the vacuum itself. Where space is essentially void, there can be virtually no heating by convection or conduction. Radiant energy emitted by the resistors in all directions is generally confined to the desired heating zone by one of four designs of radiation shielding.

Metallic Shielding. Multiple concentric layers of thin heat-resisting sheet metal reflect the energy back into the heating

chamber of the furnace. Approximately 6 mm ( 14

in.) spacing is maintained between these sheets. Thin wire coil springs

are sometimes used as spacers between adjacent surfaces. Inner shields are sometimes fabricated from molybdenum because of its superior heat-resisting qualities; the next two layers may be stainless steel, and the outer layers are composed of nickel steel or plain carbon steel, depending on the maximum operating temperature of the furnace. As the number of shield layers (N) increases, the efficiency of the shielding increases; however, the insulating effect of each added shield decreases. Consequently, no more than five or six shields or layers are used. Approximately 85% efficiency can be achieved in this manner, even with a material of 0.5 emissivity. The effect on efficiency of the number of insulating layers and the emissivity of the sheeting material is shown in Fig. 13. Efficiency declines as the originally bright, clean metal surfaces become affected by the deposition of oxide or sooty films. This decline must be considered during the design of the furnace system.

Fig. 13 Relationship between heat shield efficiency and emissivity of sheet metal for various numbers (N) of sheets

Slightly faster evacuation rates, higher ultimate vacuum levels, and cleaner work products are usually obtainable with designs incorporating radiation shields. These advantages are realized only if the furnace is subsequently used for a singular process application or is scrupulously maintained. These advantages readily revert to disadvantages when radiation-shielded units are used for multiprocess purposes or are not cleaned.

Sandwich construction, which resembles the wall of a conventional low-temperature furnace, has proved to be an efficient design. One or more layers of graphite felt or high-temperature refractory fiberfill are placed between the inner and outer sheet metal walls. For instance, the inner wall may be molybdenum, and the outer wall stainless steel. A vacuum between the fibers reduces the thermal conductivity of the wool packing to a fraction of the thermal conductivity in any atmosphere. Thermally insulated cold-wall vacuum furnaces are better suited to conventional vacuum heat treating involving multiple process and temperature requirements.

Pump rates for sandwich construction furnaces are slower than for conventional heat-treating multiple-process furnaces. Ultimate vacuum levels are not as high. The residual-gas load is not as clean as with radiant-shield designs because of higher internal outgassing loads. However, for most vacuum processing, thermally insulated vacuum furnace design is adequate. It is certainly less expensive in original cost, maintenance, thermal efficiency, and power consumption than conventional vacuum heat-treating multiple-process furnaces. Typically a radiant shield furnace with identical pump system runs a decade vacuum level lower than does a similar insulated furnace, that is, 1.3 mPa (1 × 10-5 torr) compared to 13 mPa (1 × 10-4 torr).

Multilayer Graphite. In this design, multiple layers of high-purity graphite felt are fastened to an outer cage of refractory metal by molybdenum clips. This material has a density of only about 0.08 g/cm3 (0.003 lb/in.3), and it functions as a thermal insulator in much the same way as refractory fiber fill. Because it is a felted material with inherent cohesion, it does not require support by sheet metal walls, as does the fiber-fill. Its emissivity is near unity--about 0.98 or better. Graphite felt is more economical than molybdenum and is easy to replace. This method, like sandwich construction, relies on an insulating material with low conductivity in a wall heated from one side.

Carbon-Bonded Carbon Fiber. This insulation material uses blocks or molded cylinders to form a structural lining. It has a density of approximately 0.19 g/cm3 (0.0069 lb/in.3) and is used to make a refractory lining that minimizes the use of molybdenum and is resistant to high-velocity erosion in gas quenching applications.

Insulation Maintenance. All types of insulation are subject to accidental contamination. In many cases, foreign materials, such as lubricants used in deep-drawing operations, volatile metals, or even cotton gloves, may be inadvertently carried into vacuum furnaces with the workload. Contamination can result in extensive damage to radiation shields and

considerable down-time if it necessitates the mechanical cleaning of the shields. In an insulated furnace that does not depend on reflectivity, a thorough outgassing normally returns the furnace to operating condition within several hours.

The efficiency of multiple radiation shields decreases rapidly as the shields become dirty and nonreflective and as emissivity increases. One method of cleaning molybdenum shields consists of heating the furnace to above the operating temperature after backfilling with dry hydrogen to a pressure of 130 Pa (1 torr).

An effective method of protecting the innermost radiation shield is to overlay the molybdenum with a layer of graphite foil. This material is essentially pure graphite that has been rolled into a thin sheet by a special process with dry hydrogen. It is held in place by mechanical fasteners. Any contamination from the splattering of brazing filler metal or the evaporation of volatile materials is deposited on this inert material rather than on the molybdenum shield. Graphite foil can be applied to almost any other metal lining surface as well.

Gas Quenching

The inability to obtain the high cooling rates necessary for many metallurgical operations proved to be a severe limitation to early hot-wall vacuum furnaces. The development of a retort that could be transferred from the hot furnace to a special cooling stand provided a somewhat inconvenient solution.

With the development of the cold-wall vacuum furnace, the cooling problem has been overcome to some extent by radiation shielding made of multiple thin metal sheets, thermal insulation of very low mass such as refractory fiber, special foam refractory brick, or graphite fiber felt. Because of the low heat content of these insulating media, most of the heat to be dissipated is contained in the work load itself. Cooling the work continues to be quite a slow process because of the absence of convective heat transfer in a vacuum. Backfilling the chamber with an inert gas is necessary to promote the conductive transfer of heat from the work to the water-cooled shell. Fans made from stainless steel or other heat-resisting alloys can be installed above the insulated work area to promote the convective circulation of the backfill atmosphere. Usually baffles or bungs of insulating material must be moved from above and below the work load to improve the access of the cooling atmosphere to the hot work. Finned copper coils cooled by circulating water can be installed between the insulating medium and the shell of the furnace to facilitate heat transfer, to avoid total reliance on the cold furnace shell as the transfer surface.

Circulation by an internal fan is often used because it provides the fastest gas cooling, but it is not as likely to provide uniform cooling throughout the load.

Cooling rate uniformity is not entirely due to the system used. It also depends on the size, shape, quantity, and distribution of the load being processed. Care must be exercised when loading a furnace to optimize the placement of parts for heating and cooling.

An inert-gas circulating system with an internal manifold is shown in Fig. 14. To increase the rate of heat removal, the mass-flow coefficient must be increased. Mass flow is the product of the mass of gas moved times its velocity. To increase this coefficient, either the velocity of the gas circulating through the load or the gas pressure in the system can be increased so that a greater number of gas molecules are present per unit volume.

Fig. 14 Inert-gas recirculating system

Internal pressures used during inert gas quenching range from 13 to 86 kPa (100 to 650 torr) on furnaces that do not have positive-pressure clamps on the doors. For even more rapid gas cooling with the proper furnace construction, positive pressure up to six times atmospheric pressure have been used.

Because of technical advances in fan motors, door-sealing methods, and related equipment, it has become common to cool loads with gas fans at pressures up to 170 kPa (1275 torr) in conventional units and up to several times atmospheric pressure in furnaces with turbine recirculation systems. Because the cooling rate is proportional to absolute pressure, it is increased considerably by positive-pressure cooling.

By backfilling to pressures above atmospheric pressure, for example, 70 kPa (10 psi) gage, rather than slightly below atmospheric pressure, the cooling time can be shortened by as much as 30% (see Fig. 15). This development has greatly enhanced the use of vacuum equipment for air-hardening tool steels with decreased overall cycle time.

Fig. 15 Comparison of cooling rates and backfill pressures

The proportional increase of quenching speed with increasing gas pressure up to two times atmospheric pressure also applies to 500 kPa (5 bar) if the time delay in reaching this pressure and starting rapid gas circulation is minimal. Fast backfilling is promoted by a compact design and a special turbine system that builds up the forced circulation very rapidly. Figure 16 shows the effect of pressure on quenching speeds from a hardening temperature to 1010 °C (1850 °F). An increase in pressure from 100 to 200 kPa (1 to 2 bar) will decrease cooling time by 60 s, or 50%. An increase from four to five times atmospheric pressure would give a cooling-time decrease of 6 s, or 20%.

Fig. 16 Relationship between gas pressure and quenching rate

Liquid Quenching

When even faster cooling rates are needed, furnaces with liquid quench capabilities may be used. This usually requires an arrangement in which the quenching is done in a chamber isolated from the heating chamber. In one such design, a horizontal loading furnace has a rear heating chamber and a forward cooling chamber, with a vacuum sealing door between them. When the heating cycle is completed, the work is transferred to the front chamber. After the sealing door is closed, the work load is lowered into the quench tank by an elevator. The quenching liquid is agitated vigorously by propellers, and the heat absorbed by the quenching medium is removed by a heat exchanger similar to that used in many atmosphere heat-treating furnaces. The chamber that houses the quench tank can also be equipped with a circulating fan and cooling coils to provide forced convection gas cooling or liquid quenching. Quenching is usually completed at slightly below atmospheric pressure within an inert-gas backfill.

Precautions must be taken to ensure that the trays and fixtures that carry the work load into the heating chamber are degreased before reuse. Organic residues on trays and work load will outgas and make the pump-down time longer and give the interior of the furnace, and possibly the work, an undesirable appearance.

Gases Used for Backfilling

Gases used for backfilling vacuum furnaces are argon, nitrogen, helium, hydrogen, and natural gas. Argon and nitrogen can be obtained as compressed gases or as condensed liquids stored in cryogenic containers. Helium and hydrogen are readily available only as compressed gases. If appreciable amounts of argon or nitrogen are to be used, it is much less

expensive to purchase the gas in liquid form. Very high-purity gases are available, and some typical analyses of these gases are listed in Table 4.

Table 4 Typical analysis of backfill gases

Impurity, ppm Dew point

Thermal conductivity (k) at 0 °C (32 °F)

Gas Purity, %

O2 N2 CO2 CO H2 Hydro- carbons

Carbon- aceous gas

°C °F W/m · K Btu · in./ft2· h · °F

Relative cooling rate(a)

Argon 99.9995 2 2 1 . . . . . .

1 . . . -79

-110

5.77 32.3 0.74

Nitrogen 99.9993 3 . . . . . . . . . . . .

. . . 1 -79

-110

8.65 48.4 1.0

Helium 99.998 1 10 . . . . . . 1 1 . . . -62

-80 49.0 274 1.03

Hydrogen 99.9 10 1500 1 2 . . 25 . . . - -75 60.6 339 1.4

(a) Relative to nitrogen as 1.0

When a gas is purchased in a liquid state, it is stored under pressure in a large tank with a safety pressure relief value that allows gas to escape to the atmosphere if the internal pressure exceeds a set value. Liquid gas leaves the tank through a vaporizer, where ambient heat supplies the necessary heat for vaporization. Frequently, this vaporizer is a series of finned coils used to increase heat transfer. Because of the excellent insulation provided for the tank, usually very little gas is lost by venting. Sometimes a reservoir or storage tank of appreciable volume is used to store the vaporized gas at a nominal pressure to avoid sudden surges of pressure in the cryogenic system and to ensure that sufficient volume is available when backfilling is in progress.

Work Load Support

The work load in most vacuum furnaces is placed on a tray or in a basket to facilitate loading. Because such fixtures are heated and cooled at fairly rapid rates during processing, the material and design must allow for the cycles of thermal stresses. Molybdenum is often used for these support fixtures. If moderately high temperatures are used, as with many tool and die steels, austenitic stainless steel trays may be used. However, work baskets of Inconel alloys are often used in the processing of high-speed tool steels.

The work-supporting tray usually rests on a graphite or metallic hearth. Frequently this hearth consists of three or four horizontal molybdenum or graphite bars supported by heat-resisting piers from the furnace shell below. Some hearths use ceramic bars, and others use ceramic alumina wheels on which the work tray can be rolled.

It is essential that possible reactions between the hearth support and the work baskets be taken into consideration. Graphite hearths react with stainless steel and Inconel alloys to form a eutectic melting at approximately 1125 °C (2060 °F). To prevent a graphite hearth from reacting with a work basket or work load placed directly on the hearth, a thin sheet of a ceramic material capable of resisting high temperatures can be placed between the hearth and the workpiece. Some graphite hearth blocks have a longitudinal groove at the top. An alumina ceramic rod or tube is placed in the groove and supports the work or the work basket, thus separating the graphite and the metallic workpiece.

Molybdenum reacts with nickel to form a eutectic melting at approximately 1315 °C (2400 °F). Therefore, it is necessary to separate nickel-bearing alloys from a molybdenum hearth at this temperature. Slabs of honeycomb alumina ceramic are available for this purpose. Nickel and titanium form a eutectic melting at approximately 955 °C (1750 °F), and alloys of these metals must be kept separated if temperatures this high are contemplated. Table 5 is a listing of metal combinations and their compatibility.

Table 5 Maximum temperatures at which selected pure metals and metallic oxides are compatible in a 13 to 1.3 mPa (10-3 to 10-4 mm Hg) vacuum

Temperature of workpiece material

W Mo Al2O3 BeO MgO SiO2 ThO2 ZrO2 Ta Ti Ni Fe C

Support fixture material

°C °F °C °F °C °F °C °F °C °F °C °F °C °F °C °F °C °F °C °F °C °F °C °F °C °F

W 2540 4600 1925 3500 1815 3300 1760 3200 1370 2500 1370 2500 2205 4000 1595 2900 . . . . . . . . . . . . 1260 2300 1205 2200 1480 2700

Mo 1925 3500 1925 3500 1815 3300 1760 3200 1370 2500 1370 2500 1900 3450 1900 3450 1925 3500 1260 2300 1260 2300 1205 2200 1480 2700

Al2O3 1815 3300 1815 3300 1815 3300 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1815 3300 . . . . . . . . . . . . . . . . . . . . . . . .

