Effect of workpiece geometry on the uniformity of nitrided layers

Effect of workpiece geometry on the uniformity of nitrided layers

Surface and Coatings Technology 139 Ž2001. 1᎐5 Effect of workpiece geometry on the uniformity of nitrided layers C. Alves Jr.a,U , E.F. da Silvab, A...

800KB Sizes 0 Downloads 13 Views

Surface and Coatings Technology 139 Ž2001. 1᎐5

Effect of workpiece geometry on the uniformity of nitrided layers C. Alves Jr.a,U , E.F. da Silvab, A.E. Martinelli b a

Uni¨ ersidade Federal do Rio Grande do Norte, Department of Physics, Campus Uni¨ ersitario-Lagoa No¨ a, Natal RN 59072-970, Brazil ´ b Uni¨ ersidade Federal do Rio Grande do Norte, Department of Mechanical Engineering, Campus Uni¨ ersitario-Lagoa No¨ a, ´ Natal RN 59072-970, Brazil Received 3 November 1999; accepted in revised form 30 October 2000

Abstract The growth behavior of plasma-nitrided layers on workpieces with complex geometry was systematically investigated. AISI 316 stainless steel pellets with different heights were nitrided under a mixture of N2 ᎐80% H 2 at different temperatures Ž673, 773 and 843 K. and pressures Ž100 and 500 Pa.. Significant differences in thickness and hardness of the resulting nitrided layers were observed as a function of nitriding parameters. The thickness of nitrided layers increased with sample height, excepted those nitrided at 843 K. The diameter of eroded rings, commonly observed on nitrided samples, varied with coupon height. Changes in both layer thickness and eroded ring diameter are presently addressed based on the thermal balance and charge density that take place near the edges of the samples. 䊚 2001 Elsevier Science B.V. All rights reserved. Keywords: Plasma nitriding; Glow discharge sputtering; Stainless steel; Erosion

1. Introduction Plasma nitriding is an advanced surface modification technology which has experienced substantial industrial development over the past 30 years w1x. The process consists of generating an electrical discharge in a gas mixture containing nitrogen under low pressure. Ions and active neutral species are formed and impinge onto the surface of steels and cast irons resulting in the formation of nitride layers. Typical nitriding processes require discharge potentials from 300 to 600 V and pressures between 100 and 1000 Pa, resulting in ionization rates as low as 10y6 ions per molecule w2x. The ionic species are accelerated toward the cathode of the

U

Tel.: q55-84-215-3800; fax: q55-84-215-3791. E-mail address: [email protected] ŽC. Alves..

reactor that functions as a sample holder. The interaction between ions and surface results in different effects, mainly sample heating and sputtering, corresponding to the transference of kinetic energy and momentum from the impinging species to the sample surface, respectively w3x. The geometry of a specimen affects its nitriding mechanism inasmuch as local temperatures may vary as the bombardment rate may not be uniform along a surface. In this sense, geometry controls the temperature distribution and, consequently, the kinetics of layer formation along a workpiece. It has been suggested w4x that a workpiece behaves uniformly with respect to the thermal input. The input power per unit area, Pi , is the product between discharge potential, Vd , and current density on the workpiece, Jd . Part of this power heats the workpiece whereas the remaining is dissipated by conduction, convection and, mainly, in the form of

0257-8972r01r$ - see front matter 䊚 2001 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 0 . 0 1 1 4 6 - 4

C. Al¨ es et al. r Surface and Coatings Technology 139 (2001) 1᎐5

2

radiation w4x. At steady state, this thermal balance can be expressed as the sum of two interdependent terms, such as Pi s Vd Jd s

␳c p Ž ⌬Tr⌬ t . q ␴␧ Ž Tp4 y Tw4 . Ar ¨

Ž1.

where ␳ is the density of the substrate, c p is the specific heat of the substrate, ⌬Tr⌬ t is the heating rate of the substrate, Ar¨ is the surface area to volume ratio, ␴ is the Stefan᎐Boltzmann constant Ž5.67= 10y8 Wrm2 K 4 ., ␧ is the surface emissivity coefficient, Tp is the local surface temperature and Tw is the wall chamber temperature. Eq. Ž1. shows that, under stable operating conditions, there is a relationship between sample heating and Ar¨ , which determines the temperature at each element of surface area of the workpiece. It becomes clear that, although ideally the temperature should be uniform along a specimen surface, its geometry can originate significant thermal gradients. Surfaces distinctively shaped dissipate heat at different rates w5,6x, which implies that components with different geometries are subjected to different kinetics if nitrided simultaneously. Surface sputtering also affects the kinetics of nitride formation, since the deposition and erosion rates are intimately related. The latter can be expressed as w7x: R s 62.3

Jd YMa ␳

Ž2.

