Mechanical properties and tribological behaviour of electroless Ni-P-Cu coatings on corrosion-resistant alloys under ultrahigh contact stress with sprayed nanoparticles

Threaded components manufactured from corrosion resistant alloys (CRA’s) are vulnerable to galling. This paper develops a test matrix to systematically investigate the mechanical properties and tribological performance of electroless nickel phosphorous coatings on CRA’s when subjected to high contact stress. Samples manufactured from 28Cr stainless steel were shot-peened for various periods prior to being electo/electroless coated. The coefficient of friction (CoF) of different coating systems was evaluated via sliding cross-pin method. Various wet and dry lubricants were utilised to examine tribological performance, furthermore the adhesion strength of the coatings was investigated by a bond and pull-off method. The study has shown a significant reduction in CoF for electroless nickel phosphorous coatings with prior shot-peening treatment and sprayed nanoparticles.


Introduction
Iron-based corrosion resistance alloys (CRA's) such as 28Cr stainless steel contain high levels of chromium (Cr), nickel (Ni) and molybdenum (Mo), which provide long term resistance to corrosion for many components such as valves, tubes, vessels and heat exchangers exposed to challenging environments where high temperature/pressure combined with CO2, H2S, sulphur and chlorides [1][2][3]. In such hazardous environments other materials subject to pitting and crevice corrosion easily, for instance, carbon steels present very high corrosion rates [4,5].
High content of nickel within CRA's ensures an excellent resistance to stress-corrosion cracking (SCC), furthermore CRA's resistance to environmental corrosion is the result of the passivation of chrome forming a transparent oxide film on the surface. In general, CRA's having higher chrome content present lower corrosion rate because this passive film is selfrepairing if it is scratched or removed. In the previous publications [6,7], Craig provided guidance for selection of different types of CRA's for specific environments.
Although threaded components manufactured from CRA's exhibit excellent corrosion resistance, they have a higher galling propensity when compared with components manufactured from carbon steel. There are ways to reduce the risk of galling, e.g. (1) making the hardness difference between the contact pair a preferred range; (2) controlling the surface roughness of contact surfaces --highly polished surfaces (Ra<0.25 µm) or very rough surfaces (Ra >1.5 µm) tend to have a higher galling propensity and (3) reducing friction by selecting a suitable lubricant.
To address these problems, various coatings and lubricants have been developed, of which the electrolytic copper plating combined with API dope (wet lubricant) has been widely used on threaded connections. API dope was originally developed to form a seal in threaded connections, further use confirmed the heavy metal content worked well as an anti-galling medium particularly when used in conjunction with electrolytic copper plating. Such antigalling systems work well when used on materials with a high galling propensity. The heavy metal content within previous API dope formulations was considered to have an adverse effect on the environment which has led to the development of other none toxic lubricants for the assembly of various equipment.
Electroless nickel phosphorous (Ni-P-Cu) coating is well known for its corrosion resistance, particularly when copper is added enhancing the coatings resistance to hazardous environments, such as high concentration of sodium chloride (NaCl) or hydrochloride acid (HCl) [8,9]. The electroless deposition is an autocatalytic method without the use of an Sensitivity: Confidential external electric power source, therefore it is possible to deposit a uniform even coating on substrates with complex geometry since there is no variation in current density. This is beneficial for certain components such as threaded connections that are designed with very close tolerances.
The electroless process relies on the presence of a reducing agent (sodium hypophosphite or borohydride), which reduces the nickel ions at a relatively high temperature (e.g. 70-90 °C).
Currently, there is no agreement to explain the chemical reaction mechanisms of the electroless nickel coating [10,11], however the most accepted mechanisms are that the atomic hydrogen (Hads) is released as the result of the catalytic dehydrogenation of hypophosphite molecule adsorbed at sample's surface, while the adsorbed active hydrogen reduces nickel ion at the surface of the catalyst.
(1) Free hydrogen ions (H + ) are produced during this chemical reaction, therefore organic salts (i.e. sodium citrate, ammonia acetate) are added as a buffer to prevent the PH value from decreasing too quickly. The organic salts also act as a complexing agent, maintaining a proper amount of free nickel ions in the solution to make the deposition rate controllable.
Apart from the anti-corrosion properties, electroless nickel coating exhibits excellent tribological properties. According to a statistic [12], the primary uses of electroless nickel coatings are due to their anti-corrosion (30%) and wear resistance (25%) properties. Many tribological tests for electroless nickel coatings have been conducted, such as pin on disc [13], ball on disc [14], block on ring [15], and ring on ring [16]. In the previous reports the electroless nickel coatings were applied on either magnesium/aluminium alloys or mild steel which were benefitted by the superior corrosion resistant properties, however there seems to be lack of systematic investigation of electroless nickel deposition on CRA's. This paper is aimed to investigate the effect of surface pre-treatment, bath composition and activation methods on the adhesion and mechanical properties of electroless plated nickel phosphorous coatings on 28Cr stainless steel and then to evaluate the tribological performance under intermediate to ultrahigh contact stress in sliding with dry and wet lubricants using a laboratory galling and friction test rig.

