PULSE PLASMA NITROCARBURISING OF GAS SHOCK ABSORBER TUBES FROM STEEL W.No. 1.0116 NITROKARBURIRANJE CEVI PLINSKEGA BLAŽILCA IZ JEKLA W.No. 1.0116, V PULZIRAJOČI PLAZMI VOJTEH LESKOVŠEK1, M. DOBERŠEK, A. RODIČ Vacuum Heat Treatment and Surface Engineering Centre, Institute of Metals and Technology, Lepi pot 11, 1000 Ljubljana, Slovenia Prejem rokopisa - received: 1997-10-01; sprejem za objavo - accepted for publication: 1997-10-21 Gas shock absorber tubes from steel W.No. 1.0116 vvere pulse plasma nitrocarburised at 560 and 580°C in atmosphere containing carbon dioxide additions. Variations in compound layer structure, thickness, porosity, hardness-depth profiles and stability of the process vvere investigated. It vvas found that the formation of predominantly £ phase compound layer structure is promoted by high nitrogen atmosphere. The compound layer thickness vvas found to increase vvith increasing nitrogen content, temperature, and pulse as vvell as decreasing carbon dioxide content. The experiments shovved that by use of high nitrogen atmosphere no pores appear in compound layers on steel W.No. 1.0116. Proper process control and the addition of carbon dioxide to the atmosphere vvith high content of nitrogen result in a reasonab!y thick compound layer from predominantly E phase. Key vvords: pulse plasma nitrocarburising, compound layer. structure, porosity. hardness-depth profile, thickness, stability of the process Nitrokarburiranje cevi plinskega blažilca iz jekla W.No. 1.0116 v pulzirajoči plazmi je bilo izvršeno pri temperaturi 560°C in 580°C v atmosferi z dodatkom ogljikovega dioksida. Raziskali smo mikrostrukturo spojinske plasti, debelino, poroznost ter profil trdote nitrokarburirane plasti in stabilnost procesa. Ugotovili smo, da se lahko z uporabo atmosfere, ki vsebuje visok odstotek Ni, poveča tvorbo spojinske plasti, ki je pretežno sestavljena iz E faze. Ugotovili smo, da debelina spojinske plasti raste z vsebnostjo N2, temperaturo in frekvenco pulziranja ter z zniževanjem vsebnosti ogljikovega dioksida v atmosferi. Eksperiment je pokaži, da visoka vsebnost N2 ne vpliva na nastajanje por v spojinski plasti, ki se tvori na jeklu W.No. 1.0116. Pravilno vodenje procesa ter dodatek ogljikovega dioksida v atmosfero z visoko vsebnostjo N2 omogoča doseganje sprejemljive debeline spojinske plasti, ki je sestavljena pretežno iz E faze. Ključne besede: nitrokarburiranje v pulzirajoči plazmi, spojinska plast, struktura, debelina, poroznost, profil trdote, ponovljivost procesa 1 INTRODUCTION Pulse plasma nitrocarburising has grown to a thermo-chemical heat treating process of particular technological and economic importance for improving the surface characteristics of components for mechanical engineering. The most important advantages are greatly improved wear, fatigue, and corrosion resistance. As a low temperature process it minimises distortion and volume change of parts treated. Pulse plasma nitrocarburising produces at the surface compound layer of a few mi-crometers in thickness consisting mostly of e phase, and a significantly thicker diffusion zone. As opposed to con-ventional salt bath and gas processes, nitrocarburising in glow discharge pulse plasma affets several advantages. In addition to its increasingly important environmental acceptability, plasma nitrocarburising because of the pos-sibility of freely selection at a wide variety of process parameters also makes it possible to match the properties of the layers to the specific load profile of the applica-tion. In the present paper a short survey of theoretical as-pects of pulse plasma nitrocarburising is given and its significance as industrial process is assessed. The experi-ments were carned out on two charges of gas shock absorber tubes vvith different load configurations and process parameters to produce at the surface an up to 10 |im thick compound Iayer consisting predominantly of e phase. 