- Email: [email protected]
Initiation and formation of electroless nickel–boron coatings on mild steel: Effect of substrate roughness
Materials Science and Engineering B 175 (2010) 266–273
Contents lists available at ScienceDirect
Materials Science and Engineering B journal homepage: www.elsevier.com/locate/mseb
Initiation and formation of electroless nickel–boron coatings on mild steel: Effect of substrate roughness V. Vitry a,∗ , A.-F. Kanta b , F. Delaunois a a b
Université de Mons, Service de Métallurgie, Place du Parc, 20, 7000 Mons, Belgium Université de Mons, Service de Science des Matériaux, Place du Parc, 20, 7000 Mons, Belgium
a r t i c l e
i n f o
Article history: Received 11 March 2010 Received in revised form 6 July 2010 Accepted 28 August 2010 Keywords: Film deposition Metals Surface morphology Surface roughnesses Electroless nickel–boron
a b s t r a c t The initial deposition and growth of electroless nickel–boron deposits on mild steel were studied: the films were prepared in an electroless plating bath using sodium borohydride as reducing agent. Samples were immersed in the plating solution for times from 5 s to 1 h and the morphological evolution of the deposit was followed by scanning electron microscopy (SEM) observation of the surface and prepared cross sections. Energy dispersive X-ray spectrometry (EDX) and glow discharge optical electro spectroscopy (GDOES) analyses were used to obtain information about the chemistry of the deposits and their results were correlated with the morphology of the coating. The initiation mechanism of electroless deposition on mild steel was identified. The effects of substrate roughness variation on the morphology and growth rate of the coatings were investigated by reproducing the experiment on samples with various surface preparation (grinding) states. We observed that the increase of substrate roughness favors the deposit initiation: the density of nickel nodules increases with increasing roughness of the substrate. Longer immersions in the bath lead to homogenization and densification of the coating and the nodules are clearly distinguishable. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Electroless coating is a well-established surface engineering process that was developed by Brenner and Riddel in 1946 [1]. It involves deposition of a metal–metalloid alloy coating on various substrates (including dielectric materials) by electrochemical reactions in aqueous solution. Electroless nickel deposits possess a number of interesting properties such as uniform thickness, high hardness, good corrosion resistance, etc. [2–5]. Electroless nickel deposits are usually classified according to the nature of the reducing agent. Nickel–phosphorous deposits (based on reduction by the hypophosphite ion) are the most studied and used but the properties of nickel–boron deposits are of very great interest for several industrial applications like aeronautics, petrochemical industry, food industry, firearms, etc.: their hardness is higher than nickel–phosphorous, and their electrical and tribological properties are very promising [6–14]. Most of the papers focused on electroless nickel and particularly on nickel–boron describe the optimization of plating parameters, the coating properties or the effect of heat treatment. There are very few papers dealing with deposit formation [7,8].
∗ Corresponding author. Tel.: +32 65 37 44 38; fax: +32 65 37 44 36. E-mail address: [email protected] (V. Vitry). 0921-5107/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.mseb.2010.08.003
Electroless plating occurs by the reduction of nickel ions at the surface of the active substrate immersed into the plating solution and further growth of the coating is due to a catalytic action of the deposit itself [15,16]. There are 3 known ways of inducing the initiation of electroless nickel deposition. Either the substrate is spontaneously catalytic for the oxidation of the reducing agent, or it is more easily oxidizable than nickel, in which case a thin catalytic layer of nickel is spontaneously deposited, or finally it can be activated by dipping in a solution of catalytically active metal salts (like Pd) or by galvanic coupling [4]. The few studies dedicated to the initiation mechanism of electroless deposition focused on systems in which catalysts, such as Pd are used [17–22]. This means that if no catalyst is used, the actual mechanism that allows initial deposition on a determined substrate is rarely known and is at best proposed by assumptions. The initiation and the growth of nickel–boron coatings have a non-negligible influence on their properties. In the case of experimental, unreplenished baths, the structure and the composition of the coating are not homogeneous over the coating thickness but change during the deposition, as was shown in the extreme case of a non-agitated bath by Rao et al. [23]. However, the observation of the formation of nickel–boron deposits has not been studied extensively yet, opposite to the growth of nickel–phosphorous coatings that has been studied on various substrates with and without catalytic activation [17–22,24–27].
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
267
Table 1 Deposition time for the observation of the initiation of electroless deposits. Sample
1
2
3
4
5
6
7
8
9
10
Deposition time
5s
15 s
30 s
60 s
90 s
4 min (240 s)
7 min (420 s)
10 min (600 s)
30 min (1800 s)
60 min (3600 s)
It is also well known by the electroless platers that the state of the substrate has a great influence not only on the plating process, but also on the coating properties [28,29]: surface roughness can for example affect the coating appearance. Although this may appear to be the most trivial of the problems caused by surface roughness, it is often important in terms of potential lost revenue [30]. It is thus clear that controlling surface roughness is important in terms of quality and functionality. The substrate roughness is a parameter that is sometimes difficult to control in an industrial process: most parts are supplied by the client without particular surface preparation. They also often have complicated geometries (because the plating of this kind of parts is a major application of electroless plating). For this reason, mechanical preparation (such as grinding) may be difficult to implement. As such, it is of great interest for the final application of the process to gain the most extensive knowledge of the influence of the substrate surface condition on the initiation. The aims of this work were to observe and describe the initiation and formation of electroless nickel–boron deposits on mild steel, to relate the properties of the coating to this formation and to evaluate the effects of substrate roughness on the formation of the deposit. To study the influence of roughness on the initiation, samples were submitted to different mechanical preparation processes, using varying grades of SiC grinding paper. 2. Experimental details 2.1. Sample preparation The morphological and chemical evolution of the deposited material during the plating was investigated on the same substrate (St 37 steel). Samples (steel sheets with a thickness of 1 mm) were prepared for deposition by mechanical grinding, acetone degreasing, acid etching (activation) in 30% hydrochloric acid and deionised water rinse. They were then immersed in the electroless nickel–boron bath. The bath was based on nickel chloride and sodium borohydride, with lead tungstate as a stabilising agent. More details on the bath composition have been given by Delaunois et al. [31,32]. The same bath composition was used for the whole process. The samples were immersed in the plating solution from 5 s to 60 min (Table 1). After immersion in the electroless bath, the samples were rinsed with deionised water and then dried in hot air. A specific experiment, using a bath without reducing agent, was designed for the identification of the initiation mechanism. To study the influence of roughness on the initiation, mild steel (St 37) samples were submitted to different mechanical preparation processes (with varying grades of SiC grinding paper) before the chemical step of the surface preparation (Table 2).
