ISSN 1063-7842, Technical Physics, 2018, Vol. 63, No. 2, pp. 206–210. © Pleiades Publishing, Ltd., 2018. Original Russian Text © S.A. Linnik, A.V. Gaidaichuk, V.V. Okhotnikov, 2018, published in Zhurnal Tekhnicheskoi Fiziki, 2018, Vol. 63, No. 2, pp. 214–218.
PHYSICAL SCIENCE OF MATERIALS
Influence of Cobalt on the Adhesion Strength of Polycrystalline Diamond Coatings on WC–Co Hard Alloys S. A. Linnik*, A. V. Gaidaichuk, and V. V. Okhotnikov National Research Tomsk Polytechnic University, ul. Lenina 30, Tomsk, 634034 Russia *e-mail:
[email protected] Received May 24, 2017
Abstract—The influence of cobalt on the phase composition and adhesion strength of polycrystalline diamond coatings has been studied using scanning electron microscopy, Raman spectroscopy, and X-ray microanalysis. The coatings have been deposited on WC–Co hard alloy substrates in glow discharge plasma. It has been found that the catalytic amorphization of carbon only takes place during the direct synthesis of the diamond coating, when the cobalt vapor pressure over the substrate is high and the cobalt-related degradation of the synthesized diamond is absent. DOI: 10.1134/S1063784218020226
CVD-grown polycrystalline diamond is currently considered to be the most promising coating material for improving the wear resistance and operational performance of cutting tools used in processing carbon plastics, graphite, alumosilicic alloys, wood, etc. [1, 2]. This is because diamond coatings offer a number of unique properties, such as high hardness and wear resistance, extremely high thermal conductivity, and low coefficient of friction. Hard-alloy tools based on tungsten carbide and cobalt (WC–Co) offers the highest wear resistance [3]; however, the widespread application of diamond coatings is hindered by their poor adhesion to the substrate [4, 5]. A number of companied (Cemecon, SP3 Diamond Technologies, Mitsubishi, etc.) have developed effective techniques (which are kept as know-how) for multistep conditioning of substrates before diamond coating deposition, which to a great extent solve the problem of adhesion. In most publications, however, conditioning of the hard alloy surface prior to diamond deposition is reduced to removing a surface cobalt layer [5–7], which catalyzes diamond graphitization [8, 9]. Works are known in which cobalt-induced graphitization was observed not only during diamond deposition, but also in the diamond deposit [9, 10]. In other works, the cobalt diffusion along diamond grain boundaries and the detrimental effect of cobalt on the growth of even upper layers of the coating were noted [10]. However, mechanisms of cobalt-induced diamond amorphization at different stages escaped the notice of researchers. Therefore, it is necessary to refine and
generalize experimental data to gain deeper insight into the physics of the occurring processes. In this work, we study the influence of cobalt on the phase composition and adhesion strength of polycrystalline diamond coatings deposited on hard alloy (WC–Co) substrates in the glow discharge plasma and discuss the experimentally supported cobalt-related mechanism of amorphous carbon formation.
206
1332 (Diamond )
3000 2000
TPA + G band
1000 ND VDOS
1336 (Diamond ) TPA G band
Intensity, arb. units
INTRODUCTION
0 a
D band (graphite) G band (graphite)
b c 400
800 1200 1600 Raman shift, cm−1
2000
Fig. 1. X-ray spectra taken of the diamond coating sides adjacent to the substrate. Coatings were deposited on decobalted WC–Co substrates at a temperature of 700 (spectra a and b) and 1100°C (spectrum c). Spectrum b was taken for a sample annealed in vacuum at 1200°C.
