4.9 Heat Treatment P. Mayr, H. Klümper-Westkamp, Stiftung Institut für Werkstofftechnik IWT, Bremen, Germany 4.9.1 Introduction After mechanical manufacturing, heat treatment is mainly applied in one of the final steps of production, to adjust the workpiece properties to the later mechani- cal, tribological, and corrosive load. The lifetime of these components is defined through the raw material used, the construction geometries, and the quality of heat treatment. The great demands for quality control and the documentation of process param- eters have led to the increasing importance of sensor applications in heat treat- ment. Furthermore, increasing automation combined with the idea of integrating the heat treatment into the production line needs many different sensors to make the whole process as reliable as possible. 4.9.2 Temperature Monitoring The primary process variable to be measured and controlled in heat treatment is the temperature, which directly influences component properties. There are a very large number of different temperature-dependent physical properties that can serve, at least in principle, as the basis for thermal sensors. Only a few of them are used in heat treatment plants to measure the temperature. Thermocouples are the most commonly used temperature sensors in heat treat- ment. They consist of two conductors of different metallurgical composition, which are made up into an electrical circuit with two junctions. If the two junc- tions are brought to a different temperature, a thermoelectric potential is created. The magnitude depends on the composition of the conductors chosen and the temperature difference [1]. Their simple construction allows sensors to be made which are able to provide reliable and accurate data under extreme measurement conditions. With thermo- couples a range from –270 to 2000 8C can be covered, depending on the material used. The accuracy depends on the temperature being measured and ranges from 0.1 K at room temperature to ± 10 K at very high temperature. The characteristics of various commonly used thermocouples are illustrated in Figure 4.9-1. Interna- tional, eg, IEC 584, and national standards, eg, DIN 43710 in Germany, have been set up. NiCr-Ni is the most frequently employed thermocouple in industrial plants up to 1100 8C. In most applications they are sheathed. The thermocouple wires are embedded in an isolating powder (eg, MgO, Al 2 O 3 ) and surrounded by a me- tal sheet (eg, Inconel, stainless steel). At higher temperatures, PtRh-Pt thermocou- ples are used. They should be protected by ceramic tubes [2]. 4 Sensors for Process Monitoring326 Sensors in Manufacturing. Edited by H.K. Tönshoff, I. Inasaki Copyright © 2001 Wiley-VCH Verlag GmbH ISBNs: 3-527-29558-5 (Hardcover); 3-527-60002-7 (Electronic) According to Planck’s law, the spectral radiation of a blackbody at a particular wavelength varies only with the blackbody temperature. The optical pyrometer makes use of this principle by measuring ratios of spectral radiances. The radia- tion characteristic of a blackbody is shown in Figure 4.9-2. The wavelength spec- trum utilized for radiation thermometers usually lies between 0.5 and 20 lm. The use of radiation thermometers is simple when dealing with surfaces showing blackbody condition. Unfortunately, in most applications deviations from black- 4.9 Heat Treatment 327 Fig. 4.9-1 Thermal emfs of commonly used thermocouples [2] Fig. 4.9-2 Characteristic radiation of a blackbody [3] 10 4 Wcm ±2 l m ±1 sr ±1 Spectral radiance L k S 10 3 10 2 10 1 10 0 10 ±1 10 ±2 10 ±3 10 ±4 10 ±5 6000 K 3000 K 1000 K 500 K 300 K 200 K 0.1 0.2 0.4 0.6 1 2 4 6 10 20 40 l m 100 Wavelength k body condition occur as a result of absorption, emissivity, and reflection effects. The most difficult problems arise with unknown and variable emissivity. Optical pyrometers can be subdivided into four groups depending on the wave- length and wavebands used: · total-radiation thermometers; · single-waveband thermometers; · ratio thermometers; · multi-waveband thermometers. The very common single-waveband thermometers with silicon or germanium de- tectors use short wavelengths between 0.5 and 2 lm. For very low temperatures longer wavelengths are used because of a higher rate of change of energy with temperature. Ratio thermometers measure the ratio of energy in two narrow wave- bands. The advantage is that their readings will be independent of target size, ob- scurity of optical path, and emissivity, provided that they are the same at the two wavelengths [2]. The electric resistance of some metals increases with increase in temperature in a nearly linear fashion. This property has been applied for temperature measure- ment. Platinum is the most precious metal used for thermometry. Platinum can be drawn in to very thin wires of high purity. Thick- and thin-film sensors are also available. Ruggedness and relatively low costs are combined in platinum thick-film ther- mometers. They are