BeO 1760 3200 1760 3200 . . . . . . 1760 3200 1370 2500 . . . . . . 1760 3200 1760 3200 1595 2900 . . . . . . . . . . . . . . . . . . 1760 3200

MgO 1370 2500 1370 2500 . . . . . . 1370 2500 1370 2500 . . . . . . 1370 2500 1370 2500 1370 2500 . . . . . . . . . . . . . . . . . . 1370 2500

SiO2 1370 2500 1370 2500 . . . . . . . . . . . . . . . . . . 1370 2500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1370 2500

ThO2 2205 4000 1900 3450 . . . . . . 1760 3200 1370 2500 . . . . . . 1980 3600 2205 4000 1900 3450 . . . . . . . . . . . . . . . . . . 1980 3600

ZrO2 1595 2900 1900 3450 . . . . . . 1760 3200 1370 2500 . . . . . . 2205 4000 2040 3700 1595 2900 . . . . . . . . . . . . . . . . . . 1595 2900

Ta . . . . . . 1925 3500 1815 3300 1595 2900 1370 2500 . . . . . . 1900 3450 1595 2900 2345 4250 1260 2300 1260 2300 1205 2200 1925 3500

Ti . . . . . . 1260 2300 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1260 2300 1260 2300 925 1700 1040 1900 1260 2300

Ni 1260 2300 1260 2300 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1260 2300 925 1700 1260 2300 1205 2200 1260 2300

Fe 1260 2300 1205 2200 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1205 2200 1040 1900 1205 2200 1205 2200 1095 2000

C 1480 2700 1480 2700 . . . . . . 1760 3200 1370 2500 1370 2500 1980 3600 1595 2900 1925 3500 1260 2300 1260 2300 1205 2200 2205 4000

Note: Although various materials may be compatible at a given temperature, a particular material may be unstable at this range of vacuum.

Vacuum Chambers

The prime requisites in the design, fabrication, and operation of a vacuum chamber are the prevention of leakage from the outside atmosphere and the assurance that the planned material processing will be completed without damage to the chamber or product. Leakage is a serious consideration even though the pressure within the vessel may be held at the required vacuum because the continual entrance of air into the evacuated chamber can be harmful to the product or internal components. If a high-capacity pump is used, the pump may well be able to maintain a low pressure inside the furnace regardless of a substantial air leak that causes a constant influx of oxygen, which can react with the surface of the work load. Many operators of vacuum equipment routinely conduct leak tests before energizing the heating elements. This is particularly useful when the work load consists of very expensive material. For example, a furnace is evacuated to 13 mPa (10-4 torr) for at least 1 h. The vacuum valve connecting the furnace to the pump is closed, and the rise in internal furnace pressure in a specified time interval is checked. It is difficult to set specific limits on leakage tolerance because this tolerance depends on the type of material being processed, the dew point of the air, and the length of time the furnace door is open. Common industrial practice with vacuum furnaces is to establish permissible leakage rates in terms of microns or micrometers per hour. A rate of rise between 1.3 and 3.3 Pa (10 and 25 μm Hg) per hour is generally an acceptable specification for most industrial work, but more stringent applications may require a lower rate. A more dependable method for discovering air leakage prior to or even during a run is to use a residual-gas analyzer. A residual-gas analyzer will indicate whether the leak is air, quench gas, water, solvent, or oil.

Generally, equipment manufacturers specify rate-of-rise measurements of 1.3 Pa (10 μm Hg) per hour or less for an empty, clean, cold, and thoroughly outgassed system. This measurement is usually obtained by checking the furnace after a gas cooling or vacuum cooling from a heat-up to a temperature sufficient to dry out or bake out the internal components. This procedure almost completely eliminates the out-gassing effects of work load and furnace components on rate-of-rise measurements.

Once a vacuum furnace has been properly installed and put into operation, the heat treater usually need only be concerned with preventing or correcting the leakage of air at seals around the doors, sight ports, pumping ports, electrical and water feed lines, mechanical rotating and sliding seals for introducing mechanical force and motion into the chamber, and any other point on the vessel that is subject to opening and closing.

Unless a vacuum furnace has been built for service at vacuum levels of 0.13 mPa (10-6 torr) or less, typical elastomer seals used in most applications will be worn out if subjected to overheating, to repeated openings and closings, or to rotating or sliding motions. When seals are replaced, all flanges must be properly aligned to ensure complete seating of the seal, and sharp edges or burrs must be removed to prevent physical damage. A good grade of vacuum grease lubricant should be used sparingly to facilitate placement of the seals and to obtain optimum sealing.

The best and most expensive vacuum chambers are constructed of an oxidation-resistant material such as a 300-series stainless steel. This material may be used for vacuum service at approximately 0.13 mPa (10-6 torr) pressure or less or when the vacuum chambers must operate as a hot-wall or retort unit. A less expensive carbon steel, adequately epoxy coated, is effective in cold-wall chambers for most heat-treating applications. However, unprotected carbon steel should not be used in the construction of a vacuum chamber. When carbon steel is exposed to a vacuum, it can be so thoroughly degassed that oxidation occurs very rapidly when reexposed to atmospheric conditions. Oxidation that occurs in a carbon steel chamber is permanent and progressive because the chamber is completely water-jacketed and does not achieve the temperatures necessary for surface cleaning.

The efficiency of the pumping system can be affected by the material used to build the vacuum chamber. For example, a stainless steel chamber with a total volume of 0.3 m3 (10 ft3) will require 10 min for evacuation to a pressure of 1.3 mPa (10-5 torr) by a 152 mm (6 in.) diffusion pump. A chamber with the same volume constructed of unprotected carbon steel can require evacuation times of up to 30 min to achieve the same pressure. If the carbon steel surfaces are oxidized, evacuation times of up to 2 h can be required.

Pumping Systems

Vacuum vessels are evacuated by various types of pumping systems that depend, to a great extent, on the pressure range needed for processing. An adequate vacuum pumping system must attain the specified pressure and must have sufficient capacity to handle the processing gas load, not only at the ultimate pressure but at all intermediate pressures during the pump-down cycle. Pumping systems are usually divided into two subsystems: the roughing pump and the high-vacuum pump. For certain requirements, a single pumping system is sufficient for the entire range and cycle. Pumps are usually

classified as mechanical or diffusion pumps. The choice of pump or combination of pumps depends largely on the pressure and gas volume or pumping rates required for a specific process and size of vacuum vessel. The vacuum system shown in Fig. 17 consists of a mechanical forepump that can be connected directly to the vacuum vessel by closing the high-vacuum valve, opening the roughing valve, and closing the foreline valve. This procedure isolates the diffusion pump from the rest of the system. The diffusion pump interior can then be pumped free of air by its mechanical holding pump.

Fig. 17 Typical vacuum system containing mechanical and vapor pumps with interconnected valving and piping

When the pressure as shown on vacuum gage 2 has been reduced to a level at which the diffusion pump can operate efficiently, the roughing valve is closed and the foreline valve and high-vacuum valve are opened. The residual gas in the vacuum vessel can then expand continuously into the region of the diffusion pump, then through the foreline to the forepump, and finally to the atmosphere. The water-cooled baffle or other optically dense trap on top of the diffusion pump prevents diffusion pump oil from diffusing backward into the vacuum system. Vapors from the system are condensed in the liquid nitrogen cryogenic trap, if one is used, when a very high vacuum (very low pressure) of 0.13 mPa (10-6 torr) or below is desired. Vacuum gage 1 measures the pressure in the vacuum vessel. This gage should be more sensitive than vacuum gage 2. The release valve controls backfilling of the system to atmospheric pressure up to the high-vacuum valve.

Mechanical pumps operate on the fluid-flow principle and are primarily positive-displacement pumps with suitable seals to permit operation at low pressures. Piston pumps or rotary blowers in various pumping-speed ratings are available. Vacuum system levels down to 3.3 Pa (25 μm Hg) can be obtained with oil-sealed rotary mechanical pumps. Depending on the application, they may be called roughing or forepumps. They can discharge directly into the air against normal atmospheric pressure. A portion of air in the closed system expands into and is trapped within a chamber of the pump. This volume of air is then compressed by the movement of vanes or a piston in the interior of the pump and is expelled through a port equipped with a check valve. This process is repeated and, with each cycle, a portion of the remaining air volume in the closed system is expelled. When the back leakage through the pump or the leaks in the wall of the container equal the rate of air removal by the pump, the closed-vessel internal pressure remains constant.

A nonmechanical limitation on the ultimate pressure of a mechanical pump is the vapor pressure of the oil itself. Most commercial pump oils are controlled grades of petroleum type SAE 20, which have a vapor pressure of about 130 mPa (1 μm Hg).

When air containing moisture is compressed in the interior of the pump, water may condense and contaminate the oil, affecting the attainable ultimate pressure. One method used to prevent condensation is to introduce enough pure, dry air at the beginning of the cycle to ensure that the resulting moisture level of the trapped air is below the level at which compressive condensation can occur. Air is introduced during the pump compression cycle to prevent condensation of water vapor to water and to revaporize the water during the expansion cycle. During the expansion, the water may be pumped out as a vapor but not as a liquid. Called gas ballasting, this method alleviates the moisture condensation problem.

Diffusion Pumps. When the pressure in the vacuum chamber becomes so low and the molecules so few that the path typically traversed by a gas molecule exceeds the dimensions of the chamber, the remaining gas molecules collide more often with the walls than with each other. At higher pressures, the constant and frequent collisions of adjacent gas molecules and the resulting elastic rebounding effectively scatter and expand the gas such that it quickly fills any new volume created. At lower pressures, this effect nearly disappears, and the remaining gas is difficult to pump with fluid-flow positive-displacement mechanical pumps.

At these pressures, it is necessary to allow molecules to diffuse randomly into the throat of the pump, and to impart a preferred direction of motion to the molecules by momentum transfer. For pumping at a vacuum system level below 130 mPa (10-3 torr), a vapor diffusion pump is generally used. Pumping action is directed by a high-velocity stream of heavy molecules in the form of a pump fluid, usually oil. The heavy molecules strike the gas molecules and push them in the desired downward direction toward the outlet of the pump. To be effective, the inlet pressure of the diffusion pump should be below 130 mPa (1 μm Hg) so that the vapor stream is operating in nearly empty space except for the occasional diffusing gas molecule.

A diagram of a three-nozzle vapor diffusion pump is shown in Fig. 18. Vapor from a liquid held in a closed boiler heated at the bottom is forced upward inside the boiler. The vapor passes quickly at a downward angle through narrow, circumferential openings in the nozzles. Molecules of gas that stray from the vacuum chamber above the pump toward the vapor jet streaming from the nozzles encounter the downward-directed stream of heavy molecules. The overall effect is to compress the gas molecules and force them downward to a point at which they can be removed by the mechanical forepump.

Fig. 18 Oil vapor diffusion pump

Pump vapor is condensed on the cooled inner walls of the pump and returns as a liquid to the boiler. Maximum velocity is imparted to the gas molecules by using a liquid composed of heavy molecules in the boiler. Efficiency is improved by using several nozzles in line, one above the other. Special highly stable liquids with very low vapor pressure are required for such diffusion pumps.

Backstreaming is the movement of the molecules of the pump fluid above the inlet flange of the diffusion pump and in the direction of the vacuum chamber. The rate of backstreaming increases rapidly as the inlet pressure exceeds 130 mPa (1 μm Hg) and depends on the size and type of diffusion pump. Backstreaming can also be caused by exceeding the fore pressure limit, which is about 65 Pa (500 μm Hg), of most diffusion pumps. Backstreaming can be reduced by interposing opaque baffles between the diffusion pump throat and the vacuum chamber. Most backstreaming originates at the top jet of the pump and can be reduced by placing a water-cooled cap above the jet. The oil vapor condenses on the cap and drips back into the pump.

By using a cold trap placed above the throat of the diffusion pump, it is possible to reduce oil backstreaming further. The cold trap consists of an arrangement of baffles and water-cooled or refrigerant-cooled walls that provide an opaque path through the trap. A refrigerant such as liquid nitrogen may be used. Molecules of oil condense on the surface of the cooled baffle and remain trapped. Such traps also attract condensable vapors that may be present in the vacuum chamber.

Water vapor, the most common contaminant present in high-vacuum systems, may be removed successfully in this way. The capability of the trap to serve as a pump for condensable vapors increases as the temperature of the refrigerant decreases.

Other Pumping Systems. Many types of vacuum pumps may be used, depending on the pressure to be maintained for a given vacuum process. The rotary mechanical pump and the oil vapor diffusion pump are typically used for most vacuum metallurgical processes. Other types of pumps include steam ejector, oil booster, liquid ring, turbomolecular, and cryogenic pumps. These pumps are used in vacuum processing for such functions as degassing and drying in conjunction with a rotary pump.

Steam ejector pumps have an operating range from 0.13 to 1.3 Pa (1 to 10 μm Hg). Although steam ejector pumps eliminate the need for a mechanical pump, they require a large volume of steam to operate. Because their principal advantage is the capability of removing large volumes of gas vapor at low vacuum levels, they are suitable for laboratory work and for very large vacuum vessels, especially if noxious vapors are to be pumped.

Oil booster pumps have an operating range from 130 to 0.13 Pa (1000 to 1 μm Hg) and require a backing pump. These pumps tend to introduce the excessive backstreaming of oil vapor into the vacuum vessel and are used mainly when high pumping capacities are desired above 1.3 Pa (10 μm Hg). They can be used in conjunction with an oil diffusion pump to achieve lower pressures while maintaining high pumping capacities.