INFLUENCE OF COBALT ON THE ADHESION STRENGTH
EXPERIMENTAL Substrates for diamond coatings were 5-mm-thick polished disks 20 mm in diameter prepared from Sandvik Coromant H10F polycrystalline fine-grained (a grain size of about 0.9 μm) hard alloy (10% Co). They were subjected to abrasive–waterjet processing by SiC particles to increase the surface area and remove the oxide layer. Then, the substrates were decobalted in an HNO3 : H2O solution. Diamond nucleation centers were introduced by processing the substrates in an ultrasonic bath filled with a water suspension of diamond nanopowder, after which the substrates were rinsed in acetone. Diamond coatings were deposited in a plasma CVD reactor filled with a 99.9999% pure hydrogen + 99.99% pure methane mixture, in which an ac anomalous glow discharge was initiated. The design of the reactor was detailed elsewhere [11, 12]. During deposition, the pressure in the reactor was monitored by a Pfeiffer vacuum CMR 372 precision membrane capacitive sensor and flow rates were controlled by Bronkhorst EL-FLOW mass flow controllers. The temperature of the samples was determined using a ULIRvision TI170 IR thermal imager. The samples were thermally annealed in a vacuum furnace at a residual pressure of 5 × 10–6 Torr. In this way, their oxidation and contamination were excluded. The adhesion strength was estimated by the standard method [13, 14], in which an indentation made by a conical diamond indenter (which is usually used to measure the Rockwell hardness) is analyzed. In this work, the indentation load was 590 N (60 kgf). Scanning electron micrographs of the sample surface were taken under a Philips SEM515 scanning electron microscope (SEM). The elemental composition of the surface was examined using an EDAX ECON IV X-ray microanalyzer. The surface rough-
207
ness was measured using an NTEGRA Prima NTMDT atomic force microscope (AFM). Finally, the phase composition of different sides of the diamond coatings was examined by the method of Raman spectroscopy on a Centaur I HR complex, which makes it possible to take Raman spectra of tiny samples. Sample preparation and processing conditions are summarized in Table 1. RESULTS AND DISCUSSION When studying the influence of cobalt on the phase composition and adhesion strength of the diamond coatings, we considered the influence of cobalt (i) directly in the course of deposition (in the course of film formation) and (ii) on the already synthesized film. To study the former mechanism, diamond films were deposited at different temperatures (in the range of 600–1200°C) on hard-alloy substrates subjected to abrasive machining and also on substrates with decobalted surfaces (Table 1). After thicknesses at which partially delaminated diamond coatings (5–25 μm) were achieved, the phase composition of the side of the coating that was adjacent to the substrate was analyzed by taking Raman spectra. Raman spectrum a in Fig. 1 was taken on the lower side of the film deposited on the decobalted substrate at 700°C. This spectrum contains an intense narrow line of diamond at 1332 cm–1, as well as traces of nondiamond carbon in the form of graphite (band G, 1560 сm–1) and trans-polyacetylene (TPA, 1450 cm–1). This spectrum, which is characteristic of polycrystalline diamond films with a small fraction of nondiamond carbon [15], indicates the absence of graphitization. However, spectrum c (Fig. 1) taken from the lower side of the coating deposited on the same substrate at 1100°C is typical of nondiamond carbon
Table 1. Preparation and processing conditions for H10F hard alloy Abrasive machining
Sample preparation
Diamond deposition
Annealing
SiC particles (FEPA400)
Decobaltization
Diamond seeding
Reagent
Temperature
Time
Reagent
Activator
HNO3 : H2O –60 : 40 (by weight)
70 ± 1°C
5 min
Way of gas activation
Power
Pressure
Atmosphere
Substrate temperature
Deposition rate
Anomalous glow discharge
3 ± 0.1 kW
150 ± 1 Torr
H2 : CH4 = 100 : 5
600−1200°C
0.5−5 μm/h (depending on substrate temperature)
Nanodiamond Ultrasound (~5 nm) suspension 50 kHz, 50 W (0.1 wt %) in H2O
Pressure
Temperature
Time
5 × 10−6 Torr
600−1200°C
120 m
TECHNICAL PHYSICS
Vol. 63
No. 2
2018
Time 10 min
208
LINNIK et al.
(a)
100 µm (b)
100 µm
100 µm
(c)
Fig. 2. SEM micrographs of the indentation made by the conic diamond indenter on the diamond-coated hard-alloy substrates. The diamond coatings were deposited (a) without and (b) with predecobaltization at 700°C and (c) 1100°C with decobaltization. (1) Amorphous carbon layer and (2) diamond film.
