- Email: [email protected]
Correlation between PIII nitriding parameters and corrosion behaviour of austenitic stainless steels
Surface & Coatings Technology 200 (2005) 104 – 108 www.elsevier.com/locate/surfcoat
Correlation between PIII nitriding parameters and corrosion behaviour of austenitic stainless steels S. M7ndla,T, D. Manovaa, H. Neumanna, M.T. Phamb, E. Richterb, B. Rauschenbacha b
a Leibniz-Institut fu¨r Oberfla¨chenmodifizierung, Permoserstr. 15, 04303 Leipzig, Germany Institut fu¨r Ionenstrahlphysik und Materialforschung, Forschungszentrum Rossendorf, 01314 Dresden, Germany
Available online 5 April 2005
Abstract In this work, nitrogen plasma immersion ion implantation (PIII) treatment of austenitic stainless steels 1.4301 and 1.4571 was performed to investigate the influence of the process conditions on the corrosion properties. Short treatment, high voltage and high temperature result in a decreased corrosion potential while no correlation to layer thickness, nitrogen concentration or lattice expansion was found. Except for the possibility of small CrN agglomerates at high temperatures, no direct explanation for the results can be provided and it is argued that intrinsic stress accumulation and relaxation may be responsible. D 2005 Elsevier B.V. All rights reserved. Keywords: Expanded austenite; PIII; Corrosion; XRD; GDOS
1. Introduction Recent advances in computational material science nowadays allow the design of advanced materials, e.g., alloy steels [1,2]. Especially, the parallel optimization of several properties is possible, for instance hardness, ductility and corrosion resistance. At the same time, descriptions of macroscopic parameters in terms of the electronic band structure or related features are available [3,4]. Despite these advances in computation, material design is still hampered by missing structure–property relations on the quantum scale [5]. Recent calculations aiming to optimize the ternary Fe–Cr–Ni system, which includes important austenitic stainless steel alloys as AISI 304, 316, 317, with respect to hardness and corrosion resistance [6,7] are highly controversial [8] and can be regarded only as first attempts. Despite the large commercial interest in nitrided austenitic stainless steels, no satisfying explanation for the outstanding results is known to the authors. A
T Corresponding author. Tel.: +49 341 235 2944; fax: +49 341 235 2313. E-mail address: [email protected] (S. M7ndl). 0257-8972/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2005.02.084
hardness of up to 1200 HV is reported at nitrogen concentrations of 10–25 at.%, in conjunction with a lattice expansion of 5–10% at a process temperature of 350–380 8C [9,10]. Higher temperatures lead to the decomposition into CrN and ferrite and thus in destroying of the corrosion resistance [11]. Nevertheless, even at 380 8C and below, improved and degraded corrosion resistance is reported in the literature [12–14]. The present investigation focuses on studying the influence of process parameters on the corrosion resistance of nitrided austenite steel by plasma immersion ion implantation and attempts to find a correlation between the process parameters as voltage, time or temperature and the resulting corrosion resistance.
2. Experiment Two different austenitic stainless steel grades – DIN 1.4301 (AISI 304) and 1.4571 (316Ti) – were used in these experiments. Nitrogen plasma immersion ion implantation was performed at different parameters simultaneously in samples from both steel grades. The experiments were made at a base pressure lower than 10 3 Pa and working
S. Ma¨ndl et al. / Surface & Coatings Technology 200 (2005) 104–108
pressure of 0.2 Pa resulting from nitrogen flow rate of 50 sccm. The implantation parameters varied in a limited range to obtain representative sample encompassing the typical parameters used in energetic nitriding of austenitic stainless steel. The process temperature between 340 and 380 8C was maintained by adjustment the pulse frequency and measured by an IR pyrometer. No external heating or cooling was employed so the samples were heated only by implanted ions. The pulse voltage was 10 or 20 kV with pulse length of 15 As. The process time was 1 or 2 h resulting in implanted dose between 8.0 1017 at./cm2 and 2.5 1018 at./cm2. Additionally, samples from both steel grades nitrided for 1 h at 340 8C were post-implanted with oxygen at 15 kV for 15 min at 300 8C. X-ray diffraction (XRD) was employed in Bragg– Bretano geometry to determine the phase composition while the elemental depth profiles were obtained from glow discharge optical spectroscopy (GDOS) measurements with an accuracy of 0.1–0.2 at.% and a reproducibility of better than 0.1 at.% [15]. The corrosion properties were studied by carrying out potentiodynamic polarization measurements in 1% NaCl solution with sweep rate of 10 mV/s. The potential was measured against saturated calomel electrode (SCE). At the same time, corrosion salt bath tests in 3% NaCl solution at an elevated temperature of 40 8C were conducted for several days or until the onset of corrosion was observed.