Table 2 Mechanical preparation processes for substrate roughness influence on deposit initiation. Polishing process description NP P1 P2 P3
Unpolished substrate (as received) Polished with grade 220 SiC paper Polished with grade 1200 SiC paper Polished with grade 4000 SiC paper
2.2. Deposit analysis To assess the nucleation sites and the initiation mechanism, the progressive spreading of the deposit on the surface, and other parameters linked to the first stages of the deposition process, the surface was investigated by several analytical techniques, such as SEM, EDX analyses and roughness measurement. The roughness of the samples was measured by the mechanical stylus method, using a Zeiss Surfcom 1400D-3DF apparatus. A Jeol JSM 5900 LV scanning electron microscopy (SEM) apparatus was used to characterize the structure and superficial morphology of the coatings. Cross sections were ground to mirror polish using a diamond paste (1/4 m) and then etched using 10% nital before SEM observation. During the SEM experiment, the average nickel and iron content of the surface was measured by EDX analysis. This allowed us to obtain an approximation of the nickel coverage ratio of the surface and subsurface (EDX analysis is not limited to the surface of the sample but gives an average composition for the upper part of the sample). EDX analysis was used to determine a ‘nickel coverage ratio’ of the coating. To obtain consistent results, a surface of 1 mm2 was analyzed at a magnification of 100 times, during 60 s. The results of this analysis represent the composition measured on the surface of the sample. However, this does not represent the uppermost surface because the penetration depth of the EDX analysis is of the order of 1 m. Moreover, as boron is not detected by this technique, the nickel/iron ratio cannot be used as a quantitative tool. Nevertheless, this analysis allows detecting the complete absence of nickel on the surface and gives qualitative indications about the formation of a continuous coating. Profile composition of the coating was determined by GDOES using a Horiba-Jobin-Yvon GD-Profiler 2 apparatus. 3. Results and discussion 3.1. Initiation and formation of the deposit on polished mild steel substrates The first step of the study was the observation of the formation and morphological evolution of the nickel–boron deposit on substrates submitted to the standard surface preparation method (that was used for most of our studies: mechanical polishing of all surfaces up to the grade 4000 SiC paper [33]). All the samples studied in Sections 3.1 and 3.2 were synthesized in a complete deposition bath containing the reducing agent, as opposed to the bath used in Section 3.3. 3.1.1. Surface observation Immersion in the bath up to 15 s brought no evidence of nickel deposition detectable by SEM observation. This period can be considered, as far as the microscopic observation is concerned, as an induction period. The first nodules of electroless deposit are observed after an immersion of 30 s and are preferentially concentrated on the scratches and defects of the surface. As can be seen on Fig. 1a, they do not form a continuous layer yet and their size is in the range of 0.1–0.2 m. After 60 s, the islands have colonized the whole sample surface but they do not form a continuous layer yet (Fig. 1b). At very high magnification, it is possible to see that, on some places of the samples, several layers of nodules are present
268
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
Fig. 1. Surface of polished mild steel (St 37) samples after an immersion of 30 s (a) and 60 s (b and c).
Fig. 2. Surface of polished mild steel (St 37) samples after an immersion of 90 s (a), 4 min (240 s) (b), 7 min (c), 10 min (d), 30 min (e) and 1 h (f) in the nickel–boron deposition bath.
(Fig. 1c). The nodule size is close to 0.5 m. After an immersion of 90 s (Fig. 2a), a thickening of the nodules is observed and the surface seems to be levelled. This phenomenon is still observed after 4 min of deposition (Fig. 2b). On the sample that was plated for
7 min (Fig. 2c), the beginning of a refining of the surface texture is observed. This refining is achieved after 10 min of plating (Fig. 2d). After a deposition time of 30 min, a slight thickening of the columns and levelling of the coating are observed and the intercolumnar
Fig. 3. Cross section observation of polished mild steel (St 37) samples after an immersion of 90 s (a), 4 min (240 s) (b), 7 min (c), 10 min (d), 30 min (e) and 1 h (f) in the nickel–boron deposition bath.
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
269
Table 3 Results of EDX analysis on ‘polished’ mil steel.
Fig. 4. Evolution of deposit thickness with immersion time (on polished St 37 steel).
interstices are very pronounced. After 1 h, the interstices seem to be nearly filled out and the apparent size of the cells is very small. 3.1.2. Cross section observation and thickness For technical reasons, there are no observable cross sections for samples immersed in the bath for less than 90 s. After an immersion of 90 s, a very thin and continuous layer of nickel is observed near the interface. Over this layer, nodules of various sizes are growing (Fig. 3a). This suggests that some nickel is deposited before the formation of nodules and columns during the so-called induction period. After 4 min, the nodules have colonized the whole surface and begun to coalesce. The coating is now about 1 m thick and the porosities are beginning to close (Fig. 3b). After an immersion of 7 min, the nodules are turning into columns and a secondary germination induces the refining of the column size (Fig. 3c). The structure of the columns has a nearly fractal quality: the initial column divides into several smaller columns that can be further divided during the next stages of deposition but becomes less marked after a long deposition time because of the progressive densification of the coating (Fig. 3d and e). After an immersion of 1 h, several layers of columns, generated by the alternating germination and growth phases, are observed (Fig. 3f). Thickness of the deposits was measured, when possible, on SEM micrographs, as shown in Fig. 4. At the beginning of the process (up to 4 min), the thickness increases very quickly (with a growth rate exceeding 40 m min−1 ). However, SEM observation shows that the deposit is far from being dense. Between 4 and 7 min of plating, the thickness does not increase because the coating is progressively densifying. After this, a quasi linear thickness increase is observed for up to 1 h of immersion. The mean growth rate is 18 mm h−1 . 3.1.3. EDX analysis EDX analysis carried out on a 1-mm2 area of the coating was used to follow the beginning of the deposition process. EDX results are not representative of either the uppermost surface of the coating or the average composition because EDX has an interaction depth of the order of 1 m in the conditions we used. This technique gives only information about the presence of nickel on the surface (i.e. the presence of boron was not detected). However, as it is difficult to obtain thickness measurements for electroless coatings in the early stages of deposition, the evolution of the amount of nickel detected by EDX can be used as a qualitative indicator to detect the initiation and follow deposition rate. The EDX results are summarized in Table 3. After an immersion of 15 s, a non-negligible quantity of nickel is already measured on the surface. The induction period is thus shorter when measured in terms of nickel content on the surface
Time (s)
5
15
30
60
90
240
Fe (at.%) Ni (at.%)
100 0
94.3 5.7
81 19
42 58
26 74
7 93
than when measured by microscopic observation. It can be due to the deposition of very small nuclei on the substrate before the formation of the first nickel–boron nodules. This hypothesis is reinforced by the presence of a very thin continuous layer of nickel under the nodules (Fig. 3a). After an immersion of 30 s, the amount of nickel detected on the surface and subsurface is close to 20%, indicating the presence of a non-negligible amount of nickel outside of the nodules detected by the SEM. This is attributed to a selective growth phenomenon. This phenomenon may reflect heterogeneities in nucleation and growth rate but it is possible that it may be caused by growth inhibition on some of the nuclei that were deposited during the very first moments of the process, by lead adsorption, causing a slower growth rate and a growth that is characterized by an unmodified surface topology. On other places, the growth is not inhibited and the formation of nodules can be observed during SEM analysis. After 1 min, the amount of Ni detected on the subsurface is close to 60% meaning that a nearly continuous layer of nickel is formed. After 4 min, a continuous layer is formed, as indicated by the SEM observations. 3.1.4. Chemical heterogeneities The profile composition of the coating was measured by glowdischarge optical emission spectroscopy (GD-OES). The nickel and boron contents of the coating are relatively stable across the whole deposit. However, the evolution of the lead content is really different: it is higher near the substrate/coating interface and close to the sample surface than in the bulk of the coating. Relations between the local lead content of the coating and the morphology were established by superimposing the profile chemical analysis of a coating with its SEM cross section image, as shown in Fig. 5. The lead content of the coating is close to 0.45 wt.% at the substrate/coating interface then decreases quickly down to 0.3% in the first 2–3 m of the coating. The content stays then low (0.25–0.3 wt.%) for the next 10 m of the deposit before going up near the surface (at the end of the deposition process). The lowest lead concentration of the coating can be matched with the stage of quick growth observed after 7 min of deposition, while the high concentration detected in the coating that is first deposited can
Fig. 5. Comparison of the composition and morphology across the deposit.
