An evaluation of the tensile properties of a supersaturated carbon layer via in situ synchrotron diffraction

An evaluation of the tensile properties of a supersaturated carbon layer via in situ synchrotron diffraction

Available online at www.sciencedirect.com Scripta Materialia 63 (2010) 85–88 www.elsevier.com/locate/scriptamat An evaluation of the tensile propert...

420KB Sizes 6 Downloads 33 Views

Available online at www.sciencedirect.com

Scripta Materialia 63 (2010) 85–88 www.elsevier.com/locate/scriptamat

An evaluation of the tensile properties of a supersaturated carbon layer via in situ synchrotron diffraction N.G. Jones,a C.M. Ward-Close,a,b P.M. Browna,c and D. Dyea,* a

Department of Materials, Imperial College London, Prince Consort Road, London SW7 2BP, UK b QinetiQ Ltd., Farnborough GU14 0LX, UK c Defence Science and Technology Laboratory, Porton Down, Salisbury SP4 0JQ, UK Received 22 February 2010; revised 3 March 2010; accepted 4 March 2010 Available online 7 March 2010

Tensile properties of a supersaturated carbon layer (7.5 at.%) in AISI 304 grade stainless steel have been evaluated using in situ synchrotron diffraction. A small incident beam enabled diffraction from within the enriched layer and the unaffected material during tensile deformation. Reconstructed stress–strain behaviour indicated that layer strengths >2 GPa were attained. The elevated carbon level was also found to have significantly softened the diffraction elastic constants of the enriched layer when compared to the parent material. Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Steels; Synchrotron radiation; Tension test; Kolsterising

In many engineering applications it is advantageous for a component to have a hardened surface, to improve the wear resistance and reduce the chance of surface initiated cracking. Interstitial solid solution hardening, e.g. by carbon or nitrogen, is one of the most effective alloy hardening mechanisms. However, the extent to which this mechanism can be exploited is often restricted by the solubility limit of a given element in the metal matrix. These observations are particularly pertinent to stainless steels, which offer good ductility and corrosion resistance, but have a relatively low hardness when compared to other ferrous alloys. Alternative hardening mechanisms, such as quenching and tempering or high temperature carburization, are not possible with stainless steels due to the increased stability of the austenitic phase and the formation of chromium carbides at elevated temperatures. Recently, lower temperature (400 °C) heat treatment schedules utilizing a carbon monoxide atmosphere have been developed for austenitic steels, which produce a 20–30 lm layer with elevated carbon content, without the undesirable formation of carbides or loss of corrosion resistance [1–3]. These layers have been reported as having carbon concentrations up to 25 at.% and a hardness of >1000 Hv, * Corresponding author. Tel.: +44 2075946811; e-mail: [email protected] imperial.ac.uk

without significantly altering component dimensions [2,4]. There has been a considerable level of interest in the benefit that these layers offer to stainless steels, especially with respect to corrosion resistance and enhanced fatigue performance [4,5]. However, little difference in the stress–strain curves has been observed in previous studies investigating stainless steels with enriched layers, even when the relative fraction of enriched layer per unit area is high [4,5]. It had therefore been concluded that the enriched layer has neither a beneficial or negative effect on the tensile properties [1,4]. However, it has also been speculated that this observation may not be true when the fraction of enriched to unaltered material is further increased [1]. No attempt has been reported in the existing literature to characterize the tensile behaviour of these layers. In the present study, a novel evaluation of the effect of an enriched carbon layer on the tensile properties of 200 lm thick stainless steel sheet is reported. An extremely narrow monochromated incident X-ray beam was used to interrogate the sample whilst it was undergoing tensile deformation, allowing diffraction data to be collected independently from both the bulk material and the enriched surface layer. Two sheets of 200 lm thick AISI 304 grade stainless steel were obtained from RS Components Limited, one of which was subjected to a “Kolsterising 33” treatment at Bodycote Hardiff bv, Holland. Light microscopy was conducted following mechanical polishing and etching

1359-6462/$ - see front matter Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.scriptamat.2010.03.020

86

N. G. Jones et al. / Scripta Materialia 63 (2010) 85–88

Figure 1. Light micrographs of (a) untreated as-received 304 grade stainless steel and (b) 304 stainless steel following a Kolsterising 33 treatment.

