Phase depth distribution characteristics of the plasma nitrided layer on AISI 304 stainless steel

Phase depth distribution characteristics of the plasma nitrided layer on AISI 304 stainless steel

Surface and Coatings Technology 162 (2003) 242–247 Phase depth distribution characteristics of the plasma nitrided layer on AISI 304 stainless steel ...

1MB Sizes 0 Downloads 30 Views

Surface and Coatings Technology 162 (2003) 242–247

Phase depth distribution characteristics of the plasma nitrided layer on AISI 304 stainless steel Xiaolei Xu*, Zhiwei Yu, Liang Wang, Jianbing Qiang, Zukun Hei Institute of Metal and Technology, Dalian Maritime University, Dalian 116024, PR China Received 27 April 2002; accepted in revised form 23 August 2002

Abstract The phase depth distribution of a plasma nitrided layer on AISI 304 stainless steel at 400–420 8C was investigated by glancingangle X-ray diffraction through step-wise electrio-polishing removal. The nitrogen depth profile was measured by energydispersive X-ray. The results confirm the dominant phase in the nitrogen-rich layer as an expanded austenite. The amount of the lattice expansion relative to the substrate austenite decreases from the layer surface inwards, due to the decrease in nitrogen content. The amount of lattice expansion from planes of (2 0 0) is always larger than that from (1 1 1) at a specific depth. However, the amount of distortion does not decrease monotonically with the decrease in nitrogen content with depth. It is related to the different arrangement of the nitrogen atoms in the fcc lattice at various nitrogen contents. Observations in situ on the surface morphology were performed by scanning electron microscopy. When the strain resulting from supersaturating nitrogen is sufficient to initiate the slip system activity, the slip bands are observed on the surface of the nitrided layer. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Plasma-nitriding; Austenitic stainless steel; Expanded austenite

1. Introduction When austenitic stainless steel is nitrided by plasmanitriding, or plasma-implantation methods performed at an appropriate temperature (typically -450 8C), an expanded austenite layer can generally be expected to form w1–4x. The X-ray diffraction (XRD) peaks of such layer tend to appear at lower diffraction angles, relative to the parent austenitic substrate. Importantly, anomalously large expansion along the (2 0 0) planes always appears to occur. However, the nature of the anomalous expansion is as yet open to question, with no widespread agreement. Some authors considered that the expanded austenite consists of a mixed phase with different lattice parameters w5x. However, others suggests a slight tetragonal or triclinic distortion w6,7x. However, the apparent absence of some peaks predicted to appear if such distortion occurs makes these explanations not entirely convincing. A new structural model of expanded austen*Corresponding author. Tel.: q86-411-4729613; fax: q86-4114671395. E-mail address: [email protected] (X. Xu).

ite based on a defect-rich fcc structure was recently proposed to explain the anomalous expansion w8x. But the alternative amount of the anomalous expansion or distortion with nitrogen content in the present work suggests that the model may not be suitable for layers with higher nitrogen content. This investigation tries to explore the nature of the expanded austenite through analyzing the depth-related phase structure distribution in nitrogen-expanded austenite. 2. Experiment methods The composition of the annealed AISI 304 austenitic stainless steel used for this work was (in wt.%): C, 0.07; Si, 0.78; Mn, 0.90; S, 0.13; Cr, 19.0; Ni, 8.26; and Fe, balance. The 20=18=5 mm3 samples were prepared by electrochemical polishing in a solution of 5% perchloric acid (HClO4) and 95% acetic acid (CH3COOH). The samples were plasma nitrided at 400– 420 8C using the low-pressure plasma arc-source ion nitriding technology developed by the authors w3,4x. With ammonia as the working gas the working pressure was approximately 0.4 Pa. The arc current was 40 A

0257-8972/03/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 2 . 0 0 6 7 0 - 9

X. Xu et al. / Surface and Coatings Technology 162 (2003) 242–247

Fig. 1. Cross-sectional morphology of stainless steel plasma-nitrided for 6 h at 400–420 8C.

with a voltage of 50 V. A negative bias of 0.8–1 kV was applied to the substrates. The nitrided layer was removed by step-wise electrio-polishing in the solution mentioned above. The removed thickness was controlled within approximately 1 mm at each step and measured with a spiral micrometer (accuracy 0.5 mm). The phase structure at different layer depths was determined by glancing-angle X-ray diffraction on a Rigaku DymaxIIIA diffractometer. The radiation used was Cu Ka (ls 0.1542 nm) at an incident angle of 108. The X-ray penetration depth under these diffraction conditions was estimated to be 1–1.5 mm. Scanning electron microscopy (SEM) observations were made using a JEOL 35CF SEM. Nitrogen depth profile was obtained using energy-dispersive X-ray (EDX). 3. Results and discussions Fig. 1 shows the cross-sectional SEM morphology of a stainless steel sample plasma nitrided for 6 h at 400– 420 8C. It demonstrates that the depth of the nitrided layer is approximately 13 mm thick. From Fig. 1, the nitrided layer shows excellent corrosion-resistance when compared to the untreated austenite substrate, which exhibits pitting corrosion morphology (etching solution: CuSO4Ø5H2O (20 g)qHCl (100 ml)qH2SO2 (5 ml)q H2O (80 ml).

