Surface and Coatings Technology, 59 (1993) 129—134
Process control of plasma nitriding and plasma nitrocarburizing in industry W. Rembges and W. Oppel Klöckner lonon GmbH, Leuerkusen (Germany)
Abstract Whenever plasma processes are used under continuous production conditions to produce nitrided or nitrocarburized layers, a variety of parameters has to be controlled in order to guarantee consistency of quality. Some of these parameters must be considered during the engineering period ofan installation. It is shown that a minimum power density is required independently ofthe applied frequency to achieve saturation of the surface with nitrogen. Computer simulation demonstrates the influence ofthe effective wall temperature on the power density and temperature uniformity. Compound layers with and without porosity can be obtained using different gas compositions. Additionally, iron oxide layers can be produced in a plasma on top of a compound layer. A microprocessor system with special software is presented which fulfils the requirements of modern statistically controlled production.
1. Introduction With an increasing interest in environmental problems, plasma processes have begun successfully to approach various industries and have reached an industrial stage from the point of view of performance. They are replacing conventional processes, mainly to avoid pollution problems and to reduce production costs . Plasma processes such as plasma nitriding and plasma nitrocarburizing (PN and PNC), well known under the trade name lonitriding (registered trademark of Klöckner lonon GmbH, Leverkusen), are environmentally harmless, economical and have favourable technical properties. Several process parameters (time, temperature, pressure, gas mixture, power density) enable process control, and external parameters (material, furnace design, loading, pretreatments such as machining, cleaning, etc.) influence the final results. This paper discusses some of the parameters which are controlled by in situ and remote control techniques, and parameters which had to be determined in order to control the heat losses created by different furnace designs.
2. Easily controllable process parameters 2.1. Time, temperature, material Plasma nitriding and plasma nitrocarburizing are nitriding processes which can be applied to more or less all types of steel. The implantation of nitrogen is performed in a low pressure current intensive glow discharge by a combination of sputtering and a temperature
independent condensation process . The subsequent diffusion of nitrogen into the surface depends on the nitriding temperature and the chemical composition of the material. Figure 1 shows as an example the dependence on the square root of time of the nitriding hardness depth of three different types of steel (1%, 2.5% and 5% chromium steels). The curves show the influence of time and materials. Alloying elements such as chromium, aluminium, vanadium etc. which form nitride precipitates, influence the diffusion mechanism, for example by reducing the p05sible case depth with increasing content of alloying element. Thus the chemical compositions of steels are external parameters which have to be considered when setting the process parameters but cannot be controlled during the heat treatment process. Cas, depth Immi 1 Te~per~t~r~52O’C
:: 0.2 0
10 Square root of Time k41
Fig. 1. Nitriding hardness depths at 520 °Cas a function of the square root of time for three different types of steel.
Elsevier Sequoia. All rights reserved
W. Rembges, W Oppel
Process control in industry
Plasma nitriding can be performed in a temperature range beginning at about 400°C. Under economical conditions most heat treatable steels are nitrided at more than 500°C.Nevertheless, the treatment temperature is important for the dispersion hardening of chromium nitride precipitates within the matrix. To determine which PN temperature really gives the greatest strengthening effect of alloyed steels, hardness values were compared of samples treated to the same case depth of 0.1 mm at various temperatures. This is shown in Fig. 2, where the surface hardness is plotted against the PN
to reproduce the layer thickness in a reliable way. Of the parameters to be controlled, the gas composition is the most important. Gas compositions have to be mixed within an accuracy of 1%. For the PNC process the carbon-carrying gas should be mixed with an accuracy of more than 0.1%, independently of the actual gas consumption. This is made possible by a patented gas mixing system which allows up to four gases to be mixed with the required accuracy. Figure 3 shows a typical c-compound layer after a PNC process with a C02-containing gas mixture. The
temperature. This shows that the maximum hardness is achieved at a temperature of 450°Cand is caused by the precipitation hardening effect. The size and number of precipitates at 450 °Care the most effective for blocking the movement of dislocations, thus causing the maximum strengthening and hardening effect.
well known non-porosity is not achieved using this type of gas mixture, but the porosity is less than that produced using salt bath or gas nitrocarburizing techniques. However, carbon-free PN processed steels can be produced with no pores, as shown in Fig. 4. The enhanced corrosion resistance produced by PN and PNC is used in many applications. Tests of components for use in the glass industry have shown them to
2.2. Control ofgas composition While the diffusion layer and case depth are influenced by the time, temperature and the chemical compositions of steels, the gas mixture in the plasma mainly influences the chemical composition of the compound layer [4, 5]. Under daily PN production conditions, a composition of nitrogen and hydrogen is mostly used. Methane or carbon dioxide is added whenever PNC is applied, especially when automotive components are nitrided to build up an c-compound layer on low alloyed or unalloyed carbon steel. For PN of heat treatable, nitriding or hot working steels, comprehensive investigations with regard to ductility, wear resistance to friction etc. have shown that whenever the compound layer consists of a monophase y’-compound layer the properties are considerably enhanced 16 71 L As far as the thickness of the nitride layer is concerned, it is well known that the ductility decreases with increasing layer thickness. In daily production it is important .