with waterless Kallings reagent. The as-received material (304; Fig. 1a) had a fine equiaxed grain structure, with evidence of the extensive prior cold rolling required to produce thin sheet. The Kolsterised sheet (304K; Fig. 1b) had a similar, if slightly refined, equiaxed grain structure in the unaffected bulk section when compared to the untreated material. However, a distinct region 30 lm wide was observed on both surfaces of the sheet, corresponding to the carbon enriched layer from the Kolsterising treatment. Characterization of the layer by nano-indentation and energy dispersive X-ray microanalysis indicated a maximum hardness of 8 GPa and an increase of 9 at.% carbon from bulk to edge. In situ experiments were conducted on beamline ID15B at the European Synchrotron Radiation Facility (ESRF). A schematic representation of the diffraction geometry is shown in Figure 2a, and is similar to previously reported configurations [6]. The monochromated ˚, incident X-ray beam had a wavelength of 0.1430 A and the resulting diffraction rings were acquired by a Pixium 2D area detector. Tensile testing was conducted using an Instron 5kN servohydrualic machine, using dogbone samples (1.5 mm wide by 200 lm thick) held by bespoke fixtures with a negative sample profile to prevent slippage (Fig. 2b). Tests were conducted incrementally using position control and a strain rate of

1  104 s1. Initially an untreated 304 sample was tested using an incident beam 100 lm wide and 300 lm high. At each strain increment, several diffraction spectra, with an exposure time of 0.4 s, were acquired and averaged, improving the signal to noise ratio. The incident beam size was reduced to 10 lm wide by 5 lm high when testing 304K, so that at each strain increment the beam could be rastered across the sample from the centre to the edge, collecting diffraction data from both the unaffected bulk and the carbon enriched layer. Even with reduced beam dimensions, full Debye–Scherrer patterns were acquired, indicating that the grain size was fine enough that a significant number of grains existed within the diffraction volume. The average carbon content in the diffracting volume from the layer can be estimated using the compositional dependence of lattice parameter [7], and published data for austenite give oa/oC = 0.75 pm at.%1 [8,9]. Here, a0 ˚ , and for the 304 base material was found to be 3.594 A ˚ 3.650 A in the layer, leading to an estimate of 7.5 at.% carbon in the layer. This is consistent with energy dispersive X-ray measurements and with previously reported material exhibiting a hardness between 7 and 8 GPa [4,5]. The macroscopic stress–strain curves of the as-received and Kolsterised material are shown in Figure 2c. The untreated material shows classical stress–strain behaviour, with a yield stress of 700 MPa and a tensile strength of 900 MPa. The treated material follows a similar overall shape as the untreated material but with a significantly higher tensile strength, 1200 MPa. Examination of the 304K elastic region shows a deflection in linearity at 500 MPa, following which a lower gradient regime exists prior to yielding at a stress of 1100 MPa. Evaluation of the lattice distortion during tensile testing was achieved by peak analysis from 10° radial segments of the diffraction rings, taken parallel to the loading direction, as outlined in Ref. [10]. Radial sectioning was accomplished using Fit2D [11] and Voigt function peak fitting carried out using Wavemetrics Igor Pro. Lattice strain (ehkl) was defined as ehkl = Dd/d0 for each family of planes (Fig. 3). Sample texture, from prior forming operations, resulted in weak {2 2 0} diffraction, preventing lattice strain analysis. The stress– lattice strain behaviour of 304 is shown in Figure 3a, and is similar to that expected from austenitic steels

Figure 2. (a) Schematic representation of the diffraction configurations, (b) photograph of the dogbone sample in the servohydraulic testing machine prior to testing and (c) macroscopic stress–strain behaviour of untreated (304) material and Kolsterised (304K) material.

N. G. Jones et al. / Scripta Materialia 63 (2010) 85–88

87

Figure 3. Stress–lattice strain curves showing the response of diffracting planes in the loading direction for (a) untreated 304, (b) the unaffected bulk material in 304K and (c) the Kolsterised layer in 304K.

