ELSEVIER Surface and Coatings Technology 91-95 (1997) 513-548

Influence of the heat treatment on the abrasive wear resistance

of electroless Ni-P

Abstract

Electroless Ni-P coatings on plain carbon steel werr prepared by using an acid bath based on NiClz as source of nickel cations.
__-_ -___ Hear-_---
treatments were performed at 260’ and 300°C for 25 h and 1 h, respectively. The abrasive resistances of the coatings which suffered
different heat treatments were determined by using the modified ball cratering method, recently proposed by Staia et al. The abrasive media
employed was a mixture of 100 ml ethylene glycol and I g of 3 pm diamond paste. A RB3S.UGlOOP3 ball-bearing was used as tribological
pair. The individual wear constants for the substrate and coatings were determined from the recorded plots of the crater depth vs. sliding
distances with the aid of theory of the imposed shape wear scar reported in the literature by Kassman et al. Electron microscopy was used to
study the morphology of the worn surfaces. As a result, it was found that the heat treatment decreases the abrasive wear resistance of the
coatings. A comparison between the methods proposed by Rutherford and Hutchings and ours for assessing the wear constants is carried
out, particularly in relation to the relevance of including in the computation the value of the friction force. 0 1997 Elsevier Science S.A.

Keytrvrds: Heat treatment; Electroless Ni-P coatings; Wear resistance

1. Introduction parameters characteristic to each test. Moreover, combined

with the information regarding fundamental properties, they
In the last two decades a variety of surface engineering were able to contribute to optimization of the coating pro-
processes have been developed with the aim of enhancing cessing.
the performance of materials mainly from the wear, fatigue, Previous work conducted by us [I] has allowed us to
corrosion and biocompatibility point of view. Parallel with develop a novel test by modifying the cratering ball techni-
the development of these coatings technologies, there has que, which is a standard method employed in measuring the
been an increase in the number of methods employed to coatings thickness, in order to quantify the individual abra-
determine the main properties and characteristics of the sion wear coefficient of coatings and substrates with the aid
produced coatings and, at the same time, an increase of of the theory of the imposed shape wear scar proposed for
accelerated (laboratory tests) and in situ techniques able the first time by Kassman et al. [2]. The materials under
to evaluate their tribological performance and/or corrosion study were silicon wafers, to test the data reproducibility
behavior. Without doubt, the accelerated laboratory tests of the method, and Ni-P coatings as-deposited and heat-
used to assess the tribological behavior of the coatings, in treated. Wear test5 were conducted by using as abrasive
most of the cases, are far away from describing the tribo- medium a mixture of 3 pm diamond paste and ethylene
logical performance of the system coating/substrate in ser- glycol. Kassman et al.‘s [2] theory has proved to be a simple
vice conditions. Nevertheless, they have permitted the and reproducible method for checking the mechanical qual-
screening of the quality of these surface engineered pro- ity of a thin coating by using the dimple grinder normally
ducts, thus being able to accomplish the assessment of employed for electron microscopygple preparation. By
their performance related to the intrinsic and extrinsic applying it they have achieved a major improvement in
relation to all the existing laboratory methods for determin-
ing wear resistance [3], which until then were unable to
* Corresponding author. Tel.: +58 582 7529017; fax: ~58 582 7529017; assess the separate influence of the substrate and the coating
e-mail: [email protected] on the overall tribological performance of the system.

0257~8972/97/$17.00 Q 1997 Elsevier Science S.A. Ail rights reserved

PZZ SO257-8972(97)00463-5
544 hl.H, Shin et a/. / Surjcrrce and Coatings Terhnoiog)~ 94-95 (1997) 543-548