‘~ - ~..
Fig. 3.(Original Microstructure of an c-compound layer, PNC processed with CO2. magnification 1000: 1.)
Surface hardness [HVI
450 500 PN - Temperature [°C)
Fig. 2. Comparison ofhardness values ofPN layers ofsimilar thickness for different nitriding temperatures.
Fig. 4. Microstructurc of a pore-free ,“-compound layer, PN processed in a carbon-free gas mixture. (Original magnification 1000: 1.)
W. Rembges, W. Oppel / be better or comparable with chromium-plated components whenever PNC is applied in a C02-containing gas. A further improvement in the corrosion resistance can be achieved using a special plasma-IONOX process. Plasma-IONOX produces oxidation in a specially developed gas composition after the normal PNC process. Figure 5 shows a normal c-compound layer with an iron oxide layer about 1—2 jim thick.
~oo 400 300
freqoercy Poised ow IreguenCy C__________________________________ 0—~I I I
3. Additional controllable parameters Influence ofpower density To produce plasma reactions in a power intensive glow discharge, minimal power densities are required . Too low a density produces This is independent of the type of non-uniform power sourcelayers. used (d.c., pulsed d.c., low frequency 1—10 kHz, or high frequency above 30 kHz pulsed systems). Statistical investigations have shown this to be the case. A typical result of these investigations is shown in Fig. 6. This shows the hardness profiles of 42CrMo4 steel after PN processing with low and sufficiently high power densities at the same temperature of 550 °Cfor 12 h. In all three cases where the power density is too low the nitriding is insufficient, independent of the type of power source. In order to avoid overheating, the design of the furnace must ensure that the minimally required energy which is produced by the plasma can be radiated to the effective furnace wall. Because the pressure in a PN furnace is below 10 mbar, only radiation occurs between the load and the effective furnace wall. The radiated energy Qr can be easily calculated using Planck’s radiation law: 3.1.
Process control in industry
0.4 0.8 Distance from the Surface mm]
Fig. 6. Nitriding hardness profiles after the PN process with different power sources powerd.c.densities: ———,is 0.14W —*—, cm2. —i-— are 2 and and the pulsed power density 0.07w cm
c, are the emissivities of the load area and the effective wall surface area. F~are the surface areas of either the load or the surface area where the heat is radiated to. This area is called the effective furnace wall and is the heat shield which is placed on the inner side of the furnace which is heated directly to T~from the load temperature T~. At a given load it is obvious that the energy or heat loss which has to be generated by the plasma source is only influenced by the difference between the load temperature T~and the effective furnace wall temperature T~. 3.2. Computer simulations The following computer simulation describes the influence of the wall temperature on the heat loss and the power density at the area to be nitrided. It is assumed that round workpieces are loaded in three concentric rows in a PN furnace. At a given temperature of 570 °C
Fig. 5. Microstructure of an c-compound layer, PNC processed with an iron oxide layer of 1—2 ~sm.(Original magnification 1000: 1.)
which is regulated at the second row, the plasma power input and the temperatures at the three different rows are calculated as a function of the temperature of the effective furnace wall temperature. Six different temperatures were chosen for T~:100 °C,250 °C,420 °C,470 °C, 500°Cand 530°C. Figure 7 describes in a three-dimensional form the calculated temperature profiles at the given wall temperatures in the final stationary state. The power density required to regulate temperature of 570 °C varies 2 tothe 0.045 W cm2. The temperature from 0.25 varies W cmfrom ±40°Cat T~=100°Cto ±8°C difference at T~= 530 °C. The computer simulation program has been used to check the influence of heating by convection and/or radiation. It is clear that the final stationary profile is independent of the heating conditions. From Fig. 6 we have learned that 0.07 W cm2 is too low to achieve uniform nitriding when 42CrMo4 is used. Thus a good uniformity
W. Rembges, W Oppel
power density is less compared with higher alloyed
or better are possible.