Cryogenic pumps condense gas molecules on refrigerated surfaces at -195 °C (-320 °F) or lower. These pumps are regenerated periodically by heating the condensing surfaces to vent the accumulation of condensed gases. Cryogenic pumps are often used instead of diffusion pumps to pump high volumes of water vapor and other condensable gases in the 130 to 0.013 mPa (10-3 to 10-8 mm Hg) range. Cryogenic pumps capture the gas instead of pumping it. Consequently, they must be shut off occasionally for gas removal.

Gas Flow Valves. Gas flow of the vacuum pumping system is controlled by three specific types of valves, as shown in Fig. 19. High-vacuum valves isolate the oil diffusion pumps from the vessel, and butterfly and gate valves are used as roughing, foreline, and holding-line valves. All of these valves have apertures that provide minimal gas flow impedance and special seals for vacuum service.

Fig. 19 Typical vacuum valves

Process Control Instrumentation

Except for vacuum-gaging instruments, the instruments used in vacuum equipment processing are similar to those used in other heat-treating operations. Once vacuum control has been attained, heating cycles can be initiated, either manually or automatically, by a temperature-sensing instrument.

Temperature Control Systems

Temperature control systems consist of a primary sensing device, the control instrument, and the final control element.

Pyrometers. The total radiation pyrometer is adaptable to automatic control, but its accuracy may be impaired by intervening media such as gas, smoke, or a discolored sight-glass window. Sublimation of materials within the furnace may cause the deposition of metallic vapors on the sight-glass window, thereby reducing the radiation reaching the detector.

Thermocouples are typically used as temperature-sensing devices, although their performance varies with the environment (from a vacuum to an oxidizing atmosphere). Unless the hot junction of the thermocouple is attached to the part being measured, heat transfer to the thermocouple is based almost completely on radiation. In air or other furnace atmosphere, a thermocouple receives heat by conduction and convection. For this reason, the response time of a thermocouple in a vacuum is slower than in air. A change in air gap of 0.03 mm (0.001 in.) between the hot junction of the thermocouple and the part being measured can change the response time of the thermocouple significantly.

Control thermocouples made from nickel-nickel molybdenum are satisfactory for many heat-treating applications up to 1290 °C (2350 °F) (see Table 6 and Fig. 20). Noble metal thermocouples such as platinum-platinum rhodium are used up to 1650 °C (3000 °F) (see Table 7 and Fig. 21). Refractory metal thermocouples such as those of tungsten-rhenium are used up to 2205 °C (4000 °F) (see Table 8 and Fig. 22). Thermocouples are sometimes used with unsheathed (bare wire) hot junctions, reducing lag time tremendously. In many applications, particularly with platinum, thermocouples are sheathed in an adequate ceramic or metal protection tube. Below 1095 °C (2000 °F), sheathed base-metal thermocouples can be used to measure different locations in the load.

Table 6 Properties of 19 alloy/20 alloy nonstandard thermocouple

Thermoelement Property

Negative 19 alloy

Positive 20 alloy

Nominal composition Ni-1Co Ni-18Mo

Melting point, °C (°F) 1450 (2640) 1425 (2595)

Specific gravity 8.9 9.1

Thermal conductivity, W/m · K (Btu · in./ft2 · h · °F) at 20 °C (68 °F) 50 (350) 15 (105)

Coefficient of thermal expansion, μm/m · K (μin./in. · °F) (20 to 100 °C, or 68 to 212 °F) 13.6 (7.56) 11.9 (6.62)

Magnetic susceptibility Magnetic Magnetic

Resistivity, nΩ · m at 20 °C (68 °F) 80 1650

Temperature coefficient of resistance, μΩ/Ω · °C (20 to 100 °C, or 68 to 212 °F) 3050 290

Tensile strength, MPa (ksi) 415 (60) 895 (130)

Yield strength, MPa (ksi) 170 (25) 515 (75)

Elongation, % 35 35

Table 7 Properties of standard thermocouples

Melting point Maximum temperature

Type Thermo- elements

Base composition

°C °F

Resistivity, nΩ · m

Recommended service

°C °F

KP 90Ni-9Cr 1350 2460 700 K

KN 94Ni-Al, Mn, Fe, Si, Co 1400 2550 320

Oxidizing 1260 2300

RP 87Pt-13Rh 1860 3380 196 R

RN Pt 1769 3216 104

Oxidizing or inert 1480 2700

SP 90Pt-10Rh 1850 3362 189 S

SN Pt 1769 3216 104

Oxidizing or inert 1480 2700

B BP 70Pt-30Rh 1927 3501 190 Oxidizing, vacuum, or inert 1705 3100

Table 8 Properties of tungsten-rhenium thermocouples

Thermocouple type

W versus W-26Re

W-3Re versus W-25Re

W-5Re versus W-26Re

Nominal operating temperature range, °C (°F)(a) 2760 (5000)

2760 (5000) 2760 (5000)

Maximum short-time temperature, °C (°F) 3000 (5430)

3000 (5430) 3000 (5430)

Approximate thermoelectric output, μV/°C (μV/°F)

Mean, over nominal operating range 0 to 2315 °C (32 to 4200 °F)

16.7 (9.3) 17.1 (9.5) 16.0 (8.9)

At 2315 °C (4200 °F) 12.1 (6.7) 9.9 (5.5) 8.8 (4.9)

Nominal melting temperature, °C (°F)

Of positive thermoelement 3410 (6170)

3360 (6080) 3350 (6062)

Of negative thermoelement 3120 (5648)

3120 (5648) 3120 (5648)

Stability with thermal cycling Good Good Good

High-temperature tensile properties Good Good Good

Stability under mechanical working Fair Fair Fair

Ductility (of most brittle thermoelement) after use Poor Poor to good depending on degree of vacuum

Poor to good depending on degree of vacuum

Resistance to handling contamination Good Good Good

Extension wire Available Available Available

(a) Preferential vaporization of rhenium may occur when bare (unsheathed) couple is used at high temperatures and high vacuum. Vapor pressure of rhenium at operating temperature and vacuum should be checked before bare couple is used.

Fig. 20 Thermal electromotive force (emf) of 19 alloy and 20 alloy versus Pt-67

Fig. 21 Thermal emf curves for Instrument Society of America (ISA) standard thermocouples. Thermal emf plots are based on the International Practical Temperature Scale (IPTS) 1968 (amended 1975).

Fig. 22 Thermal emf of tungsten-rhenium thermocouples

Because rapid heating rates can be achieved in most vacuum furnaces, it is important to determine temperatures at various locations in the work load. For instance, a large die may overheat in certain locations within the furnace. Information on heating uniformity is also required when processing smaller parts loaded in baskets.

To bring thermocouple lead wires through the exterior shell of the furnace, special feed-through fixtures are provided. These fixtures are electrically insulated and vacuum tight, and they constitute a permanent portion of the furnace. The ends of the lead wires inside the furnace have ceramic-insulated quick-disconnecting terminals to which workpiece thermocouples can be connected. Type K chromel-alumel thermocouples are often used to monitor workpieces if the temperature to be measured is not above 1175 °C (2150 °F) and the time at temperatures above 980 °C (1800 °F) is not excessively long (see Table 7 and Fig. 21).

Additional information is available in the article "Furnace Temperature Control" in this Volume.

Pressure Control Systems

Instruments used to measure and record the pressure inside a vacuum processing chamber fall into two classifications: those that measure the pressure hydrostatically, and those that sense some physical characteristic of the gas that bears a definite relationship to the pressure.

Hydrostatic Measuring Devices. The Bourdon pressure gage and the McLeod gage are two hydrostatic measuring devices.

The Bourdon gage (Fig. 23) accurately and continuously indicates the pressure from 100 kPa (1 atm) to approximately 1 kPa (0.01 atm) and can be used effectively to monitor the roughing cycle and the performance of the roughing pump.

Fig. 23 Spriral Bourdon tube gage

McLeod Gage. Special, expensive diaphragm gages are available for measuring pressures continuously from atmospheric pressure down to approximately 13 mPa (10-4 torr). Pressures from a few torr down to 13 mPa (10-4 torr) can be measured periodically with a McLeod gage (Fig. 24), which samples the gas and compresses it to a small, calibrated volume. It then registers the ratio of the initial and final volumes, and this ratio is an indication of the pressure at which the gas sample was taken. McLeod gages are manual units that are usually used to monitor the accuracy of other gages.

Fig. 24 McLeod gage. (a) In filling (charging) position. (b) In measuring position

Thermal and Electrical-Conductivity Measuring Devices. The second classification of gages includes those that sense the thermal and electrical conductivity of the gas. These physical characteristics bear a direct relationship to pressure and to the specific type of gas.

The thermal-conductivity gages are widely used for most vacuum metallurgical processing because they are relatively inexpensive and can continuously monitor vacuum levels between 130 and 0.13 Pa (1 and 10-3 torr). The thermal conductivity, or convective heat transfer, of a gas is essentially constant as pressure is reduced until a pressure of about 130 Pa (1 torr) is reached. From that point, the conductivity declines until, at a pressure somewhat less than 0.13 Pa (10-3 torr), there is almost no heat transfer by the gas molecules to the surface of the gage walls, and thermal conductivity becomes virtually 0. This trend is shown for a typical common gas in Fig. 25.

Fig. 25 Relationship between the thermal conductivity of gas and pressure

Thermal conductivity varies considerably from 0.13 to 130 Pa (10-3 to 1 torr), but varies little above that pressure; this behavior constitutes the operational principle of the thermal-conductivity gage. Two gages use this principle, the thermocouple and the Pirani.

The radiation or thermocouple gage is based on the principle that, as the amount of gas in a vessel decreases, the temperature of a constantly heated wire increases because less heat is radiated to the surrounding environment. The thermocouple gage is the most commonly used measuring device used with vacuum furnaces.

Wire temperature is measured by a fine wire thermocouple attached to the midpoint of the heated wire. The maximum temperature of the wire is about 115 °C (240 °F), a temperature reached at pressures of 130 mPa (1 μm Hg) or less. A typical thermocouple gage is shown in Fig. 26. Advantages of this gage are:

• Circuitry and tubes are comparatively inexpensive • It samples temperature continuously • Because of the relatively low wire temperature, it can be exposed to air for years without damage or

danger of burnout • The signal can be used to activate relays or other remote controls

Disadvantages of this gage are:

• Because different gases have different thermal-conductivity values, the gage requires calibration for each type of gas

• The scale is markedly nonlinear; at low pressures, the marks on the scale are spread widely; in higher ranges, the marks are closely spaced (Fig. 25)

• Output depends somewhat on the ambient temperature of the tube • The useful range is from 0.13 to 130 Pa (1 to 1000 μm Hg)

Fig. 26 Thermocouple-type thermal-conductivity vacuum gage

The Pirani gage (Fig. 27) uses the same thermal conductivity principle as the radiation or thermocouple gage. Operation is based on the change of heat conductivity of a gas with pressure and the change of electrical resistance of a wire with temperature. When the wire is electrically heated with a constant current, its temperature changes with pressure, producing a voltage across the bridge network. The compensating cell corrects for changes in room temperature. The vacuum pressure is thus measured in terms of the bridge imbalance. The system is somewhat more expensive and complicated than the thermocouple gage, with essentially the same advantages and disadvantages.

Fig. 27 Schematic of a Pirani gage

Hot-Filament Ionization Gage. The thermocouple and Pirani gages do not measure pressures of less than 0.13 Pa (1 μm Hg); therefore, a different type of gage is necessary for systems operating below this level. For vacuums from 130 mPa to 0.13 μPa (10-3 to 10-9 torr), the hot-filament ionization gage can be used. The sensing element of the gage resembles a triode vacuum tube. Figure 28 shows the three elements of a hot-filament ionization gage. A heated filament emits electrons that are attracted to the loosely wound and positively charged grid. The electrons strike residual molecules of gas in the vacuum. These collisions may remove an electron from the gas molecule, resulting in a positively charged gas ion, which is attracted to the collector. The positive ion current flow from the collector to ground is a measure of the gas pressure or vacuum.

Fig. 28 Components of a hot-filament ionization gage. (a) Movement of electrons and ions in relation to filament cathode. (b) Simplified electrical circuit of the device. (c) Typical gage construction. A tungsten or thoria-coated filament cathode emits a current of approximately 5 mA. The electrons are accelerated toward a cylindrical grid operated at approximately 150 V.

The glass ionization gage, which originated in the radio tube industry, is usually calibrated with dry nitrogen. Its response varies slightly with other gases, and this must be considered when great accuracy is required. The filament can deteriorate, with a resulting loss of accuracy and subsequent burnout.

Cold-Cathode Ionization Gage. Another type of ion gage is the cold-cathode gage (Fig. 29), which depends on the measurement of an ion current produced by a high-voltage discharge. The cathode in the sensing element releases electrons that spiral through a magnetic field toward the anode. This spiralling motion lengthens the distance that electrons travel between cathode and anode, increasing the probability of collision with gas molecules and the formation of positive ions. The formation of ions varies linearly with pressure, and the ion current indirectly indicates the pressure. These rugged gages are well suited for production applications. They cannot be degassed as easily, however, and they are more readily contaminated and less accurate than hot-filament gages. Output is linear below 130 mPa (10-3 torr), and the usable range for measuring vacuum is between 1300 and 0.13 mPa (10-2 and 10-6 torr).

Fig. 29 Components of a cold-cathode discharge gage. (a) Movement of electrons in relation to magnetic field. (b) Typical gage construction showing cathode body and anode flange. PTFE, polytetrafluoroethylene

Whether the ionization gage is the hot-filament or the cold-cathode type, vacuum gas is used for vacuum pressures below 130 mPa (10-3 torr).