Where Jd is the current density, Y is the sputtering yield, and Ma and ␳ are the atomic weight and density of the material under bombardment, respectively. During a nitriding treatment, both the current density and sputtering rate depend on the ion distribution and the electrical field in the vicinity of the surface w8x. Both parameters exhibit an anomalous behavior where two conducting surfaces meet, such as the edges of intricate workpieces w9,10x. The charge density on a surface is given by Laplace’s equation w10x, q Ž r . s kr n

Ž3.

where k is a coefficient that depends on the induced electrical potential Žremote potential. at an edge, n depends on the edge angle Ž ␲r2 and 3␲r2 for smooth and sharp edges, respectively. and r is the distance from the edge. Eq. Ž3. shows that small components, where the presence of edges is significant, such as gears, moulds, needles and drilled parts, are expected to produce irregular layers due to non-uniform sputtering rates. The objective of the present study is to experimentally assess the effect of edges on the characteristics of

nitrided layers. To that end, AISI 316 stainless steel samples having different heights were plasma nitrided under different temperatures and pressures. The results obtained from a simple cylindrical geometry give an indication as to the effect of thermal flux and electric field lines on plasma nitriding and can be used to evaluate the nitriding behavior of complex parts.

2. Experimental procedure The nitriding unit used in the present study consisted of a cylindrical stainless steel d.c. plasma reactor 400 mm= 400 mm Žheight = diameter.. A detailed set-up of this unit can be found elsewhere w11x. A series of AISI 316 stainless steel rods 8.0 mm in diameter were cut to various heights: 1, 3, 5, 8 and 10 mm. The top surfaces of the rods were ground and polished in alumina slurry to a 1-␮m finish. The specimens were then washed in an ultrasonic bath with acetone and placed on the cathode of the reactor, according to the arrangement illustrated in Fig. 1. Samples were nitrided for 3 h in an atmosphere of N2 ᎐80% H 2 under 100 or 500 Pa. Nitriding temperatures were set to 673 K, 773 K or 843 K. Three specimens with the same height were grouped in each radial direction, resulting in a total number of 15 specimens nitrided for each set of temperature᎐pressure values. After the treatment, the specimens were observed with respect to their superficial aspect and later prepared for microstructural analysis and microhardness testing. In order to evaluate the heterogeneity caused by surface erosion, five different regions of the eroded area were evaluated by microhardness tests. Microstructure and uniformity of the nitrided layers were assessed using a Carl Zeiss Neophot optical microscope. Microhardness measurements were carried using a Carl Zeiss Jena mhp tester applying a load of 100 g onto the samples.

3. Results and discussions All samples, except those nitrided at 843 K, exhibited erosion rings similar to that shown in Fig. 2. Such rings were formed at relatively low nitriding temperatures Ž673 and 773 K. and could be macroscopically observed. Ring surfaces appeared as a bright contrast under an optical microscope, corresponding to a smooth surface where little or no nitrided layer was formed. This fact was later confirmed by cross-sectional layer examination and microhardness tests. Fig. 3 shows the superficial aspect of a typical ring along with plots of ring thickness ŽTr . and distance to sample edge Ž De . as a function of specimen height. The dimensions of the rings varied according to the height of the sample and the nitriding parameters. Sample height significantly

C. Al¨ es et al. r Surface and Coatings Technology 139 (2001) 1᎐5

3

Fig. 1. Schematic illustration of sample arrangement on plasma reactor cathode.

affected the aspect of the ring especially up to 5 mm. The presence of rings also affected the hardness profile of the specimen. Fig. 4 shows a cross-section view of a nitrided layer, along with the positions in the eroded rings subjected to indentation. The corresponding microhardness values are also shown. The hardness in the middle region of the rings reached values as low as 250 Hv10 , similarly to what was observed from samples with 1.0 mm height nitrided at 673 K. This value is slightly higher than the hardness of specimens prior to ion nitriding Ž; 200 Hv10 ., providing clear evidence of the deprivation of a nitrided layer in those areas. This behavior can be rationalized based on Eq. Ž3., which yields the charge distribution as a function of the distance to sample edge. Upon nitriding, an aureole could be observed on the surface of the cathode sheath and was probably associated with high ion concentra-

Fig. 2. Stainless steel sample nitrided at 773 K under a pressure of 500 Pa showing an eroded ring.