˚ This 43.218, corresponding to the spacing ds2.09 A. peak suggests that there is an ´-(Fe,Cr)2–3N phase w9x on the surface of the layer in addition to gN. The gN peaks are gradually shifted to higher angles with increasing depth into the nitrided layer suggesting as expected, that the lattice parameter of the supersaturated solid solution decreases with increased removal, due to the decrease in nitrogen content. The gN (2 0 0) peak is shifted more than the gN (1 1 1), relative to the substrate peaks, at every depth; in other words, the amount of expansion for the (2 0 0) plane is always larger than that for the (1 1 1) plane. This indicates that there seems to be a distortion relative to the fcc lattice. The lattice constants from d(1 1 1) and from d(2 0 0), measured by XRD (against depth) are shown in Fig. 3 according to the cubic lattice constant ahklsdhklyh2qk2ql2. Fig. 3 reveals clearly the expansion of the austenitic structure as a function of depth. There are presently two possible explanations for the anomalous expansion. One conclusion by other authors is that a tetragonal w6x or triclinic w7x distortion of the fcc lattice occurs, but the extra peaks predicted have not been observed. Another conclusion drawn in the literature, is that the expanded austenite is a mixed phase w5x. The structure is fcc, but grains with planes of {1 0 0} formed parallel to the surface of the sample have a larger lattice parameter than other grains. In Ref. w5x, the authors’ SEM observations show an inhomogeneous nitrided layer to verify the hypothesis of different diffusion velocity of nitrogen atoms along the different crystallographic orientations leading to different nitrogen contents in various grains. However, the cross-sectional SEM observations made in our present work show a homogeneous nitrided layer (Fig. 1). Additionally, according to Seemann–Bohlin diffraction geometry, the angle made by planes of (1 0 0) in these grains to the surface of the sample is 11.98– 12.48 in present study, which is from being parallel to the surface. The recent work of Blawert et al. w8x proposed a new structural model for expanded austenite based on a defect-rich fcc lattice. The anomalously high shift of the (2 0 0) peak in XRD was explained by assuming a high stacking fault density in gN and D(2u)8, the shift of the peaks from their fcc position uhkl due to a fault probability a, is expressed by the following equation:

3.1. XRD results The XRD patterns at various removed depths are shown in Fig. 2. The peaks of the phase induced by nitriding appear at lower Bragg angles than the peaks of the substrate austenite at every depth. Peaks from the austenite peaks are labeled as g and the phase produced by nitriding as gN. It can be seen that the nitrided layer is composed of a supersaturated nitrogen solid solution known as ‘expanded austenite’ (gN), from previous literature. It should be noted that there is a peak at 2us

243

D(2u)8s

90y3atanuhkl p2

chkl

(1)

where, a is the stacking fault density, uhkl is the peak positions of perfect fcc lattice, c111sq1y4, c200sy1y 2. In order to show a picture of the interrelationship between degree of distortion and nitrogen content, the ratio d(1 1 1)yd(2 0 0), which represents the amount of the distortion, is used as the ordinate for a plot against

244

X. Xu et al. / Surface and Coatings Technology 162 (2003) 242–247

Fig. 2. XRD patterns of the nitrided sample at different depths.

Fig. 3. Lattice constant of gN from (1 1 1) and from (2 0 0) at different depth.

Fig. 4. Relationship of distortion with depth.

X. Xu et al. / Surface and Coatings Technology 162 (2003) 242–247

245

lattice expansion. Ozturk and Williamson w5x assumed that there are 2 GP residual stress value in gN layer (0– 1 mm) and the nitrogen content predictions are found to correlate well with the EPMA analysis of the nitrogen content, which is lower than prediction from the lattice expansion without inducing average compress stress. 3.2. Surface morphology of the nitrided layer

Fig. 5. Relationship of the distortion with d(1 1 1).