105
3. Results and discussion Fig. 1a shows the nitrogen depth profiles of two samples from 1.4301 steel grade implanted with 10-kV acceleration voltage at 340 8C for 2 h and at 380 8C for 1 h, respectively. The near surface concentration differs between 40% and 25% for sample at 340 and 380 8C, respectively. In both samples, a slow initial concentration decrease starting just below the surface is followed by a steeper decrease beyond 1.6 Am, respectively at approximately 25 and 15 at.% nitrogen, indicating a concentration dependent diffusion coefficient [16]. The layer thickness, defined here as 1.5 at.% nitrogen, obtained from the depth profiles is approximately 2.2 Am for both samples. XRD spectra for the same samples, presented in Fig. 2, show different peak intensities as well as a different lattice expansion, despite the quite similar layer thickness. The relative lattice expansion estimated from XRD peaks is 4.7% and 6.5% for the (111) and (200) peaks for the sample at 340 8C, respectively, 3.3% and 4.3% for the (111) and (200) peaks at 380 8C. These two samples show typical results of the whole series, where the layer thickness is between 1 and 3 Am while the nitrogen surface concentration varies between 18 and 42 at.% and the relative lattice expansion, averaged over the (111) and (200) peaks is in the range of 2.4 to 5.5%. The corrosion current as a function of the applied potential, obtained from potentiodynamic polarization
Depth ( m) 0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
40
1.4301, 380°C, 10 kV, 1h 1.4301, 340°C, 10 kV, 2h
35 30 25 20 15 10
40
45
γ(200)
γN(200)
a)
0 45 40 35 30 25 20 15 10 5 b) 0 35
γ(111)
5
γN(111)
X-Ray Intensity
Nitrogen Conc. (at.%)
0,0 45
50
55
Angle 2θ (°) Fig. 1. (a) Nitrogen depth profile after PIII into steel 1.4301 at 340 8C and 380 8C for 2 and 1 h, respectively; (b) XRD spectra for the same samples with the peak positions for the austenite base material and the expanded austenite marked.
106
S. Ma¨ndl et al. / Surface & Coatings Technology 200 (2005) 104–108 0,1 1.4301 untreated 1.4301, 340°C, 10 kV, 2h 1.4301, 380°C, 10 kV, 1h 1.4301, 340°C, 20 kV, 1h
Corrosion current (A/cm2)
0,01
1.4571 untreated 1.4571 380°C, 10 kV, 1h
1E-3 1E-4 1E-5 1E-6 1E-7 1E-8 1E-9 -0,4
-0,2
0,0
0,2
0,4
0,6
Potential (V) Fig. 2. Potentiodynamic polarization curves of selected implanted and unimplanted samples from 1.4301 and 1.4571 steel grades after nitrogen PIII. The curves for 1.4571 steel implanted at 340 8C are close to the curves of 1.4301 steel.