Sample preparation
The tribological performance of electroless nickel phosphorous coating depend on many factors, such as surface pre-treatment, chemical composition of coating solution, bath temperature and lubricants used. This study developed a test matrix to include such parameters, covering different coating systems and lubricants, as shown in Table 1.
Electrolytic copper plating is also shown in the table as a benchmark. Samples were designed and manufactured into cylinders with a radius of 6 mm and length of 100 mm. The samples were treated with various surface pre-treatments, such as shot-peening, mechanical polishing (#600 silicon carbide paper), conventional cleaning, alkaline cleaning with Ultraclean SPX alkaline detergent (Ultrawave®), and nickel strike, then they were immersed into the chemical baths for electroless deposition. The electrolytic copper coating was prepared by the same procedure by replacing the electroless bath with an electrolytic bath.
Shot-peening with spherical austenitic stainless steel beads (Chronital®) was performed within a Sealey SB970 shot blasting cabinet (Sealey, UK) connected to an air compressor (8 bar). Nickel strike is a surface activation process in which a thin layer of pure nickel is deposited on the substrate surface. The nickel strike solution contained 10 wt% HCl and 300 g/L nickel chloride hexahydrate (NiCl2·6H2O), and a current density of 2.1 A/dm 2 was applied to the cylindrical test pins.
Three coating systems were chosen for this study, including electroless nickel phosphorous (Ni-P), electroless nickel phosphorous copper (Ni-P-Cu) and electrolytic copper (Cu). Table 2 gives detailed information of the bath composition of the three coatings.   [17]. In this study, a zinc rod was immersed into the bath and contacted to the substrate for the activation of the electroless deposition of Ni-P-Cu, the copper sulphate content varied between 0.5 to 2 g/L.

Mechanical and tribological testing methods
Various topographical, mechanical and tribological properties, including adhesion strength, surface roughness and micro-hardness, CoF with different lubricants, were examined to characterize the performance of the three coatings. The wet lubricants used in this study included API dope, and the dry lubricants included tungsten disulphide (WS2), polytetrafluoroethylene (PTFE), tin and lead. The WS2 and PTFE were in nanoparticle form and were sprayed on the coatings directly, whilst the tin and lead were deposited film thick onto the substrate by a rotating burnishing method.
A preliminary test was conducted to investigate the effects of copper content in the Ni-P-Cu coatings (copper sulphate concentration varied from 0.5-2 g/L) on the tribological properties, and it was found that such effects were negligible therefore only the results of specific Ni-P-Cu coating prepared from chemical bath with 0.75 g/L copper sulphate concentration were chosen to compare with the other two coatings.

3.1.Adhesion test
The adhesion strength between coating and substrate was measured by an Elcometer 508 digital pull-off adhesion tester (Elcometer, UK), and an acrylic adhesive (3M) was used as a glue bonding the dolly and coating together. This is a versatile and instant glue which can be used to bond a polymer to metal. In order to get the strongest bonding performance, the bonded parts were kept for 24 hours before carrying out the adhesion test. Hydraulic force was applied on a small cylindrical pin which went through the central hole of the dolly causing a relative movement between the coated sample and the dolly. Since the dolly was directly bonded to the substrate coating provided a convenient method for measuring adhesion of the coating with the substrate.