270
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
Table 4 Substrate roughness after mechanical preparation processes. Samples
Ra
NP P1 P2 P3
0.674 0.559 0.294 0.476
Rp ± ± ± ±
0.111 0.144 0.084 0.120
1.896 2.122 1.103 1.759
± ± ± ±
0.502 0.467 0.515 0.596
be linked with the plateau in growth due to densification of the coating. This suggests an explanation for the plateau phenomenon: during the first 90–250 s of deposition, the deposition rate is high but the material that is deposited is not very dense and possesses thus a very high specific surface. At this moment, an important amount of lead is adsorbed on this surface and slows significantly the deposition process until the majority of the cavities formed by the initial deposit are filled. The growth continues then at a more controlled rate because the process enters a steady state. At the end of the experiment, a lowering of the deposition rate is observed, that can likewise be linked to the higher lead content observed near the surface of the coating. Comparison of the lead content and morphology shows that the lower lead content observed in the center of the coating coincides with the presence of wider columns, while the columns appear to be smaller at the very beginning and the very end of the process, when the lead content is higher, which is in good agreement with the SEM observation of the surface of the samples. 3.2. Influence of the substrate roughness The influence of roughness on the initiation was investigated by following three same coating properties than on polished steel: the morphology, the coverage (as measured by EDX analysis), and the roughness. These were carried out on substrates submitted to different mechanical preparation processes, as described in Table 1. Roughness measurement carried out just after the surface preparation (Table 4) shows that grinding with grade 220 SiC abrasive paper (P1 ) induces a slight decrease of Ra with a non-negligible
increase of Rp . Here follow the results of ordering the samples from the rougher to the smoother according to Ra : NP, P1 , P3 , P2 ; and according to Rp : P1 , NP , P3 , P2 . The P3 sample, which has been ground only with 4000 grade paper, is rougher than the sample ground with grade 1200 paper (P2 ) because the finer grained paper, when used without pregrinding, provides a less effective surface preparation. The P1 sample has a higher Rp roughness than the NP sample but this difference is not statistically significant because the standard deviation for both values (close to 0.5) is higher than the difference between them. The parameters chosen to represent the roughness in this study are Ra (arithmetic average of the height of every point of the surface) and Rp (maximum peak height). Ra was chosen because it is the most in use roughness parameter. However, as the electroless process takes place in solution, we feel that the peaks will have a more important contribution to the deposit initiation and growth than the valleys, thus we chose to use Rp conjointly with Ra . After a plating time of 5 s, the increase of substrate roughness favors the deposit initiation, as shown in Fig. 6: the density of nickel nodules increases with increasing roughness of the substrate. Moreover, the size of the nodules present on the surface appears smaller when the roughness is higher, with a nodule size of the order of 50 nm on the unpolished surface. This shows that the nucleation of the nickel nodules on the substrate is easier when the substrate is rougher. The higher density of nodules on the rougher substrate is in accordance with the findings of Liu and Gao [29] but the formation of bigger nodules on the rougher substrate was not observed in this case. It is not surprising as the Ni–P does not form nodules on the Mg alloy substrate but small cubic crystals, which denotes a different initiation mechanism. Moreover, low magnification micrographs on the samples give the impression of bigger deposits on the rougher substrate because aggregates are already formed. After an immersion of 15 s in the plating bath, the density of nickel nodules is still decreasing with decreasing roughness (Fig. 6) but the size of the nodules appears homogeneous on all the samples and is close to 50 nm. After 30 s in the plating bath, the surface of
Fig. 6. SEM observation of the surface of samples with varying roughness immersed in the plating bath for 5 s, 15 s and 30 s.
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
271
Fig. 7. SEM observation of the surface of samples with varying roughness immersed in the plating bath for 560 s, 90 s and 4 min.
the samples is nearly completely colonized by the nickel deposit (Fig. 6). The size of the nodules is still similar for all the samples. After 60 s, all the samples are completely colonized (Fig. 7). However, the deposit appears more homogeneous on the sample submitted to the P1 surface treatment, which has the higher Rp . Longer immersions in the plating bath lead to homogenization and densification of the coating, as shown in Fig. 6. The size of the nodules is still in the same range as in the previous cases but they tend to form aggregates, which are clearly distinguishable on low magnification images while the nodules are only observable at high magnification. To assess the deposition rate on substrates with varying roughness, EDX analysis was used (qualitatively) in the first moments and thickness measurements were carried out by SEM on thicker coatings. The results of the EDX analysis are presented in Fig. 8. After 5 s, a similar amount of nickel is detected on the samples with NP, P1 and P2 preparations while the nickel is slightly lower on the P3 sam-
ple. This is in agreement with the SEM observation that roughness favors deposit initiation. After 15 s, the P3 sample has reached the same level as the NP, P1 and P2 samples. During the next time of immersion the P1 sample has a higher deposition rate (its nickel content is already close to 90% after 90 s) than all the other samples. The NP sample on the other hand has a faster initiation but grows slower and reaches similar value to the P2 sample after 90 s. From those results, it seems that the P1 surface treatment produces the best results, as far as deposit initiation is concerned. This is not surprising because the sample submitted to the P1 treatment has the higher Rp roughness. This sample has thus higher peaks that are preferred nucleation sites. The NP sample shows a good comportment at the very first stages but the deposit grows slower on this sample, probably because its higher roughness is due to valleys (the Rv roughness, that measures the depth of valleys, is 1.98 ± 0.4 for the NP sample and 1.75 ± 0.6 for the P1 sample) in which the deposition is slowed by the diffusion of reactive needed to attain them. We can conclude from those observations that the presence of peaks influences favorably the initiation process (the peaks are the preferred nucleation sites) but that the presence of grooves or valleys in the substrate decreases the growth rate because diffusion of the reactive inside the valleys is difficult. Thickness measurements at later stages of the deposition process (Fig. 9) showed that all samples had similar growth rates. It is however interesting to note that the end thickness decreases with the initial roughness of the substrate. 3.3. Initiation mechanism
Fig. 8. EDX analysis carried out on samples with varying substrate roughness, during the first stages of the plating process.
