Mechanical properties of thin hard coatings on TiC-NiMo substrates/Ohukeste kovapinnete mehaanilised omadused TiC-NiMo-kermistest alusel.
Yaldiz, Can Emrah ; Veinthal, Renno ; Gregor, Andre 等
1. INTRODUCTION
Tungsten-free TiC-NiMo cermets are very hard and have low
susceptibility to diffusion and adhesion and high oxidation and wear
resistance at elevated temperatures [1].
TiC-base cermets offer an attractive combination of high specific
mechanical properties, such as strength/density, because of their
relatively low density [2]. A general comparison of TiC-based cermets
and WC-Co hardmetals reveals their advantages and disadvantages. The
coefficient of friction against steel counter body for titanium
carbide-based cermets is approximately 1.5-2 times lower than that of
WC-based hardmetals [3]. Moreover, maximum service temperatures for TiC
are much higher and their density is 2-3 times lower. However, they
exhibit lower thermal conductivity; in addition, their thermal expansion
coefficient is about twice higher [4]. Finally, they show lower
Young's modulus, endurance limit, thermal stability, transverse
rupture strength, fracture toughness, and plasticity. Also, reprocessing
of TiC-based cermets is more complicated [4,5].
Using a hard coating on machining tools in order to improve tool
lifetime has become a standard application in industry. In order to
realize these promised benefits, the properties of coatings and
relationships between them as well as with the overall coating
performance need to be understood. Figure 1 shows some of the properties
of coating/substrate systems. Only after understanding these properties,
coating processes can be optimized in order to produce tailor-made
coatings for particular applications.
It would be very time consuming and difficult to investigate all of
these parameters to characterize a coating/substrate pair. Instead, some
fundamental properties were chosen to reflect the character of the
coating/substrate system for a wide range of applications. Among them
there are hardness and adhesion properties. The tools for testing are
discussed below in detail.
[FIGURE 1 OMITTED]
2. EXPERIMENTAL PROCEDURES
In order to compare various samples to meet the objectives of this
research, first the substrates were fabricated. After fabrication of a
sufficient number of substrates, they were coated with the chosen
coatings. When the coating deposition was completed, the samples were
analysed. Individual steps of the process are outlined below.
2.1. Sample production
Sample production consisted of two distinct steps: substrate
fabrication and coating deposition. Two different types of substrates
with varying binder content were fabricated and within each type
substrate surfaces were treated to have varying degrees of roughness
values.
Two different types of substrates were fabricated with the help of
conventional P/M routine: TiC in nickel-molybdenum matrix, with Ni: Mo
ratios of 2: 1 and 1:1. The dimensions of substrate samples were 15 x 5
x 25 mm.
The pre-sintering regime was selected according to [2]. After
pre-sintering, the samples were taken to the Sinter-HIP. In the
Sinter-HIP process, the samples were sintered in vacuum and in a
high-pressure environment, respectively. The regime consisted of various
steps, including temperature increase stepwise up to 1500[degrees]C in
vacuum, replacing vacuum with argon, pressurizing it up to 50 bar, dwell
time, cool-down and depressurizing. The manufacturing technology is
described elsewhere [5].
Further the specimen preparation included surface grinding to four
different average roughness values, namely to [R.sub.a] equal to 0.2,
0.1, 0.05 and 0.005 [micro]m. In every roughness category, five samples
of each substrate type were prepared (in total ten per a roughness
category). After sintering, every sample was first ground with the 40
[micro]m grinding paper. Then the samples were gradually polished to the
required [R.sub.a] values in the order of diamond grit sizes of 40, 20,
9, 3 and 1 [micro]m. The surface roughness was measured with the Mahr
Perthometer.
2.2. Coating deposition
Platit [pi]-80 was used as the PVD deposition system. Five
different coatings: TiN, TiCN, TiAIN, AlTiN, and
nc-(Al,Ti)N/[alpha]-[Si.sub.3][N.sub.4], ranging from monolayers to
gradient coatings and nanostructured coatings, were deposited. Coating
thickness was set to 2.0 gm, and the loading factor was set to 25% in
the coating program.
Before coating deposition, the substrate was prepared and cleaned
in an ultrasonic bath. Arc cleaning of sample surfaces was done at
450[degrees]C, sample surfaces were cleaned with a pulsed Ar glow
discharge. Moreover, the Ti cathode was also cleaned with Ar plasma,
followed by the Ti etching process.