Fig. 3. Ža. Superficial aspect of a sample 8.0-mm-high nitrided at 673 K and 500 Pa showing the presence of a typical ring; and Žb. thickness of rings and distance to the edge as a function of specimen height.

Fig. 4. Upper: cross-section view of a nitrided layer; below: microhardness as a function of relative position with respect to eroded ring.

4

C. Al¨ es et al. r Surface and Coatings Technology 139 (2001) 1᎐5

tions. The diameter of the aureole varied with specimen height and apparently determined the occurrence and dimensions of a ring. It should be further noted that rings were not observed in specimens nitrided under 500 Pa at 843 K or under 100 Pa, regardless of the nitriding temperature. The formation of a ring resulted from two opposing processes, i.e. erosion and deposition. At relatively low temperatures, the erosion rate prevailed over the deposition rate and rings were observed. Increasing the temperature improved deposition by surface diffusion, which is a thermally activated process ruled by an Arrhenius-type relationship. At relatively high pressures, there seems to be a discontinuity on the observable sheath, suggesting that it may be in fact a result of two superimposed sheaths, i.e. one formed over the cathode and another formed over the specimen. Lowering the pressure, the cathode sheath grew and, consequently masked the superimposing effect on short samples. A similar result can be obtained at higher pressures by decreasing the height of the sample. In general, erosion was significant for relatively high samples nitrided at lower temperatures. The ion nitriding mechanism involves a series of effects such as the bombardment, sputtering, internal and superficial diffusion, and chemical reactions. Whereas the first two effects act towards increasing erosion, the remainder improves deposition. Lower temperatures decrease the effect of the diffusion of neighboring species in the surface of the material Žatomic nitrogen, molecular nitrogen, and precursor nitrides. inwards to the centermost area of the ring. For short samples, nitrogen depletion can be attributed to an increase in the diffusion path of those species towards the center of the ring. Thus, thicker rings resulted from those conditions ŽFig. 3.. The thickness of the nitrided layer also depended on specimen height, mainly for samples shorter than 8 mm. Fig. 5 shows such a trend for samples nitrided at 673, 773 and 843 K either under 100 or 500 Pa. The results showed that the thickness of the nitrided layer systematically increased with sample height, except for samples 8.0 and 10.0 mm high nitrided at 843 K. These two samples exhibited similar overall behavior upon nitriding, with respect not only to layer thickness but also to eroded ring diameter, surface hardness, and phase formation. The trend observed in layer growth, which is associated with nitriding kinetics, can be explained from Eqs. Ž1. and Ž3.. The latter states that the charge density near the edges of the sample is a function of r n, where n s 1 for edges of 90⬚ or y1r3 for 270⬚. In other words, the charge density increases linearly with the distance from the point where the sample touches its holder Ž n s 1., thus increasing the local

Fig. 5. Thickness of nitrided layer as a function of nitriding temperature for samples treated at indicated pressures.

potential. However, Eq. Ž1. shows that Ar¨ decreases with sample height, originating a competing effect that determines the heating behavior of the sample. Such effects become less evident as the sample height increases, which is a consequence of the fact that both Ar¨ and Vd are essentially constant for relatively high values of h. For taller samples Ž8.0 and 10.0 mm., the effect of sample height on the thickness of the nitrided layer cannot be noticed any longer. Nitriding specimens 8 and 10 mm high at 843 K resulted in relatively thinner nitrided layers. This is probably related to the precipitation of CrN, as it could be observed from XRD results. Such trend is shown in Fig. 6 which depicts diffraction patterns obtained from samples having three different heights nitrided at the same cathode temperature. A relative increase in CrN contents could be observed from taller samples Ž10 mm., probably corresponding to higher surface temperatures compared with shorter samples, confirming similar trends reported elsewhere w12x. Improved precipitation of CrN on AISI 316 steel takes place as the temperature increases, thus inhibiting the diffusion of nitrogen. This condition may have been satisfied in the present study since temperatures on the surface of specimens were invariably higher than on the cathode, especially in taller samples nitrided at higher temperatures. Finally, cross-section views of nitrided samples revealed that the thickness of the nitrided layer was affected by the presence of an eroded ring. Fig. 7 clearly shows the thickness of a typical nitrided layer decreasing as it approached a ring. This was a direct consequence of the charge distribution imposed by Eq. Ž3. where charge density decreases with the distance to the edge.