Based on the theoretical knowledge of the electroless process [4], one can deduct that the beginning of the process (the deposition of the very first atoms of nickel on the surface) on an un-catalyzed steel substrate can only be induced by 2 phenomena: (i) the formation of an ultra thin layer of nickel by displacement (redox reaction between nickel ions and the substrate, accompanied by dissolution of the substrate material, without intervention of the reducing
272
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
4. Conclusion The present work brings a lot of answers about the formation of electroless nickel–boron coatings which is a really complex process. However the present knowledge of the process does not permit yet to describe and to understand all the phenomena that happen during this formation. This work will thus be completed by further investigations including structural characterization of the growing coating, surface analysis by XPS of the deposits at different stages of their growth and TEM observation. A study of the influence of substrate nature on the nucleation and growth (carried out on various stainless steel grades) is also underway. The following results were obtained concerning the initial deposition of Ni–B films, which were electrolessly deposited from sodium borohydride baths. Fig. 9. Thickness measurements on samples with varying substrate roughness, during the later stages of the plating process.
agent) and (ii) oxidation of the sodium borohydride due to catalytic activity of the substrate (steel) and subsequent reduction of nickel salts that form a very thin nickel layer. Since the resulting nickel layer should be similar in both cases, simple SEM observation and EDX analysis of the coatings cannot permit the identification of this mechanism during deposition. Thus, experiments dedicated to the identification of the actual initiation mechanism taking place in our bath for mild steel (St 37) substrates were developed for this study. During these experiments, mild steel (St 37) substrates were immersed in a bath exempt of the reducing agent, every other parameter (composition, temperature, pH and agitation) kept constant. The samples were immersed in this bath for varying durations (Table 1) and were then observed by SEM. No nickel was detected on any of the samples immersed in the electroless bath without reducing agent, even after an immersion of 4 min (Fig. 10). However, lead had been detected on the surface by EDX and forms small cubic crystals that are concentrated on the surface defects i.e. polishing scratches, etc. (Fig. 10). This allows excluding the displacement reaction as initiation mechanism for the electroless deposition. The formation of lead crystals could be however attributed to a displacement reaction between iron and lead, which is not really surprising because the redox potential of lead is higher than nickel and iron (−0.47 V for Fe2+ /Fe; −0.27 V for Ni2+ /Ni; −0.13 V for Pb2+ /Pb).
• The initiation mechanism on mild steel for those baths is a catalytic effect and not a displacement reaction between nickel and iron. • The different phases of the deposition process are as follows: - A very short induction period during which some nickel is deposited on the coating but not in the form of nodules. - The formation of a very thin continuous layer. - The formation of nodules. - The densification of the coating. This phase begins after 4 min in the bath, when the nodule layer is continuous. During this phase, the thickness of the coating does not increase but the amount of deposited metal does. - Several nucleation/growth phases, each one beginning by the refinement of the columnar structure. The first re-nucleation happens between 4 and 7 min of immersion. • The columns that form the coating are smaller where the lead content of the coating is higher (near the interfaces). • Initiation occurs quicker on rougher substrates but the consecutive growth of the deposit is slower. Acknowledgments One of the authors (V. Vitry) wishes to thank the FRIA (Fonds pour la formation à la recherche dans l’industrie et l’agriculture) for funding. The authors wish to thank Mr J. Dutrieux from INISMA for his help with the SEM work and the University of Udine for the GDOES analysis. References
Fig. 10. Surface observation after 4 min of a sample immersed in a bath without reducing agent.
[1] A. Brenner, G. Riddel, J. Res. Natl. Bur. Stand. 37 (1946) 31. [2] K. Hari Krishnan, S. John, K.N. Srinivasan, J. Praveen, M. Ganesan, P.M. Kavimani, Metall. Mater. Trans. 37A (2006) 1917–1926. [3] P. Peeters, G.v.d. Hoorn, T. Daenen, A. Kurowski, G. Staikov, Electrochim. Acta 47 (2001) 161–169. [4] A. Riedel, Electroless Nickel Plating, Finishing Publication LTD., London, 1991. [5] R. Elansezhian, B. Ramamoorthy, P. Kesavan Nair, J. Mater. Process. Technol. 209 (2009) 233–240. [6] T. Watanabe, Y. Tanabe, Trans. Jpn. Inst. Met. 24 (1983) 396–404. [7] M.-A. Clerc, PhD Thesis, Besanc¸on (1986). [8] P.S. Kumar, P.K. Nair, Nanostruct. Mater. 5 (1994) 183–198. [9] A. Chiba, H. Haijima, W.C. Wu, Ultrasonics 42 (2004) 617–620. [10] A. Mondal, S. Nath, A. Mondal, S. Bandopadhyay, U. Gangopadhyay, H. Saha, Mater. Res. Bull. 39 (14–15) (2004) 2187–2192. [11] Y.W. Riddle, T.O. Bailer, JOM April (2005) 40–45. [12] K. Krishnaveni, T.S.N. Sankara Narayanan, S.K. Seshadri, Surf. Coat. Technol. 190 (2005) 115–121. [13] Z. Shi, D. Wang, Z. Ding, Appl. Surf. Sci. 221 (2004) 32–68. [14] T.S.N. Sankara Narayanan, S.K. Seshadri, J. Alloys Compd. 365 (2004) 197–205. [15] B.J. Hwang, S.H. Lin, J. Electrochem. Soc. 142 (1995) 3749. [16] G.O. Mallory, J.B. Hajdu, Electroless Plating: Fundamentals and Applications, American Electroplaters and Surface Finishers Society, Orlando, FL, 1990, pp. 1–56. [17] L. Li, M. An, J. Alloys Compd. 461 (2008) 85–91.
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273 [18] S. Karmalkar, V.P. Kumar, J. Electrochem. Soc. 151 (2004) C554–C558. [19] H. Jha, T. Kikuchi, M. Sakairi, H. Takahashi, Mater. Lett. 63 (2009) 1451–1454. [20] A. Duhin, Y. Sverdlov, Y. Feldman, Y. Shacham-Diamand, Electrochim. Acta 54 (25) (2009) 6036–6041. [21] L. Li, M. An, G. Wu, Appl. Surf. Sci. 252 (2005) 959–965. [22] T. Homma, M. Tanabe, K. Itakura, T. Osaka, J. Electrochem. Soc. 144 (1997) 4123–4127. [23] Q.-L. Rao, G. Bi, Q.-H. Lu, H.-W. Wang, X.-L. Fan, Appl. Surf. Sci. 240 (2005) 28–33. [24] H. Matsubara, T. Yonekawa, Y. Ishino, H. Nishiyama, N. Saito, Y. Inoue, Electrochim. Acta 47 (2002) 4011–4018. [25] H. Matsubara, T. Yonekawa, Y. Ishino, N. Saito, H. Nishiyama, Y. Inoue, Electrochim. Acta 52 (2) (2006) 402–407.
273
[26] Z. Liu, W. Gao, Surf. Coat. Technol. 200 (2006) 3553–3560. [27] H. Matsubara, Y. Abe, Y. Chiba, H. Nishiyama, N. Saito, K. Hodouchi, Y. Inoue, Electrochim. Acta 52 (9) (2007) 3047–3052. [28] C.K. Lee, Surf. Coat. Technol. 202 (2008) 4868–4874. [29] Z. Liu, W. Gao, Surf. Coat. Technol. 200 (2006) 5087–5093. [30] R.S. William, W.C. Feist, J. Coat. Technol. 66 (1994) 109–121. [31] F. Delaunois, P. Lienard, Surf. Coat. Technol. 160 (2002) 139–148. [32] F. Delaunois, J.P. Petitjean, P. Lienard, M. Jacob-Duliere, Surf. Coat. Technol. 124 (2000) 201–209. [33] A.-F. Kanta, V. Vitry, F. Delaunois, J. Alloys Compd. 486 (2009) L21–L23.