at 10 kV show typical passivating behaviour, i.e., a very low corrosion current nearly independent of the applied potential for potential values more positive than the corrosion potential. Visualisation of the data is complicated by the fact that there are six parameters – pulse voltage, implantation time, thickness of the implanted layer and process temperature, lattice expansion and nitrogen concentration – which could influence the corrosion process. Selected two-dimensional cross-sections through this parameter space are shown in
measurements, is depicted in Fig. 2 for selected samples nitrided at different parameters as well as for non-implanted samples. The measured corrosion potentials range from 0.30 to +0.05 eV. In addition, varying corrosion currents are found. The shape of the current–voltage curves can be roughly divided into two groups: one group showing passivating behaviour and the other not. It can be clearly seen from Fig. 2 that the sample from 1.4571 steel grade implanted at 380 8C does not show a passivating behaviour. In contrary, the samples from 1.4301 steel grade implanted Voltage (kV) 0
0,1
5
10
Time (h) 15
20
0
1
2
1.4301 1.4571 1.4301 + oxide 1.4571 + oxide
0,0 -0,1
Potential (v)
-0,2 -0,3 0,1
a)
b)
c)
d)
0,0 -0,1 -0,2 -0,3 0
100
200
300
Temperature (°C)
400
0
1
2
3
4
Layer Thickness ( m)
Fig. 3. Corrosion potential of nitrogen implanted 1.4301 and 1.4571 steel samples as a function of (a) applied voltage, (b) implantation time, (c) process temperature and (d) thickness of implanted layer, as defined by a nitrogen concentration of 1.5 at.%.
S. Ma¨ndl et al. / Surface & Coatings Technology 200 (2005) 104–108
107
Fig. 4. Optical micrographs after salt bath test: (a and b) 1.4301 and 1.4571 implanted at 340 8C for 2 h after 24 h in salt bath; (c and d) 1.4301 and 1.4571 implanted at 380 8C for 1 h after 80 min in salt bath.
Fig. 3 to obtain an overview on the relevance of some parameters. Except for one data point (see below), the difference between the corrosion behaviour of the two steels grades disappears after the nitrogen implantation, i.e., nearly identical pairs of corrosion potentials are obtained. At a pulse voltage of 10 kV (see Fig. 3a), corrosion potentials across the whole range are obtained, independent of the base material whereas the set of samples at 20 kV does show a quite negative corrosion potential. The postoxidized samples, in contrast, exhibit excellent values. For the treatment time, an inverse effect may be diagnosed: longer times lead to better corrosion protection. A higher temperature leads apparently to inconclusive results (Fig. 3c), with improvement and degrading found simultaneously in the potentiodynamic measurements, the only instance where the results are strongly depending on the steel grade. When plotting the corrosion potential vs. the layer thickness as in Fig. 3d, a weak – nevertheless still existing correlation between thick layers and high corrosion potential
is found. In contrast, the average relative lattice expansion and nitrogen surface concentration show only uncorrelated results with respect to the corrosion potential (no figures shown here). Salt bath tests were additionally performed as a different diagnostic tool. Fig. 4 shows the results for samples (1.4301 and 1.4571) implanted at 340 8C and 380 8C for 2 and 1 h, respectively. Both steel grades, implanted at 380 8C started to show corrosion within 80 min, while the samples implanted at 340 8C were without corrosion after 72 h in salt bath. As a certain incubation time is needed before the onset of pitting corrosion, this could be missed in potentiodynamic measurements due to the different timescales. Resolving the apparent contradiction with Fig. 3c, it must be stated that long-term salt bath tests are a more reliable indicator of the corrosion resistance. As a result from these experiments, it was found that nitrogen insertion can improve or degrade the corrosion resistance depending on the treatment parameters while the
0.34
1.4301
1.4571 before after oxygen implantation at 300°C
0.32
Cr/Fe Ratio
0.30
0.28
0.26
0.24
0.22 0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Depth ( m) Fig. 5. Chromium depth profiles before and after subsequent oxygen implantation for samples implanted for 1 h at 340 8C.