Some other improvements, accordingly explained [ 1, 2.2. Mnterinls

51, were obtained when using the ball cratering method
instead of the dimple grinder. These are regarding two The abrasive tests were performed on plain carbon AISI
main features: firstly, the fact that the modified ball 1020 steel uncoated and coated with electroless Ni-P. All
cratering equipment permits the delivery of the a fresh samples were coated in an acid electroless deposition bath
abrasive mixture during the experiments, impeding its by employing the procedure described elsewhere [6]. All
degradation; and secondly, the possibility of collecting coated samples are practically identical, with an average
the results for different values of sliding distances with- phosphorus content of 6.67% determined by using glow
out errors which can arise from the removal of the discharge optical spectroscopy (GDEOS) and an average
sample. coating thickness of 9.3 Frn determined by using the ball
At the same time, Rutherford and Hutchings [4], have cratering method. Table I presents some of the character-
proposed the use of the ball cratering method to determine istic features of the sample under study. Values of Knoop
the individual abrasion wear coefficients for a variety of thin microhardness determined by using a load of 0.98 N for 10 s
hard films and different bulk materials by using the theory and of roughness determined by using a Mitutoyo Surftest
of the imposed wear scar shape proposed by Kassman et al. 301 profilometer are also included.
[2]. There are many differences between their method and
that proposed by us. These differences can be divided into 2.3. Main test parameters
two groups: those regarding the theoretical basis (mainly
calculation of the forces) and those regarding the test As was stated previously [l], there are three main para-
method per se as sliding velocity, nature of the abrasive meters that should to be taken into consideration: load, test
mixture and its delivery, and the way of data collecting, velocity and abrasive medium. In the following treatment
among others. only changes in load calculations will be addressed, keeping
In the calculation of the forces in this test, Rutherford and the same sliding velocity (1.11 m s-‘) and the same abrasive
Hutchings [4] have pointed out the influence of the friction mixture.
force between the sphere and the sample in determining the
normal load for a water-based slurry. Also, they have 2.4. Load cdcrhtion
assumed that the friction force is equal to zero for a more
viscous liquid (glycerol) when it was used as carrier of the When friction forces are not considered, the load exerted
abrasive particles, due to formation of a fully lubricating on the sample, N,, has been reported [I] to be equal to:
hydrodynamic film.
Therefore, the present study has been conducted with the
aim of determining whether the frictional force between the
sample and the ball could be neglected for our experimental
abrasive mixture and, if not, to determine the extent of the
variation of the independent wear coefficients for electroless
Ni-P coatings on plain carbon steel, taking into account this
influence. At the same time, it will be very interesting to
Air bed
discuss if this modification, not taken into consideration by
the authors in previous publications, is indeed necessary in
performing this test as a screening method in assessing the
--WI Abrasive + Lubricant

three body abrasion wear resistance parameters of metallur-

gical coatings.

2. Experimental details

2.1. Test equipnerzr

A general view of the instrument and its attachments is

presented in Fig. 1.
The abrasive mixture of 100 ml of ethylene glycol and 1 g
of 3 pm diamond paste, magnetically stirred, was delivered
by using a glass syringe in such a manner as to ensure roll-
ing of the ball in contact with the shaft without slip, and at
the same time to provide the mixture continuously at a
constant flow value of 2.5 ml/h. Fig. 1. The test equipment and its attachments.
Table 1 2.5. Determinatiorz of the cl~namic friction coeficient
Sample characteristics and heat treatments
The pin-on-disk method was employed for evaluating the
Sample Heat treatment Roughness, KnoopIoo dynamic friction coefficient. The mandrel-holding rod was
R, k-d) micro- substituted by a container which allows the positioning of
hardness
(kg mm-‘) the coated sample and the addition of the abrasive mixture.
The maximum sliding velocity of 0.1 m s-’ was used in
Ni-P, as-deposited - 0.40 497 order to avoid the abrasive mixture ejection. The experiment
Ni-P 260°C for 25 h, air-cooled 0.41 795
Ni-P 400°C for 1 h, air-cooled 0.41
was performed in air at 22°C and 35 F 5% humidity. The
916
pin, a ball-bearing of 6 mm diameter, was loaded with a
normal load of 1 N. The contact radius was of 4 mm and
five tests were carried out.
Nt = =~n~cosw -.- (1)
case + 7 2.6. Collection of the results fi-orn the nbrnsive w’ear test
sinwcos8
where 0 = the angle between the direction of the normal In a tested sample there are three sets of five craters each.
force N, and the weight of the ball, IV; w = the angle A wear test is made of three single tests and a single test is
between the weight of the ball, IV, and the distance, K, made of five different runs, one for each sliding distance.
between the two points of contact on the axis with the After each single run, the sample is rotated counter-clock-
ball; 0 = the angle between the reaction force on the axis, wise and another identically but longer single run is per-
N2, and the x direction. formed.
This load depends on the size of the ball-bearing, the When the test is finished, the diameters of the craters are
angle between the sample and the horizontal plane, the fric- measured in-an image analyzer. Subsequently, the worn-off
tion force between the ball-bearing and the sample in the volume (V, and the crater depth (h) are easily calculated
presence of the slurry, and the distance from the sample to from the measured crater diameters (D) for a determined
the rotational axis. sliding-distance (S) and the ball radius (R).
When friction forces are taken into account, some addi-
tional geometrical parameters have to be determined
together with the value of the dynamic friction coefficient 3. Results and discussion
for the tribological system under study. These measure-
ments are related to the angle between the direction of the 3. I. Deteeation ~~OJ the dynamic friction coeficient
friction forcefi and the x axis, which is equal to the angle
between the direction of the normal force N1 and the weight The average value of the dynamic friction coefficient, p,
of the ball (0) and the angle between the friction forcefz and was found to be approximately 0.26 and no changes of it
the x axis (w). As the ball achieves a constant velocity, in a with time were observed. With this value of ,LLI,the normal
few seconds from the start, it can be assumed that the net load N, calculated by using eqn (2) gives a value of 0.12 N,
torque acting on the ball is equal to zero, which implies that which is nearly 30% smaller than the value of the normal
fi and f2 are approximately equal. load evaluated when the friction force between the ball and
The normal force exerted on the sample (Ni ), obtained by
combining the summation of forces in the z and x directions
represented in Fig. 2, is expressed as:
N1 = (WsinQcosa)/ ( cosQ[sinO~i (2cosw - cos@]