After the initial processes, process parameters for different
coatings differed significantly. Standard Platit coating recipes were
used in order to deposit the coatings. Table 1 shows the main process
parameters for each coating type.
2.3. Coating structure and thickness
After deposition, the corresponding coating thickness was measured
using the Kalo-Max Ball-Crater test method. In this method, a hard metal
ball with diamond suspension is used to wear out a crater on the coated
surface until the substrate is exposed. Then the resulting crater
dimensions (inner and outer crater diameters) are measured under optical
microscope to calculate the coating thickness t. The calculation is done
using the following equation:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)
For structure analysis the coated samples were broken into two
pieces in order to obtain SEM cross-section images.
2.4. Indentation testing for coating adhesion
In order to test the adhesion quality of the coatings, the
well-known Rockwell adhesion test method was used [7]. A Rockwell
hardness-testing machine, conforming to the requirements of EN ISO
6508-2, was applied. Every sample was indented at four different
representative locations as a minimum. Indentations were made in a
direction perpendicular to the specimen surface. Sample surfaces were
free from dust, oil, and other contaminations. A load of 598 N (60 kgf),
i.e. Rockwell A scale, was used in order to conform to the relevant
standards [7]. The indented samples were then analysed with an optical
microscope at a magnification of 100x and results were classified into
the categories given in the CEN/TS 1071-8 standards [7].
2.5. Nanoindentation for hardness and modulus
Nanoindentation tests for measuring hardness and modulus were
conducted on a Micromaterials Nano Test platform using a standard
Berkovich indenter with tip radius of 100-200 nm.
To analyse changes in hardness over coating thickness; indentations
with different depths were necessary. Indentations of 1, 5, 10, 20, 30,
50, 75, 150 and 300 mN were performed on the sample. The initial load
was 0.03 mN, and then the indentation load was applied as a ramp that
reached the full load in 20 s. After a 10-s dwell time, the load was
relieved in another 20 s. For each load, seven different locations were
indented.
After the load vs depth data was collected, the resulting curves
were first visually observed, and some problematic indents due to
various extrinsic process parameters, such as external vibrations or
fluctuations in the input voltage etc., were removed. Once the
indentation cycle was completed and the corresponding load vs depth data
was logged, the raw data was analysed with the Oliver Pharr power-law
fitting method to determine the hardness and modulus values. Details of
the Oliver Pharr power-law fitting can be found in [8].
Because some of the indentation loads were very small, their
resulting indentation depths were also very low and at low depths the
influence of the tip geometry of the diamond indenter on the diamond
area function was included in the analysis step.
To calculate the actual diamond area function, a series of indents
with loads ranging from 0.5 to 150 mN were conducted on a fused silica
sample. Fused silica (quartz) has very high purity and extremely
homogeneous distribution of mechanical properties; therefore it is the
standard material in nanoindentation calibration tests. From the raw
depth vs load data, the software calculates the best diamond area
function in order to compensate for the diamond tip geometry for both
low and high loads and indentation depths. A good fit was found for
A = 2393.17d + 21.61[d2.sup.2], (2)
where d denotes the indentation depth and A equals the projected
diamond area.
3. RESULTS AND DISCUSSION
3.1. Coatings structure
Figure 2 shows the SEM cross-section images of the broken samples
for TiN and TiCN coatings. In both samples, a columnar microstructure is
observed. Some droplet formation was also observed; however, the size of
the droplets is smaller than that of the coatings with aluminium
content. Coating thickness was also calculated from the pictures with an
image analysis software that was found to be 1.61 and 1.85 gm for TiN
and TiCN coatings, respectively.
For coatings consisting aluminium, higher rates of droplet
formation were observed. The melting temperature of aluminium is lower
than that of the other components in the coating and the substrate, and
because of this property, aluminium cathodes are more susceptible to
droplet formation. As seen in Fig. 3, TiAIN and AlTiN coatings have
droplets on the surface, with diameters close to or more than 1 gm. The
nACo[R] coatings also have some droplet formation on the surface, but
their sizes are considerably smaller than those of TiAIN and AlTiN
coatings.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
TiAIN and nACo[R] coatings were deposited as multilayered, also
verifiable from the SEM pictures (Fig. 3). In both coatings, below the
alternating multilayered structure, the top layer is a gradient layer
with thickness higher than that of any other layer below it. A1TiN, on
the other hand, was deposited as a gradient single-layer coating.