Contents lists available at ScienceDirect
Materials Science and Engineering B journal homepage: www.elsevier.com/locate/mseb
Initiation and formation of electroless nickel–boron coatings on mild steel: Effect of substrate roughness V. Vitry a,∗ , A.-F. Kanta b , F. Delaunois a a b
Université de Mons, Service de Métallurgie, Place du Parc, 20, 7000 Mons, Belgium Université de Mons, Service de Science des Matériaux, Place du Parc, 20, 7000 Mons, Belgium
a r t i c l e
i n f o
Article history: Received 11 March 2010 Received in revised form 6 July 2010 Accepted 28 August 2010 Keywords: Film deposition Metals Surface morphology Surface roughnesses Electroless nickel–boron
a b s t r a c t The initial deposition and growth of electroless nickel–boron deposits on mild steel were studied: the films were prepared in an electroless plating bath using sodium borohydride as reducing agent. Samples were immersed in the plating solution for times from 5 s to 1 h and the morphological evolution of the deposit was followed by scanning electron microscopy (SEM) observation of the surface and prepared cross sections. Energy dispersive X-ray spectrometry (EDX) and glow discharge optical electro spectroscopy (GDOES) analyses were used to obtain information about the chemistry of the deposits and their results were correlated with the morphology of the coating. The initiation mechanism of electroless deposition on mild steel was identified. The effects of substrate roughness variation on the morphology and growth rate of the coatings were investigated by reproducing the experiment on samples with various surface preparation (grinding) states. We observed that the increase of substrate roughness favors the deposit initiation: the density of nickel nodules increases with increasing roughness of the substrate. Longer immersions in the bath lead to homogenization and densification of the coating and the nodules are clearly distinguishable. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Electroless coating is a well-established surface engineering process that was developed by Brenner and Riddel in 1946 [1]. It involves deposition of a metal–metalloid alloy coating on various substrates (including dielectric materials) by electrochemical reactions in aqueous solution. Electroless nickel deposits possess a number of interesting properties such as uniform thickness, high hardness, good corrosion resistance, etc. [2–5]. Electroless nickel deposits are usually classified according to the nature of the reducing agent. Nickel–phosphorous deposits (based on reduction by the hypophosphite ion) are the most studied and used but the properties of nickel–boron deposits are of very great interest for several industrial applications like aeronautics, petrochemical industry, food industry, firearms, etc.: their hardness is higher than nickel–phosphorous, and their electrical and tribological properties are very promising [6–14]. Most of the papers focused on electroless nickel and particularly on nickel–boron describe the optimization of plating parameters, the coating properties or the effect of heat treatment. There are very few papers dealing with deposit formation [7,8].
∗ Corresponding author. Tel.: +32 65 37 44 38; fax: +32 65 37 44 36. E-mail address: [email protected] (V. Vitry). 0921-5107/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.mseb.2010.08.003
Electroless plating occurs by the reduction of nickel ions at the surface of the active substrate immersed into the plating solution and further growth of the coating is due to a catalytic action of the deposit itself [15,16]. There are 3 known ways of inducing the initiation of electroless nickel deposition. Either the substrate is spontaneously catalytic for the oxidation of the reducing agent, or it is more easily oxidizable than nickel, in which case a thin catalytic layer of nickel is spontaneously deposited, or finally it can be activated by dipping in a solution of catalytically active metal salts (like Pd) or by galvanic coupling [4]. The few studies dedicated to the initiation mechanism of electroless deposition focused on systems in which catalysts, such as Pd are used [17–22]. This means that if no catalyst is used, the actual mechanism that allows initial deposition on a determined substrate is rarely known and is at best proposed by assumptions. The initiation and the growth of nickel–boron coatings have a non-negligible influence on their properties. In the case of experimental, unreplenished baths, the structure and the composition of the coating are not homogeneous over the coating thickness but change during the deposition, as was shown in the extreme case of a non-agitated bath by Rao et al. [23]. However, the observation of the formation of nickel–boron deposits has not been studied extensively yet, opposite to the growth of nickel–phosphorous coatings that has been studied on various substrates with and without catalytic activation [17–22,24–27].
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
267
Table 1 Deposition time for the observation of the initiation of electroless deposits. Sample
1
2
3
4
5
6
7
8
9
10
Deposition time
5s
15 s
30 s
60 s
90 s
4 min (240 s)
7 min (420 s)
10 min (600 s)
30 min (1800 s)
60 min (3600 s)
It is also well known by the electroless platers that the state of the substrate has a great influence not only on the plating process, but also on the coating properties [28,29]: surface roughness can for example affect the coating appearance. Although this may appear to be the most trivial of the problems caused by surface roughness, it is often important in terms of potential lost revenue [30]. It is thus clear that controlling surface roughness is important in terms of quality and functionality. The substrate roughness is a parameter that is sometimes difficult to control in an industrial process: most parts are supplied by the client without particular surface preparation. They also often have complicated geometries (because the plating of this kind of parts is a major application of electroless plating). For this reason, mechanical preparation (such as grinding) may be difficult to implement. As such, it is of great interest for the final application of the process to gain the most extensive knowledge of the influence of the substrate surface condition on the initiation. The aims of this work were to observe and describe the initiation and formation of electroless nickel–boron deposits on mild steel, to relate the properties of the coating to this formation and to evaluate the effects of substrate roughness on the formation of the deposit. To study the influence of roughness on the initiation, samples were submitted to different mechanical preparation processes, using varying grades of SiC grinding paper. 2. Experimental details 2.1. Sample preparation The morphological and chemical evolution of the deposited material during the plating was investigated on the same substrate (St 37 steel). Samples (steel sheets with a thickness of 1 mm) were prepared for deposition by mechanical grinding, acetone degreasing, acid etching (activation) in 30% hydrochloric acid and deionised water rinse. They were then immersed in the electroless nickel–boron bath. The bath was based on nickel chloride and sodium borohydride, with lead tungstate as a stabilising agent. More details on the bath composition have been given by Delaunois et al. [31,32]. The same bath composition was used for the whole process. The samples were immersed in the plating solution from 5 s to 60 min (Table 1). After immersion in the electroless bath, the samples were rinsed with deionised water and then dried in hot air. A specific experiment, using a bath without reducing agent, was designed for the identification of the initiation mechanism. To study the influence of roughness on the initiation, mild steel (St 37) samples were submitted to different mechanical preparation processes (with varying grades of SiC grinding paper) before the chemical step of the surface preparation (Table 2).