the depth, or d(1 1 1), which represents the nitrogen content at different depths (Figs. 4 and 5). It is well known that d(1 1 1)yd(2 0 0) for an undistorted fcc structure is 1.15. From Figs. 4 and 5, the amount of distortion does not appear to decrease monotonically with depth, (or, conversely, increases monotonically with nitrogen content). With lower nitrogen contents the distortion increases (i.e. the anomalous expansion of (2 0 0) increases) with increasing nitrogen content. According to Eq. (1) the increase in D(2u) for (2 0 0) resulting from stacking faults is always greater than that for (1 1 1) and increases with the density of the stacking faults (a). The stacking fault density should increase with nitrogen content, due to the higher stress resulting from interstitial supersaturation by nitrogen. So the anomalous expansion should increase monotonically with nitrogen content according to Eq. (1). In fact, the ˚ ˚ with microstructure of gN (2.08 A-d(1 1 1)-2.19 A) a higher density of stacking faults was observed w3x. ˚ However, at higher nitrogen contents (2.19 A˚ the distortion decreases with increasd(1 1 1)-2.28 A) ing nitrogen content, such that the ratio d(1 1 1)y d(2 0 0) tends to approach to the theoretical value of 1.15 (Figs. 4 and 5), which cannot be explained by the equation above. We suggest that the nitrogen atoms preferably occupy at the octahedral interstitial sites in the middle of the N2 0 0M cubic edges in the fcc lattice. With increasing nitrogen content the octahedral interstices at N2 0 0M would become almost full, at which point the octahedral interstices at N0 2 0M or N0 0 2M would then begin to be occupied; thus the overall lattice distortion would tend to decrease. Of course, other effects in the layer will also influence on the peak position, such as residual stress. The nitrogen depth profiles by EDX are shown in Fig. 6. From Fig. 6, it can be seen that the nitrogen content at different depth is lower than that predicted from the

SEM observations in situ on the surface layer are shown in Fig. 7, which shows that the change in surface morphology induced by nitriding. The original substrate austenite exhibits typical annealed microstructure, such as annealed twin (at arrows in Fig. 7a, c, e). Compared with the surface morphology of the original substrate austenite, the nitriding process did not modify the grain size, but the high slip band density can be observed on the nitrided surface. In Fig. 7d, the included angle between the traces of the two directions of the slip bands is approximately 708, which should be the intersection lines of the slip planes {1 1 1} with the surface. The strain resulting from nitrogen supersaturation is responsible for the formation of the slip bands in the nitrided layer. Serial observations of surface morphology with different accumulation times in nitriding process indicate that the slip bands do not appear until after some incubation period (4 h, in present work). Also the number of the slip bands does not appear obviously to increase with further nitriding time. It is therefore apparent that the strain would be great enough to activate the {1 1 1} slip system only when the nitrogen content in the layer reaches to some critical value. The critical value should be determined by observations in situ on layer surface morphology nitrided for different time or different removal depths, combined with measuring nitrogen depth profile.

Fig. 6. Nitrogen content depth profile by EDX.

246

X. Xu et al. / Surface and Coatings Technology 162 (2003) 242–247

Fig. 7. SEM observation in situ on the change in surface morphology induced by nitriding (a), (c), (e) untreated; (b), (d), (e) nitrided for 4 h.

4. Conclusions (1) The plasma nitrided layer on austenitic stainless steel is composed mainly of expanded austenite (gN). The amount of the lattice expansion decreases with depth. There is an anomalously large expansion for (2 0 0) planes relative to that of (1 1 1), but the lattice

distortion which this creates does not increase monotonically with increasing nitrogen content. With lower nitrogen contents, the amount of distortion increases with increasing nitrogen, but subsequently decrease with increasing high nitrogen content. This is attributed to the arrangement of the nitrogen atoms in the octahedral interstices. The preferable occupation at the octahedral

X. Xu et al. / Surface and Coatings Technology 162 (2003) 242–247

interstitial sites in the middle of the N2 0 0M cubic edges in the fcc lattice is assumed. (2) A high slip band density can be observed on the nitrided surface when the strain resulting from nitrogen supersaturation is great enough to activate primary {1 1 1} the slip system. References w1x R. Wei, J.J. Vajo, J.N. Matossian, P.J. Wilbur, J.A. Davis, D.L. Williamson, Surf. Coat. Technol. 83 (1996) 235–242. w2x T. Czerwiec, N. Renevier, H. Michel, Surf. Coat. Technol. 131 (2000) 267–277.

247

w3x X.L. Xu, L. Wang, Z.W. Yu, Z.K. Hei, Surf. Coat. Technol. 132 (2–3) (2000) 270–274. w4x L. Wang, B. Xu, Z. Yu, Y. Shi, Surf. Coat. Technol. 124 (2000) 93–96. w5x O. Ozturk, D.L. Williamson, J. Appl. Phys. 77 (8) (1995) 3839–3850. w6x K. Marchev, R. Hidalgo, M. Landis, R. Vallerio, C.V. Cooper, B.C. Giessen, Surf. Coat. Technol. 112 (1999) 67–70. w7x M.P. Fewell, D.R.G. Mitchell, J.M. Priest, K.T. Short, G.A. Collins, Surf. Coat. Technol. 131 (2000) 300–306. w8x C. Blawert, H. Kalvelage, B.L. Mordike, et al., Surf. Coat. Technol. 136 (2001) 181–187. w9x M.K. Lei, Z.L. Zhang, J. Mater. Sci. Lett. 16 (1997) 1567–1569.