[12,13]. Grains in the {1 1 1} diffraction condition are observed to have the highest modulus but yield at the lowest strain, indicated by the cessation of lattice strain accumulation. The {3 1 1} and {2 0 0} diffraction conditions exhibited lower elastic moduli than the {1 1 1} but continued to accumulate lattice strain after the {1 1 1} had yielded, evidence of load partitioning. Diffraction elastic constants (DECs) for each plane family were determined from the gradient of the elastic region and compare favourably with those measured for a 316 grade stainless steel using neutron diffraction [13] (Table 1). Typical uncertainties for the DECs were ±5 GPa. The stress–lattice strain for the bulk material in the Kolsterised sample is shown in Figure 3b. Macroscopic observations suggested that at a stress close to 500 MPa the unaffected bulk material yielded and load partitioned to the Kolsterised layer (Fig. 2c). Load transfer can clearly be seen in the lattice strain accumulation of the bulk material (Fig. 3b), where increasing load between 500 and 1000 MPa results in little change to the magnitude of lattice strain. DECs obtained from the initial elastic section of graph are shown in Table 1 and are similar to those for both 316 and the as received 304. The stress–lattice strain behaviour of the Kolsterised layer is shown in Figure 3c. The enriched layer was able to accommodate significantly greater lattice strains than the bulk material, and possessed lower DECs (shown in Table 1). The carbon enrichment dramatically decreased the variation in elastic moduli between different families of planes in the layer; however, the observed order in which they yield (namely {1 1 1}, {3 1 1}, {2 0 0}) remained the same. The reduction of elastic moduli by increasing the concentration of interstitial carbon atoms is consistent with previous bulk measurements in austenitic steels [14], and when testing in the austenitic phase field [15]. These observations are also consistent with Eshelby’s [16] analysis, which predicts that elastic properties soften when lattice defects cause an increase in volume. Table 1. Measured DECs for untreated 304 and Kolsterised material from the current work and previously published DECs for a 316 grade stainless steel from neutron diffraction data (all in GPa).

{1 1 1} {2 0 0} {3 1 1}

304

304K Bulk

304K Layer

316 [13]

251 134 184

239 148 193

96 82 87

246 147 183

However, the extent of the observed reduction in Young’s modulus (E) is in excess of what could be realized from linear trends determined at either (i) room temperature with limited interstitial carbon (<1.3 at.%) [14] or (ii) high temperature with compositions up to 7 at.% [15]. A nonlinear dependence of E with respect to carbon content has been reported at room temperature in martensitic Fe–C alloys, with compositions reaching 2.3 at.% carbon [17]. The general form of this relationship is E[C] = 1.91[C]2  1.70[C] + 208 (where [C] is at.% carbon), which, when extrapolated to concentrations of 7–8 at.%, depresses the modulus to a similar extent to those of the Kolsterised layer. Thus, whilst a linear relationship between carbon content and E may exist for low interstitial solute concentrations, the current data suggests that further enrichment results in a non linear dependence in face-centred cubic Fe at room temperature. The lower modulus of the Kolsterised layer is significant in terms of the nominal yield stress of the untreated base material, which was lower (500 MPa) than that of the as-received material (700 MPa). It can be inferred that, because of the lower modulus of the layer, the stress borne by the base material is elevated in the elastic region, and therefore the actual strength of the base material is greater than 500 MPa, i.e. its strength is broadly consistent with the untreated 304 material. Existing work considering the tensile properties of carburized materials has been limited to measurements of the macroscopic behaviour, rather than that of the enriched material and thus the properties of the layer are largely unknown. A simple rule of mixtures calculation, based on the volume fraction of the enriched layer and macroscopic stress behaviour (Fig. 2c), leads to an estimated tensile strength of 2 GPa for the Kolsterised layer. However, by combining the experimentally measured lattice strains and the corresponding DECs, it is possible to produce an estimated value of stress (r ffi Ehklehkl) for each measured strain increment, as is conventional for stress measurements from lattice strains in components [18]. Thus it becomes possible to construct a stress–strain curve for each family of planes in both the bulk material and the Kolseterised layer. Figure 4 shows the stress as a function of macroscopic strain for the {1 1 1}, {2 0 0} and {3 1 1} diffraction orientations of the layer and bulk material in 304K.

88

N. G. Jones et al. / Scripta Materialia 63 (2010) 85–88

Figure 4. Reconstructed stress–strain curves for the unaffected bulk material and the enriched Kolsterised layer, showing the average stress for grains in each diffraction orientation.