Table 2 Mechanical preparation processes for substrate roughness influence on deposit initiation. Polishing process description NP P1 P2 P3
Unpolished substrate (as received) Polished with grade 220 SiC paper Polished with grade 1200 SiC paper Polished with grade 4000 SiC paper
2.2. Deposit analysis To assess the nucleation sites and the initiation mechanism, the progressive spreading of the deposit on the surface, and other parameters linked to the first stages of the deposition process, the surface was investigated by several analytical techniques, such as SEM, EDX analyses and roughness measurement. The roughness of the samples was measured by the mechanical stylus method, using a Zeiss Surfcom 1400D-3DF apparatus. A Jeol JSM 5900 LV scanning electron microscopy (SEM) apparatus was used to characterize the structure and superficial morphology of the coatings. Cross sections were ground to mirror polish using a diamond paste (1/4 m) and then etched using 10% nital before SEM observation. During the SEM experiment, the average nickel and iron content of the surface was measured by EDX analysis. This allowed us to obtain an approximation of the nickel coverage ratio of the surface and subsurface (EDX analysis is not limited to the surface of the sample but gives an average composition for the upper part of the sample). EDX analysis was used to determine a ‘nickel coverage ratio’ of the coating. To obtain consistent results, a surface of 1 mm2 was analyzed at a magnification of 100 times, during 60 s. The results of this analysis represent the composition measured on the surface of the sample. However, this does not represent the uppermost surface because the penetration depth of the EDX analysis is of the order of 1 m. Moreover, as boron is not detected by this technique, the nickel/iron ratio cannot be used as a quantitative tool. Nevertheless, this analysis allows detecting the complete absence of nickel on the surface and gives qualitative indications about the formation of a continuous coating. Profile composition of the coating was determined by GDOES using a Horiba-Jobin-Yvon GD-Profiler 2 apparatus. 3. Results and discussion 3.1. Initiation and formation of the deposit on polished mild steel substrates The first step of the study was the observation of the formation and morphological evolution of the nickel–boron deposit on substrates submitted to the standard surface preparation method (that was used for most of our studies: mechanical polishing of all surfaces up to the grade 4000 SiC paper [33]). All the samples studied in Sections 3.1 and 3.2 were synthesized in a complete deposition bath containing the reducing agent, as opposed to the bath used in Section 3.3. 3.1.1. Surface observation Immersion in the bath up to 15 s brought no evidence of nickel deposition detectable by SEM observation. This period can be considered, as far as the microscopic observation is concerned, as an induction period. The first nodules of electroless deposit are observed after an immersion of 30 s and are preferentially concentrated on the scratches and defects of the surface. As can be seen on Fig. 1a, they do not form a continuous layer yet and their size is in the range of 0.1–0.2 m. After 60 s, the islands have colonized the whole sample surface but they do not form a continuous layer yet (Fig. 1b). At very high magnification, it is possible to see that, on some places of the samples, several layers of nodules are present
268
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
Fig. 1. Surface of polished mild steel (St 37) samples after an immersion of 30 s (a) and 60 s (b and c).
Fig. 2. Surface of polished mild steel (St 37) samples after an immersion of 90 s (a), 4 min (240 s) (b), 7 min (c), 10 min (d), 30 min (e) and 1 h (f) in the nickel–boron deposition bath.
(Fig. 1c). The nodule size is close to 0.5 m. After an immersion of 90 s (Fig. 2a), a thickening of the nodules is observed and the surface seems to be levelled. This phenomenon is still observed after 4 min of deposition (Fig. 2b). On the sample that was plated for
7 min (Fig. 2c), the beginning of a refining of the surface texture is observed. This refining is achieved after 10 min of plating (Fig. 2d). After a deposition time of 30 min, a slight thickening of the columns and levelling of the coating are observed and the intercolumnar
Fig. 3. Cross section observation of polished mild steel (St 37) samples after an immersion of 90 s (a), 4 min (240 s) (b), 7 min (c), 10 min (d), 30 min (e) and 1 h (f) in the nickel–boron deposition bath.
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
269
Table 3 Results of EDX analysis on ‘polished’ mil steel.
Fig. 4. Evolution of deposit thickness with immersion time (on polished St 37 steel).
interstices are very pronounced. After 1 h, the interstices seem to be nearly filled out and the apparent size of the cells is very small. 3.1.2. Cross section observation and thickness For technical reasons, there are no observable cross sections for samples immersed in the bath for less than 90 s. After an immersion of 90 s, a very thin and continuous layer of nickel is observed near the interface. Over this layer, nodules of various sizes are growing (Fig. 3a). This suggests that some nickel is deposited before the formation of nodules and columns during the so-called induction period. After 4 min, the nodules have colonized the whole surface and begun to coalesce. The coating is now about 1 m thick and the porosities are beginning to close (Fig. 3b). After an immersion of 7 min, the nodules are turning into columns and a secondary germination induces the refining of the column size (Fig. 3c). The structure of the columns has a nearly fractal quality: the initial column divides into several smaller columns that can be further divided during the next stages of deposition but becomes less marked after a long deposition time because of the progressive densification of the coating (Fig. 3d and e). After an immersion of 1 h, several layers of columns, generated by the alternating germination and growth phases, are observed (Fig. 3f). Thickness of the deposits was measured, when possible, on SEM micrographs, as shown in Fig. 4. At the beginning of the process (up to 4 min), the thickness increases very quickly (with a growth rate exceeding 40 m min−1 ). However, SEM observation shows that the deposit is far from being dense. Between 4 and 7 min of plating, the thickness does not increase because the coating is progressively densifying. After this, a quasi linear thickness increase is observed for up to 1 h of immersion. The mean growth rate is 18 mm h−1 . 3.1.3. EDX analysis EDX analysis carried out on a 1-mm2 area of the coating was used to follow the beginning of the deposition process. EDX results are not representative of either the uppermost surface of the coating or the average composition because EDX has an interaction depth of the order of 1 m in the conditions we used. This technique gives only information about the presence of nickel on the surface (i.e. the presence of boron was not detected). However, as it is difficult to obtain thickness measurements for electroless coatings in the early stages of deposition, the evolution of the amount of nickel detected by EDX can be used as a qualitative indicator to detect the initiation and follow deposition rate. The EDX results are summarized in Table 3. After an immersion of 15 s, a non-negligible quantity of nickel is already measured on the surface. The induction period is thus shorter when measured in terms of nickel content on the surface
Time (s)
5
15
30
60
90
240
Fe (at.%) Ni (at.%)
100 0
94.3 5.7
81 19
42 58
26 74
7 93
than when measured by microscopic observation. It can be due to the deposition of very small nuclei on the substrate before the formation of the first nickel–boron nodules. This hypothesis is reinforced by the presence of a very thin continuous layer of nickel under the nodules (Fig. 3a). After an immersion of 30 s, the amount of nickel detected on the surface and subsurface is close to 20%, indicating the presence of a non-negligible amount of nickel outside of the nodules detected by the SEM. This is attributed to a selective growth phenomenon. This phenomenon may reflect heterogeneities in nucleation and growth rate but it is possible that it may be caused by growth inhibition on some of the nuclei that were deposited during the very first moments of the process, by lead adsorption, causing a slower growth rate and a growth that is characterized by an unmodified surface topology. On other places, the growth is not inhibited and the formation of nodules can be observed during SEM analysis. After 1 min, the amount of Ni detected on the subsurface is close to 60% meaning that a nearly continuous layer of nickel is formed. After 4 min, a continuous layer is formed, as indicated by the SEM observations. 3.1.4. Chemical heterogeneities The profile composition of the coating was measured by glowdischarge optical emission spectroscopy (GD-OES). The nickel and boron contents of the coating are relatively stable across the whole deposit. However, the evolution of the lead content is really different: it is higher near the substrate/coating interface and close to the sample surface than in the bulk of the coating. Relations between the local lead content of the coating and the morphology were established by superimposing the profile chemical analysis of a coating with its SEM cross section image, as shown in Fig. 5. The lead content of the coating is close to 0.45 wt.% at the substrate/coating interface then decreases quickly down to 0.3% in the first 2–3 m of the coating. The content stays then low (0.25–0.3 wt.%) for the next 10 m of the deposit before going up near the surface (at the end of the deposition process). The lowest lead concentration of the coating can be matched with the stage of quick growth observed after 7 min of deposition, while the high concentration detected in the coating that is first deposited can
Fig. 5. Comparison of the composition and morphology across the deposit.