C. Al¨ es et al. r Surface and Coatings Technology 139 (2001) 1᎐5

Fig. 6. XRD of nitrided surface samples 10.0 mm, 5.0 mm and 1.0 mm height, nitrided at 843 K.

5

1. Specimens nitrided at 673 and 773 K under 500 Pa depicted localized erosion, observed in the form of rings. Ring thickness and distance to sample edge decreased with increased height Žexcept for samples 1.0 mm tall., remaining constant for samples over 8.0 mm tall. 2. Impaired growth kinetics implied in a decrease in hardness in regions characterized by eroded rings. This effect was more significant in relatively short samples. 3. The thickness of nitrided layers produced at 673 and 773 K varied directly with sample height. For specimens 8.0 and 10.0 mm high the difference in thickness was negligible. 4. Samples 8.0 and 10.0 mm tall depicted a decrease in layer thickness upon nitriding at 843 K, departing from the anticipated linear behavior. This can be attributed to the precipitation of CrN, which inhibited nitrogen diffusion. 5. From a general viewpoint, samples 8.0 and 10.0 mm high behaved in a similar way with respect to growth kinetics and features of the eroded rings. This suggests a negligible effect of the sample holder for such sample heights.

Acknowledgements This work was sponsored by PADCTrFINEP and CAPES-Brazil.

Fig. 7. Optical micrograph of 5.0 mm high sample nitrided at 500 Pa and 773 K showing variation of layer thickness with distance to ring.

4. Conclusions Nitrided layers formed on samples characterized by the presence of edges depicted considerably wide ranges of hardness and thickness. Edges also contributed to the formation of eroded rings. Such effects were systematically investigated measuring hardness profiles and thickness of nitrided layers on cylindrical pieces 8.0 mm in diameter cut to different heights. The distribution of charge density near the edges of those samples significantly affected the features of the nitrided layers. For shorter samples, charge distribution in the intersection of sample and sample holder affects the charge density at the edge of the top surface of the sample, leading to the formation of nitrided layers whose growth kinetics depends on sample height. The following conclusions could be drawn and may be used to assess different geometries.

References w1x T. Bell, P.A. Dearnley, Surf. Eng. 10 Ž2. Ž1994. 123᎐128. w2x J.W. Coburn, IEEE Trans. Plasma Sci. 19 Ž6. Ž1991. 1048᎐1062. w3x A. Grill, Cold Plasma in Materials Fabrication ᎏ From Fundamentals to Applications, New York, IEEE Press, 1993, p. 256. w4x C. Ruset, Heat Treat. Metals 3 Ž1993. 81᎐83. w5x R. Grun, R.G. Gunther, Mater. Sci. Eng. A1 140 Ž1991. 435᎐441. w6x C. Ruset, Heat Treat. Metals 3 Ž1994. 90᎐92. w7x N.A.G. Ahmed, Ion Plating Technology ᎏ Developments and Applications, New York, Wiley, 1987, p. 171. w8x H. Michel, T. Czerwiec, M. Gantois, D. Ablitzer, A. Ricard, Surf. Coat. Technol. 72 Ž1995. 103᎐111. w9x A. Ricard, J.E. Oseguera-Pena, L. Falk, H. Michel, M. Gantois, International Congress on Heat Treatments of Materials, 6, 1998. Proceedings. Metals, ASM International Treatments of Materials, 6, Chicago, 1988. Proceedings, ASM International, Metals Park, 1998, pp. 55᎐66. w10x J.-P. Boeuf, J. Appl. Phys. 63 Ž5. Ž1988. 1342᎐1349. w11x R.S. de Souza, C. Alves, Jr., J. Mater. Eng. Perform. 6 Ž3. Ž1997. 300᎐302. w12x C.F.M. Borges, S. Hennecke, E. Pfender, Surf. Coat. Technol. 123 Ž2000. 112᎐121.