3.2.Surface roughness measurement
The surface profile for various peening periods was evaluated by a confocal laser scanning microscopy (Olympus LEXT OLS3000). The microscope reconstructs 3D structures from the obtained images by collecting sets of images at different depths for the evaluation of surface texture. The surface roughness was measured by a stylus type surface roughness tester (SRT6210, HUATEC, China) which records the position of a diamond probe along a straight path with approximately 4 mm of travel. Three samples were measured to evaluate the surface roughness for all samples and the average value (Ra) was calculated.

3.3.Micro-hardness test
The hardness testing was performed on a Buehler Omnimet Automaic MHT System (Buehler, UK), which is based on Vickers scale. Samples with coatings were subjected to a 10 g weight and the imprints on the coatings were evaluated for the calculation of micro-hardness. The asmachined substrate was tested as a reference using a 200 g weight. Three measurements were performed and the average value was calculated.

3.4.Friction and galling test
Friction tests were performed to determine the CoF and anti-galling properties of various combinations of wet/dry lubricants and coatings. Fig.1 shows an image of the test rig and a schematic of the contact imprint of the crossed pins. The rig was designed so that the two test A small initial normal force was applied between the top and bottom test pins by tightening the nuts, and the normal force increased whilst the platform translated via a linear actuator.
The sliding speed was controlled at a constant speed of 3 mm/s by LabVIEW (National Instrument, UK), and the total sliding distance of one stroke was 45 mm (one cycle is 90 mm). Three cycles (270 mm) were performed continuously for each test. Tangential and normal forces were recorded continuously by data acquisition to record CoF. The configuration and instrument setup has been described previously [18].

Fig.1. Friction and galling rig test setup
Two loading levels were applied to investigate the frictional and galling properties of the coating / lubricant systems. Hertz theory of contact between elastic bodies was used to  [19]. For crossed cylinders of equal radius R, the maximum and average contact pressure P0, Pa are given by: Where F is the normal force, E is elastic modulus and ν is the Poisson's ratio.

4.1.Shot-peening and surface roughness
The shot-peening was applied for various periods (0, 1, 3, 5 minutes) with a fixed distance of 10 mm between compressor nozzle and pin surface using a shot-peening system. Fig.2 shows  The peening process can not only remove contamination on the surface but also induces surface residual stresses which improves galling resistance. The references [20,21] have discussed the relationship between shot-peening, surface residual stress and substrate fatigue Sensitivity: Confidential performance. The craters generated from peening also act as a convenient lubricants trap during sliding contact reducing the amount lubricant being removed from the surface. Fig.3 shows the surface roughness (Ra) for samples with different peening periods. The respective standard variation for each peening condition is also shown in the chart. This measurement indicates an optimized peening configuration for surface modification, therefore 4-minute shot-peening time was chosen for the tribological investigation in this study.  Table 3. The Ni-P-Cu coating was prepared by adding various amounts of copper sulphate in the plating bath, from 0.5 to 1.5 g/L. In the figure, 0 g/L represents Ni-P coating. It can be seen from the figure that the Ni-P showed relatively smooth surface profile and very rare coating grains. More coating grains appeared on the surface when the copper sulphate concentration increased from 0 to 1 g/L but decreased again when copper sulphate exceeded 1 g/L.