270
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
Table 4 Substrate roughness after mechanical preparation processes. Samples
Ra
NP P1 P2 P3
0.674 0.559 0.294 0.476
Rp ± ± ± ±
0.111 0.144 0.084 0.120
1.896 2.122 1.103 1.759
± ± ± ±
0.502 0.467 0.515 0.596
be linked with the plateau in growth due to densification of the coating. This suggests an explanation for the plateau phenomenon: during the first 90–250 s of deposition, the deposition rate is high but the material that is deposited is not very dense and possesses thus a very high specific surface. At this moment, an important amount of lead is adsorbed on this surface and slows significantly the deposition process until the majority of the cavities formed by the initial deposit are filled. The growth continues then at a more controlled rate because the process enters a steady state. At the end of the experiment, a lowering of the deposition rate is observed, that can likewise be linked to the higher lead content observed near the surface of the coating. Comparison of the lead content and morphology shows that the lower lead content observed in the center of the coating coincides with the presence of wider columns, while the columns appear to be smaller at the very beginning and the very end of the process, when the lead content is higher, which is in good agreement with the SEM observation of the surface of the samples. 3.2. Influence of the substrate roughness The influence of roughness on the initiation was investigated by following three same coating properties than on polished steel: the morphology, the coverage (as measured by EDX analysis), and the roughness. These were carried out on substrates submitted to different mechanical preparation processes, as described in Table 1. Roughness measurement carried out just after the surface preparation (Table 4) shows that grinding with grade 220 SiC abrasive paper (P1 ) induces a slight decrease of Ra with a non-negligible
increase of Rp . Here follow the results of ordering the samples from the rougher to the smoother according to Ra : NP, P1 , P3 , P2 ; and according to Rp : P1 , NP , P3 , P2 . The P3 sample, which has been ground only with 4000 grade paper, is rougher than the sample ground with grade 1200 paper (P2 ) because the finer grained paper, when used without pregrinding, provides a less effective surface preparation. The P1 sample has a higher Rp roughness than the NP sample but this difference is not statistically significant because the standard deviation for both values (close to 0.5) is higher than the difference between them. The parameters chosen to represent the roughness in this study are Ra (arithmetic average of the height of every point of the surface) and Rp (maximum peak height). Ra was chosen because it is the most in use roughness parameter. However, as the electroless process takes place in solution, we feel that the peaks will have a more important contribution to the deposit initiation and growth than the valleys, thus we chose to use Rp conjointly with Ra . After a plating time of 5 s, the increase of substrate roughness favors the deposit initiation, as shown in Fig. 6: the density of nickel nodules increases with increasing roughness of the substrate. Moreover, the size of the nodules present on the surface appears smaller when the roughness is higher, with a nodule size of the order of 50 nm on the unpolished surface. This shows that the nucleation of the nickel nodules on the substrate is easier when the substrate is rougher. The higher density of nodules on the rougher substrate is in accordance with the findings of Liu and Gao [29] but the formation of bigger nodules on the rougher substrate was not observed in this case. It is not surprising as the Ni–P does not form nodules on the Mg alloy substrate but small cubic crystals, which denotes a different initiation mechanism. Moreover, low magnification micrographs on the samples give the impression of bigger deposits on the rougher substrate because aggregates are already formed. After an immersion of 15 s in the plating bath, the density of nickel nodules is still decreasing with decreasing roughness (Fig. 6) but the size of the nodules appears homogeneous on all the samples and is close to 50 nm. After 30 s in the plating bath, the surface of
Fig. 6. SEM observation of the surface of samples with varying roughness immersed in the plating bath for 5 s, 15 s and 30 s.
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
271
Fig. 7. SEM observation of the surface of samples with varying roughness immersed in the plating bath for 560 s, 90 s and 4 min.
the samples is nearly completely colonized by the nickel deposit (Fig. 6). The size of the nodules is still similar for all the samples. After 60 s, all the samples are completely colonized (Fig. 7). However, the deposit appears more homogeneous on the sample submitted to the P1 surface treatment, which has the higher Rp . Longer immersions in the plating bath lead to homogenization and densification of the coating, as shown in Fig. 6. The size of the nodules is still in the same range as in the previous cases but they tend to form aggregates, which are clearly distinguishable on low magnification images while the nodules are only observable at high magnification. To assess the deposition rate on substrates with varying roughness, EDX analysis was used (qualitatively) in the first moments and thickness measurements were carried out by SEM on thicker coatings. The results of the EDX analysis are presented in Fig. 8. After 5 s, a similar amount of nickel is detected on the samples with NP, P1 and P2 preparations while the nickel is slightly lower on the P3 sam-
ple. This is in agreement with the SEM observation that roughness favors deposit initiation. After 15 s, the P3 sample has reached the same level as the NP, P1 and P2 samples. During the next time of immersion the P1 sample has a higher deposition rate (its nickel content is already close to 90% after 90 s) than all the other samples. The NP sample on the other hand has a faster initiation but grows slower and reaches similar value to the P2 sample after 90 s. From those results, it seems that the P1 surface treatment produces the best results, as far as deposit initiation is concerned. This is not surprising because the sample submitted to the P1 treatment has the higher Rp roughness. This sample has thus higher peaks that are preferred nucleation sites. The NP sample shows a good comportment at the very first stages but the deposit grows slower on this sample, probably because its higher roughness is due to valleys (the Rv roughness, that measures the depth of valleys, is 1.98 ± 0.4 for the NP sample and 1.75 ± 0.6 for the P1 sample) in which the deposition is slowed by the diffusion of reactive needed to attain them. We can conclude from those observations that the presence of peaks influences favorably the initiation process (the peaks are the preferred nucleation sites) but that the presence of grooves or valleys in the substrate decreases the growth rate because diffusion of the reactive inside the valleys is difficult. Thickness measurements at later stages of the deposition process (Fig. 9) showed that all samples had similar growth rates. It is however interesting to note that the end thickness decreases with the initial roughness of the substrate. 3.3. Initiation mechanism
Fig. 8. EDX analysis carried out on samples with varying substrate roughness, during the first stages of the plating process.
