Available online at www.sciencedirect.com
Applied Surface Science 254 (2008) 2276–2280 www.elsevier.com/locate/apsusc
Characterization of the nitrided layers of XC38 carbon steel obtained by R.F. plasma nitriding M. Keddam * De´partement de Sciences des Mate´riaux, Faculte´ de Ge´nie Me´canique et Ge´nie des Proce´de´s, USTHB, B.P N832, 16111, El-Alia, Bab-Ezzouar, Alger, Algeria Received 19 August 2007; received in revised form 5 September 2007; accepted 5 September 2007 Available online 11 September 2007
Abstract XC38 carbon steel was nitrided in a low-pressure R.F. plasma using a mixture of 60% N2-40% H2 without cathodic bias on the samples. The experiments were carried out at different temperatures for various time durations. The generated nitride layers were characterized by SEM observations, XRD and GDOS analyses. These analyses indicate that the compound layer was composed of the g0 -Fe4N phase with a surface content of N close to 6 wt%. An approach was used to study the growth kinetics of the compound layer at 500 8C. Furthermore, it was concluded that its kinetics follows a power law, which deviates from the classical parabolic growth. # 2007 Elsevier B.V. All rights reserved. Keywords: R.F. plasma; Iron nitride; Kinetics; Nitriding; Compound layer
1. Introduction Plasma nitriding is a surface modification technique used for improving wear, corrosion and fatigue resistance of steels and metallic materials. Depending on the plasma nitriding parameters and the base material composition, the so-called white or compound layer consisting of e and/or g0 iron nitrides can be formed at the surface . In the diffusion zone, the ferrite is saturated due to the nitrogen diffusion in Armco or pure iron. A precipitation of metallic nitrides can occur in case of low or high alloy steels within this zone. The build up of the nitrided layer starts from nitrogen absorption and the layer develops by nitrogen diffusion into the bulk material. The thickness and the composition of a surface layer that develops during the nitriding process are affected by the type of chemical reactions occurring at the sample surface as well as the diffusivity of nitrogen in the treated material. Studies have shown that the microstructures of the surface layer and the nitrided layer can be influenced by changing the
* Tel.: +213 21 24 79 19; fax: +213 21 24 79 19. E-mail address: [email protected]
0169-4332/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2007.09.012
process parameters such as temperature, time and gas mixture ratio [2–5]. Details of the plasma nitriding process and its parameters have been well described by Kovacs and Russell . In an industrial practice, the plasma nitriding process is preferred to both processes: the bath and gas nitriding. The bath nitriding poses a major problem related to toxicity of the cyanide–cyanate salt. The gas nitriding requires several hours to get the desired case depth. In contrast, the plasma nitriding has many advantages (short treatment time, minimal distortion, clean samples and low energy use) in comparison to the gas nitriding process [7–9]. In order to reduce nitriding time and enhance nitriding efficiency, many technologies have been developed for plasma nitriding such as plasma immersion ion implantation, pulsed glow discharge plasma nitriding and radio frequency (R.F.) plasma nitriding, etc. For indication, the low-pressure R.F. plasma may have the advantage that the thickness of the treated layer falls more slowly with lowering temperatures than with other nitriding media . The advantages of the low R.F. plasma nitriding over the gas nitriding are the possibility of using lower temperatures with a minimal distortion of the treated workpiece. This work presents experimental results obtained and discusses the response of XC38 carbon steel to R.F. plasma
M. Keddam / Applied Surface Science 254 (2008) 2276–2280
nitriding in terms of nature of the nitrided layers and its nitriding kinetics. 2. Experimental details The carbon steel studied here (XC38) possesses the following composition (wt%): 0.38% C; 0.66% Mn; 0.25% Cr; 0.02% Mo; 0.27% Si; 0.02% Ni; 0.02% P; 0.016% S; Fe balance. The samples to be treated were machined in the form of 10 mm diameter and 2 mm thick discs. The surfaces of the substrates were mechanically polished sequentially by 600, 800, 1200 and 2400 grit wet SiC emery paper, followed by fine polishing with alumina slurry. The samples were cleaned in alcohol and acetone before loading into the sample holder. Afterwards, they were nitrided with a mixture of 60% N240% H2 at a working pressure of 750 Pa in a URANOS plasma type system, which has been described in ref. . The cold plasma was generated in a quartz tube by a frequency of 13.56 MHz with a working power of 700 W. In this process, the ion energy is mainly controlled by the wall furnace temperature and the sample surface is not subjected to the sputtering effect. The cross-sections and the surface topography of the samples were characterized by scanning electron microscopy (SEM-JOEL 5600LV). The penetration depth of nitrogen into the steel was estimated by glow discharge optical spectrometry (GDOS) using JOBIN YVON PROFILER analyzer. X-ray diffraction (XRD) analysis (X Pert Philips X-ray diffractometer using Cu Ka with PC ADP, Diffraction software) was applied for phase identification. The angular range 30–908 2u was scanned with a step size of 0.018 2u, covering the major reflections of the expected phases in the nitrided layer. 3. Experimental results The R.F. plasma nitriding of XC38 carbon steel was carried out at different temperatures for various time durations. The compound layer thickness depending on the process parameters is listed in Table 1. Each layer thickness presented is the average value obtained from 10 independent measurements. The treatment of the compound layer is less than 1 mm when the nitriding treatment was carried out at 350 8C for 4 h, while its thickness attains a value close to 5.5 mm at 570 8C for the same time duration. This indicates that its growth is affected by Table 1 The change of the compound layer thickness with R.F. plasma parameters Process parameters Temperature (8C)
Compound layer thickness (mm)
3 4 8 16
1.1–3.3 3.5–4.5 3.8–4.9 4–6.5
Fig. 1. Variation of the g0 layer thickness vs. time.