Fig.4. Typical SEM images of electroless nickel coating by various copper sulphate concentration
Because copper is not a catalytic element for such an electroless bath, the Ni-P-Cu bath became increasingly difficult to control when copper content increased within the coating film. The coating film appeared to be a pure copper layer if the electroless Ni-P-Cu bath contained more than 2.5 g/L copper sulphate, and no Ni-P-Cu coating was obtained on the CRA's substrate regardless of activation methods.  CH CH COOH ) in the chemical bath, and their chemical properties in aqueous solution are altered when they combine with these complexing agents [22,23]. These two types of ions compete for the electrons provided by the reducing agent (i.e. hypophosphite). As shown in chemical reactions in Eq. 1, the reduction of Ni 2+ is always accompanied by the reduction of 22 H PO  producing P. However, Cu 2+ is much easier to reduce due to a much stronger electrode potential. This is in agreement with disproportional increase in copper content shown in Table 3. Copper can also be reduced by 22 H PO  without co-deposition of P leading to the decrease in P content in the coating [24].
The industry normally classifies an electroless nickel coating with over 10 wt% phosphorous content as a high phosphorous coating, while the phosphorous coating is mainly controlled by pH value (the higher pH the lower phosphorous content). The electroless nickel coating becomes amorphous and presents relatively lower hardness when phosphorous content is higher than 7 wt% [25], and this controllable hardness can be carefully designed to match the hardness of the contact pair to provide optimized wear and anti-galling properties. Due to the physical barrier of phosphorous and copper within the nickel matrix, coating with high content of phosphorous and copper tends to have better corrosion resistant properties [26]. In literature [27], Parkinson has provided a detailed discussion on properties and applications of electroless nickel coatings.
Coatings with poor adhesion cannot provide sufficient protection to the substrates. The adhesion between coating and the substrate strongly depends on a catalytic layer (nickel strike). It was found that both electroless and electrolytic coatings appeared to have no adhesion without the presence of a catalytic nickel layer. Carrying out repeated pull-off adhesion tests in this study showed a maximum value of 20 MPa for all the three coatings (Ni-P, Ni-P-Cu and Cu) until the adhesive failed. The pull-off adhesion test clearly showed that the adhesive bond between the coating and adhesive consistently failed, consequently the coatings always remained on the substrate during the friction test indicating good adhesion Sensitivity: Confidential between substrate and coating. Indeed, the surface free energy of coating decreases with the increase of copper content [28], which reduces the bonding between adhesive and coating, leading to failure of adhesive.

Micro-hardness
The micro-hardness of the coatings was measured using the Vickers scale and the results are given in Fig.5. Each coating was measured at three positions and the standard deviation of each coating is also shown on top of the corresponding bar. In the figure, the measurement for Ni-P-Cu coating was presented as the variable of copper concentration in the chemical bath, whilst 0g/L represents Ni-P. As a benchmark, the CRA's substrate is also presented in the figure, showing a value around 317 HV0.2 (corresponding to HRC34).  helped to reduce the CoF significantly when shot-peening was applied to the test pins. This is because the surface profile had been modified by the shot-peening process so that the thin burnished metal layer acted like a semi-fluid film to prevent the direct hard contact between the contact pair during sliding as discussed in [29]. Due to the limit of the burnishing process, it is unlikely to obtain a thick metal layer, consequently, the burnished layer was easily worn.
Indeed, the thickness of the burnished layer was thinner than the measurement capacity of a micro-meter (1 µm).
The presence of nanoparticles showed similar friction behaviour when compared to the burnishing process. The PTFE and WS2 nanoparticles were dispersed as an aerosol and contained within a spray bottle. Nanoparticles remained on the surface just a few second after the solvent evaporated. It has been found that the friction performance was strongly dependant on the total amounts of PTFE and WS2 nanoparticles on the surfaces of contact pair --the nanoparticles can be removed from the surface after a very short sliding distance, particularly at higher contact pressure. Fig.7 shows a progressive increase in CoF when testing electroless Ni-P-Cu coating under high Hertzian contact stress (Pa =2 GPa) lubricated by PTFE and WS2 nanoparticles.

Both the top and bottom samples were shot-peened
There are nine sliding cycles in Fig.7, in which the nanoparticles lubricated the contact surfaces sufficiently at the very beginning of sliding under high contact pressure. However, the CoF increased gradually whilst the sliding distance increased due to the removal of nanoparticles. It is interesting to note that the friction performance at high contact pressure showed an obvious improvement by increasing the amount of WS2 nanoparticles but there was no change having increased PTFE nanoparticles. A closer inspection to the test pins Sensitivity: Confidential indicated that the excessive PTFE nanoparticles coagulated and were removed completely from the contact region, whilst the WS2 nanoparticles distributed uniformly on the surface.
It has been noticed that the shot-peening had helped to improve the friction behaviour of electroless nickel coatings, particularly when lubricated by nanoparticles, i.e. PTFE and WS2.
It has been found that the friction performance was highly dependent on the adhesion and distribution of the nanoparticles. Fig.8 shows the effects of tin burnishing prior to the application of nanoparticles on the CoF of the two electroless coatings. The PTFE nanoparticles formed solid condensed powder after a few seconds of spraying, which exhibited very poor adhesion on the tin burnished surface, therefore the PTFE condensed powder was mostly removed during sliding. As a result, the CoF was between 0.11 and 0.13.
The application of sprayed WS2 nanoparticles attached very well on the burnished surfaces so that galling was reduced to minimum, as discussed below, generating a much lower CoF. In terms of CoF, there was no fundamental difference between electroless Ni-P and Ni-P-Cu, considering the friction tests under both intermediate and high contact pressures, however the corrosion resistance of Ni-P-Cu has been reported to be considerably better than Ni-P [30,31].