108
S. Ma¨ndl et al. / Surface & Coatings Technology 200 (2005) 104–108
influence of the base material seems to be negligible. However, neither the layer thickness, nor the nitrogen concentration or the lattice expansion, characterising the implanted layers, can explain the corrosion behaviour, so that a different factor must be pursued. Following mechanisms may be present in the surface layer determining the corrosion behaviour: chromium immobilisation, change of morphology and stress accumulation. Cr immobility can be excluded at 340 8C as GDOS measurements of the sample after additional oxygen implantation show a clear migration of Cr towards the surface (see Fig. 5; it has to be kept in mind that a surface recess of about 100–200 nm occurs during the oxygen implantation due to sputtering, hence the chromium mobility is more pronounced than visible in the figure), which means that, despite reports of Cr–N bonds from XPS measurements [17], mobile Cr does still exist. It must be pointed out that a Cr/N ratio of close to 1 (cf. Fig. 1a and Ref. [18]) is a result obtained only in special circumstances for the presented PIII implantations. XRD data show no peaks beside expanded austenite and the base material, nevertheless the formation of Cr–N-aggregates at 380 8C may occur at 380 8C, leading to the compromised corrosion resistance. As the size would be below the XRD detection limit, the conventional assignment of CrN is preferred, albeit Cr2N cannot be ruled out. No change in the grain size was observed in any sample, thus enhanced or reduced grain boundary corrosion is excluded. Stress accumulation through dislocations and stacking faults within the nitrided layer could be an explanation for the modified corrosion properties. As the expanded austenite shows a gradual variation of the lattice expansion from the surface towards the end of the layer, no unequivocal identification of these defects from XRD data is possible. However, then it has to be assumed that a defect-free layer has excellent corrosion properties, independent of the base material, and any degradation occurs due to the defect density specific to the respective treatment. This may help to explain why lower voltages and longer times improve the corrosion as less defects are produced and more defects can be annealed respectively.
4. Summary and conclusions The influence of different process parameters during nitrogen PIII on the corrosion behaviour of two austenitic
steel grades was investigated. No influence of the steel grade on the corrosion resistance after the implantation was found with a corrosion potential of the best samples slightly better than that of untreated 1.4571 and much better than untreated 1.4301. Short implantations, high voltages or higher temperatures all lead to a degradation of the corrosion resistance. For the latter one, CrN agglomerates can be envisaged as a cause, whereas no immediate mechanism can be presented for the former two. A correlation with stress-induced defects is perhaps possible. Several film properties, i.e., lattice expansion, layer thickness and nitrogen concentration are not correlated at all with the corrosion properties.
Acknowledgements This work was supported by the S7chsische Aufbaubank project SMWA 6204/947.
References [1] K. Ohno, K. Esfarjani, Y. Kawazoe, Computational Materials Science, Springer-Verlag, Berlin, 1999. [2] D. Raabe, Computational Material Science, Wiley-VCH Verlag, Weinheim, 1998. [3] M.L. Cohen, Mater. Sci. Eng. A105/106 (1988) 11. [4] F. Willaime, Adv. Eng. Mater. 3 (2001) 283. [5] M.E. Eberhart, D.P. Clougherty, Nat. Mater. 3 (2004) 659. [6] L. Vitos, P.A. Korzhavyi, B. Johansson, Adv. Eng. Mater. 4 (2004) 228. [7] L. Vitos, P.A. Korzhavyi, B. Johansson, Nat. Mater. 2 (2003) 25. [8] L. Weber, Adv. Eng. Mater. 6 (2004) 786. [9] M. Samandi, B.A. Shedden, D.I. Smith, G.A. Collins, R. Hutching, J. Tendys, Surf. Coat. Technol. 59 (1993) 261. [10] B. Larisch, U. Brusky, H.-J. Spies, Surf. Coat. Technol. 116/119 (1999) 199. [11] A. Leyland, D.B. Lewis, P.R. Stevenson, A. Matthews, Surf. Coat. Technol. 62 (1993) 608. [12] S. Mukherjee, P.M. Raole, A. Kumar, I. Chattoraj, K.R.M. Rao, I. Manna, Surf. Coat. Technol. 186 (2004) 282. [13] M. Samandi, B.A. Shedden, D.I. Smith, G.A. Collins, R. Hutchings, J. Tendys, Surf. Coat. Technol. 59 (1993) 261. [14] S. M7ndl, R. Gqnzel, E. Richter, W. Mfller, Surf. Coat. Technol. 100/ 101 (1998) 372. [15] T. Asam, Surf. Coat. Technol. 116/119 (1999) 310. [16] S. M7ndl, B. Rauschenbach, J. Appl. Phys. 91 (2002) 9737. [17] J.P. Riviere, M. Cahoreau, P. Meheust, J. Appl. Phys. 91 (2002) 6361. [18] S. M7ndl, F. Scholze, H. Neumann, B. Rauschenbach, Surf. Coat. Technol. 174/175 (2003) 1191.