+ sinwcosa[cos0 +p, (sin0 + 2sino)j)

(2)
Since the S and K distances are measured on the equipment,
the radius I’ is known, the angle 8 is set, the angles Q, 0 and
w can be readily found. From trigonometric considerations
and the measured values for K = 16.25 mm and S =23.92
mm, the angles Q = 25.48”, u = 80.56” and w = 3.46”have
been determined. Knowing that 0 = 80”, r = 19.05 mm and
W = 0.225 Kgf, a value of the normal load of 0.17 N was
obtained on this specific setup, when using eqn (1).
If eqn (2) is employed to calculate the normal load N1 the
dynamic friction coefficient, p, has to be determined. Fig. 2. Schematic diagram for load calculation.
546 M. H. Stain et al. / Srtrjke rind Coarings Techdog): 94-95 (1997) 543-548

the sample is not considered and therefore, a variation in

both individual wear coefficients of coating and substrate is 5 (hi-t)3
i= 1
expected. KS= (6)

where N represents the number of experimental data.

As reported by Kassman et al. [2] by using the special When dealing with only one homogeneous material as
case of a spherical cap shaped wear scar produced on a steel, eqn (5) can be applied for the case corresponding to
coated material with a coating thickness, t, the approximate h 2 t. For the heat-treated nickel electroless samples at 260”
sliding distance, SapPr.,is given by: and 4OO”C, the equation corresponding to h > t was used.
In all cases, as indicated previously [l], a high degree of
correlation was obtained between the experimental and the-
(3)
oretical points and, as an example, Fig. 3 presents the var-
iation of the crater depth vs. sliding distance for the Ni-P

( >
whereas:
coating heat-treated at 400°C for 1 h, irrespective of the
value of the normal load.

[ 1
IIhr h-i
SVPT. +h-t ifh>r The calculated values of the wear constants for every
14)
=-r 2t- Kc Ks material under study with and without considering the fric-
tion force N1 are presented in Table 2.
where h represents the depth of the crater. From the values As can be noticed, there is an increase of 30% in the
experimentally determined of S, h and t the constants Kc values of the wear constants for each material under study
and K, can be readily determined by means of the least- when the normal load is calculated by taking into account
square method. the friction force. This increase, without doubt, cannot be
Thus: considered insignificant. However, as was mentioned
above, the proposed experiment is a good accelerated
laboratory test which ensures reliability, as demonstrated
rb !gl 11:
by the experimental results. This method can be used for
I&=----- (5)
optimizing the deposition conditions, and to quantify the
L ‘5 Si relative wear constant of different coatings and substrates
i=l
produced by the same process under different conditions.
and
Nevertheless, this test is unable to simulate the real operat-

M-P
Heat treated (400°C x 1h)

300 400 500 600 700 800 900

Sliding Distance Im]