Wear scar
Typical images of wear scar on the coating (bottom test sample) and indentation on the contact counterpart (bottom test sample) for three coatings are shown in Fig.10. The appearance of indentation at the contact area of the counterpart of Ni-P and Ni-P-Cu coatings looked similar (around 0.86mm in diameter), as a contrast, the diameter of indentation corresponding to copper coating was considerably smaller (0.66mm) due to a much lower micro-hardness. The wear scar on the coatings showed an opposite trend: the harder the coating, the lighter wear scar. The characteristic of anti-galling property of coatings should be the balance of wear scar and indentation. where F is the normal force; R is the radius; and E and ν are elastic modulus and Poisson's ratio.  A measurement for radius of indentation was carried out by an optical microscope following sliding wear, and the results showed a much larger value compared to the prediction by Equation 3. Fig.11 gives a comparison of the radius of wear scar (r, bar chart) and the depth of indentation (d, scatter balls) for the three coatings. The predicted radius of wear scar is also shown in the figure as flat dash line. The measurement of the wear scar radius against electroless Ni-P and Ni-P-Cu coatings was about 40% higher than the prediction of Hertzian model. Though wear scar in the test against Cu coating was considerably smaller, the absolute value was still 18% higher than the prediction. This indicates a higher wear rate of the top sample against Ni-P and Ni-P-Cu coatings than much softer Cu coating. Note that the Hertzian model is based on the static condition and does not take into account the effects of the coating's micro-hardness. Considering the material hardness and sliding condition, the depth of indentation dw is calculated by abrasive wear mechanics [33], Where K is non-dimensional constant; Pa is contact stress; L is sliding distance; and H is the material hardness. In this equation, the depth of indentation is proportional to the sliding distance. It is inferred that K constant for tests against Ni-P and Ni-P-Cu is higher than that against Cu. However, after a careful inspection by optical microscope, no obvious change of Sensitivity: Confidential depth of indentation and diameter of wear scar was observed following 3, 6 and 9 cycles of sliding (corresponding to 270mm, 540mm, and 810mm). One probable reason is that the asperities of both contact pair were burnished at the very beginning of sliding, leading to a stable CoF in the following sliding cycles (refer to section 4.3).

Conclusions
This paper has investigated the effects of surface pre-treatment, bath composition and activation method on the topographical, mechanical and tribological properties of electroless Ni-P and Ni-P-Cu coatings on CRA's. Friction and galling tests were carried out at intermediate (1.5GPa) and high (2GPa) Hertzian contact stress, with various surface pretreatments and lubricating conditions using a cross cylinder friction and galling rig. Based on the experimental investigation some new findings can be concluded from this study: (1) An electrolytic Ni coating is essential to achieve excellent coating adhesion on CRA substrate; (2) Maximum micro-hardness was obtained with 1 g/L of copper sulphate in Ni-P-Cu bath; (3) The binary and ternary electroless coatings (Ni-P, Ni-P-Cu) presented similar behaviours such as CoF and wear scar, regardless of copper content within the coating alloy, which was very different from their anti-corrosion properties of the two coatings; (4) The CoF of sliding cycles were decomposed as forward/backward strokes, which were classified as 'makeup' and 'breakout' regimes. It was found that CoF increases with increasing contact stress for both regimes indicating a change from adhesive friction or abrasive friction.
(5) The wear scar and depth of indentation were considerably higher (approximate 40%) than the prediction of static Hertzian contact model. These parameters of electroless coatings stabilised after a small sliding distance while there was an increasing trend with Cu/API system, indicating a shorter 'running in' period for electroless coatings.
(6) Shot-peening showed a promising improvement to the tribological performance of the electroless coatings, but had a negative effect on the commercial coating system, i.e. electrolytic copper/API dope (Cu/API). Minimum CoF of 0.05 was achieved with electroless Ni-P or Ni-P-Cu plating coated with an ultrathin tin and WS2 nanoparticles.