2 SOME THEORETICAL ASPECTS OF PULSE PLASMA NITROCARBURISING The pulse plasma nitrocarburising process is charac-terised by a glovv discharge surrounding the workpieces surface, vvhich appears at lovv pressures vvhen a voltage is applied betvveen the vvorkpieces and the furnace wall. The vvorkpieces are held at a negative potential (cathode) vvhile the furnace beli is positive (anode). The furnace must initially be pumped dovvn to a lovv pressure and subsequently refilled vvith a suitable gas mixture to a pressure of -1-10 hPa. A voltage of 200-1000 V is then applied, vvhereupon the gas species in the reaction vessel becomes ionised. By controlling the povver input the vvorkpieces can be heated to the nitrocarburising temperature, but better control of the magnitude and uni-formity of the temperature is obtained vvith auxiliary convection heating. Another way to control the heat input to a greater precision is to apply a pulsating voltage and supply the furnace vvith heating elements. In this manner the vvorkload temperature can be control led vvith the retention of a high degree of glovv discharge control. The problems vvith are formation, local overheating, and the hollow cathode effect can thus be minimised. The pulse frequency that could be used ranges betvveen DC and 33 kHz. To obtain uniform nitriding depths and uniform compound layer thickness it is important to empha-sise that process parameters - applied voltage, pulsing frequency, load configuration, etc,- must be selected em-pirically. Load configuration is very important. If components are placed too close together the hollovv cathode effect can cause severe problems. The compound layer structure and the depth of the diffusion zone can be controlled by varying the gas mix-ture, pulsing frequency and temperature. Normally the compound layer contains only the y phase. To achieve a compound layer consisting of e phase on steels a gas mixture of nitrogen, hydrogen and carbon dioxide is used. The base for understanding the events during the formation of the compound layer and the diffusion zone is provided by the ternary Fe-N-C phase diagram. The tvvo significant binary systems making up the sides of the ter-nary system, Fe-C and Fe-N1, are vvell knovvn. Hovvever, many observations are in disagreement vvith this dia-grams. These deviations are often explained in terms of kinetics or orientational relations betvveen the lattices of the various phases prevailing over the pure thermody-namic equilibrium in the formation of these phases. For these reasons a nevv ternary phase diagram has been pro-posed by J. Slycke et al.2. This diagram eliminates ali ambiguity regarding the interpretation of hovv structures have evolved, since ali observations can be explained by the local equilibrium approach. The major difference betvveen the nevv diagram and that published by Naumann et al.1 is that it allovvs the frequently observed direct contact betvveen ferrite (a) and e phase. The nevv Fe-N-C phase diagram shovvn in Figure 1 is characterised by tvvo Figure 1: Schematie ternary Fe-N-C phase diagram at ~580°C suggested by J. Slycke et al.2 Slika 1: Shematski ternemi Fe-N-C fazni diagram pri ~580°C po J. Slycke et al.2 Figure 2: Schematie illustration of the evolution of the diffusion zone and compound layer during nitrocarburisig of carbon steels3 Slika 2: Shematska ponazoritev razvoja difuzijske in spojinske plasti pri nitrikarburiranju ogljikovih jekel3 three phase equilibrium (a + e + Y and a + Fe3C + e) and one ternary tvvo phase field (a + e). The e phase field extends tovvards lovver nitrogen contents vvith increasing carbon content as consequence of the stabilising effect of carbon on the e phase, vvhich can exist at a lovver nitrogen content (-5% N) than the y phase (5.9% N). The different sequences during the nitrocarburising process of lovv carbon steels3 are shovvn schematically in Figure 2. During this process e phase and y phase are formed vvithin the evolving diffusion zone, follovved by the formation of y and e phases in the fully developed and grovving compound layer. 