Correlation between PIII nitriding parameters and corrosion behaviour of austenitic stainless steels S. M7ndla,T, D. Manovaa, H. Neumanna, M.T. Phamb, E. Richterb, B. Rauschenbacha b
a Leibniz-Institut fu¨r Oberfla¨chenmodifizierung, Permoserstr. 15, 04303 Leipzig, Germany Institut fu¨r Ionenstrahlphysik und Materialforschung, Forschungszentrum Rossendorf, 01314 Dresden, Germany
Available online 5 April 2005
Abstract In this work, nitrogen plasma immersion ion implantation (PIII) treatment of austenitic stainless steels 1.4301 and 1.4571 was performed to investigate the influence of the process conditions on the corrosion properties. Short treatment, high voltage and high temperature result in a decreased corrosion potential while no correlation to layer thickness, nitrogen concentration or lattice expansion was found. Except for the possibility of small CrN agglomerates at high temperatures, no direct explanation for the results can be provided and it is argued that intrinsic stress accumulation and relaxation may be responsible. D 2005 Elsevier B.V. All rights reserved. Keywords: Expanded austenite; PIII; Corrosion; XRD; GDOS
1. Introduction Recent advances in computational material science nowadays allow the design of advanced materials, e.g., alloy steels [1,2]. Especially, the parallel optimization of several properties is possible, for instance hardness, ductility and corrosion resistance. At the same time, descriptions of macroscopic parameters in terms of the electronic band structure or related features are available [3,4]. Despite these advances in computation, material design is still hampered by missing structure–property relations on the quantum scale [5]. Recent calculations aiming to optimize the ternary Fe–Cr–Ni system, which includes important austenitic stainless steel alloys as AISI 304, 316, 317, with respect to hardness and corrosion resistance [6,7] are highly controversial [8] and can be regarded only as first attempts. Despite the large commercial interest in nitrided austenitic stainless steels, no satisfying explanation for the outstanding results is known to the authors. A
T Corresponding author. Tel.: +49 341 235 2944; fax: +49 341 235 2313. E-mail address: [email protected] (S. M7ndl). 0257-8972/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2005.02.084
hardness of up to 1200 HV is reported at nitrogen concentrations of 10–25 at.%, in conjunction with a lattice expansion of 5–10% at a process temperature of 350–380 8C [9,10]. Higher temperatures lead to the decomposition into CrN and ferrite and thus in destroying of the corrosion resistance [11]. Nevertheless, even at 380 8C and below, improved and degraded corrosion resistance is reported in the literature [12–14]. The present investigation focuses on studying the influence of process parameters on the corrosion resistance of nitrided austenite steel by plasma immersion ion implantation and attempts to find a correlation between the process parameters as voltage, time or temperature and the resulting corrosion resistance.
2. Experiment Two different austenitic stainless steel grades – DIN 1.4301 (AISI 304) and 1.4571 (316Ti) – were used in these experiments. Nitrogen plasma immersion ion implantation was performed at different parameters simultaneously in samples from both steel grades. The experiments were made at a base pressure lower than 10 3 Pa and working
S. Ma¨ndl et al. / Surface & Coatings Technology 200 (2005) 104–108
pressure of 0.2 Pa resulting from nitrogen flow rate of 50 sccm. The implantation parameters varied in a limited range to obtain representative sample encompassing the typical parameters used in energetic nitriding of austenitic stainless steel. The process temperature between 340 and 380 8C was maintained by adjustment the pulse frequency and measured by an IR pyrometer. No external heating or cooling was employed so the samples were heated only by implanted ions. The pulse voltage was 10 or 20 kV with pulse length of 15 As. The process time was 1 or 2 h resulting in implanted dose between 8.0 1017 at./cm2 and 2.5 1018 at./cm2. Additionally, samples from both steel grades nitrided for 1 h at 340 8C were post-implanted with oxygen at 15 kV for 15 min at 300 8C. X-ray diffraction (XRD) was employed in Bragg– Bretano geometry to determine the phase composition while the elemental depth profiles were obtained from glow discharge optical spectroscopy (GDOS) measurements with an accuracy of 0.1–0.2 at.% and a reproducibility of better than 0.1 at.% [15]. The corrosion properties were studied by carrying out potentiodynamic polarization measurements in 1% NaCl solution with sweep rate of 10 mV/s. The potential was measured against saturated calomel electrode (SCE). At the same time, corrosion salt bath tests in 3% NaCl solution at an elevated temperature of 40 8C were conducted for several days or until the onset of corrosion was observed.