Heat-Resistant Materials for Heat-Treating Furnace Parts, Trays, and Fixtures

Revised by G.Y. Lai, Haynes International, Inc.

Introduction

TRAYS AND FIXTURES made of heat-resistant alloys are among the many parts used in industrial heat-treating furnaces that operate at temperatures from 540 to 1200 °C (1000 to 2200 °F). Coverage of heat-resistant structural materials may be found in Volume 1 of ASM Handbook, formerly 10th Edition Metals Handbook, together with room-temperature and high-temperature mechanical properties.

Basic Metallurgy and Product Forms

A partial list of typical products can be divided into two categories: The first consists of parts that go through the furnaces and are therefore subjected to thermal and/or mechanical shock; these include trays, fixtures, conveyor chains and belts, and quenching fixtures. The second comprises parts that remain in the furnace with less thermal or mechanical shock; these include support beams, hearth plates, combustion tubes, radiant tubes, burners, thermowells, roller and skid rails, conveyor rolls, walking beams, rotary retorts, pit-type retorts, muffles, recuperators, fans, and drive and idler drums.

These heat-resistant alloys are supplied in either wrought or cast forms. In some situations, they may be a combination of the two. The properties and costs of the two forms vary, even though their chemical compositions are similar. Because

there are many foundries and fabricators experienced in the design and application of these products, it is important to seek their advice when purchasing high-alloy parts.

Five types of heat-resistant alloys are listed in Vol 1 of ASM Handbook, formerly 10th Edition Metals Handbook:

• Iron-chromium alloys • Iron-chromium-nickel alloys • Iron-nickel-chromium alloys • Nickel-base alloys • Cobalt-base alloys

The great majority of heat-treating furnaces use only the second and third types because the iron-chromium alloys do not have sufficient high-temperature strength to be useful. Some iron-chromium alloys (more than 13% Cr) are susceptible to so-called 475 °C (885 °F) embrittlement. Because of increasing temperatures (for example, >980 °C, or 1800 °F), more and more applications use nickel-base alloys because of their improved creep-rupture strengths and oxidation resistance. Cobalt-base alloys are generally too expensive except for very special applications. Therefore, this discussion will be limited to the use and properties of the iron-chromium-nickel, iron-nickel-chromium, and nickel-base alloys.

Room-temperature mechanical properties have limited value when selecting materials or designing for high-temperature use, but they may be useful in checking the quality of the alloys. These properties are shown in Vol 1 and also may be found in ASTM specification A 297. The useful high-temperature properties of these alloys are summarized in Table 1 for castings and Table 2 for wrought products. The tables include nominal composition of the alloys and the stress required to produce 1% creep in 10,000 h and rupture in 10,000 h and 100,000 h, at temperatures of 650, 760, 870, and 980 °C (1200, 1400, 1600, and 1800 °F). A design stress figure commonly used for uniformly heated parts not subjected to thermal or mechanical shock is 50% of the stress to produce 1% creep in 10,000 h, but this should be used carefully and should be verified with the supplier.

Table 1 Composition and elevated-temperature properties of selected cast heat-resistant alloys

Approximate composition, %

Temperature Creep stress to produce 1% creep in 10,000 h

Stress to rupture in 10,000 h

Stress to rupture in 100,000 h

Grade UNS number

C Cr Ni °C °F MPa ksi MPa ksi MPa ksi

Iron-chromium-nickel alloys

650 1200 124 18.0 114 16.5 76 11.0

760 1400 47 6.8 42 6.1 28 4.0

870 1600 27 3.9 19 2.7 12 1.7

HF J92603 0.20-0.40 19-23 9-12

980 1800 . . . . . . . . . . . . . . . . . .

650 1200 124 18.0 97 14.0 62 9.0 HH J93503 0.20-0.50 24-28 11-14

760 1400 43 6.3 33 4.8 19 2.8

870 1600 27 3.9 15 2.2 8 1.2

980 1800 14 2.1 6 0.9 3 0.4

650 1200 . . . . . . . . . . . . . . . . . .

760 1400 70 10.2 61 8.8 43 6.2

870 1600 41 6.0 26 3.8 17 2.5

HK J94224 0.20-0.60 24-28 18-22

980 1800 17 2.5 12 1.7 7 1.0

Iron-nickel-chromium alloys

650 1200 . . . . . . . . . . . . . . . . . .

760 1400 . . . . . . . . . . . . . . . . . .

870 1600 43 6.3 33 4.8 22 3.2

HN J94213 0.20-0.50 19-23 23-27

980 1800 16 2.4 14 2.1 9 1.3

650 1200 . . . . . . . . . . . . . . . . . .

760 1400 55 8.0 58 8.4 39 5.6

870 1600 31 4.5 26 3.7 16 2.4

HT J94605 0.35-0.75 15-19 33-37

980 1800 14 2.0 12 1.7 8 1.1

650 1200 . . . . . . . . . . . . . . . . . .

760 1400 59 8.5 . . . . . . . . . . . .

870 1600 34 5.0 23 3.3 . . . . . .

HU . . . 0.35-0.75 17-21 37-41

980 1800 15 2.2 12 1.8 . . . . . .

650 1200 . . . . . . . . . . . . . . . . . . HX . . . 0.35-0.75 15-19 64-68

760 1400 44 6.4 . . . . . . . . . . . .

870 1600 22 3.2 . . . . . . . . . . . .

980 1800 11 1.6 . . . . . . . . . . . .

Note: Some stress values are extrapolated.

Table 2 Composition and elevated-temperature properties of selected wrought heat-resistant alloys

Approximate composition, % Temperature Creep stress to produce 1% creep in 10,000 h

Stress to rupture in 10,000 h

Grade UNS number

C Cr Ni Other °C °F MPa ksi MPa ksi

Iron-chromium-nickel alloys

650 1200 48 7.0 . . . . . .

760 1400 14 2.0 . . . . . .

870 1600 3 0.5 10 1.45

309S S30908 0.08 max 22-24 12-15 . . .

980 1800 . . . . . . 3 0.5

650 1200 63 9.2 . . . . . .

760 1400 17 2.5 . . . . . .

870 1600 9 1.3 13.5 1.95

310S S31008 0.08 max 24-26 19-22 . . .

980 1800 . . . . . . 4 0.6

Iron-nickel-chromium alloys

760 1400 25 3.6 30 4.4

870 1600 13 1.9 12 1.8

RA 330 N08330 0.08 max 17-20 34-37 . . .

980 1800 3.5 0.52 4.5 0.65

760 1400 47 6.8 54 7.8 RA 330 HC . . . 0.4 max 17-22 34-37 . . .

870 1600 18 2.6 18 2.6

980 1800 5 0.7 5 0.7

760 1400 43 6.2 65 9.4

870 1600 21 3.1 21 3.1

RA 333 N06333 0.08 max 24-27 44-47 3 Mo, 3 Co, 3 W

980 1800 6 0.9 7 1.05

760 1400 19 2.8 23 3.3

870 1600 4 0.61 12 1.7

Incoloy 800 N08800 0.1 max 19-23 30-35 0.15-0.60 Al, 0.15-0.60 Ti

980 1800 1 0.23 6 0.8

760 1400 83 12.0 79 11.5

870 1600 30 4.4 33 4.8

Incoloy 802 N08802 0.2-0.5 19-23 30-35 . . .

980 1800 8 1.1 11.5 1.65

Nickel-based alloys

760 1400 28 4.1 41 6.0

870 1600 14 2.0 16 2.3

Inconel 600 N06600 0.15 max 14-17 72 min . . .

980 1800 4 0.56 8 1.15

760 1400 28 4.0 42 6.1

870 1600 14 2.0 19 2.7

Inconel 601 N06601 0.10 max 21-25 58-63 1.0-1.7 Al

980 1800 5.5 0.79 8 1.2

In general, these materials contain iron, nickel, and chromium as the major alloying elements. Carbon, silicon, and manganese also are present and affect the foundry pouring and rolling characteristics of these alloys, as well as their properties at elevated temperature. Nickel influences primarily high-temperature strength and toughness. Chromium increases oxidation resistance by the formation of a protective scale of chromium oxide on the surface. An increase in carbon content increases strength.

Since the mid to late 1970s, a number of heat-resistant wrought alloys have been developed and are now being used in the heat-treating industry. Some of these alloys, such as Haynes alloys 230 (UNS N06230) and 556 (UNS R30556) and Inconel alloy 617 (UNS N06617), were originally developed for gas turbines, which require alloys with high creep-rupture strengths, good oxidation resistance, good fabricability, and good thermal stability. These alloys, commonly

referred to as solid-solution-strengthened alloys, use molybdenum and/or tungsten for strengthening. The alloys are also strengthened by carbides. Another high creep strength alloy, originally developed for gas turbine combustors, is Incoloy alloy MA 956, which is strengthened by oxide dispersion. This alloy is produced by a mechanical alloying process, using the high-energy milling of metal powders. These wrought heat-resistant alloys, along with chemical compositions and major characteristics, are tabulated in Table 3.

Table 3 New heat-resistant wrought alloys developed from about 1975 to 1990

Composition, wt% Alloy UNS number

Fe Ni Co Cr Mo W C Other

Major characteristics

253 MA(a) S30815 Bal 11 . . . 21 . . . . . . 0.08 1.7 Si, 0.17 N, 0.04 Ce

Oxidation resistance

RA85H(b) S30615 Bal 14.5 . . . 18.5 . . . . . . 0.2 3.6 Si, 1.0 Al Carburization resistance

Fecralloy A(c) . . . Bal . . . . . . 15.8 . . . . . . 0.03 4.8 Al, 0.3 Y Oxidation resistance

HR-120(d) . . . Bal 37 . . . 25 . . . . . . 0.05 0.7 Nb, 0.2 N Creep-rupture strength

556(d) R30556 Bal 20 18 22 3 2.5 0.1 0.6 Ta, 0.2 N, 0.02 La

Creep-rupture strength

HR-160(d) . . . 2 Bal 29 28 . . . . . . 0.05 2.75 Si Sulfidation resistance

214(d) . . . 3 Bal . . . 16 . . . . . . 0.05 4.5 Al, Y (present) Oxidation resistance

230(d) N06230 . . . Bal . . . 22 2 14 0.1 0.005 B, 0.02 La Creep-rupture strength/oxidation resistance

Inconel 617(e) N06617 1.5 Bal 12.5 22 9 . . . 0.07 1.2 Al Creep-rupture strength/oxidation resistance

Incoloy MA . . . Bal . . . . . . 20 . . . . . . . . . 0.5 Y2O3, 4.5 Al, Creep-rupture strength/oxidation

(a) 253 MA is a registered trademark of Avesta Jernverks Aktiebolag.

(b) RA85H is a registered trademark of Rolled Alloys, Inc.

(c) Fecralloy A is a trademark of UK Atomic Energy.

(d) HR-120, HR-160, 556, 214, and 230 are trademarks of Haynes International, Inc.

(e) Inconel and Incoloy are registered trademarks of Inco family of companies.

All of the alloys commonly used in castings for furnace parts have essentially an austenitic structure. The iron-chromium-nickel alloys (HF, HH, HI, HK, and HL) may contain some ferrite, depending on composition balance. If exposed to a temperature in the range of 540 to 900 °C (1000 to 1650 °F), these compositions may convert to the embrittling σ phase. This can be avoided by using the proper proportions of nickel, chromium, carbon, and associated minor elements. Chromium and silicon promote ferrite, whereas nickel, carbon, and manganese favor austenite. Use of the iron-chromium-nickel types should be limited to applications in which temperatures are steady and are not within the σ-forming temperature range. Transformation from ferrite to σ phase at elevated temperature is accompanied by a change from ferromagnetic material and from a soft to a very hard, brittle material. All heat-resistant alloys of the iron-nickel-chromium group are wholly austenitic and are not as sensitive to composition balance as is the iron-chromium-nickel group. Also, the iron-nickel-chromium alloys contain large primary chromium carbides in the austenitic matrix and, after exposure to service temperature, show fine, precipitated carbides. The iron-nickel-chromium alloys are considerably stronger than the iron-chromium-nickel alloys and may be less expensive per part if the increased strength is considered when designing for a known load.

The life expectancy of trays and fixtures is best measured in cycles rather than hours, particularly if the parts are quenched. It may be cheaper to replace all trays after a certain number of cycles to avoid expensive shutdowns caused by wrecks in the furnace. Chains or belts that cycle from room temperature to operating temperature several times a shift will not last as long as stationary parts that do not fluctuate in temperature. Parts for carburizing furnaces will not last as long as those used for straight annealing.

Finally, alloy parts represent a sizable portion of the total cost of a heat-treating operation. Alloys should be selected carefully, designed properly, and operated with good controls throughout to keep costs at a minimum.

Material Comparison for Heat-Resistant Cast and Wrought Components

The selection of a cast or fabricated component for furnace parts and fixtures depends primarily on the operating conditions associated with heat-treating equipment in the specific processes, and secondarily on the stresses that may be involved. The factors of temperature, loading conditions, work volume, rate of heating, and furnace cooling or quenching need to be examined for the operating and economic trade-offs. Other factors that enter into the selection include furnace and fixture design, type of furnace atmosphere, length of service life, and pattern availability or justification.

Some of the factors affecting the service life of alloy furnace parts, not necessarily in order of importance, are alloy selection, design, maintenance procedures, furnace and temperature control, atmosphere, contamination of atmosphere or work load, accidents, number of shifts operated, thermal cycle, and overloading. High-alloy parts may last from a few months to many years, depending on operating conditions. In the selection of a heat-resistant alloy for a given application, all properties should be considered in relation to the operating requirements to obtain the most economical life.