3 EXPERIMENTAL PROCEDURE 3.1 Material and process of pulse plasma nitrocarburising The gas shock absorber tubes 28 x 175 mm used in the present vvork are from steel W.No. 1.0116 vvith the chemical composition given in Table 1. Table 1: Chemical composition of steel W.No. 1.0116, (in wt-%) Tabela 1: Kemijska sestava jekla VV.No. 1.0116 (v ut.%) C Si Mn P S Cr Mo Ni Al Cu WNr 1.0116 0,15 0,013 0,53 0,011 0,008 0,023 0,006 0,019 0,049 0,031 The pulse plasma furnace used vvas a GP 1000/80 M nitriding unit manufactured by Metaplas-Ionon GmbH, Figure 3. The furnace is equipped vvith a convection heating system and vvith an internal gas/vvater heat-ex-changer for quick cooling. The gas shock absorber tubes in load configuration as shovvn in Figure 4 vvere pulse plasma nitrocarburised at 560 and 580°C at 5,2 hPa and 2,8 hPa pressure, using a total gas flovvrate of 100 and 67 1 h1. The gas shock ab- Figure 3: Pulse plasma nitriding furnace GP 1000/80 M Slika 3: Peč za nitriranje v pulzirajoči plazmi GP 1000/80 M sorber tube's temperature was measured with two chromel-Alumel termocouples embedded in a tubes on two different levels (top and bottom) in the first and the third circuit. Gas atmospheres vvith nitrogen contents 70,5 and 87%, carbon dioxide contents 2,5 and 2% and hydrogen contents 27 and 11% respectively, vvere employed. The pulse frequency used vvas 2 and 2,5 kHz, respectively. Convection and plasma heating to process temperature took appr. 4h and the isothermal treatment lentghs vvere 10 and 4 hr respectively, follovved by forced cooling in a flovv of nitrogen. 4 RESULTS AND DISCUSSION 4.1 Compound layer structures In pulse plasma nitrocarburising, the large number of freely definable treatment parameters make it possible to control precisely the structure, composition, and grovvth characteristics of the compound layer vvithout impairing the formation of the diffusion layer. The tubes nitricarburised for 10 and 4 hr in the pulsed plasma mode, considering the 0,480 ms glovv-on time, 0,020 ms glow-of time, the 0,200 ms glovv-on time, and 0,200 ms glovv-of time, respectively vvere transver-sally sectioned at the middle, metallographically prepared, and etched vvith 3% nital for the determination of the thickness of the compound layer on an optical micro-scope. The thickness of the compound layer vvas taken by averaging five measurements for each tube. Figure 5 shovvs the microstructure of the nitrocarbur-ised layer on the outer surface of the 5 tubes taken from the third circuit from 5 different levels of the charge no. 1 (560°C). The charge no.l contain 685 tubes in 5 circuits vvith 20 mm intercircuit distance, processed for 10 hr in pulsed plasma mode, considering the 0,480 ms glovv-on time and 0,020 ms glow-of time. The compound layers thickness appears to be 5 to 8 |im and the grovvth rate 0,5 Figure 4: Gas shock absorber tubes in load configuration Slika 4: Razporeditev cevi plinskega blažilca v sarži to 0,8 (im/hr. Metallographical analysis (Figure 5) indi-cates that the compound layer on tubes on the top and bottom (sample 1 and 5) obtained by the above pulse plasma nitrocarburising parameters mainly consisted of Y phase - Fe4(N,C)i-x. While the compound layers of the tubes betvveen two (samples 2 - 4) consisted beside of Y phase also of e phase- Fe2-3(N,C)i.x and carbide particles. In the compound layers grain boundaries are parallel to the diffusion direction and thus perpendicular to the surface. Under compound layers of these same tubes (samples 2 - 4) a 15 to 25 |am thick fringe of pearlite is found, vvhile this fringe is not found under the compound layer on top and bottom tubes (samples 1 and 5). •>: .