105
3. Results and discussion Fig. 1a shows the nitrogen depth profiles of two samples from 1.4301 steel grade implanted with 10-kV acceleration voltage at 340 8C for 2 h and at 380 8C for 1 h, respectively. The near surface concentration differs between 40% and 25% for sample at 340 and 380 8C, respectively. In both samples, a slow initial concentration decrease starting just below the surface is followed by a steeper decrease beyond 1.6 Am, respectively at approximately 25 and 15 at.% nitrogen, indicating a concentration dependent diffusion coefficient [16]. The layer thickness, defined here as 1.5 at.% nitrogen, obtained from the depth profiles is approximately 2.2 Am for both samples. XRD spectra for the same samples, presented in Fig. 2, show different peak intensities as well as a different lattice expansion, despite the quite similar layer thickness. The relative lattice expansion estimated from XRD peaks is 4.7% and 6.5% for the (111) and (200) peaks for the sample at 340 8C, respectively, 3.3% and 4.3% for the (111) and (200) peaks at 380 8C. These two samples show typical results of the whole series, where the layer thickness is between 1 and 3 Am while the nitrogen surface concentration varies between 18 and 42 at.% and the relative lattice expansion, averaged over the (111) and (200) peaks is in the range of 2.4 to 5.5%. The corrosion current as a function of the applied potential, obtained from potentiodynamic polarization
Depth ( m) 0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
40
1.4301, 380°C, 10 kV, 1h 1.4301, 340°C, 10 kV, 2h
35 30 25 20 15 10
40
45
γ(200)
γN(200)
a)
0 45 40 35 30 25 20 15 10 5 b) 0 35
γ(111)
5
γN(111)
X-Ray Intensity
Nitrogen Conc. (at.%)
0,0 45
50
55
Angle 2θ (°) Fig. 1. (a) Nitrogen depth profile after PIII into steel 1.4301 at 340 8C and 380 8C for 2 and 1 h, respectively; (b) XRD spectra for the same samples with the peak positions for the austenite base material and the expanded austenite marked.
106
S. Ma¨ndl et al. / Surface & Coatings Technology 200 (2005) 104–108 0,1 1.4301 untreated 1.4301, 340°C, 10 kV, 2h 1.4301, 380°C, 10 kV, 1h 1.4301, 340°C, 20 kV, 1h
Corrosion current (A/cm2)
0,01
1.4571 untreated 1.4571 380°C, 10 kV, 1h
1E-3 1E-4 1E-5 1E-6 1E-7 1E-8 1E-9 -0,4
-0,2
0,0
0,2
0,4
0,6
Potential (V) Fig. 2. Potentiodynamic polarization curves of selected implanted and unimplanted samples from 1.4301 and 1.4571 steel grades after nitrogen PIII. The curves for 1.4571 steel implanted at 340 8C are close to the curves of 1.4301 steel.
