Fig. 1. Cooling mechanism for a polymer quenchant

Induction heat treating can be performed on a wide range of steels and cast irons. Acceptable hardening responses can be achieved with medium carbon steels (0.35-0.5% carbon) with or without alloying additions [1]. Although steels with higher carbon content can be induction heat treated, they are typically susceptible to increased risk of cracking [2]. Induction hardening processes can also be used in conjunction with thermochemical processes such as carburizing and ferritic nitrocarburizing to obtain even greater enhancement of properties [2,3].

Fig. 2. Wetting process of cylindrical CrNi-steel probe (25 × 100 mm) quenched in oil at 60˚C, no agitation

Induction heat treating often is the process of choice [2,4] because:

• It can be used to reduce energy consumption in certain applications (e.g., where furnaces are used for localized heating, such as heating the end of shafts prior to hardening)
  
• It can be readily incorporated in in-line or cell-type manufacturing processes
  
• It eliminates the need for stop-off paints and copper plating used with conventional heat treating
  
• It permits targeted heating operations (e.g., localized hardening to strengthen a part at critical points)
  
• No preheating of parts is necessary
  
• No harmful emissions are produced
  
• It may provide optimal distortion control
  
• It only uses energy on demand
  
• Faster localized cooling rates may produce superior hardness compared with conventional furnace hardening
  
• It results in deeper hardening (typically 0.5 to 10 mm case depth) than that typically achieved with thermochemical hardening processes
  
• The process can be automated

Fig. 3. Wetting process of a cylindrical CrNi-steel probe (15 × 45 mm) quenched in water at 60˚C, no agitation

Process basics

The depth of hardening (reference depth = d) is controlled by the material properties of the component, such as hardenability, primary carbon content of the steel [1] and the frequency of the current ,and is not dependent on the shape of the component. The reference depth may be calculated using the following equation [4]: δ(cm)=[ρ/(πμ°')]½ where r is the electrical resistivity of the material, µ is relative (dimensionless) magnetic permeability of the component being heat treated and °' is the electrical frequency. Permeability varies with the power density for magnetic materials, and has a value of 1 for nonmagnetic materials and carbon steel above the Curie point (760°C) [5].

Hardening is accomplished by the use of a suitable quenchant, which provides the required cooling rate to achieve the desired hardness. Typical quenchants include forced air, water, water mists, oil-in-water emulsions, aqueous polymer solutions, and, in some cases, oil [1, 6-8]. In some cases, intensive quenching using water can be used to replace petroleum oil [7,9]. The selection of a specific quenchant depends on material, shape, crack-sensitivity, hardness pattern and the possibility of a nonuniform quench [8]. Water, if it can be used, is the preferred quenchant because it is clean, inexpensive, nontoxic and fire-safe [9]. Although forced air has been used, it is generally only applicable to deep-hardening steels and distortion and crack-sensitive component shapes. Oil-in-water emulsions are not discussed because they are seldom used currently.

Fig. 4. Heat transfer coefficient for a water spray quench as a function of surface temperature; Fig. 5. Heat transfer coefficient for an oil spray quench and an oil immersion quench as a function of surface temperature

How quenchants work

A key element in the induction heat treating process design is quenchant selection. Actual photographs of commonly used induction spray quenching processes have not been published to date. However, cooling rate mediation using different vaporizable quenchants such as water, aqueous polymer solutions and petroleum oil have been studied.

Three different stages of heat removal when quenching in vaporizable liquids are film boiling or vapor blanket cooling, nucleate boiling and convective cooling (Fig. 1). In film boiling, surface temperatures are sufficiently high to produce a stable vapor film around the part. Cooling rates during film boiling are characteristically low due to the slow heat transfer through the vapor barrier. The transition temperature from vapor blanket cooling to nucleate boiling is the Leidenfrost temperature. As the surface temperature decreases, the vapor film collapses and nucleate boiling begins. During nucleate boiling, the liquid contacts the hot surface and evaporates, producing vapor bubbles. Nucleate boiling produces strong convective currents resulting in high heat transfer rates. When the surface temperature cools below the boiling point of the liquid, the surface is permanently wetted by the fluid resulting in convective cooling. Convective cooling rates are relatively low and are primarily a function of the viscosity of the fluid.