-s?' t-s;-:.*,-*^, t,- • *<«;:<• • M. _' - J , ^ 0 J- , • ' V - > \ ; '"T - * ■ ' v,,,. v* V - ."t / ■ "V * A* * »- i? fr/t^T z***-1 J-\- "V , "i'/-- ^■ 'x "t- -V v * - -r^v: - ^ Figure 5: Microstructure of the nitrocarburised layer at the outer surface of tubes (200x) Slika 5: Mikrostruktura nitrokarburirane plasti ob zunanji površini cevi (200x) From the obtained microstructures it is possible to conclude that the temperature within the charge vvas not equalised or the distribution of gas mixture because of charge configuration was not fulfilled. The analysis of the process parameters actually shovvs that in spite of the relative high pressure (5,2 hPa) the temperature differ-ence between the two thermocouples was too large (~25°C), because of the hollovv cathode effect caused lo-cal overheating. This effect called the hollovv cathode can occur when the cathode drops to a dimension equal to the distance between tubes. Nitrocarburising depth on 5 tubes from top to bottom is shown in Table 2. Table 2: Metallographically determined nitrocarburising depth on tubes from charge no. I Tabela 2: Metalografsko določene globine nitrokarburirane plasti na ceveh iz sarže 1 No. of tube Nitrocarburising depth in mm 1 0,43 2 0,48 3 0,46 4 0,43 5 0,47 Microhardness profiles HV0,3 of the nitrocarburised layers for the same 5 tubes (top and middle of the tube) are shown in Figure 6. From the figures 5 and 6, it can be seen that the microstructure and hardness of diffusion layer on tubes from different levels are affected by the hollovv cathode efffect, vvhich causes also the irregular thickness and the type of compound layer vvhich is also too thin. As mentioned above the goal of the present vvork vvas to obtain at the surface of the gas shock absorber tubes a compound layer consisting predominantly of e phase vvith a thickness of > 10 |Ltm. In the čase of the gas shock absorber tubes the thickness of diffusion zone is not so very important. Such compound layer produced by pulse plasma nitrocarburisig process improves corrosion and HV 0.3 Figure 6: Microhardness profiles HV0,3 of the nitrocarburised layers Slika 6: Profil mikrotrdote HV0.3 nitrokaburirane plasti vvear resistance and after fine polishing one can obtain also satisfactory level of decorativeness. In order to avoid the hollovv cathode effect causing local overheating of tubes it vvas necessary to rearrange the load configuration in such a way that the distance betvveen tubes and each next circuit is appr. 1,3 time the distance betvveen tubes. With nevv load configuration charge no. 2 processed at 580°C contains 552 tubes in 6 levels. The thickness of the compound layer produced at 580°C vvith a 4 h treatment depends not only on the ni-trogen and carbon dioxide levels in the treatment atmos-phere, but also on the pulse frequency used. The results indicate that increasing the pulse frequency and the ni-trogen level in the atmosphere an increased compound layer thickness > 10 and a surface hardness in excess of 320 HV1 are obtained. The grovvth rate becomes greater than 2,5 pm/hr. This tendency is mostly due to the increased nitrogen "activity" in the plasma. Figure 7 shovvs the typical microstructure of the nitrocarburised layer on the outer surface of the shock absorber tubes vvhich consists predominantly of e phase -Fe2-3(N,C)i.x. The compound layer at the surface consists of 2 pm thick monophase range of e - Fe2-3(N,C)i.x, and it is follovved by a tvvo-phase field (e + y) up to 6 p.m thick. In the compound layer grain boundaries are paral-lel vvith the diffusion direction and thus perpendicular to the surface. The diffusion zone belovv the compound layer consisted of the eutectoid constituent braunit (a + Y) from the binary Fe-N1 system and some islands of pearlite and needles of y. The present experiments have confirmed that the use of high nitrogen atmospheres pro-duces no pores in the compound layers developed on steel W.No. 1.0116. 