an increase of the temperature. In order to study the layer growth kinetics of the g0 phase grown on the XC38 steel’s substrate, it is assumed that the g0 layer grows according to the power law given by Eq. (1): x ¼ a tn
The parameters a and n are the material constants depending on the nitriding conditions. The averaged experimental data obtained at 500 8C for various time durations are plotted in Fig. 1. By fitting the data with Eq. (1), the material constants were estimated as: a = 2.47 and n = 0.288 with a correlation factor = 0.957. It is concluded that the growth kinetics of the g0 phase does not follow a classical parabolic law, where n = 0.5. It is due to the fact that the nitrogen diffusion generates a nonuniform thickness of the compound layer, so it is difficult to make its measurement. It may also be the result of the preferential nucleation and growth of the g0 phase along grain boundaries. Fig. 2(a) shows X-ray diffraction patterns on the samples of XC38 steel nitrided at 500 8C for different time durations and on the unnitrided substrate. a-Fe and g0 -Fe4N were detected over the whole sample surface. Fig. 2(a) shows the existence of diffraction peaks of g0 -Fe4N from (1 1 1), (2 0 0) and (3 1 1) planes, whereas the a-Fe phase diffracts from (1 1 0), (2 0 0) and (2 1 1) planes. The same phases, as shown in Fig. 2(b), were also detected for the XC38 steel nitrided for 4 h at different temperatures. In both cases, the compound layer is only consisted of g0 -Fe4N as an iron nitride. Fig. 3 presents the cross-sectional SEM photomicrographs, slightly etched with 2% nital, of XC38 steel nitrided at 500 8C for 3 h and 570 8C for 4 h, respectively. In Fig. 3(a), it is clearly seen that the compound layer has a non-uniform thickness for a treatment of 500 8C for 3 h, whereas its thickness is regular and continuous (Fig. 3b) for a treatment of 570 8C during 4 h. In addition, the compound layer formed on the surface (Fig. 3a) does not contain micropores. Fig. 4(a) shows the surface of original matrix to be nitrided. In Fig. 4(b) and (c), an examination of the microstructures
M. Keddam / Applied Surface Science 254 (2008) 2276–2280
Fig. 3. SEM photomicrographs; (a) nitrided 3 h at 500 8C, (b) nitrided 4 h at 570 8C.
Fig. 2. (a) XRD patterns of XC38 carbon steel nitrided at 500 8C for different time durations. (b) XRD patterns of XC38 carbon steel nitrided for 4 h at different temperatures.
reveals that the surface is covered with a light network, which outlines the grain boundaries. This morphology may be the consequence of the preferential nucleation and growth of the iron nitrides along grains boundaries. Fig. 5 shows the surface topography of XC38 carbon steel treated at 570 8C for 4 h. Many small asperities are visible on the surface. The left part of the surface was covered by a thin foil made of silicon to avoid the R.F. plasma nitriding, while the right part shows its effect is covered by small asperities. The observed surface roughness results in the competition between the surface kinetics and bulk diffusion. The nitrogen profiles versus depth were reported in Fig. 6. They were obtained by GDOS. The nitrogen concentration indicates a highest value close to 6 wt% near the surface and decreases gradually towards the layer/core interface. This result confirms well the XRD analysis.