500 600 300
045 W/cm2 .08 W/cm2 10 w/cm2 W/crn2 25 W/cm2 .22 Wicm2
Wail 1.R0w2.Row3.R0w3.Row2.Rowl.Row Wall Potltior of load
Fig. 7. Temperature profiles at different effective furnace wall temperatures and corresponding power densities,
of temperature will result in non-uniformity of the hardness profile. The only way to achieve uniform nitriding is to increase the heat loss by choosing a wall temperature of 420 °C. In this case a power density of 0.14 W cm2 is achieved and a temperature uniformity of ±24°C. In some cases a temperature distribution of ±24°C is acceptable, but in other cases it is not. One way to improve the conditions is to reduce the load surface area by loading fewer workpieces. In this case the process is less economical. Thus, another possibility has to be found to improve the temperature uniformity without losing the minimum power density which is desirable economically. Figure 8 shows the result of a computer simulation where a central internal anode with a cooling device was installed. For a wall temperature of 420 °Cthe power density is slightly increased to 0.16 W cm2 but the temperature profile has been reduced to ±8°Cwhich is fully acceptable. For lower alloyed materials where the minimum Temperature
Process control in industry
3.3. Plasma furnaces with controlled and uncontrolled wall temperatures In several publications the above conditions are materials, improved temperature uniformities of ±5 °C described as “hot wall” or “cold wall” furnaces. These descriptions are misleading because a furnace is called a hot or cold wall when the outside wall which can be touched operatorair is hot or cold. Typical wall furnaces by aretheunforced cooled furnaces or hot furnaces with little insulation. Typical cold wall furnaces are forced cooled furnaces, either by water cooling or by air cooling. Assuming the same loading configuration in all described cases, the same nitriding results will be achieved whenever the inside temperature of the effective wall (i.e. the heat shield) is the same. This important phenomenon allows two different types of furnace to be constructed. The first type has an uncontrolled effective wall temperature; the second type allows the effective wall temperature to be controlled or regulated. Plasma furnaces with uncontrolled effective wall temperatures Plasma furnaces with uncontrolled effective furnace walls are vessels with one or several heat shields on the inside minimizing the heat loss of the load. On the outside of these vessels a cooling system, using either water or air, can prevent the outside wall becoming too hot. Furnaces of the same type are vessels which have no heat shield on the inside but are insulated either on the inside or on the outside by special dam material of low heat capacity. Whenever the furnace is cool (less than 40°C)on the outside and the “cold wall” furnace is equipped with heat shields on the inside the innermost heat shield reaches a temperature of 300—450 °Cdepending on the load temperature. This means it can only be determined by experience what the energy consumption will be. These types of furnace are normally equipped with only two heat shields which means a slightly higher energy consumption but greater production security because of superior, non-critical power density. Furnaces of this construction are difficult to overload with respect to maximum possible surface area. 3.3.1.
06 W; 0012 09 W/cm2
12 W10m2 W’cfll2 .25 W’cm2 .28 W,om2
Wall 1,R0w2.RowS.Row IA 3.RowS.Rowl.Row Wall Potition of Load
Fig. 8. Temperature profiles at different effective furnace wall temperatures and corresponding power densities after installation of an additional central internal anode,
Furnaces with insulation either on the outside or on the inside normally havedescribed a reduced energy consumption. According to the above effect of power density, these furnaces allow less surface area to be treated and can be overloaded, resulting in less uniform nitriding. 3.3.2. Plasma furnaces with controlled effective wall temperatures When the temperature of the surface area next to the load is controlled and regulated by a resistance or
W. Rembges, W. Oppel
plasma heater (cathode heater), the heat loss and the temperature uniformity of the load can be controlled. Resistance heaters are installed either on the inside or on the outside. An outside resistance heater can be insulated by dam material. If there is a space in between, cooled air can be blown through in a controlled way, allowing the power density to be increased if required. Improved temperature uniformity can be achieved by separating the heating zones into several sections. A similar method can be used for furnaces with a cathode heater. Separate heat shields are connected to the power source and are independently switched on and off according to the measured temperature distribution. Figure 9 shows a tandem installation equipped with the described regulation system. It is running in an automotive company producing components for automatic gearing systems. Depending on the component size and possible loading configurations, up to 12000 pieces can be processed reproducibly using PNC. During run off tests the furnace was equipped with ten thermocouples to monitor the uniformity. Figure 10 shows the maximum and minimum values of the normally installed four thermocouples which are controlled by a microprocessor, and the maximum and minimum values of the six additional thermocouples. The maximum deviation was measured to be ±8°C while the average was determined to be ±6°C which is very satisfactory. This low spread has turned out to be reproducible since the furnace is running under continuous production at the automotive company. 3.4. Tele diagnostic and process documentation Modern lonitriding installations are controlled by microprocessor systems such as industrially used PCs with a 386 or 486 processor. The software has to be
Process control in industry
540 T ~C]
530 520 510
500 ~ I 19:12
I I I
-B- M:rr L1-L4
I I 0-12
Fig. 10. Temperature distribution in a load of about 12 000 workpieces during PNC: TI—T6 test thermocouples; LI—L4 load thermocouples.