Fig. 3. Crater depth vs. sliding distance for the Ni-P coating heat-treated at 3OOT for 1 h. Velocity I. 1I m s-’ for both normal load values of0.17 N and 0.12
N.
 517

Table 2

Calculated values for the wear constant of the samples under study for different values of the normal load A’, depending whether (with f.f.) or not (no f.f.) the
friction force was considered in its calculation

Sample Load (N) K, (m’&I) K,~$& R_’

Ni-P as-deposited No f.f. 0.17 3.57 x 10-11 - 0.97

With f.f. 0.12 5.06 x 10-l” -- 0.97
Ni-P heat-treated (260°C x 26 h) No f.f. 0.17 8.00 x IO“’ 1.85 x IO-” 0.95
With f.f. 0.12 11.39 x 10-14 ~ 2.62 x lo-” 4.95
Ni-P heat-treated (400°C x 1 h) No f.f. 0.17 1.00 x lo“:’ 5.17 x 10-I” 0.91
With f.f. 0.12 1.42 x 10-13 7.74 x 10-I” 0.91

ing conditions and cannot ensure that the coating will have As was indicated by Rutherford and Hutchings [A-], the
the same behavior. From the above discussion it is possible action of the abrasive particles within the contact will
to conclude three important facts: (a) the inclusion of the depend on the ratio between the film thickness of a lubricant
friction force in the computation of the wear constants give between the sample and the ball and the abrasive mean
rise to a significant increase in the values; (b) however, in particle diameter, d. They proposed the following equation
order to take into account such a force it is required to in order to evaluate the lubricant film thickness, h,. This
evaluate two additional geometrical parameters as well as equation represents the solution to Reynolds equation in
to determine the dynamic friction coefficient, which renders two dimensions for a sphere of radius R, sliding against a
the method much more complicated that the simpler and plane in the presence of a viscous fluid and is expressed as:
easier method earlier proposed by us; (c) since for the
screening of the wear properties of different materials
only relative values of the wear constants are required, the
inclusion of the friction force in the calculation is irrelevant. where IVY is the normal load, ~1is the sliding speed and r
However, it must be specified wether or not such a force was represents the fluid viscosity.
taken into account into the calculation if a comparison In the present investigation for values of V = 1.11 ms-.!,
between results of different laboratories is to be made. R = 0.01905 m, qEthY]e,,e slycOl= 4.3 x IO-’ N m-?-S and
Thus, this specification will become an intrinsic parameter N, = 0.12, the film thickness has been determined to be
of the test in the same manner as is, for example, the abra- about 0.3 1 pm, that is to say smaller than the mean abrasive
sive mixture employed or the diamond tip radius used in the particle diameter. In these conditions, the ratio hJd is equal
scratch test to determine adhesion [6]. The inclusion of the to 0.1, indicating that the abrasive particle has a major con-
friction force implies additional experiments in order to tribution to the wear process. This can be corroborated by
evaluate the dynamic friction coefficient for the system analyzing the scanning electron micrographs of the worn
under study, which makes the use of this simple technique surfaces of the Ni-P coating, heat-treated at 4OO”C, which
dependent on other equipments with their inherent experi- are presented in Fig. 4a and b. Well-defined wear scars of
mental errors. nearly the same width as the me-anva1u.e of the abrasi-ve
The relative values of the wear constants obtained for particle diameter are-observed, indicat%lg- a pure abrasion
different conditions of electroless Ni-P indicate that there wear mechanism, which results in the material removal due
is not a direct relationship between the coating hardness and to the cutting performed by the abrasive particles. Similar
abrasion wear resistance, since other factors are injluencing morphologies of the worn surfaces were obtained for all the
this behavior, underlying the fact that hardness could not be samples under study.
considered as a key guide to the coating abrasive wear. It was considered [8] that when the tribosystem is com-
As can be observed, the abrasive wear constant, Kc, for posed of a two triboelements and lubricant,it~qossible to
the Ni-P coating heat-treated at 400°C (a microhardness of obtain an indication of the dominating lubrication or wear
916 HKrm) is higher than the wear constant obtained when mechanism by monitoring suitable test parameters such as
the heat treatment was conducted at 260°C (microhardness friction, lubricant film thickness, temperature, wear and sur-
of 795 HKr,), results which contradict the explanation pre- face roughness. In these conditions, three main lubrication
sented by Duncan [7] which attributed the improvement of regimes have been identified [9] as a function of the varia-
abrasive resistance of Ni-P coatings to the Ni3P particle tion of the friction coefficient, ,LL,and the film thickness-to-
coarsening, as the heat treatment temperatures increase. roughness ratio, A. In the present work a value of film thick-
Also, it has to be mentioned that due to the difference ness-to-roughness ratio, X, equal to 0.77 has been deter-
between the thermal coefficients of the coating and the mined, which together with a friction coefficient of 0.26
steel substrate, the coating heat-treated at 400°C presented calculated previously possibly could place our system in a
cracking at the surface which could negatively influence its boundary lubrication mode, were the tribological behavior
abrasive resistance behavior. is governed by solid-solid friction and wear processes.
548 MH. Slain er al. / Surfiire cd Coatings Technology 94-95 11997) 543-548