The R.F. plasma nitriding of XC38 steel was done in a gas mixture consisting of 60% N2-40% H2. This gas composition was optimized in previous works carried out by L. Marot in the framework of his PhD thesis . This gas mixture, used to form the plasma, generated a compound layer composed of g0 phase as an iron nitride on the substrate’s surface. The XRD analysis showed the presence of this phase, and this result is consistent with GDOS analysis (Fig. 6). It is also found that the nitrogen surface content reaches a value close to 6 wt%, which corresponds to the composition range of the g0 phase. In a recent paper by Hirsch et al. , an experimental work dealing with plasma nitriding of AISI 1045 carbon steel was performed in a temperature range (450–560 8C) with variable gas composition. It was reported that a single phase g0 nitride has been detected for a gas mixture of 5% N2-95% H2, when plasma nitrided the AISI 1045 carbon steel containing 0.47 wt% C at a temperature of 520 8C (with a compound layer thickness of 1–3 mm) for 3 h. In our experiment, a treatment of 500 8C for 3 h by R.F. plasma of the XC38 carbon steel in a gas mixture of 60% N2-40% H2, gave rise also to a single compound layer (g0 iron nitride) with a thickness of 1.1–3.3 mm, which is in the same order as in a previous study . Furthermore, another experimental result obtained by Ashrafizadeh  proved that the compound layer was single g0 phase (3 mm for 5 h without specifying the treatment temperature and the used gas composition) during the plasma nitriding of Ck45 carbon steel at 0.46 wt% C. However, a gas mixture of 25% N2-75% H2 generates both iron nitrides e and g0 at 540 and 560 8C, respectively. The nitriding plasma experiment, performed on the AISI 1010 carbon steel by Simon et al. , indicates also the formation of two iron nitrides e and g0 by use of a mixture of 60% N2-40% H2 for 0.25 and 0.5 h at a temperature of 447 8C. According to Zhan et al. , it is mentioned that a ratio of N2/
M. Keddam / Applied Surface Science 254 (2008) 2276–2280
Fig. 5. The surface topography of XC38 carbon steel treated at 570 8C during 4 h. The left part of the surface was covered by a thin foil made of silicon to avoid the R.F. plasma nitriding while the right part shows its effect (appearance of small asperities).
deviates from a parabolic diffusion law [17–18] due to the sputtering effect (in R.F. plasma nitriding, this effect is not observed due to the use of cold plasma). Depending on the used techniques, different kinetic behaviors of the nitrided case (compound layer and diffusion zone) were reported in the literature. So, a kinetic analysis regarding the plasma nitriding of the AISI 304 stainless steel has shown that the growth of nitrided layer, consisting of expanded austenite: gN and a diffusion zone, follows a parabolic law at 420 8C . In the case of pulse plasma nitriding, a parabolic growth of both compound layer and diffusion zone was also observed experimentally on the Armco iron samples [20,21] and AISI C1043 steel . Another experimental work, conducted by Salas et al. , using the plasma nitriding of Armco iron samples, has explained that an appearance of the non-planar interface (compound layer/diffusion zone) was attributed to a growth law that surpasses the parabolic regime. Furthermore, Metin and Inal  have observed a similar behavior in their nitriding experiments. On the contrary, our study shows that the compound layer exhibits a growth law lower than that of the parabolic growth, i.e. a sub-parabolic law (see Eq. (1)). A deviation from the parabolic growth was also observed while plasma nitriding the EN19 steel at 450 and 530 8C due to sputtering effect .
Fig. 4. The surface morphology of XC38 steel before and after R.F. plasma treatment. (a) The original surface of the sample: (b) nitrided 3 h at 500 8C, (c) nitrided 4 h at 570 8C.
H2 = 9/1 gives rise to the e iron nitride, whereas a ratio of N2/ H2 = 3/7 forms on the surface g0 iron nitride. It is then proved experimentally that the gas composition of the plasma nitriding represents a key parameter that affects directly the microstructural configurations of the compound layer. The R.F. plasma nitriding is recommended as the feasible process to reduce the cost of the treatment and to get a thin compound layer suitable to some industrial applications. In the case of plasma nitriding, the growth rate of the compound layer
Fig. 6. Depth profiles of nitrogen element by GDOS analysis.
M. Keddam / Applied Surface Science 254 (2008) 2276–2280
In this current work, an approach about the nitrogen diffusion phenomenon was proposed. The growth kinetics of the g0 phase was assumed to follow a power law (Eq. (1)). A series of experiments were carried out to study its kinetic behavior at the nitriding temperature of 500 8C for various time durations. Due to the non-flatness of the compound layer/ diffusion layer interface for certain nitriding conditions (at exception for a treatment at 570 8C during 4 h, Fig. 3b), it was difficult to measure the thickness of the g0 phase, although 10 measurements were done. Consequently, the material parameters of the power law (estimated at 500 8C) were identified. This result shows a deviation from a classical parabolic growth. This experimental evidence can be ascribed to the preferential nucleation and growth of the g0 phase along grain boundaries. The compound layer forms on the surface of the sample, where its growth is controlled by nitrogen diffusion through this layer. The nucleation starts in some energetically preferred places such as grain boundaries and structural defects [26,27]. At some points along the interface, the compound layer extends into the grains boundaries  (see Fig. 3a). The formation of this iron nitride is then affected by internal stresses produced by the distortion of the lattice due to dissolved nitrogen, where the nitrogen diffusion is promoted by the presence of H2 in the gas mixture . Swelling effect was evidenced by an examination of the surface topography of the treated sample (Fig. 5). The formation of the iron nitride on the surface causes a modified topography.