used in industrial applications up to 450 8C when encased in ceramic, glass, or other protective materials. Atmospheres containing carbon or hydrogen are poisonous for platinum and lead to unstable resistance. Resistance thermometers have the advantage of high accuracy measurement. Typically, a tem- perature resolution of ± 10 –3 8C can be achieved up to 500 8C whereas thermocou- ples are limited to ± 0.5 8C. Special geometries are available for surface tempera- ture measurement. In special applications other metals such as copper, nickel, and iridium are used [2]. A number of physical effects have been checked to realize fiber-optic temperature sensors [4]. The main fiber-optic thermometers today use the temperature depen- dence of the fluorescence decay time. This effect has been utilized in commercial instruments by Luxtron and Sensycon. The main advantage of this temperature measurement method is the capability to allow measurements in severely hostile environments. They can be made in the presence of intense radiofrequency and mi- crowave fields as well as very high voltages and strong magnetic fields. In the range from –200 to 450 8C, an accuracy of ± 0.1 8C can be reached at the calibration point. In continuously working furnaces, temperature measurements can be made by furnace tracker systems. A data logger for thermocouples is positioned in a ther- mal protection unit made of inert and maintenance-free material. During the fur- nace travel the data logger stores the measured temperatures, which can be read later. Such a furnace tracker works up to temperatures of 1200 8C. Such systems have been developed and are sold by DATAPRO and Stoppenbrink. 4 Sensors for Process Monitoring328 4.9.3 Control of Atmospheres Heat treatment is done, depending on the aim of the treatment, in different fur- naces with different pressures, atmospheres, and plasma application. The scale in pressure ranges from vacuum up to high pressures of several hundred bar. De- pending on the aim of heat treatment, different gases and gas mixtures are ap- plied. When there are no requirements on the workpiece surface, heat treatment can be done in air, which means that oxidation takes place during heat treatment. To avoid this oxidation, inert atmospheres with low oxygen partial pressure must be used and controlled. Another group of processes is the thermochemical heat treatment processes. The purpose is to modify the workpiece by specific reactive atmospheres, so that carbon (carburizing) and/or nitrogen (nitriding) diffuses into the near-surface re- gion and parallel oxidation can take place at the surface. To control the oxygen partial pressure, the oxygen probe was the first developed in situ sensor in ther- mochemical heat treatment. With the knowledge of the oxygen partial pressure, many other process parameters can be evaluated. Some different constructions of oxygen probes are discussed in [5]. 4.9.4 Carburizing For controlling the carburizing process, the chemical equilibrium between the car- burizing gas and steel is used. Under the assumption of unchanged basic compo- nents CO and H 2 , the simplified value ‘CO 2 ’ or ‘dew point’ is derived, which is the main controlled variable for the carbon potential. It has been well established for more than 25 years that the partial pressure of oxygen measured by oxygen probes [6] can be used as an exact indicator of the carbon potential. This permits considerable progress with regard to accuracy, re- sponsiveness, reproducibility, and technical handling, so that this sensor is being increasingly used. The basic principle is an oxygen ion-conducting electrochemical cell of stabi- lized ZrO 2 . The measured potential E depends on the difference in oxygen partial pressure P O 2 at the two electrodes. P 0 O 2 is the known reference oxygen pressure, normally of air. In Figure 4.9-3 a schematic diagram of such a sensor is shown. The temperature T is measured in kelvin [7]. E mV0:0496 T logP O 2 =P 0 O 2 4:9-1 In oxygen-containing carburizing atmospheres, the following chemical reaction is the main indicator for the carburizing reaction: CO C 1 2 O 2 4:9-2 4.9 Heat Treatment 329 Under the assumption of thermodynamic equilibrium, the carbon activity a c is given by log a c logP CO =P O 2 0:5 À5927=T À 4:545 4:9-3 With known pressure of carbon monoxide P CO and temperature T, the carbon activity can be calculated from the measured oxygen partial pressure. The carbon activity of the carburizing atmosphere defines the dissolved carbon in austenite under the assumption of thermodynamic equilibrium. More information about oxygen-free carburizing atmospheres can be found in [7]. A further innovation is the application of oxygen probes in heat treatment, pri- marily known from automobiles for measuring the lambda point of combustion. When they are used in heat treatment plants they are normally located outside the furnace and measure the oxygen partial pressure on the basis of an electro- chemical