If either cast or wrought alloy fabrications can be used practically, both should be considered. Similar alloy compositions in cast or wrought form may have varying mechanical properties, different initial costs, and inherent advantages and disadvantages. Castings are more adaptable to complicated shapes, and fabrications to similar parts, but a careful comparison should be made to determine the overall costs of cast and fabricated parts. Initial costs, including pattern or tooling costs; maintenance expenses; and estimated life are among the factors to be included in such a comparison. Lighter-weight trays and fixtures will use less fuel in heating. Cast forms are stronger than wrought forms of similar chemical composition. They will deform less rapidly than wrought products, but may crack more rapidly under conditions of fluctuating temperatures. Selection should be based on the practical advantages, with all facts considered.

General Considerations. Both cast and wrought alloys are well accepted by the designers and users of furnaces requiring high-temperature furnace load-carrying components. Heating elements are also manufactured in either cast or wrought form (Fig. 1).

Fig. 1 Cast grid heating element (top) and ribbon heating element (bottom). Courtesy of the Electric Furnace Company

There are certain advantages for each type of manufactured component; often, the compositions are similar, if the carbon and silicon levels in the castings versus the wrought material are ignored. In general, the specifications of the wrought grades have carbon contents below 0.25%, and many are nominally near 0.05% C. In contrast, the cast alloys have from 0.25 to 0.50% C. This difference has an effect on hot strength. The difficulty in hot working the higher-carbon alloys accounts for their scarcity in the wrought series. Castings and fabricated parts are not always competitive; each product has advantages, which include:

Advantages of cast alloys

• Initial cost: A casting is essentially a finished product as-cast; its cost per pound is frequently less than that of a fabricated item

• Strength: Similar alloy compositions are inherently stronger at elevated temperatures than are wrought alloys

• Shape: Some designs can be cast that may not be available in wrought form; also, even if wrought material is available, it may not be able to be fabricated economically

• Composition: Some alloy compositions are available only in castings; they may lack sufficient ductility to be worked into wrought material configurations

Advantages of wrought alloys

• Section size: There is practically no limit to section sizes available in wrought form • Thermal-fatigue resistance: The ductility of the fine-grain microstructure of wrought alloys may

promote better thermal fatigue resistance • Soundness: Wrought alloys are normally free of internal or external defects; they have smoother

surfaces that may be beneficial for avoiding local hot spots • Availability: Wrought alloys are frequently available in many forms from stock

Shape, complexity, and number of duplicate parts (eventually affecting cost) usually determine the choice between casting or wrought part. Where section thickness and configuration permit, castings are usually cheaper. The cost per pound of the casting metal is comparable to that of a fabricated part. The total projected cost of the fabrication is usually higher because the cost of forming, joining, and/or assembling must be added to the cost of the material. However, when only one or two types of parts are to be made, the pattern cost precludes the use of a casting.

In energy-intensive heat-treating industries, the use of wrought fabrications allows fuel savings through reduced heat-treating time cycles. At the present level of energy costs, wrought fabrications may be economically preferable because of improvements in thermal efficiency.

Fabrications are preferred for thin sections and for parts where less weight or greater heat transfer may be required. Where thick walls are necessary for strength or where heavy loads are transported or pushed, the cost of fabricated sections may be prohibitive. Wrought materials have a greater degree of acceptance in fabricated baskets used under carburizing or carbonitriding conditions.

A factor that must be considered in evaluating castings and fabrications is the importance of good welding techniques, particularly for parts that are used in case hardening atmospheres. Castings have replaced fabricated products because of weld failures in multiwelded fabrications.

Although cast alloys exhibit greater high-temperature strength, it is possible to place too much emphasis on this characteristic in materials selection. Strength is rarely the only requisite and frequently is not the major one. More failures are due to brittle fracture from thermal fatigue than from stress rupture or creep. However, high-temperature strength is important where severe thermal cycling is required.

Specific Applications. Recommended alloy applications for parts and fixtures of various types of heat-treating furnaces, based on atmosphere and temperature, are summarized in Tables 4, 5, and 6. Where more than one alloy is recommended, each has proved adequate, although service life varies in different installations because of differences in exposure conditions.

Table 4 Recommended materials for furnace parts and fixtures for hardening, annealing, normalizing, brazing, and stress relieving

Retorts, muffles, radiant tubes Mesh belts Chain link Sprockets, rolls, guides, trays

Wrought Cast Wrought Wrought Cast Wrought Cast

595-675 °C (1100-1250 °F)

430 HF 430 430 HF 430 HF

304 304 304 304

675-760 °C (1250-1400 °F)

304 HF 309 309 HF 304 HF

347 HH HH 316 HH

309 309

760-925 °C (1400-1700 °F)

309 HH 309 314 HH 310 HH

310 HK 314 RA 330 HC HL RA 330 HK

253 MA HT 253 MA 800H/800HT HT 800H/800HT HL

RA 330 HL RA 330 HR-120 HR-120 HT

800H/800HT HW

HR-120

600

925-1010 °C (1700-1850 °F)

RA 330 HK 314 314 HL 310 HL

800H/800HT HL RA 330 RA 330 HC HT RA 330 HT

HR-120 HW 600 802 HX 601 HX

600 HX 601 601 617

601 214 617 X

617 X 556

X 556 230

214 230

556

230

1010-1095 °C (1850-2000 °F)

601 HK 80-20 80-20 HL 601 HL

617 HL 600 617 HT 617 HX

X HW 601 X HX X

HX 214 556 214

556 NA22H 230 556

230 230

1095-1205 °C (2000-2200 °F)

601 HL 601 601 HX 601 HL

617 HU 214 617 617 HX

230 HX 230 230

Table 5 Recommended materials for parts and fixtures for carburizing and carbonitriding furnaces

815-1010 °C (1500-1850 °F) Part

Wrought Cast

RA 330 HK

800H/800HT HT

HR-120 HU

600

601

617

X

214

556

Retorts, muffles, radiant tubes, structural parts

230

HX

RA 330

800H/800HT

Pier caps, rails

HR-120

HT

600

601

RA85H HT

RA 330 HT (Nb)

800H/800HT HU

HR-120 HU (Nb)

600

601

Trays, baskets, fixtures

617

HL X

556

214

HX

230

HX

Table 6 Recommended materials for parts and fixtures for salt baths

Process and temperature range

Electrodes Pots Thermocouple protection tubes

Salt quenching at 205-400 °C (400-750 °F) Low-carbon steel Low-carbon steel Low-carbon steel, 446

Tempering at 400-675 °C (750-1250 °F) Low-carbon steel, 446, 35-18(a)

Aluminized low-carbon steel, 309

Aluminized low-carbon steel, 446

Neutral hardening at 675-870 °C (1250-1600 °F)

446, 35-18(a) 35-18(a), HT, HU, Ceramic. 600, 556

446, 35-18(a)

Carburizing at 870-940 °C (1600-1720 °F) 446, 35-18(a) Low-carbon steel(b), 35-18(a), HT

446, 35-18(a)

Tool steel hardening at 1010-1315 °C Low-carbon steel(c), 446 Ceramic 446, 35-18(a), ceramic

(1850-2400 °F)

Note: Where more than one material is recommended for a specific part and operating temperature, each has proved satisfactory in service. Multiple choices are listed in order of increasing alloy content (except ceramic parts).

(a) A series of alloys generally of the 35Ni-15Cr type or modifications that contain from 30 to 40% Ni and 15 to 23% Cr and include RA 330, 35-19, Incoloy, and other proprietary alloys.

(b) Immersed electrode furnaces only.

(c) Low-carbon steel is recommended for completely submerged electrodes only.

Typical Applications

Trays and Grids. Many parts to be heat treated are irregular in shape and as such must be conveyed through the continuous-heat-treating furnaces or loaded and unloaded from the batch furnaces on grids or trays (Fig. 2). These trays or grids must withstand exposure to the same furnace conditions as the product: They are subjected to repeated heating and cooling, as well as repeated compression and tensile loading. Heat-resistant alloys are used extensively for these parts, although there are instances in which dispensable carbon or low-alloy-steel fabricated trays are employed. In such instances, the choice is based on the economics of the particular situation, taking into account the cost of materials and the expected service life.

Fig. 2 Articulated tray for roller rail furnace

Two-thirds of the approximately 15 common heat-resistant alloy compositions find application in the heat-treating industry. Of these, half are recommended for use in trays and grids. The particular alloy chosen should be selected on the basis of required strength at temperature, ductility, and oxidation corrosion resistance.

Trays and grids made of austenitic stainless steels containing approximately 10% Ni may find an application at furnace temperatures of 650 to 870 °C (1200 to 1600 °F), but as the service temperature goes up, for example, to 1040 to 1150 °C (1900 to 2100 °F), an alloy with twice as much nickel would probably be selected. If the tray or fixture is to be subjected to the thermal shock of rapid heating and cooling, an even higher nickel content would probably be selected. The particular atmosphere surrounding the trays necessitates the consideration of varied amounts of chromium addition to enhance resistance to oxidation or high-temperature corrosion. If trays are to be used in an atmosphere with very high sulfur, an alloy with rather high chromium and moderate nickel would be selected. Some alloys contain relatively large amounts of silicon to fortify against carburization in carburizing applications (Fig. 3).

Fig. 3 Typical HT alloy carburizing furnace trays. Dimensions given in inches

Families of commercially available heat-resistant alloys provide sufficient selection so that an alloy that is optimized for each application and use can be specified. Alloy producers as well as vendors of trays and grids, both cast and fabricated, are an invaluable source of information regarding service life, design considerations, and fabrication. Generally, a tray or grid should be of sufficient section size to provide reasonable service life under specified loading conditions. An overly heavy tray may prolong service life, but the added energy cost to heat the tray through each cycle may offset any cost savings realized through added life. It is sometimes possible to combine materials in trays to provide sufficient strength yet maintain minimum weight. For example, in an articulated tray used in an extremely long pusher furnace, the tray grid that is subjected to the compressive force of the pusher bar is of a higher nickel content than are the vertical load supports that must bear the compressive load on a per-tray basis. This dual-alloy tray represents a compromise between weight, cost, and service life. In addition, service life is greatly affected by the tray-cooling process, and, in general, uniform section size throughout the tray is highly desirable to minimize thermal contraction/expansion stresses during cooling and heating.

All service conditions should be considered when selecting an alloy for trays and grids. Unlike furnace structural parts, a tray is subject to alternate heating and cooling during each cycle. The cooling can be rapid, as in quenching, or relatively slow, as in furnace-cooling applications. The selection of a proper alloy ensures adequate service life if all service conditions are known and considered.

Baskets and Fixtures. In many situations, parts being heat treated are of a size that does not permit them to be loaded directly on a furnace hearth, tray, or grid. They require some type of container, such as a basket. The design of these baskets varies because each product is developed for a specific application and loading and must function with a specific type of furnace equipment.

Baskets and fixtures can be produced from cast or wrought alloys. Fabricated parts are used in light-to-medium loading applications, intricate designs, complex shapes, and generally with lighter metal sections. Typically, these include the bar frame-type basket (Fig. 4) or corrugated box or shroud. In applications involving heavy loading and/or simple shapes and designs, cast alloys are commonly selected; typically they are large-pit-furnace baskets (Fig. 5).

Fig. 4 Bar frame-type basket

Fig. 5 Large-pit-furnace basket

In specific applications, a part may require special positioning. This is accomplished by using a fixture that is generally adaptable to an existing tray or grid or, in some instances, placed directly into a basket or container. These components can range from simple shapes, such as round, square, rectangular, or fluted bars, to extremely intricate shapes. Figures 6,

7, and 8 are examples of such fixtures. Figure 6 is a tray/fixture assembly used for carburizing pinions. Figure 7 was designed for heat treating lawn mower blades, and Fig. 8 was designed for heat treating shafts.

Fig. 6 Tray/fixture assembly for carburizing pinions

Fig. 7 Fixture designed for heat treating lawn mower blades

Fig. 8 Fixture designed for heat treating shafts

In most applications involving operating temperatures of 790 to 1010 °C (1450 to 1850 °F), the product is generally manufactured with a material having a nominal composition of 35Ni-15Cr, which provides a fully stable austenitic structure virtually free from any embrittling phases. In addition, it provides a reasonable cost-to-life ratio in applications involving endothermic, exothermic, and inert atmospheres even with properly controlled enrichments of natural gas, air, or ammonia, typically used for gas carburizing or gas carbonitriding. For quenching, a 35Ni-15Cr alloy provides acceptable life; however, in applications of severe quenching, higher-nickel alloys may be considered, depending on the cost-to-life ratio of the product. If applications involve higher temperatures, excessive oxidation, or carburization, consideration should be given to increasing the nickel-chromium content of the alloy. For nitriding, a higher nickel content provides the best cost-to-life ratio.

In vacuum furnace applications, various heat-resistant alloys are available, depending on operating temperatures. The principal controlling factor in this case is the creep-rupture strength of the alloy. A combination of distortion and warpage is generally the major failure mode. Care should be taken to prevent the vaporization of any element within these alloys. If a specific application has operating parameters that will not allow the use of a conventional alloy, molybdenum fabrications may be used, providing that air and oxygen are absent because catastrophic oxidation may become significant at higher temperatures.

For baskets and fixtures that may be restructured to lower-temperature operations of 260 to 595 °C (500 to 1100 °F), materials such as 304, 309, and 310 stainless steel may be acceptable. If the application involves temperatures between 595 and 815 °C (1100 and 1500 °F), caution should be taken because of the potential formation of σ phase, primarily in types 309 and 310 stainless steel. In addition, when type 304 is exposed to this temperature range, some embrittling from carbide participation results. Therefore, if the operating temperature is between 595 and 815 °C (1100 and 1500 °F), iron-nickel-chromium alloys, such as 35Ni-15Cr and 35Ni-20Cr, are generally suitable.