specially developed in order to guarantee reliable and reproducible processing . Figure 11 shows a typical diagram as visualized on a screen. All important data such as the voltage, current, pressure, four temperatures, disturbances, process time and date etc. are summarized on these data files. For quality control the process data are copied onto hard discs and can be analysed at any time. Examination of these data under magnified conditions enables detailed analyses to be carried out, even many years after the process has been performed. Tele diagnostic systems are available to control several units within the same plant at the supervisor’s office. Connections via a telephone modem allow trouble shooting thousands of kilometres away from the installation. This system means that expensive production stops can be avoided or minimized.
L cl __ __— ___
Fig. 9. Tandem lonitridinginstallation for PNC of automotive components, equipped with temperature uniformity control by plasma heaters: useful height 1800mm, useful diameter 1200mm. The furnace bell top (1) can be lifted and moved to a second operation position by the lifting device (2).
___ op. .‘ .~‘
Fig. I I. Process diagram of a complete cycle as visualized on a PC screen.
W. Rembges, W Oppel
4. Conclusions Plasma processes are environmentally clean, economical and allow surfaces with favourable technical properties to be produced. They have reached the industrial stage from the point of view of performance. Today they are substituting for conventional processes, mainly to avoid pollution problems and to reduce production costs . Compared with conventional surface hardening techniques, plasma processes have several significant advantages. To achieve these advantages in contract heat treatment shops or in factories under continuous production conditions for automotive components, equipment has to be installed which can be maintained by in situ process control and remote control techniques. In addition to the time and temperature, the material has a significant influence on the surface hardness and case depth. Variations in the plasma gas composition are responsible for the type and construction of the cornpound layer. Using these processes for daily production, the product quality can be influenced by the engineering and design of the furnace. Thermal insulation of the furnace reduces the flexibility of the process and limits the surface area which can be plasma nitrided or plasma nitrocarburized, and hence the profitability. Computer simulations show how to influence the temperature uniformity with regard to achieving a sufficiently high power density. Furnaces with controlled effective wall temperature offer superior production security. No
Process control in industry
advantage can be gained by varying the type of power source used in these systems. New developments are showing that the process can be used to produce oxide layers. Materials such as titanium or aluminium can also be plasma nitrided.
References 1 W. Rembges and J. Lühr, Plasma (ion) carburizing, applications and experiences, Proc. 2nd mt. Ion Nitriding—Carburizing Conf., Cincinnati, OH, 1989, American Society for Metals, Metals Park, OH, pp. 235—243. 2 J. Kölbel, Die Nitridschichtbildung bei der Glimmentladung, Forsch. Ber. d. Ld. NRW Nr.:1555, 1965. 3 B. Edenhofer and T. J. Bewley, Low temperature ion nitriding: nitriding at temperatures below 500°C for tools and precision machine parts, Proc. 16th mt. Heat Treatment Conf., 1976. 4 T. Bell and P. A. Dearnley, Plasma surface engineering, Proc. mt. Seminar on Plasma H.T., Senlis, 1987, pp. 13—53. ~ W. Rembges, Ion nitriding applications grow for automotive components, Heat Treat., March (1990) 27-30. 6 M. Week and K. SchlOtermann, Plasmanitriding to enhance gear properties, Metallurgia, 8 (1984) 328—332. 7 A. Roelandt, J. Elwart and W. Rembges, Plasma nitriding of gear wheels in mass production, Surf. Eng., 3 (1985) 187—191. 8 W. Rembges, W. Oppel and F. Hombeck, Plasma (ion) nitriding and plasma (ion) nitrocarburizing, its units and its applications, Proc. Conf. on Plasma Surface Engineering, DGM, Oberursel, Vol. 1, 1988, pp. 277—287. 9 W. Rembges and W. Oppel, Mikroprozessorsteuerung für die Warmebehandlung, Gas War. mt., 33(6—7) (1984) 349—353. 10 W. Rembges, Fundamentals, applications and economical considerations of plasma nitriding, Proc. mt. Conf. on Ion Nitriding, Cleveland, OH, 1986, American Society for Metals, Metals Park, OH, p. 189.