unit at temperatures around 500 8C. The measured value has to be recal- culated to the furnace temperature and additionally it should be mentioned that owing to the outside mounting, more uncontrolled influences can change the value. As the measured oxygen potential can be regarded in most cases as integral information for the whole furnace, this system can also be useful for controlling the degree of sealing of the furnace. Because of the mass production of this lamb- da probe it is much cheaper but in many applications not as reliable as the stan- dard in situ oxygen probe. A direct method for measuring the carbon uptake in almost every atmosphere is the so-called wire sensor [8]. This sensor is based on the resistivity change of a thin, pure iron wire with increasing amount of dissolved carbon, as shown in Fig- 4 Sensors for Process Monitoring330 Fig. 4.9-3 Mode of action of an oxygen probe [7] ure 4.9-4. Periodically the wire is decarburized with hydrogen so that it can be used again. This in situ sensor is mostly applied for periodic checking of the abso- lute carbon potential in furnaces. Continuous control of the carbon potential in a furnace is not practicable because of problems with oxidation, formation of car- bides, and changing surface of the wire. This in situ sensor is sold by Process- Electronic. 4.9.5 Nitriding Gas nitriding is usually carried out in ammonia-containing atmospheres. Since Lehrer’s paper [9], it is known that the nitriding potential K N is the driving force of nitriding and defines the phase composition of the white layer. The nitriding potential K N is defined by the well-known ratio of the partial pressures of ammo- nia, P NH 3 , and hydrogen, P H 2 : 2NH 3 N 2 3H 2 4:9-4 K N P NH 3 =P H 2 0:5 4:9-5 In the case when pure decomposition of ammonia dominates the nitriding pro- cess, this nitriding potential presents a reliable indicator for the nitriding process. In nitriding atmospheres little is known about the basic level of the partial pres- sure of oxygen. Established process control have used external gas analyzers based on infrared absorption or heat conduction for many years. Recently, in situ sen- sors have been developed which are able to measure directly the nitriding poten- tial inside the furnace. They are based on oxygen probes containing a solid elec- trolyte of stabilized zirconia. The experimental approach to measuring the nitrid- ing potential with oxygen probes is different to that for the control of the carbon 4.9 Heat Treatment 331 Fig. 4.9-4 Probe tip of the wire sensor (Werkfoto PE-Essen) potential. For a precise function the concentration of water vapor present in the nitriding atmosphere should never drop below 1 vol.%. The measuring principle is as follows. The nitriding atmosphere with its resi- dual ammonia passes over the outer zirconia electrode of the sensor. Subse- quently it is dissociated totally into hydrogen and nitrogen by a 1000 8C nickel cat- alyst. This atmosphere with altered composition and lower oxygen potential reaches the inner electrode of the sensor. Between the two electrodes a potential difference according to the ammonia content in the nitriding atmosphere can be measured and the nitriding potential can be calculated [10]. DU mV0:0992 T log1 3 2 P NH 3 =P H 2 4:9-6 This method also works in nitrocarburizing atmospheres provided that no addi- tional gas components dissociate in the heated catalyst, changing the thermody- namic equilibrium of the water-gas reaction. Hydrogen, for instance resulting from decomposition of additional methane, shifts the water-gas and therefore in- fluences the oxygen partial pressure. H 2 O H 2 O 2 4:9-7 The utilization of this sensor for nitrocarburizing atmospheres should also work, provided that the inlet gas flow and composition are presicely measured. Some different types of construction are sketched in Figure 4.9-5. The main dif- ference is that the nickel catalyst is positioned in a second steel tube. Whereas types 2 and 3 pump the reference gas outside the furnace from the catalyst to the second electrode, type 4 works as two oxygen probes measuring two partial pres- sures (the dissociated and the normal atmosphere) against air. Some practical results of a complete measuring and control system for nitriding and nitrocarburizing processes were presented in [11], which demonstrate that the system is capable of reliable and reproducible operation. Important values of the nitrocarburizing process are ready for documentation to accomplish the claims of DIN/ISO 9000ff. Another method for measuring the hydrogen partial pressure in heat treatment atmospheres is described in [12]. Special membranes of palladium and palladium- silver are permeable only for hydrogen. With a pressure gage the hydrogen partial pressure can be directly measured beneath the membrane. 