at 10 kV show typical passivating behaviour, i.e., a very low corrosion current nearly independent of the applied potential for potential values more positive than the corrosion potential. Visualisation of the data is complicated by the fact that there are six parameters – pulse voltage, implantation time, thickness of the implanted layer and process temperature, lattice expansion and nitrogen concentration – which could influence the corrosion process. Selected two-dimensional cross-sections through this parameter space are shown in
measurements, is depicted in Fig. 2 for selected samples nitrided at different parameters as well as for non-implanted samples. The measured corrosion potentials range from 0.30 to +0.05 eV. In addition, varying corrosion currents are found. The shape of the current–voltage curves can be roughly divided into two groups: one group showing passivating behaviour and the other not. It can be clearly seen from Fig. 2 that the sample from 1.4571 steel grade implanted at 380 8C does not show a passivating behaviour. In contrary, the samples from 1.4301 steel grade implanted Voltage (kV) 0
0,1
5
10
Time (h) 15
20
0
1
2
1.4301 1.4571 1.4301 + oxide 1.4571 + oxide
0,0 -0,1
Potential (v)
-0,2 -0,3 0,1
a)
b)
c)
d)
0,0 -0,1 -0,2 -0,3 0
100
200
300
Temperature (°C)
400
0
1
2
3
4
Layer Thickness ( m)
Fig. 3. Corrosion potential of nitrogen implanted 1.4301 and 1.4571 steel samples as a function of (a) applied voltage, (b) implantation time, (c) process temperature and (d) thickness of implanted layer, as defined by a nitrogen concentration of 1.5 at.%.
S. Ma¨ndl et al. / Surface & Coatings Technology 200 (2005) 104–108
107
Fig. 4. Optical micrographs after salt bath test: (a and b) 1.4301 and 1.4571 implanted at 340 8C for 2 h after 24 h in salt bath; (c and d) 1.4301 and 1.4571 implanted at 380 8C for 1 h after 80 min in salt bath.
Fig. 3 to obtain an overview on the relevance of some parameters. Except for one data point (see below), the difference between the corrosion behaviour of the two steels grades disappears after the nitrogen implantation, i.e., nearly identical pairs of corrosion potentials are obtained. At a pulse voltage of 10 kV (see Fig. 3a), corrosion potentials across the whole range are obtained, independent of the base material whereas the set of samples at 20 kV does show a quite negative corrosion potential. The postoxidized samples, in contrast, exhibit excellent values. For the treatment time, an inverse effect may be diagnosed: longer times lead to better corrosion protection. A higher temperature leads apparently to inconclusive results (Fig. 3c), with improvement and degrading found simultaneously in the potentiodynamic measurements, the only instance where the results are strongly depending on the steel grade. When plotting the corrosion potential vs. the layer thickness as in Fig. 3d, a weak – nevertheless still existing correlation between thick layers and high corrosion potential
is found. In contrast, the average relative lattice expansion and nitrogen surface concentration show only uncorrelated results with respect to the corrosion potential (no figures shown here). Salt bath tests were additionally performed as a different diagnostic tool. Fig. 4 shows the results for samples (1.4301 and 1.4571) implanted at 340 8C and 380 8C for 2 and 1 h, respectively. Both steel grades, implanted at 380 8C started to show corrosion within 80 min, while the samples implanted at 340 8C were without corrosion after 72 h in salt bath. As a certain incubation time is needed before the onset of pitting corrosion, this could be missed in potentiodynamic measurements due to the different timescales. Resolving the apparent contradiction with Fig. 3c, it must be stated that long-term salt bath tests are a more reliable indicator of the corrosion resistance. As a result from these experiments, it was found that nitrogen insertion can improve or degrade the corrosion resistance depending on the treatment parameters while the
0.34
1.4301
1.4571 before after oxygen implantation at 300°C
0.32
Cr/Fe Ratio
0.30
0.28
0.26
0.24
0.22 0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Depth ( m) Fig. 5. Chromium depth profiles before and after subsequent oxygen implantation for samples implanted for 1 h at 340 8C.