Fig. 6. Wetting process of cylindrical silver sample (15 × 45 mm) quenched in a 10% aqueous solution of a polymer quenchant at 25˚C, no agitation

If fluid convection is sufficient, such as might occur with spray quenching, the vapor film may be ruptured by the convection either reducing or eliminating film boiling on the surface. However, if there is poor quenchant bath temperature control, the film boiling process may be nonuniform, leading to unsatisfactory hardness gradients, an increase in residual stress and possibly even cracking or unacceptable distortion control.

For example, it is generally believed to be crucial to maintain induction quench bath temperature to within ±1-2°C. Also, for aqueous polymers and water quenches, it is often required that the quenchant temperature not exceed 30°C (85°F), because the film-boiling regime starts to become more stable above this temperature, leading to a nonuniform quench with increased problems with distortion control.

For water and oil quenching processes illustrated in Figs. 2 and 3, all three cooling mechanisms may coexist on the cooling surface simultaneously [10-12]. This strongly affects the wetting processes and surface heat transfer occurring during quenching and the resulting thermal gradients that can form contribute to an undesirable nonuniform quenching process.

Nemkov and Goldstein showed the difference in heat transfer coefficients (a) for a water spray quench (Fig. 4), an oil spray quench and immersion quenching into agitated oil bath (Fig. 5) [6]. Water spray had the greatest heat transfer coefficient between 150 and 250°C (300 and 480°F) surface temperature during cooling. The maximum heat transfer-coefficient value for oil occurred in the range of 350 to 600°C (660 and 1110°F). The heat transfer coefficient was significantly greater for the oil spray quench than for the oil immersion quench [6].

Fig. 7. Effect of polymer quenchant concentration on surface cooling rate

Aqueous polymer quenchants mediate heat transfer by the formation of a film around the hot metal surface (Fig. 6), which provides uniform heat transfer relative to the moving vapor front observable for oil and water [10-12]. Upon initial immersion, the polymer film encapsulates the vapor formed around the hot metal. At approximately the Leidenfrost temperature, the vapor blanket explosively ruptures, resulting in a pseudo-nucleate boiling process. The thickness of this insulating film is dependent on the concentration of the polymer quenchant (Fig. 7) [13], as well as the quenchant fluid temperature. Typically, quench severity decreases with increasing polymer quenchant concentration (increased film thickness), increasing quenchant temperature and decreasing flow rate. If agitation is excessive, the polymer film will simply be blown off the surface without significant cooling rate remediation.

Fig. 8. Comparison of cooling curves measured at different positions in a cylindrical CrNi-steel probe (25 × 100 mm) during slow wetting (water) and during sudden wetting (aqueous polymer solution) at the center (a) and close to probe surface (b) at three indicated heights

Relatively little useful quantitative information regarding the critical surface wetting process is readily available to the designer. For example, Fig. 8 shows the core temperatures for water and a dilute solution of a polymer quenchant, such as that typically used for induction heat treating [10-12]. Cooling curve data provided at the center of the cylindrical probe suggest that both quenchants have the same quench severity. However, near-surface cooling curve data measured at three axial positions along the cylindrical probe show that the wetting process is clearly different for both quenchants. This example illustrates the necessity of using surface, or near-surface cooling data for induction system design, and it also shows that a polymer quenchant has the capability of providing significantly greater quench uniformity than water alone.

Because induction hardening is a case-hardening process, proper system design requires having available either time vs. surface or near-surface (1-2 mm, or 0.04-0.08 in., below the surface) cooling curves or heat transfer (or heat flux) as a function of surface temperature for the quench media being considered.

Look for Part 2 of this article in the December 2005 issue of IH.