It should be noted that in the applications of baskets and fixtures, periodic straightening and rewelding can greatly enhance product life and improve the cost-to-life ratio.

Skid Rails, Hearth Components, and Rollers. Certain furnace parts are subjected to an additional service condition that must be considered when opting for a particular design or alloy selection. This group of parts includes components of the conveyance system in a continuous furnace that is subjected to wear as a result of interfacing with product or trays. Furthermore, this interfacing or wear occurs at elevated temperatures where alloy strength is diminished. The proper selection of an alloy for a specific high-temperature service involves consideration of many factors. One important factor is to avoid selecting the same composition for components that have sliding or rolling contact in order to minimize the possibility of galling or seizing. For example, when selecting an alloy to make skid rails (Fig. 9), it is necessary to consider whether the rail will be cooled and, if so, by what method; whether adequate expansion space has been specified; the amount of contact area present at the interface; and how the rail will be supported and at what

intervals. Thus, it can be inferred that the design of skid rails and selection of the alloy are an integral part of furnace design; the same principles apply to rollers and hearth components.

Fig. 9 Water-cooled skid pipe with welded strips

Perhaps the greatest single factor affecting a roller in a heat-treating-furnace application is the actual bearing or roller support of the roller. In roller hearth furnaces, the rollers protrude through the furnace walls, and the roller bearing can operate in a relatively reduced ambient temperature (Fig. 10). However, in some roller tray furnaces, the individual rollers operate within the furnace heated area, and the roller spindle or shaft must rotate on a roller support without aid of a precise, lubricated bearing. Hearth components are usually nonrotating or nonmoving parts and, in most situations, are well supported by refractory piers and/or ledges. Hearth components are almost always subjected to compressive loading, although they could on occasion be subjected to lateral thrust and/or bending. When selecting an alloy for these applications, it is necessary to consider the elevated-temperature mechanical properties required for the anticipated loading.

Fig. 10 Thin-walled furnace roller

Belting. Conveyor belts are used extensively in the design of furnaces used for the brazing, sintering, and hardening of carbonitriding applications. Woven belts or mesh belts are commonly used for light-duty loading, whereas cast link belts are designed for heavy-loading requirements. Figure 11 shows an assembled conveyor belt with a 100 mm (4 in.) pitch and the drive drum ready for installation.

Fig. 11 Conveyor belt assembled with 100 mm (4 in.) pitch and drive drum ready for installation

When mesh belting is required for applications between 260 and 790 °C (500 and 1450 °F), medium-carbon steel (grades 1040 and 1055) can be selected for application up to 540 °C (1000 °F). For higher temperatures, materials containing 1 to 5% Cr may be selected, or type 430 stainless is acceptable. Types 304, 309, and 316 stainless steel tend to be susceptible to carbide participation or the formation of σ phase within this temperature range and therefore are not frequently selected. If stainless steel is required, type 347, which is stabilized with niobium and virtually free from carbide participation, may be selected.

Alloys commonly used for mesh belts in the temperature range of 790 to 1205 °C (1450 to 2200 °F) are 35Ni-15Cr; 80Ni-20Cr; type 314 stainless steel; and alloy 600, alloy 601, and 214 alloy, with the latter three nickel-base alloys servicing the high end of the temperature range (that is, 980 to 1205 °C, or 1800 to 2200 °F) The selection of the proper alloy is based on temperature, atmosphere, possible process contaminants, and cost-to-life ratios of the application. In addition to material selection, other key considerations for mesh belt applications are belt support, drive system, proper tension, and control of side travel.

In applications involving heavier loading, the cast link belt is often used. These applications tend to be in the temperature range of 790 to 1095 °C (1450 to 2000 °F) and not in the low-temperature range, 260 to 790 °C (500 to 1450 °F). Materials, therefore, are similar to the high-temperature mesh belting alloys, except 35Ni-15Cr is the most popular alloy. It provides acceptable service in most conventional heat-treating applications, such as hardening, gas carburizing, and gas carbonitriding. The cast links are generally assembled using a wrought 35Ni-15Cr alloy with a higher carbon level. In the application of cast link belts, consideration should be given to support, drive systems, tension, and side travel.

Radiant tubes can be manufactured from cast alloys or fabricated with wrought alloys and, in most applications, can be selected interchangeably depending on cost-to-life ratios. Fabrications may be selected because of the direct savings in fuel resulting from reductions in weight (fabricated tubes can weigh as much as one-quarter of the equivalent cast tubes). Also, the smooth surface of a fabricated tube is beneficial in avoiding focal points of concentrated or accelerated corrosion. The sound, smooth interior of the wrought tube permits optimum design stresses and helps to prevent the buildup of soot deposit. Figure 12 shows a typical U-shaped radiant tube used in carburizing furnaces. Some furnaces use a straight radiant tube.

Fig. 12 U-shaped radiant tube

Wrought alloys for radiant tubes include type 309, type 310, RA 330 alloy, alloy 800H, alloy 601, 230 alloy, and 214 alloy. Most radiant tubes are fired with natural gas in the inner diameter. The inner diameter of the radiant tube is subject to oxidation. The outer diameter of the tube is exposed to the furnace atmosphere; thus, the furnace atmosphere can also influence the selection of alloys. For example, nickel-base alloys are preferred for nitriding atmospheres. However, nickel-base alloys are not recommended for use with gases having a high sulfur content.

In addition to temperature and atmosphere, consideration should be given to tube design for proper expansion and contraction, support for horizontal mounting, and burner positioning to prevent flame impingement. These considerations, as well as dissipation rates, affect service life as severely as the material selection.

Pots. Furnace design is the most important consideration in the selection of material for pots holding molten lead or salt. Externally heated pots act as a muffle or barrier between the heating and work zones. This type of service is severe because of the great difference between outside and inside temperatures, especially while the furnace is being heated to the operating temperature, when the outside of the pot is subjected to maximum heat input and the lead or salt it contains is still solid.

When the furnace is heated by immersed or submerged electrodes, the pot is completely sealed from the outside air, and the inside of the pot is protected by the molten bath. A pot in this type of installation lasts much longer than an externally heated pot. For environmental reasons, salt operations, such as those using cyanide salts, have diminished greatly. The most popular operations remaining generally involve neutral salt and lead. The specific alloy selected for pots used in salt operations is directly related to salt composition.

Pots are available in both cast alloys and fabricated wrought alloys. However, because the availability of cast pots has become somewhat limited, fabricated pots are more widely used. Carbon steel pots can be used within a temperature range of 260 to 540 °C (500 to 1000 °F). For applications between 540 and 815 °C (1000 and 1500 °F), type 309 stainless, 35Ni-15Cr, and higher-nickel alloys can be applied.

Electrodes. The choice of heat-resistant alloys used for electrodes depends chiefly on the type of furnace in which they are used. The most popular alloy for neutral salt pot electrodes is type 446 stainless steel. Immersed electrodes deteriorate rapidly along the line where the surface of the salt bath comes in contact with them. This is known as air-line attack. Submerged electrodes, entering the bath through the side of the furnace, are never exposed to air and last much longer. This type of electrode is used only with ceramic pots.

Electrodes deteriorate badly at the salt line during the start-up period. Better service is obtained by maintaining them at a temperature just above the freezing point of the salt during short pauses in operation. This practice not only prolongs the life of electrodes, but also eliminates the tedious task of starting a cold bath. Very little power is required to hold a well-insulated, unused furnace at about 700 °C (1300 °F).

Retorts and muffles are used in heat-treating furnaces to separate materials being heated from the products of combustion and, in some instances, to contain atmospheres that would otherwise escape through more porous containment vessels. In most situations, a muffle may be made either of metallic or nonmetallic materials. A typical D-shaped muffle with internal hearth is shown in Fig. 13.

Fig. 13 Typical fabricated D-shaped muffle with internal hearth

Muffles are treated as a separate category of HT alloy applications. An important and different set of constraints apply because the heat necessary to raise the inside of a muffle to the proper process temperature is applied from without. Materials and designs must be selected that will not only withstand the rigors of furnace temperature and corrosion conditions, but will also not significantly prevent heat transfer. Designs must provide for expansion and contraction, be atmosphere tight, and provide maximum area for radiating surfaces because most muffles do not include internal recirculation features. For this reason, many cast or fabricated muffles are corrugated in design. This corrugated construction increases the radiating area while assisting in accommodating expansion and contraction as the muffle is cycled to and from operating temperature. Heat is transmitted by conduction to the inner-wall radiating surface of a muffle. In order to transfer heat, there must be a temperature drop across the wall of the muffle. The temperature drop is directly proportional to the thickness of the muffle wall. With heavy wall construction, the outside temperature must be raised to effect a given temperature within the muffle. Muffle material should be selected to provide a balance between alloy content (which represents strength), cost, and wall thickness.

Cost of any specific furnace part or fixture increases as the alloy content increases, although not necessarily in the same proportion as the base cost of the alloy. Some cost items will be approximately the same regardless of the type of alloy used.

To be meaningful, computations of cost for furnace parts and fixtures must be based on the number of hours of operation. In many instances, the more expensive alloys prove to be more economical. For example, service comparisons show that HU may be less expensive than HT for oil-quenched carburizing trays, and HW may be less expensive than HT for oil-quenched carburizing fixtures. On the other hand, some examples, such as brazing belts, show that the alloy of lower initial cost may also be less expensive when judged by cost per service hour.

From a practical standpoint, even cost-per-service hour data may be incomplete. Other factors should be considered for some components, notably the labor cost of replacement, the loss of productivity during downtime, and the possibility of damage to other components when failure occurs.

Nonmetallic Material Radiant Tubes (Ref 1, 2)

Silicon-silicon carbide composite radiant tubes having a density of 2.80 g/cm3 (0.088 lb/in.3) have undergone field trials in gas-fired indirect heating applications in a variety of atmospheres. One manufacturer installed these composite tubes in a pusher-type carburizing furnace in place of previously used mullite tubes in 1988 and the silicon-silicon carbides tubes were still operational after 26 months of continuous operation (24 1-h cycles/day at 980 °C (1800 °F). Another manufacturer indicates that the life expectancy of the silicon-silicon carbide tubes averaged 16 months while the life expectancy of the mullite tubes averaged 1 month.

These ceramic radiant tubes have been used in the following heat-treating processes:

• Annealing

• Carburizing • Carbide solution treating • Neutral hardening • Carbonitriding • Ferritic nitrocarburizing

Furnace atmospheres have included:

• Endothermic (both lean and rich) • Carbon-enriched gases • Ammonia-enriched gases • Nitrogen • Mixed endothermic and ammonia (50/50)

The composite radiant tubes are produced by using a moving hot zone (induction coil) to progressively melt the silicon and cause the particulate silicon to infiltrate the coarse-grained silicon carbide. The end product is a material that contains reaction-bonded silicon carbide grains in a silicon matrix. Tube composition is 53 wt% C and 47 wt% Si. An exothermic chemical reaction yields a composite material tube having excellent oxidation resistance, creep resistance, thermal shock resistance, and heat transfer properties.

Figure 14 demonstrates the results of a compression creep test run on tube sections of the silicon-silicon carbide composite material and a Ni-Cr-Fe alloy (8.51 g/cm3, or 0.307 lb/in.3, density). The Inconel 600 tube section actually melted after 1 h at 1350 °C (2460 °F) while the composite material showed no effect even after 2 weeks at 1350 °C (2460 °F). When 50 mm (2 in.) long samples of both materials were tested in compression for 2 weeks at 1200 °C (2190 °F), Inconel 600 showed 1.6% (0.79 mm, or 0.031 in.) creep while the silicon-silicon carbide composite creep was negligible at <0.03 mm (<0.001 in.).

Fig. 14 Comparison of high-temperature (1350 °C, or 2460 °F) creep testing of radiant tube sections. (a) Silicon-silicon carbide composite after 360 h. (b) Ni-Cr-Fe alloy after <1 h

Table 7 provides strength and fracture toughness data for the silicon-silicon carbide composite material at selected temperatures.

Table 7 Strength and fracture toughness properties of silicon-silicon carbide composite for radiant tubes used in gas-fired indirect heating applications

Strength(a) Temperature

C-ring configuration O-ring configuration

Fracture toughness

MPa ksi MPa ksi MPa m ksi in

RT(b) 58.7 8.52 59.8 8.67 2.09 1.90

1000 59.4 8.62 77.7 11.27 2.78 2.53

1250 69.6 10.10 92.6 13.43 . . . . . .

1350 72 10.49 90.5 13.13 . . . . . .

(a) Fracture stress is calculated on the basis of elastic beam theory, which is an overestimate of the actual fracture stress.

(b) RT, room temperature

References cited in this section

1. M.C. Kasprzyk, Silicon-Silicon Carbide Composite Radiant Tubes Produced by New Thermal Processing Method Promising in Furnace Operations, Ind. Heat., June 1990

2. B. Vinton, Ceramic Radiant Tube System Speeds Batch Furnace Recovery, Heat Treat., Vol 21 (No.2), Feb 1989, p 24-27

Energy-Efficient Heat-Treating Furnace Design and Operation

S. Lampman, ASM International

Introduction

ENERGY EFFICIENCY AND COST of heat-treating operations are related concerns that are substantially affected by the method of converting stored energy into molecular kinetic energy (temperature) of the workpiece(s). In electric furnaces, for example, high relative efficiencies between 85 and 100% are feasible, but the cost of electric energy has a substantially higher rate than that of fuels such as natural gas. Consequently, gas-fired furnaces are often more economical than electric heating, even though their efficiencies (which can range from 5 to 70% depending on operating temperatures and furnace design) are generally lower than electric heating methods.