From Figure 4 it can be observed that the {2 0 0} diffractions in the Kolsterised layer have a tensile strength in excess of 2 GPa, and the {3 1 1} diffractions have a strength of 1900 MPa, both of which are in good agreement with the rule of mixtures calculation. The base material strength is observed to be just less than 600 MPa for all of the measured orientations. It is speculated that the remaining decrease in strength compared to the as-received material (100 MPa) is due to the relaxation of the dislocation network, induced by cold rolling, during the Kolsterising treatment. Current estimates for the relationship between diamond indenter hardness and yield stress (ry in MPa) for steels suggest ry = 2.876 Hv  90.7 [19]. Therefore, the 8 GPa measured hardness implies a yield stress of 2.27 ± 0.2 GPa, as compared to maximum 1.8 GPa measured here for {2 0 0}. This overprediction could suggest that hardness–strength correlations cannot be reliably extended to austenitic steels with high interstitial concentrations. However, the reconstructed stresses represent an average across the diffracting volume, and are representative of the average carbon content in the layer. Therefore, the strengths attained by the most heavily enriched material might be even higher than the reconstructed values. The significant increase in strength resultant of the Kolsterising process would suggest that such a treatment could be combined with other strengthening mechanisms, such as precipitate hardening and grain size reduction, as part of a multi-length scale approach to high strength alloys. In summary, it has been shown that a heavily enriched (7.5 at.%) carbon layer can cause significant enhancement of the tensile properties of a stainless steel if it is present in a sufficient area fraction (30% in the current work). High energy synchrotron diffraction using an incident beam focused to a size of 10 lm by 5 lm enabled diffraction spectra to be obtained from within the layer whilst undergoing tensile deformation. The DECs of the enriched layer were considerably softened in comparison to the parent material, consistent with previous reports [14]. The stress–strain behaviour of the enriched layer was constructed from the diffraction data and showed significant strengthening of all lattice planes, and strengths >2 GPa were observed.

The authors would like to thank Thomas Buslaps and Jerome Andrieux on ID15B for their assistance in setting up the diffraction experiment and the ESRF for funding experiment MA862. Finn Giuliani at Imperial College performed the nano-indentation measurements. This work was supported from the materials and structures research programme delivered by Team MAST for the Defence Technology and Innovation Centre, part of the UK Ministry of Defence. N.G.J. was also supported by the EPSRC PhD Plus scheme and D.D. by EPSRC under EP/H004882/1. [1] O. Rey, P. Jacquot, Surface Engineering 18 (2002) 412. [2] Y. Cao, F. Ernst, G.M. Michal, Acta Materialia 51 (2003) 4171. [3] P.C. Williams, S.V. Marx (Swagelok Company) US Patent 6547888, 2003. [4] G.M. Michal, F. Ernst, H. Kahn, Y. Cao, F. Oba, N. Agarwal, A.H. Heuer, Acta Materialia 54 (2006) 1597. [5] N. Agarwal, H. Kahn, A. Avishai, G. Michal, F. Ernst, A.H. Heuer, Acta Materialia 55 (2007) 5572. [6] R.J. Talling, R.J. Dashwood, M. Jackson, S. Kirumoto, D. Dye, Scripta Materialia 59 (2008) 669. [7] L. Vegard, Zeitschrift fu¨r Physik 5 (1921) 17. [8] M. Onink, C.M. Brakman, F.D. Tichelaar, E.J. Mittemeijer, S. van der Zwaag, J.H. Root, N.B. Konyer, Scripta Materialia 29 (1993) 1011. [9] L. Cheng, A. Bottger, T.H. Dekeijser, E.J. Mittemeijer, Scripta Materialia 24 (1990) 509. [10] A.M. Korsunsky, K.E. Wells, P.J. Withers, Scripta Materialia 39 (1998) 1705. [11] A.P. Hammersley, S.O. Svensson, A. Thompson, Nuclear Instruments & Methods 346 (1994) 312. [12] D. Dye, H.J. Stone, R.C. Reed, Current Opinion in Solid State and Materials Science 5 (2001) 31. [13] B. Clausen, T. Lorentzen, M.A.M. Bourke, M.R. Daymond, Materials Science and Engineering A259 (1999) 17. [14] H.M. Ledbetter, M.W. Austin, Materials Science and Engineering 70 (1985) 143. [15] W.J. Arnoult, R.B. McLellan, Acta Materialia 23 (1975) 51. [16] J.D. Eshelby, Journal of Applied Physics 25 (1954) 255. [17] G.R. Speich, A.J. Schwoeble, W.C. Leslie, Metallurgical Transactions 3 (1972) 2031. [18] D. Dye, K.T. Conlon, R.C. Reed, Metallurgical and Materials Transactions 35A (2004) 1703. [19] E.J. Pavlina, C.J. Van Tyne, Journal of Materials Engineering and Performance 17 (2008) 888.