4.9.6 Oxidizing Oxidizing is of increasing interest. Since it became recognized that post-oxidation, directly conducted after nitrocarburizing, significantly increases the corrosion resistance of nitrocarburized steel parts, this combination is widely used today [13, 14]. 4 Sensors for Process Monitoring332 Pre-oxidation is also a process under discussion in order to prepare the surface after mechanical manufacturing for thermochemical treatment such as nitrocar- burizing and carburizing. Finally, oxidation as a stand-alone process for increasing corrosion and wear resistance can be a cheap alternative instead of nitrocarburiz- ing or chrome plating. The oxidizing atmosphere can consist of water, air, CO, CO 2 , and additional gas components. If the oxidation is performed in an uncontrolled manner, the type and thickness of the oxidized layer can vary over wide ranges and thus result in different surface properties. In order to provide the workpiece with an optimum magnetite layer on the surface, a certain oxidation potential has to be established. The main controlled variable for oxidation is the partial pressure of oxygen. This variable is easily measured by an oxygen probe containing a solid electrolyte of stabilized zirconium oxide. New developments in thin-film technology make oxygen sensors based on the change of resistivity much cheaper. The sensing material is TiO 2 [15]. Further materials are still under investigation. 4.9 Heat Treatment 333 Fig. 4.9-5 Different types of nitriding gas sensors (Werkfoto PE-Essen) 4.9.7 Control of Structural Changes In order to establish proper thermodynamic conditions for the desired heat treatment result, measurement of temperature and atmospheric constituents are necessary. In addition to the thermodynamic considerations, the kinetics play an important role. This aspect is monitored by a group of sensors which are able to measure the result of mass transfer directly. Such sensors have been developed making use of the eddy current technique [16]. Nitriding treatment not only influences the mechanical and chemical properties of a steel but also modifies the electrical and magnetic characteristics in the near- surface regions. For this reason, measurement of electrical and magnetic changes will yield substantial information about the progress of nitriding and nitrocarbur- izing. As shown in Table 4.9-1, the electrical resistance, the magnetic permeability, and the Curie temperature are changed. The so-called KiNit sensor is shown in Figure 4.9-6. The main components are the coil and the exchangeable specimen mounted as the coil core. The coil and the mounting unit are specially designed so that temperature and the nitriding atmosphere will not change the electromag- netic properties, even after prolonged exposure. The complex impedance of this configuration is measured and correlated with the nitriding and nitrocarburizing parameters. For each steel group a calibration is necessary to obtain quantitative information about white layer thickness and nitrided case depth. Using tempera- ture-dependent measurements the phase composition of the white layer can also be detected. Many problems in nitrocarburizing are connected with passivation. A thin layer in the surface prevents mass transfer into the material. Most passivation thin films originate unintentionally during manufacturing and finishing processes or are present due to immediate oxidation on high-chromium containing steels. An example is shown in Figure 4.9-7. At the beginning, after heating up, the sensor signal becomes constant with time, but after chemical activation of the surface the eddy current loss decreases, which means that nitrocarburizing takes place. After 7 h of nitrocarburizing a precipitation layer of about 170 lm is measured. 4 Sensors for Process Monitoring334 Tab. 4.9-1 Electrical and magnetic properties of iron and iron compounds [17] Material Specific electrical resistance (10 –8 Xm) Curie temperature (8C) Relative magnetic permeability at 500 8C Fe 10 770 10 2 –10 4 Fe with 1 wt% N +19 ? Reduced Fe 4 N Similar to Fe 3 N 480–510 1 Fe 3 N 4500 294–299 1 Fe 3 (N, C) Similar to Fe 3 N 320–388 1 Fe 3 C 90 213–223 1 Another application is nitrocarburizing with a well-defined surface layer thick- ness. Figure 4.9-8 shows an example of how it is possible to nitrocarburize a 6.5 lm thick white layer with any even larger case depth. After the sensor signal shows the defined layer thickness, the nitriding potential is decreased to that level, so that no further white layer grows. In this way, any combination of white layer thickness and case depth can be precisely obtained. Even nitrocarburizing without the occurrence of a white layer is possible and reproducible by using this eddy current sensor. 4.9 Heat Treatment 335 Fig. 4.9-6 KiNit sensor (Werkfoto Ipsen-Kleve) Fig. 4.9-7 KiNit sensor signal during nitrocarburizing of a passivated probe (X40Cr13) [16] Treatment time (h) . intense radiofrequency and mi- crowave fields as well as very high voltages and strong magnetic fields. In the range from –200 to 450 8C, an accuracy of ± 0.1