3.2. Coating thickness analysis
Average results for different coating types are shown in Table 2.
Coating thickness results from the ball crater test were verified with
the SEM cross-section pictures and image analysis.
SEM image results are consistently about 1%-10% lower than the ball
crater test results. This is related to the slightly inclined
positioning of the samples on the specimen stub. If the inclination
angle is [theta] deg, the measured thickness is [t.sub.m] and the real
thickness is [t.sub.r], then the relationship between the three can be
expressed as
[t.sub.r] = [t.sub.m]/cos[theta], (3)
which explains consistent difference between the SEM image and the
ball crater test results.
3.3. Coating adhesion
The results of the Rockwell adhesion tests revealed a common trend
between all coatings, regardless of the substrate binder ratio. For
samples with fine surface topographies, such as [R.sub.a] equal to 0.005
or 0.05 gm, coating adhesion was found to be sufficient for all coating
types (Class I--cracking adhesive delamination of the coating, Fig. 4,
left). As the substrate surface roughness increased to [R.sub.a] equal
to 0.1 or 0.2 gm, coating adhesion quality dropped to poor levels and
showed full delamination around the indent (Class III--complete adhesive
delamination, Fig. 4, right).
[FIGURE 4 OMITTED]
3.4. Nanoindentation for hardness and modulus measurements
After the function A(d) was found, the indentation data was
analysed following the Oliver and Pharr method [8], and the mechanical
properties were calculated. The resulting hardness values at different
depths (as a percentage of the coating thickness) are shown in Fig. 5.
[FIGURE 5 OMITTED]
Figure 5 demonstrates that at low load ranges 1-5 mN, corresponding
to plastic indentation depths up to 3% of the total coating thickness,
the results are consistently lower than at higher loads. The reason of
this lower hardness value is attributed to the fact that no further
polishing was done on the coated samples, and the effects of the
roughness profile were very pronounced at the lower load regions.
Depending on the part of the roughness profile the indenter hits a
particular value (i.e. peak, valley or in-between), usually the results
tend to be lower than the actual values. With the increasing load, the
effects of the roughness profile diminish due to the increased indenter
depth, and the hardness values increase. After passing 10% of the
thickness limit as the plastic indentation depth, the influence of the
substrate starts to affect hardness measurements, and hardness starts
dropping.
As a rule of thumb for healthy hardness measurements, the plastic
indentation depth should not be higher than 10% of the coating thickness
in order to avoid the influence of the relatively softer substrate. As
shown in Fig. 5, the very low load region is also affecting the overall
result due to the roughness profile. Therefore, to quantify the core
hardness of the coatings, measurements with 20 mN were chosen as the
most suitable ones. Because the resulting indent is deep enough to be
free from the influence of the roughness profile, but still below the
10% coating thickness depth, influence of the substrate is eliminated.
Table 3 shows the hardness and modulus measurement results.
4. CONCLUSIONS
The coating thickness measurements revealed that coating thickness
was slightly below the set value of 2.0 gm, which was due to the
selected loading factor of 25%. Coating thickness values were consistent
and even among the measured samples.
Nanoindentation results from the 1 and 5 mN hardness measurements
were lower than expected due to the surface roughness effects, more
pronounced at low indentation loads. Polished samples should be used for
measurement in order to minimize the effects of the roughness profile
and to increase the accuracy of core hardness measurements.
Concerning core hardness measurements, nACo[R] coatings showed the
highest hardness value. These coatings also have a lower modulus value,
which makes them the best candidate for wear resistant applications. The
TiCN coatings have the second highest hardness value, but they exhibited
poor adhesion qualities, especially with higher roughness values, as
delamination was present even directly after the coating process without
any load. This is assumed to be related to the higher inherent residual
stresses of TiCN coatings with TiC-NiMo cermets. Further research should
be conducted in order to analyse and optimize the adhesion properties of
this type of coating-substrate pairs.
According to the Rockwell Adhesion Testing, coatings on substrates
with surface roughness 0.05 and 0.005 gm show good adhesion
characteristics. However, with increasing substrate roughness values,
adhesion drops dramatically and therefore should be controlled carefully
in design and manufacturing of the tools to be coated.
doi: 10.3176/eng.2009.4.10
ACKNOWLEDGEMENTS
Dr. Marcus Morstein from Platit AG is acknowledged for his advice.