   
Through this work, an experimental study about R.F. plasma of XC38 carbon steel was performed and some concluding points are drawn as follows: (1) The R.F. plasma nitriding of XC38 carbon steel with a mixture of 60% N2-40% H2 leads to the development of a single nitride layer. The compound layer was found to be composed of g0 iron nitride, result confirmed by XRD analysis. (2) The layer thickness of g0 phase, grown over the material surface, has no perfect flat interface according to certain nitriding conditions (i.e. non-uniform thickness along the nitrided layer). (3) It was possible to get a thin layer of the g0 phase (less than 1 mm) at 350 8C for a treatment time of 4 h via this nitriding process. (4) The R.F. plasma nitriding leads to the dimensional change on the top surface of the sample. Swelling effect is
confirmed by an examination of the surface topography of the treated sample. (5) The layer growth kinetics of g0 phase can be described by a power law. The material parameters of the power law (estimated at 500 8C) show that the nitriding kinetics of g0 phase is lowered compared to the classical parabolic law. Acknowledgments The author would like to thank all members of the Laboratory of Physical Metallurgy (University of Poitiers, France) for their help for carrying out the R.F. plasma experiments and performing some characterizations. Prof. J. P. Rivie´re is especially thanked. References      
                
T. Bell, Y. Sun, Surf. Eng. 6 (2) (1990) 133. A. Alsaran, A. Celik, C. Celik, Surf. Coat. Technol. 160 (2002) 219–226. L. Marot, L. Pichon, M. Drouet, A. Straboni, Mater. Lett. 44 (2000) 35–38. A. Alsaran, Mater. Charact. 49 (2003) 171–176. M. Sahara, T. Sato, S. Ito, K. Akashi, Mater. Chem. Phys. 54 (1998) 123– 126. W. Kovacs, W. Russell, in: T. Spalvins (Ed.), Ion Nitriding, ASM, Metals Park, OH, 1987, pp. 9–17. F. Ashrafizadeh, Surf. Coat. Technol. 174/175 (2003) 1196. S. Karaoglu, Mater. Charact. 49 (2003) 349. P. De la Cruz, M. Oden, T. Ericsson, Mater. Sci. Eng. A 242 (1998) 181. M.P. Fewell, J.M. Priest, M.J. Baldwin, G.A. Collins, K.T. Short, Surf. Coat. Technol. 131 (2000) 284–290. J. Perrie`re, J. Siejka, N. Re´mili, A. Laurent, A. Straboni, B. Vuillermoz, J. Appl. Phys. 59 (1986) 2752–2759. L. Marot, The`se de Doctorat, Universite´ de Poitiers, 2000 (in French). T. Hirsch, T.G.R. Clarke, A. da Silva Rocha, Surf. Coat. Technol. 210 (2007) 6380–6386. F. Ashrafizadeh, Surf. Coat. Technol. 173/174 (2003) 1196–1200. G. Simon, M.A.Z. Vasconcellos, C.A. dos Santos, Surf. Coat. Technol. 102 (1998) 90–96. R.J. Zhan, C. Wang, X. Wen, X. Zhu, Surf. Coat. Technol. 105 (1998) 72. A. Marciniak, Surf. Eng. 1 (1985) 283–288. V.I. Dimitrov, J. D’Haen, G. Knuyt, C. Quaeyhaegens, L.M. Stals, Comp. Mater. Sci. 15 (1999) 22–34. L. Wang, S. Ji, J. Sun, Surf. Coat. Technol. 200 (2006) 5067–5070. M. Yan, J. Yan, T. Bell, Modell. Simul. Mater. Sci. Eng. 8 (2000) 491–496. M. Yan, Q. Meng, J. Yan, J. Mater. Sci. Technol. 19 (2003) 164–166. Z. Durisic, A. Kunosic, J. Trifunovic, Surf. Eng. 22 (2) (2006) 147–152. O. Salas, U. Figueroa, J.L. Bernal, J. Oseguera, Surf. Coat. Technol. 163/ 162 (2003) 339–346. E. Metin, O.T. Inal, J. Mater. Sci. 22 (1981) 2783. Y. Sun, T. Bell, Surf. Eng. 11 (2) (1995) 146–148. H. Du, N. Lange, J. Agren, Surf. Eng. 11 (4) (1995) 301. W.D. Jentzsch, S. Bohmer, Kristall und Technik 14 (5) (1979) 617. B. Edenhofer, T.J. Bewley, Heat Treatment ’76, The Metals Society, London, 1996, 7-13.