108
S. Ma¨ndl et al. / Surface & Coatings Technology 200 (2005) 104–108
influence of the base material seems to be negligible. However, neither the layer thickness, nor the nitrogen concentration or the lattice expansion, characterising the implanted layers, can explain the corrosion behaviour, so that a different factor must be pursued. Following mechanisms may be present in the surface layer determining the corrosion behaviour: chromium immobilisation, change of morphology and stress accumulation. Cr immobility can be excluded at 340 8C as GDOS measurements of the sample after additional oxygen implantation show a clear migration of Cr towards the surface (see Fig. 5; it has to be kept in mind that a surface recess of about 100–200 nm occurs during the oxygen implantation due to sputtering, hence the chromium mobility is more pronounced than visible in the figure), which means that, despite reports of Cr–N bonds from XPS measurements [17], mobile Cr does still exist. It must be pointed out that a Cr/N ratio of close to 1 (cf. Fig. 1a and Ref. [18]) is a result obtained only in special circumstances for the presented PIII implantations. XRD data show no peaks beside expanded austenite and the base material, nevertheless the formation of Cr–N-aggregates at 380 8C may occur at 380 8C, leading to the compromised corrosion resistance. As the size would be below the XRD detection limit, the conventional assignment of CrN is preferred, albeit Cr2N cannot be ruled out. No change in the grain size was observed in any sample, thus enhanced or reduced grain boundary corrosion is excluded. Stress accumulation through dislocations and stacking faults within the nitrided layer could be an explanation for the modified corrosion properties. As the expanded austenite shows a gradual variation of the lattice expansion from the surface towards the end of the layer, no unequivocal identification of these defects from XRD data is possible. However, then it has to be assumed that a defect-free layer has excellent corrosion properties, independent of the base material, and any degradation occurs due to the defect density specific to the respective treatment. This may help to explain why lower voltages and longer times improve the corrosion as less defects are produced and more defects can be annealed respectively.
4. Summary and conclusions The influence of different process parameters during nitrogen PIII on the corrosion behaviour of two austenitic
steel grades was investigated. No influence of the steel grade on the corrosion resistance after the implantation was found with a corrosion potential of the best samples slightly better than that of untreated 1.4571 and much better than untreated 1.4301. Short implantations, high voltages or higher temperatures all lead to a degradation of the corrosion resistance. For the latter one, CrN agglomerates can be envisaged as a cause, whereas no immediate mechanism can be presented for the former two. A correlation with stress-induced defects is perhaps possible. Several film properties, i.e., lattice expansion, layer thickness and nitrogen concentration are not correlated at all with the corrosion properties.
Acknowledgements This work was supported by the S7chsische Aufbaubank project SMWA 6204/947.
References [1] K. Ohno, K. Esfarjani, Y. Kawazoe, Computational Materials Science, Springer-Verlag, Berlin, 1999. [2] D. Raabe, Computational Material Science, Wiley-VCH Verlag, Weinheim, 1998. [3] M.L. Cohen, Mater. Sci. Eng. A105/106 (1988) 11. [4] F. Willaime, Adv. Eng. Mater. 3 (2001) 283. [5] M.E. Eberhart, D.P. Clougherty, Nat. Mater. 3 (2004) 659. [6] L. Vitos, P.A. Korzhavyi, B. Johansson, Adv. Eng. Mater. 4 (2004) 228. [7] L. Vitos, P.A. Korzhavyi, B. Johansson, Nat. Mater. 2 (2003) 25. [8] L. Weber, Adv. Eng. Mater. 6 (2004) 786. [9] M. Samandi, B.A. Shedden, D.I. Smith, G.A. Collins, R. Hutching, J. Tendys, Surf. Coat. Technol. 59 (1993) 261. [10] B. Larisch, U. Brusky, H.-J. Spies, Surf. Coat. Technol. 116/119 (1999) 199. [11] A. Leyland, D.B. Lewis, P.R. Stevenson, A. Matthews, Surf. Coat. Technol. 62 (1993) 608. [12] S. Mukherjee, P.M. Raole, A. Kumar, I. Chattoraj, K.R.M. Rao, I. Manna, Surf. Coat. Technol. 186 (2004) 282. [13] M. Samandi, B.A. Shedden, D.I. Smith, G.A. Collins, R. Hutchings, J. Tendys, Surf. Coat. Technol. 59 (1993) 261. [14] S. M7ndl, R. Gqnzel, E. Richter, W. Mfller, Surf. Coat. Technol. 100/ 101 (1998) 372. [15] T. Asam, Surf. Coat. Technol. 116/119 (1999) 310. [16] S. M7ndl, B. Rauschenbach, J. Appl. Phys. 91 (2002) 9737. [17] J.P. Riviere, M. Cahoreau, P. Meheust, J. Appl. Phys. 91 (2002) 6361. [18] S. M7ndl, F. Scholze, H. Neumann, B. Rauschenbach, Surf. Coat. Technol. 174/175 (2003) 1191.