This work was supported by the Estonian Ministry of Education and
Research (target financed project SF01400091) and the Estonian Science
Foundation (grant No. 7227).
Received 3 September 2009, in revised form 7 October 2009
REFERENCES
[1.] Kubarsepp, J., Klaasen, H. and Pirso, J. Behaviour of TiC-base
cermets in different wear conditions. Wear, 2001, 249, 229-235.
[2.] Kubarsepp, J., Pirso, J. and Klaasen, H. Tribological
characterization of TiC-base cermets. In Proc. 2000 Powder Metallurgy
World Congress. Kyoto, 2000, Part 2, 1633-1636.
[3.] Pirso, J., Viljus, M. and Letunovits, S. Friction and dry
sliding wear behaviour of cermets. Wear, 2006,260,815-824.
[4.] Ellis, J. L. and Goetzel, C. G. ASM Handbook, Vol. 2,
Properties and Selection: Nonferrous Alloys and Special-Purpose
Materials, 1991, ASM.
[5.] Juhani, K. Reactive Sintered Chromium and Titanium
Carbide-Based Cermets. PhD Thesis, TUT Press, 2009.
[6.] Platit Technical Specifications: www.platit.com
[7.] Technical Specification Guide CENTS 1071-8:2004 of the
European Committee for Standardization, "Advanced technical
ceramics--Methods of test for ceramic coatings, Part 8: Rockwell
indentation test for evaluation of adhesion".
[8.] Oliver, W. C. and Pharr, G. M. Measurement of hardness and
elastic modulus by instrumented indentation: Advances in understanding
and refinements to methodology. J. Mater. Res., 2004,19,3-20.
Can Emrah Yaldiz (a), Renno Veinthal (b), Andre Gregor (b) and
Kyriakos Georgiadis (c)
(a) Faculty of Mechanical Engineering, RWTH Aachen, Templergraben
55, 52056 Aachen, Germany; can.emrah.yaldiz@rwth-aachen.de
(b) Department of Materials Engineering, Tallinn University of
Technology, Ehitajate tee 5, 19086 Tallinn, Estonia; {renno.veinthal,
andre.gregor}@ttu.ee
(c) Fraunhofer Institute for Production Technology, Steinbachstr.
17, 52074 Aachen, Germany; Kyriakos.Georgiadis@ipt.fraunhofer.de
Table 1. Process parameters for coating deposition [6]
Coating Bias Pressure,
voltage, mbar
V
TiN -75 ... -120 8 x [10.sup.-3]
TiCN -60 ... -120 (5-7) x [10.sup.-3]
TiAlN -60 ... -150 8 x [10.sup.-3]-1.5 x [10.sup.-2]
AlTiN -60 ... -150 4 x [10.sup.-3]-1.2 x [10.sup.-2]
nACo[R] -75 ... -150 9 x [10.sup.-3]-1.2 x [10.sup.-2]
Ar/[N.sub.2];
Coating Ti/Al/AlSi Tempera- [C.sub.2][H.sub.2]
cathode arc ture, flow,
current, A [degrees]C sccm
TiN (100-125) 450 6/200
TiCN (120-130) 450 6/(165-180);
7/39
TiAlN (85-125)/(65-115) 475 6/200
AlTiN (60-125)/(52-130) 430-450 6/(150-200)
nACo[R] (82-125)/(65-100) 435-475 6/200
Table 2. Average coating thickness of different coating types
Coating type Average thickness, [micro]n
Ball crater test SEM
AlTiN 1.75 1.61
TiCN 1.98 1.85
TiAlN 1.68 1.51
TiN 1.71 1.62
nACo[R] 1.51 1.49
Table 3. Nanoindentation results
Coating type Hardness, GPa Modulus, GPa H/E
TiN 25.86 [+ or -] 1.55 357 [+ or -] 21.4 0.072
TiCN 31.04 [+ or -] 2.82 345 [+ or -] 31.3 0.090
TiAlN 30.17 [+ or -] 1.52 342 [+ or -] 17.2 0.088
AlTiN 29.00 [+ or -] 3.51 325 [+ or -] 39.3 0.089
nACo[R] 46.18 [+ or -] 4.95 320 [+ or -] 34.3 0.144