4.2 Reliability survey The reproducibility of the process is shovvn in figure 8, vvhich presents the results of the processing of 23 S ■8 +y' s phase - Fe2_3(N,C)1. f phase - Fe4(N,C)1.x B - braunit - (cc+y) P - perlit Figure 7: Microstructure of the nitrocarburised layer on the outer surface of the shock absorber tubes Slika 7: Mikrostruktura nitrokarburirane plasti na zunanji površini cevi plinskega blažilca fluctuations in the hardness of the core materials, i.e. 150-230 HV1 and used UIC (Ultrasonic Contact Imped-ance) hardness testing method4 can affected the nitrocar-burising results and are not taken into account in the sta-tistical control. Clearly, improved control over the variations experienced in core hardness vvould be ex-pected to reduce the amount of scatter observed in the nitriding results. Fortunately, the quality requirements are already satisfied to sueh an extent that these effects may be disregarded. 5 CONCLUSIONS Figure 8: Reproducibility of the pulse plasma nitrocarburising process on 23 charges Slika 8: Ponovljivost procesa nitrokarburiranja v pulzirajoči plazmi pri 23 saržah Mn,..: 3ZZ.00 I1ax..: 521.DO Range.: 199.00 Xq v : 4BB.B2 SO« t.: 28.95 VB ir;: 1174,79 |) ± lir: 67.5 Z j> ± Za95.1 X C ±31: 95.3 X P ±>3ir: 0.7 Z Cp....: 1.065 CpK: 1.046 p - Nean value - Toleranca -3c ~Zr -lff li »lr *Zc »3a _ — — 1 1 -- — -ff I -- BFl i 1 1 3Z2 346 379 334 41B 442 466 Surface hardness HV1 „ Figure 9: Statistical consideration of surface hardness HV1 data on 23 charges Slika 9: Statistična obdelava podatkov površinske trdote HV1 pri 23 saržah On the basis of the present experiments, it is con-firmed that the atmosphere consisting of 87% N2 + 2% CO2 + 11 % H2 and inereased pulse frequency are re-quired in the glow diseharge to produce a compound layer thickness > 10 |im without pores and vvith a pre-dominantly e phase strueture. From the results is evident that higher nitrogen con-tent in atmosphere, higher temperature and inereased pulse frequency strongly influenced the compound layer grovvth rate vvhich is by the second charge nearly three times higher than in by the charge no. 1 processed in a less suited atmosphere. With a nevv load configuration charge no. 2 and inereased pulse frequency the phenome-non called the hollovv cathode effect, vvhich causes local overheating among the tubes, vvas also avoided. Beside this, it is very important that pulse plasma ni-trocarburisig quickly overcomes its sometimes stili unfa-vourable image and that this modern and progressive technology is recognised as a reliable čase hardening process. What good are ali the remarkable technical, eco-nomical and environmentally clean properties of the process if it cannot be justified in practical use under in-dustrial conditions? The results presented in this vvork prove that sueh a demand can be fulfilled. charges. A typical charge may contain up to 552 tubes vvith a net vveight of around 94 kg. Statistically confirmed quality control data for surface hardness HV1 are shovvn in figure 9. These distributions satisfy customer's ± 3o (standard deviation) quality requirements, that mean 95% of ali pulse plasma nitrocarburised tubes have to meet this specification. Hovvever, it should be pointed out that 6 REFERENCES 'F. K. Naumann and G. Langenscheid: Arch. Eisenhiittenwes., 36 (1965) 677-682 2 J. Slycke, L. Sproge, and J. Agren: Scand. J. Metali., 17 (1988) 122-126 3 J. SIycke and L. Sproge: Surface Engineering, 5 (1989) 2, 125-140 4 Krautkramer: Technical Reference and Operating Manual, 1992, Ident. No. 28894