The sources of heat loss in fuel-fired furnaces result from flue losses and furnace losses (Fig. 1). Flue losses, which can be the most significant component of energy inefficiency, occur from either incomplete combustion or the loss of sensible heat due to the discharge of hot flue gas. Furnace losses depend on the particular furnace design and method of operation. For example, batch processing can be less efficient than continuous processing because the heating and cooling cycles require expenditure of unused energy. Other typical types of furnace losses are illustrated in Fig. 1. All forms of heat loss become more significant as furnace temperature increases.

Fig. 1 Sankey diagram of heat loss in a fuel-fired furnace

This article briefly reviews some of the methods used to improve the efficiency of gas-fired furnaces. The methods are classified into three areas:

• Combustion control • Waste-heat recovery • Furnace design and operation

Of these methods, recovery of flue gas heat can have a significant impact on efficiency, particularly at higher temperatures. For example, if flue gases are discharged at 1300 °C (2400 °F) from a natural gas burner, about 70% of the energy input is lost in flue discharge. If incomplete combustion or an improper air/gas ratio occurs, then losses will be even higher. These losses can be reduced by recovery of heat from flue gases and/or improved combustion control.

At high furnace temperatures ( ≥ 1040 °C, or 1900 °F), electric heating may also become more effective because electric heating elements may require less maintenance than metallic radiant tubes. The higher relative efficiency of electrical heating may also be an advantage at higher temperatures. However, with advances in the technology of gas-fired heating, gas is becoming more competitive at higher temperatures. In particular, ceramic radiant tubes (reaction-bonded SiC) with recuperative burners can operate up to 1200 °C (2200 °F) with efficiencies of about 60%. High-alloy metallic tubes with regenerative burners also can provide economic life-cycle costs at higher temperatures.

Combustion Control

Combustion control is a basic factor in reducing flue losses, which are closely related to the fuel/air ratio prior to combustion. Too much excess air decreases efficiency because the energy expended to heat this air is wasted up the stack. Too little excess air results in fuel being unburned and also wasted up the stack. Consequently, the best efficiency theoretically occurs when the fuel/air ratio is closest to stoichiometric combustion conditions. In practice, however, some level of excess air is required to completely burn the fuel because of imperfect fuel/air mixing conditions in any commercial burner.

In most processes, efficiency is lost because the fuel/air ratio is too lean (too much excess air). The effect of this type of loss is shown in Fig. 2 in which the percent fuel required is plotted as a function of excess oxygen for several values of exhaust gas temperature. It is clear from this figure that the potential savings are greatest in high-temperature processes. For example, a reduction of excess oxygen from 6 to 5% for a process with 815 °C (1500 °F) exhaust gases produces a fuel savings of more than 5%.

Fig. 2 Efficiency losses due to excess air in a typical gas-fired combustion process. Source: Ref 1

Control of Fuel/Air Ratio. There are two well-known methods of determining fuel/air ratios in gas-fired furnaces:

• Metering of gas and air flow rates into the furnace or burner unit • Determining O2 and CO content of flue gases

Currently flue gas analysis is becoming increasingly attractive with the availability of lower cost instruments, particularly oxygen analyzers. The sensor most widely favored for measuring O2 uses a solid electrochemical zirconium oxide cell, which has an output directly related to the product of the absolute temperature and the logarithmic difference in the partial pressure of O2 across the cell. One side of the cell is connected to a reference air supply that provides a quantified supply of oxygen ions. There are several methods of determining CO content, including a catalytic method that can be incorporated into the same assembly as the O2 sensor. Because O2 sensors have been prone to maintenance problems, flue gas analyzers are often supplemented with flow meters for backup.

The control of fuel/air ratios depends, in part, on whether the furnace provides direct heating of the workpiece or indirect heating with radiant tubes. Because direct heating involves the exposure of workpieces to the hot flue gases, the control of fuel/air ratios in direct heating furnaces may be slightly on either the oxidizing side or the reducing side of stoichiometric conditions depending on the application. In the iron and steel industry, for example, slab reheating furnaces are supposed to be run at about 1% excess oxygen to avoid tight scaling on the surface. Other processes must be run slightly on the

reducing side to avoid oxidation. For heat treating of steel and copper, direct heating furnaces are controlled at or near stoichiometric conditions to avoid oxidation.

Depending on the specific combustion process, the optimum oxygen content in terms of efficiency is 1 to 3% excess O2. The optimum value of CO is in the range of 200 to 250 ppm. (As a regulated pollutant, however, 60 ppm of CO is an industry standard.) For combustion on the reducing side, efficiency drops off more rapidly than for combustion on the oxidizing side for an equal change in the fuel/air ratio (Fig. 3). Therefore, if a reducing atmosphere is required, then indirect heating with radiant tubes may be more efficient because the combustion process could be run more efficiently near the oxidizing side of stoichiometric conditions.

Fig. 3 Effects of excess air or fuel on combustion efficiency. (a) General effect on efficiency. (b) Effect on percent of fuel wasted

Heat transfer from the combustion flame is due to a combination of convection and radiation in the flue gases. With natural gas flames, for example, heat transfer by radiation is low because of the transparent nature of natural gas flame. Therefore, convection of flue gases is an important factor in both direct-fired and indirect-fired gas furnaces. Efforts to increase the heat transfer of combustion flames also include increasing the flame temperature with a preheat of combustion air (see the sections on recuperators and regenerative burners in this article).

High-velocity burners represent a recent development to improve convective heat transfer in direct-fired furnaces. The use of high-velocity burners create vigorous circulation of gases in the furnace and thereby promote uniform heat transfer. A small number of high-velocity burners therefore provides an efficient alternative to burner arrays, which require individual adjustment of a large number of burners so as to achieve uniform heating.

A potential limitation of high-velocity burners is that the circulating flow diminishes when the burners are turned down. Using excess air (to reduce temperature and increase the velocity of circulating) also leads to a reduction in thermal efficiency. Therefore, pulse firing can be useful in achieving lower temperatures and adequate circulation without resorting to the use of excess air (see the section "Pulse Firing" in this article).

Heat Transfer in Radiant Tubes. When a controlled furnace atmosphere is required, indirect heating is performed with radiant tubes. Inside the tube, heat transfer to the tube surface is governed by radiation and convection. Improvements in heat transfer inside radiant tubes can be attained by the use of axial fins (Ref 2).

Pulse techniques include two distinct methods known as pulse firing and pulse combustion. Pulse firing involves on-off cycling of burners, while pulse combustion is a resonant technique that depends on oscillations in a suitably designed combustion container.

Pulse firing is a method of controlling temperature by cycling the fuel or fuel/air supply so that a burner is either off or operating at full capacity. Temperature is thus controlled by varying the frequency of the on-off cycle, which typically has a period of about 3 to 6 s. This method of frequency modulation differs from conventional combustion control, which involves the control of heating by varying the amplitude of the flame.

Pulse firing has several advantages such as:

• Improved efficiency because fuel burns more effectively with maximum flame velocity • Turndown ratios as high as 20:1 or even 30:1 as compared to a typical turndown ratio of 8:1 in an

amplitude controlled system • Improved convective heat transfer from the turbulence created by pulsing

Pulse firing also has other potential benefits, depending on the particular firing practice and furnace. In direct-fired furnaces, for example, pulse firing allows the reduction in the operating temperature of high-velocity burners without the addition of excess air. Pulse firing thus allows improved energy efficiency, which is the main motivation for its use in high-velocity burners.

In other applications, pulse firing can have other advantages that are more important than improved energy efficiency. In radiant tubes, for example, the pulsing provides more uniform heating of the tube and promotes the extension of tube life. The high turndown ratios also provide better furnace control. If energy efficiency of radiant tubes is a concern, pulse firing can be incorporated with regenerative burners (see Example 1 in this article).

Pulse Combustion. The basic principle of pulse combustion involves the enhancement of heat transfer by creating oscillation of gas flow within the combustion system. This oscillation follows a pattern of periodic combustion, which in many cases can have a cycling frequency up to 150 Hz. Specific advantages of pulse combustors include: ability to burn various fuels; high combustion intensities; low NOx formation; low excess air requirements; and self aspiration, which eliminates the need for compressors or fans to pump the air and combustion products through the system. Additionally, the presence of pulsations in the exhaust flow enhances the rates of mass, momentum, and heat transfer in the process.

To date, pulse combustors have been utilized in such applications as drying, steam raising, water heating, and domestic space heating (Ref 3). In heat treatment furnaces, however, there are some potential drawbacks of pulsed combustion. One disadvantage is that the furnace (or combustion chamber) must be designed around the burner so that oscillations are sustained adequately. The heat-transfer enhancement resulting from the pulsations will also decrease as distance from the pulse combustor discharge is increased. The greatest heat transfer enhancement occurs within the combustor and is related to the acoustic intensity. The pressure and velocity fluctuations decrease as the gases expand outward from the discharge of the system. There is not an extensive body of data on the pressure and velocity fluctuations downstream from the combustor, but there are some indications that the pressure levels may decease by 3 to 5 dB within 3 m (10 ft) of the outlet (Ref 4).

Nevertheless, radiant tube burners might be a potential application of pulse combustion technology. In this application, the benefit of pulse combustion would be to improve convective heat transfer from the burner to the tube. A theoretical estimate relates heat transfer enhancement with tube efficiency (Ref 5). Another possible advantage is reduced NOx emissions, which may be a more important factor with future environmental restrictions.

References cited in this section

1. N. Burk and G. Woolbert, Technologies for Low Cost Combustion Control, in Industrial Combustion Technologies, American Society for Metals, 1986, p 213-220

2. A.C. Thekdi et al., "Development of an Indirect Gas-Fired High Temperature Heating System," 1984 International Gas Conference, 1984, p 709-718

3. B.T. Zinn, Applications of Pulse Combustion in Industry, in Industrial Combustion Technologies, American Society for Metals, 1986, p 55-61

4. W.A. Thrasher, "Development of a Pulse Combustor Space Heater," GRI83/0061, Gas Research Institute, 1983

5. J.M. Corliss and A.A. Putnam, Heat-Transfer Enhancement by Pulse Combustion in Industrial Processes, in Industrial Combustion Technologies, American Society for Metals, 1986, p 47

Recovery of Waste Heat

Recovery of waste heat is a basic objective that can be implemented in a variety of ways. Waste heat can be used to heat water, or it can even be used to generate electricity. The main objective, however, is the most efficient utilization of the waste heat for the given energy needs of a particular plant.

This section focuses exclusively on the use of recuperators and regenerative burners for the preheating of combustion air with hot flue gases. Preheating combustion air improves combustion efficiency (Fig. 4) and is common to many fuel-fired furnaces. Other recovery methods depend on the specific plant operations. For example, the heat stored in a batch furnace can be partially recovered to preheat another batch furnace or another load.

Fig. 4 Effect of air preheat on conbustion efficiency. Source: Ref 6

Recuperators preheat combustion air by transferring heat from the hot flue gases to the inlet air. In a direct-fired furnace, for example, tubes can carry the combustion air directly through the main flow of flue gases. This type of recuperator design transfers heat by convection. Convective recuperators, which are typically used with flue temperatures up to about 1000 °C (1850 °F), can preheat combustion air up to about 450 °C (850 °F). This would provide a fuel savings of about 25% (Fig. 4).

Conventional recuperators are also designed with inlet air passages placed alongside flue ducts (Fig. 5). Conventional recuperators, which can be based on metallic or ceramic tube design, are used with flue temperatures up to about 1500 °C (2700 °F) with ceramic materials (Ref 7). Wrought metallic radiant tubes with recuperation can be used up to about 1000 °C (1850 °F), but the tube life is limited to less than six months at these temperatures. Cast radiant tubes, which do not require a tradeoff between strength and workability, can operate satisfactorily at temperatures in excess of 1200 °C (2200 °F) (Ref 8).

Fig. 5 Radiant tube recuperator systems. (a) External recuperation with U tubes. (b) Single-ended recuperation with inner and outer tubes made of reaction-bonded silicon carbide

The amount of air preheat from conventional recuperation depends on the design of the heat exchanger and flue gas temperatures. For an external recuperator (Fig. 5a) with wrought metallic radiant tubes, the combustion air is commonly preheated up to about 590 °C (1100 °F). A single-ended ceramic tube with recuperation (as shown in Fig. 5b), which has a maximum operating temperature of 1260 °C (2300 °F), can preheat combustion air to 650 °C (1200 °F) (Ref 9). A special two-stage recuperator with a metallic stage and a ceramic stage (Ref 10) preheated combustion air up to 1100 °C (2000 °F). This can provide fuel savings comparable to that of regenerators (Fig. 4).

Self-Recuperative Burners. Because conventional recuperators (Fig. 6) are generally large and confined to large-scale plants, self-recuperative burners have been developed. Self-recuperative burners are compact units that combine the function of burning and recuperation into a single integral unit (Fig. 6).

Fig. 6 Self-recuperative burner system

Self-recuperative burners are widely used in Europe and can provide fuel savings up to about 30% as compared to cold-air combustion. In a furnace with direct heating, the waste gases are drawn from the furnace through a series of ports that surround the burner quarl by means of an air ejector. The arrangement obviates the need for insulated hot air mains and

simplifies the necessary control system. Self-recuperative burners are also used successfully with radiant tube burners for indirect heating.

Regenerative burners, which operate on the same principle as the large regenerators used in steel and glass industries since the 1850s, provide a second method of preheating combustion air with hot flue gases. Regenerative burners typically operate at higher temperatures than recuperators and thus provide larger improvements in fuel savings (Fig. 4). Regenerative burners are also more reliable than recuperators when dirty fuel is used.