Based on the theoretical knowledge of the electroless process [4], one can deduct that the beginning of the process (the deposition of the very first atoms of nickel on the surface) on an un-catalyzed steel substrate can only be induced by 2 phenomena: (i) the formation of an ultra thin layer of nickel by displacement (redox reaction between nickel ions and the substrate, accompanied by dissolution of the substrate material, without intervention of the reducing
272
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273
4. Conclusion The present work brings a lot of answers about the formation of electroless nickel–boron coatings which is a really complex process. However the present knowledge of the process does not permit yet to describe and to understand all the phenomena that happen during this formation. This work will thus be completed by further investigations including structural characterization of the growing coating, surface analysis by XPS of the deposits at different stages of their growth and TEM observation. A study of the influence of substrate nature on the nucleation and growth (carried out on various stainless steel grades) is also underway. The following results were obtained concerning the initial deposition of Ni–B films, which were electrolessly deposited from sodium borohydride baths. Fig. 9. Thickness measurements on samples with varying substrate roughness, during the later stages of the plating process.
agent) and (ii) oxidation of the sodium borohydride due to catalytic activity of the substrate (steel) and subsequent reduction of nickel salts that form a very thin nickel layer. Since the resulting nickel layer should be similar in both cases, simple SEM observation and EDX analysis of the coatings cannot permit the identification of this mechanism during deposition. Thus, experiments dedicated to the identification of the actual initiation mechanism taking place in our bath for mild steel (St 37) substrates were developed for this study. During these experiments, mild steel (St 37) substrates were immersed in a bath exempt of the reducing agent, every other parameter (composition, temperature, pH and agitation) kept constant. The samples were immersed in this bath for varying durations (Table 1) and were then observed by SEM. No nickel was detected on any of the samples immersed in the electroless bath without reducing agent, even after an immersion of 4 min (Fig. 10). However, lead had been detected on the surface by EDX and forms small cubic crystals that are concentrated on the surface defects i.e. polishing scratches, etc. (Fig. 10). This allows excluding the displacement reaction as initiation mechanism for the electroless deposition. The formation of lead crystals could be however attributed to a displacement reaction between iron and lead, which is not really surprising because the redox potential of lead is higher than nickel and iron (−0.47 V for Fe2+ /Fe; −0.27 V for Ni2+ /Ni; −0.13 V for Pb2+ /Pb).
• The initiation mechanism on mild steel for those baths is a catalytic effect and not a displacement reaction between nickel and iron. • The different phases of the deposition process are as follows: - A very short induction period during which some nickel is deposited on the coating but not in the form of nodules. - The formation of a very thin continuous layer. - The formation of nodules. - The densification of the coating. This phase begins after 4 min in the bath, when the nodule layer is continuous. During this phase, the thickness of the coating does not increase but the amount of deposited metal does. - Several nucleation/growth phases, each one beginning by the refinement of the columnar structure. The first re-nucleation happens between 4 and 7 min of immersion. • The columns that form the coating are smaller where the lead content of the coating is higher (near the interfaces). • Initiation occurs quicker on rougher substrates but the consecutive growth of the deposit is slower. Acknowledgments One of the authors (V. Vitry) wishes to thank the FRIA (Fonds pour la formation à la recherche dans l’industrie et l’agriculture) for funding. The authors wish to thank Mr J. Dutrieux from INISMA for his help with the SEM work and the University of Udine for the GDOES analysis. References
Fig. 10. Surface observation after 4 min of a sample immersed in a bath without reducing agent.
[1] A. Brenner, G. Riddel, J. Res. Natl. Bur. Stand. 37 (1946) 31. [2] K. Hari Krishnan, S. John, K.N. Srinivasan, J. Praveen, M. Ganesan, P.M. Kavimani, Metall. Mater. Trans. 37A (2006) 1917–1926. [3] P. Peeters, G.v.d. Hoorn, T. Daenen, A. Kurowski, G. Staikov, Electrochim. Acta 47 (2001) 161–169. [4] A. Riedel, Electroless Nickel Plating, Finishing Publication LTD., London, 1991. [5] R. Elansezhian, B. Ramamoorthy, P. Kesavan Nair, J. Mater. Process. Technol. 209 (2009) 233–240. [6] T. Watanabe, Y. Tanabe, Trans. Jpn. Inst. Met. 24 (1983) 396–404. [7] M.-A. Clerc, PhD Thesis, Besanc¸on (1986). [8] P.S. Kumar, P.K. Nair, Nanostruct. Mater. 5 (1994) 183–198. [9] A. Chiba, H. Haijima, W.C. Wu, Ultrasonics 42 (2004) 617–620. [10] A. Mondal, S. Nath, A. Mondal, S. Bandopadhyay, U. Gangopadhyay, H. Saha, Mater. Res. Bull. 39 (14–15) (2004) 2187–2192. [11] Y.W. Riddle, T.O. Bailer, JOM April (2005) 40–45. [12] K. Krishnaveni, T.S.N. Sankara Narayanan, S.K. Seshadri, Surf. Coat. Technol. 190 (2005) 115–121. [13] Z. Shi, D. Wang, Z. Ding, Appl. Surf. Sci. 221 (2004) 32–68. [14] T.S.N. Sankara Narayanan, S.K. Seshadri, J. Alloys Compd. 365 (2004) 197–205. [15] B.J. Hwang, S.H. Lin, J. Electrochem. Soc. 142 (1995) 3749. [16] G.O. Mallory, J.B. Hajdu, Electroless Plating: Fundamentals and Applications, American Electroplaters and Surface Finishers Society, Orlando, FL, 1990, pp. 1–56. [17] L. Li, M. An, J. Alloys Compd. 461 (2008) 85–91.
V. Vitry et al. / Materials Science and Engineering B 175 (2010) 266–273 [18] S. Karmalkar, V.P. Kumar, J. Electrochem. Soc. 151 (2004) C554–C558. [19] H. Jha, T. Kikuchi, M. Sakairi, H. Takahashi, Mater. Lett. 63 (2009) 1451–1454. [20] A. Duhin, Y. Sverdlov, Y. Feldman, Y. Shacham-Diamand, Electrochim. Acta 54 (25) (2009) 6036–6041. [21] L. Li, M. An, G. Wu, Appl. Surf. Sci. 252 (2005) 959–965. [22] T. Homma, M. Tanabe, K. Itakura, T. Osaka, J. Electrochem. Soc. 144 (1997) 4123–4127. [23] Q.-L. Rao, G. Bi, Q.-H. Lu, H.-W. Wang, X.-L. Fan, Appl. Surf. Sci. 240 (2005) 28–33. [24] H. Matsubara, T. Yonekawa, Y. Ishino, H. Nishiyama, N. Saito, Y. Inoue, Electrochim. Acta 47 (2002) 4011–4018. [25] H. Matsubara, T. Yonekawa, Y. Ishino, N. Saito, H. Nishiyama, Y. Inoue, Electrochim. Acta 52 (2) (2006) 402–407.
273
[26] Z. Liu, W. Gao, Surf. Coat. Technol. 200 (2006) 3553–3560. [27] H. Matsubara, Y. Abe, Y. Chiba, H. Nishiyama, N. Saito, K. Hodouchi, Y. Inoue, Electrochim. Acta 52 (9) (2007) 3047–3052. [28] C.K. Lee, Surf. Coat. Technol. 202 (2008) 4868–4874. [29] Z. Liu, W. Gao, Surf. Coat. Technol. 200 (2006) 5087–5093. [30] R.S. William, W.C. Feist, J. Coat. Technol. 66 (1994) 109–121. [31] F. Delaunois, P. Lienard, Surf. Coat. Technol. 160 (2002) 139–148. [32] F. Delaunois, J.P. Petitjean, P. Lienard, M. Jacob-Duliere, Surf. Coat. Technol. 124 (2000) 201–209. [33] A.-F. Kanta, V. Vitry, F. Delaunois, J. Alloys Compd. 486 (2009) L21–L23.