without ensuring the reproducibility of the real operating

conditions. A simple indication of the fact that the friction
force was not taken into account when reporting the wear
constants, as an intrinsic parameter, will be enough to allow
comparisons to be made with the work performed in other
laboratories.
Also, it has to be pointed out the simplicity of the test
proposed by us in reference to: {a) use of a mnch simpler
expression to calculate the load which involves parameters
that can be easily determined, (b) easy measurement of the
craters diameters for different sliding distances without the
misfit produced when moving the sample or the complicated
setup to measure crater in situ, and (c) practical and low-
cost setup for the slurry delivery attachment.

Acknowledgements

The authors wish to acknowledge the financial support of

National Council for Scientific and Technological Research
of Venezuela (CONICIT) through the project Sl-2612, to
the Council for Scientific and Humanistic Development
(CDCH) of the Central University of Venezuela through
the project O&17-2789/95, and to the Organization of the
American States (OAS) through the Multinational Project
on Materials. The financial support of the Postgraduate Stu-
dies Council of Central University of Venezuela is also
acknowledged.

References

[1] M.H. Staia, C. Enriquez, ES. Puchi, B. Lewis and M. Jeandin, in

Fig. 3. (a and bj Scanning electron micrographs of the worn surface of the T.S. Sudarshan, M. Jeandin and M. Khor (eds.), S&zce Modificntion
heat-treated Ni-P coating [at 300°C ~for 1 h) at different magnifications. Technologies X, Inst. of Materials, UK, 1997, p. 130.
[2] A. Kassman. L. Erickson, M. Olsson, P. Hedenqvist, S. Jacobson and
4. Conclusions S. Hogmark, Su$ Cont. Technoi., 50 (1991) 75.
[3] B. Bhushan, in W.B.Harding and G.A. DiBari teds.), O~rrvie~~ of
The analysis carried out in the present work has shown Comings Mareriais, Su&ce Trearmenrs md Screening Techniques
that our initially proposed method is able to quantify the for Tribological Appiicntions, Part 2: Screening Techniques, Testing
of Mefallic and inorganic Coatings, STP-947, 1987, p. 310.
relative abrasive wear resistance of electroless Ni-P coat-
[3] K.L Rutherford and I.M. Hutchings, &lj-: Coot. Technol., 79 (1996),
ings which have undergone different post-deposition treat- 231.
ment conditions. Although a significant increase in the [5] M.H. Staia. E. Castillo, E.S. Puchi, D. B. Lewis and H.E. Hinter-
values of the individual abrasive wear constants of the coat- mann, Surj Cont. Technol., 86-87 (1996) 598.
ings were obtained when the friction force was taken into [6] P.A. Steinmann and H.E. Hintermann, J. Var. Sci. Techno/., A7 (3)
( 1989) 2257.
account, in our opinion, this approach will make this simple
[7] R. Duncan, Meral finishing, March (1990) 11.
test troublesome. The inclusion of the friction force implies [8] K. Holmberg and A. Matthews, in D. Dowson (ed.), Coating Tribol-
additional experiments in order to evaluate the dynamic ogy, Tribology Series, 28, Elsevier, Amsterdam, 1994, p. 307.
friction coefficient for the system under study, which 191 H. Czichos, Basic Tribological Parameters, ASM Handbook vol. 18,
makes this test costly and dependent on other equipment ASM International, 1992, 474.