A regenerator basically consists of two chambers, each containing a permeable storage bed constructed of firebrick or another refractory shape. Flue gas gives up its heat to the refractory as it flows through one chamber, while combustion air flows through the other chamber, absorbing the heat stored in it during the preceding half of the "regenerative" firing cycle. After a certain length of time--about 20 min in a traditional system--the flows are switched.

Traditional regenerators often are larger than the furnaces they serve, and they also require large exhaust fans. To create a combination burner-regenerator compact enough to be mounted on a furnace, changes in bed material and cycle timing have been made. The surface-to-volume ratio of the refractory was increased 100-fold by switching to a granular refractory, which results in a large increase in heat-storage capacity per unit volume. Adoption of microprocessor-based controls shortens the switching time from 20 min to 20 s, which reduces the amount of heat that needs to be stored during each cycle. The regenerator bed is composed typically of ceramic spheres. Advantages of metals over ceramics in this application include higher thermal conductivity and density.

Regenerative burners are used on direct-fired furnaces and indirect-fired furnaces. In direct-fired units, exhaust fans draw flue products to the regenerator bed. In an indirect-fired unit, regenerative burners are placed at each end of a radiant tube (Fig. 7). While the regenerative burner at one end of the radiant tube is firing, the flapper valve, or damper, on its eductor is closed, which forces air through the hot refractory bed and into the burner for combustion. At the same time, the flapper of the other burner is open, which causes hot exhaust gases to be sucked through the refractory bed of the burner. Eductor suction at the nonfiring burner also maintains a negative pressure inside the radiant tube. Preheated-air temperatures are within 55 °C (100 °F) of waste-gas temperatures.

Fig. 7 Schematic of a regenerative radiant-tube burner system shows key components and principle of operation. Burner at right is firing using preheated combustion air, while burner at left reclaims heat from flue gases. Burners switch roles every 20 s. Burners are actually mounted side by side on the furnace wall, with the U-shaped radiant tube extending into the furnace.

The regenerative burner is probably the most significant advance in combustion technology in recent years. Successful heat-reclamation applications for regenerative burners are found in melting of nonferrous metals (aluminum reverberatory furnaces, for example), heat treating and forge heating (box, slot, car-bottom, and tunnel furnaces), continuous glass melters, and steel-mill pusher reheat furnaces, ladle preheaters, and continuous annealing furnaces. A specific application is described below.

Example 1: Fuel Savings with Regenerative Burners on an Annealing Furnace.

One of the first applications for regenerative radiant-tube burners, sponsored by the Gas Research Institute, was in 1984 on the annealing furnace for a continuous galvanizing line at the Inland Steel Company works in Gary, IN. The 15-zone galvanneal furnace uses 64 radiant tubes (128 regenerative burners) to heat steel strip prior to zinc coating. Annealing is performed under a protective hydrogen-nitrogen atmosphere to ensure good steel surface quality. Zone temperatures range from 815 to 980 °C (1500 to 1800 °F). The furnace originally was heated by conventional radiant-tube burners with U-shaped metallic tubes. Energy balances developed for the furnace both before and after conversion revealed that the use of regenerative burners reduced the requirement for purchased energy by more than 48%.

A microprocessor-based sequencer controls burner cycling. It serves primarily as a timer, switching fuel and air between each of the two burners of the tube at 20-s intervals (Fig. 7). Instead of modulating heat input, burners are turned on and off. Flame length is constant, which produces a uniform tube temperature from end to end.

Success of the regenerative system hinges on reliable combustion-air and gas switching valves. Under normal operating conditions, valves cycle approximately 432 × 103 times/y. In tests, these valves have survived more than 5 × 106 cycles.

Conventional radiant-tube burners create a hot spot a few feet from the burner, which is usually the point of tube burnout. Pulse firing of each regenerative burner equalizes temperatures in both tube legs. There were initial concerns that the short switching time associated with compact regenerators would induce temperature oscillations, which could cause thermal fatigue. However, because the mass of the radiant tube is large, and switching times are short, temperature fluctuations are negligible.

References cited in this section

6. D.F. Hibberd, Recent Developments in Reheating and Heat-Treatment Furnaces, Metallurgia, Vol 53 (No. 2), Feb 1986, p 52-58

7. W.R. Laws, The Developing Role of Ceramic Heat Exchangers for Industry, Modern Practice in Reheating and Heat Treatment Furnaces, Institute of Energy, 1985

8. D. Marchant, Technological Developments in Radiant Alloy Tubes for Fuel-Fired Furnaces, Ind. Heat., Vol 53 (No. 8), Aug 1986, p 44, 47-48

9. B. Vinton, Ceramic Radiant Tube Speeds Batch Recovery, Heat Treat., Vol 21 (No. 2), Feb 1989, p 24-27 10. M. Peltier, Ind. Heat., Vol 53 (No. 10), 1986, p 16-17 Furnace Design and Operation

Although flue losses often constitute the major source of energy loss in a typical fuel-fired furnace (Fig. 1), energy efficiency is also affected by wall losses, opening losses, and conveyor system losses (that is, heating and cooling of trays, fixtures, and/or skids). The energy needed to heat a furnace to temperature can also be classified as a source of energy loss. Therefore, if intermittent operation of a batch furnace is necessary, the relative costs of idling a furnace versus furnace reheating after shutdown should be considered.

Once a furnace reaches a steady-state operating condition, the furnace losses from the furnace walls, openings, and conveyor system will remain constant provided that the operating conditions are unchanged, regardless of whether the furnace is being operated empty or with a capacity load. This constant loss reduces the relative efficiency of a furnace when it is operated at less than rated capacity. Therefore, operating below the heating rate capacity of a given furnace results in a decrease of efficiency. This effect is shown in Fig. 8 for a pusher-type furnace operating at 950 °C (1750 °F). Bringing the workpiece to temperature as quickly as possible thus can provide energy savings. Energy can also be saved by:

• Reducing the extent of opening losses • Selecting effective materials that reduce wall and conveyor losses • Improving heat transfer to the workpiece • Reducing treatment temperatures and hold times, if possible

Fig. 8 Relationship between furnace efficiency and heating rate. Data are based from the metering of gas use in a pusher-type furnace operating at 950 °C (1750 °F).

Opening losses occur primarily from radiation losses, which are proportional to the area of the opening and the fourth power of temperature (T) in degrees kelvin (T4).

Convective heat losses are also proportional to the effective area of the door opening. However, this loss can be accentuated when both doors are opened at the same time on a continuous furnace, because a tunnel effect is created. Convective losses also result around closed doors and through other openings in fuel-fired furnaces. To conserve energy, all unintentional openings should be sealed. The number of intentional openings should be reduced to the absolute minimum, and these openings should be sealed when not in use. Even when not permitting convective losses, openings in furnaces can permit infiltration of cold air, which must be heated to exit-gas temperature. Proper control of furnace pressure will help prevent these convective losses through necessary openings.

Ceramic fiber lining has replaced conventional brickwork in many batch and continuous heat treatment furnaces. Ceramic linings reduce heat loss by conduction through the walls and decrease furnace heat-up time because of their low thermal conductivity and low heat storage mass. Their low mass, reduced heat capacity, and good insulation properties are ideal for intermittent furnace operation. The low mass of ceramic linings also has structural advantages in the design of new furnaces. In some cases, the rapid thermal response of ceramic liners can be a problem.

Ceramic fiber linings are available in three basic product forms:

• Ceramic fiber blanketing • Ceramic fiber veneer or tiles • Sprayed-on ceramic fiber

Each of these product forms is based on the use of alumina and silica fibers containing a small amount of reducible metal oxides. Alumina contents are increased for higher operating temperatures. For temperatures above 1250 °C (2280 °F), it is necessary to use ceramic fibers with 95% alumina (such as Saffil). Saffil has a maximum recommended temperature of 1600 °C (2900 °F).

Ceramic fiber blanketing is limited to temperatures up to about 1000 °C (1830 °F) because the blanketing is susceptible to splitting from shrinkage at higher temperatures. To minimize shrinkage problems, the continuous operating

temperature of the furnace should be at least 200 °C (360 °F) below the maximum temperature rating of the fiber. Fiber blankets are also unsuitable in highly reducing atmospheres because of the disintegrating action of the atmosphere at typical temperatures.

Ceramic fiber veneer has a modular tile construction with some advantages over ceramic fiber blanketing. However,

because reliable anchoring can be a concern, a safety lining of 25 to 40 mm (1 to 1 12

in.) blanket behind the veneer can be

desirable. Veneer is useful in reducing atmospheres or when shrinkage from high temperatures (<1000 °C, or 1830 °F) is a concern.

Ceramic Fiber and High-Emissivity Coatings. Because veneer can have adherence or anchoring problems, sprayed-on ceramic fiber is used in veneer installation. High-emissivity coatings (for example, Pilbrico) are used to improve emissivity, which can be low (0.4 to 0.5) for some ceramic fiber products. In one application, a high emissivity coating allowed an initial 10 °C (5 °F) reduction in furnace temperature and a further 50 °C (25 °F) reduction after aging for a number of months (Ref 11).

Equipment Replacement or Modification. Proper application of refractories in reworked furnaces often can reduce or eliminate constant heat losses that may occur in water-cooled members. Water-cooled skid pipes can be replaced with alloy load supports if maximum temperature of operation permits, or with skid blocks made of high-strength, high-temperature refractories. Walking beam rails sometimes can be topped with refractory shoes rather than with noninsulating alloy shoes. Refractory materials in use include silicon nitride ceramics. Silicon nitride provides a good combination of excellent high-temperature strength along with resistance to oxidation and thermal shock.

Addition of alloy fans to existing furnaces sometimes can change a stagnant atmosphere into a high-velocity stream. This increase in ambient velocity breaks up boundary layers of furnace gases that surround the workpieces and shortens the heating time. This reduced heating time is a result of the change from heating only by radiation heat transfer to a combined radiation and convection transfer. The energy savings accrue through reduced furnace time required per cycle.

When the opportunity arises, much energy can be saved by altering the gases used in atmosphere heat treating. In some cases, an inert carrier gas can be added to the normal working gas.

Other energy-saving opportunities may be realized by improving the insulation systems in heat-treating furnaces. The energy lost through a furnace wall is a function of area, operating temperature, and composition of the insulation. The first two factors are fixed, but heat flow through and heat storage in the insulating system can be reduced by addition of insulation. Heat storage, which is the amount of heat contained in the wall, can be greatly reduced by using newly developed insulating materials, principally ceramic fibers and mineral wools. At the same time, the thickness of a wall for a given heat flow (loss) also can be changed significantly. One major heat-treating firm not only reduced energy requirements by using ceramic-fiber insulating materials, but also was able to heat treat larger rolls because of the decrease in required wall thickness, which in turn provided greater furnace work-zone width.

Energy Savings with Improved Quality Control. Energy savings can sometimes be achieved by improving heat-treating quality controls. For example, pyrometers can be used to measure surface temperature of workpieces in the furnace and thereby provide information for assessing required hold times in the furnace. This approach requires some relatively simple modeling of heat transfer within the workpiece. The potential savings accrue from knowledgeable reductions in soak times instead of relying on general specifications of soak times. Improved quality control also reduces scrap from improper treatment.

Furnace Design and Modeling. Another aspect of energy use is furnace design, which affects the transfer of heat from the burners to the workpiece. Direct-fired furnaces can provide efficient heat transfer depending on the design and disposition of the burners and flow patterns within the furnace. High-velocity burners, as mentioned earlier, are effective in improving heat transfer in direct-fired furnaces.

Improvements in furnace design depend on the particular application, and often improvements are based on experimental judgment and good furnace instrumentation. In some cases, however, modeling techniques such as those described in Ref 12 are also useful in avoiding costly furnace trials. Those techniques are summarized below.

Experimental modeling techniques utilize air, water, or perspex plastic models to study convective flow and establish the type and disposition of burners. Another method is the acid-alkali technique (Ref 13), which allows the

shape and size of diffusion flames to be observed directly in a perspex model. The use of this technique has led to burner designs that have achieved up to 30% fuel savings on glass melting furnaces (Ref 14).

Mathematical models provide a more quantitative assessment of the effects of firing conditions on furnace performance. Commonly used techniques are based on methods developed several decades ago by Hottel and coworkers (Ref 15). These methods are described by Tucker (Ref 16, 17) in more recent reviews. Modeling is also used to optimize or improve the analysis of energy use (Ref 18, 19, 20, 21, 22, 23, 24).

Energy Requirements of Different Furnace Types. As noted in the introduction of this article, energy efficiency and cost are related concerns. Some furnaces are very energy efficient, but the associated costs may not warrant their use. This is particularly true of electric furnaces, which are more efficient than gas-fired furnaces, but are often more expensive to operate because of the higher cost of electricity. An energy-cost comparison of various gas and electric furnaces is given in Fig. 9. Additional comparisons between electric and fuel-fired furnaces are discussed in the article"Types of Heat-Treating Furnaces" in this Volume.

Fig. 9 Energy and cost comparisons of various furnace types with a soak temperature of about 850 °C (1560 °F). Data includes energy to heat furnace to temperature. R, recuperators fitted to burners. Source: Ref 25

References cited in this section

11. H.C. Hay, Energy Savings in BSC Stainless Steel, Metal Mater., Vol 4 (No. 1), Jan 1988, p 18-23 12. N. Fricker, Effective Use of Gas on High Temperature Furnaces, Metallurgia, Vol 53 (No. 12), Dec 1986, p

544, 546, 550, 553 13. N.K. MacFadyen and M.W. Page, Verification of the Acid/Alkali Flame Modelling Technique by