(graphite bands D, 1355 cm–1, and G, 1560 cm–1) [16]. The spectrum taken on the upper side of this coating is similar to spectrum a. This suggests that cobalt prevents the formation of the diamond phase at early stages of growth. The formation of a diamond film on the nondiamond carbon implies that the nondiamond carbon layer adjacent to the substrate prevents cobalt diffusion. Figure 2c shows the SEM micrographs taken of the film covering the given substrate after indentation by a diamond indenter (Rockwell method). The film is seen to be bilayer (Fig. 2c; layers 1, 2). Layer 1 is nondiamond carbon 0.5–0.9 μm thick with a cobalt content in the range of 1–4 wt %, and layer 2 is a diamond film. The adhesion of the diamond layer to the nondiamond layer and the substrate is extremely weak. Table 2 shows the surface phase composition of the diamond films deposited at different temperatures on the cobaltized and decobalted hard-alloy substrates. It can be seen that, at temperatures above 1000°C, the lower layer of the coating consists of nondiamond carbon, even if the surface was predecobalted, whereas without decobaltization, it is amorphous, even at low deposition temperatures (710°C). Certainly, the growth of the amorphous (not diamond) layer is associated with the negative effect of cobalt. For substrates with cobalt-free surfaces, this effect is only observed at high temperatures, which means that cobalt emerges on the surface. Since the melting point of cobalt, i.e., the temperature at which cobalt atoms become free to move in between tungsten carbide grains, is 1495°C, which is much higher than the temperature of diamond deposition, cobalt may only emerge on the surface at temperatures above 1000°C by means of sublimation at a subatmospheric pressure; the cobalt saturation pressure at 1069°C reaches 10–3 Torr [17]. The presence of cobalt vapor over the sample explains why the growth of the crystal diamond phase is hindered (cobalt cannot be embedded in the diamond lattice). The reason for the growth of diamond above amorphous carbon is that this amorphous layer prevents the emergence of cobalt vapor. Figure 2 shows the SEM micrographs of indentations made by the diamond indenter on the diamond-
covered hard-alloy surface. The thickness of the samples was the same (7 μm), but their surface conditioning and deposition temperatures differed. It can be seen that adhesion is only good for the predecobalted coating deposited at 800°C (Fig. 2b). The coating on the sample that was not subjected to decobaltization and was deposited at the same temperature (800°C) has much weaker adhesion (Fig. 2a). The Raman spectrum taken for the lower side of the film covering this sample is identical to that shown in Fig. 1, spectrum с, i.e., the spectrum of nondiamond carbon, but the thickness of the coating here does not exceed 0.5 μm. The adhesion of the decobalted coating deposited at 1100°C remains very poor for the same reason (in this case, the thickness of the amorphous layer reached 0.9 μm). The above data suggest that without decobaltization the cobalt vapor pressure over the substrate is fairly high, even at low temperatures of diamond growth (600–700°C). In the case of decobaltization, the cobalt vapor pressure over the substrate only reaches the same values at high temperatures (above 1000°C). The surface elemental composition and surface roughness of the samples before deposition, as well as their surface micrographs, are presented in Figs. 3a and 3b. It can be seen that the cobalt content on the surface after decobaltization drops from 8.6% to zero. Table 2. Surface composition of diamond films adjacent to the hard-alloy substrate versus the deposition temperature with and without decobaltization Deposition temperature, °C
With decobaltization
Without decobaltization
710
X X X X O O
O O O O O O
820 900 960 1050 1110
X—diamond, O—nondiamond carbon. TECHNICAL PHYSICS
Vol. 63
No. 2
2018
INFLUENCE OF COBALT ON THE ADHESION STRENGTH
(a)
(b)
(с)
×3500 Element Co W C
×3500 wt % Ra, nm 8.6 50.6 74 40.8
(d)
5 µm
5 µm
5 µm
×3500
Element wt % Ra, nm Co 0 W 54.5 115 C 45.5
209
Element Co W C
5 µm ×3500
wt % Ra, nm 3.6 52.9 101 43.5
Element Co W C
wt % Ra, nm 50.5 27.3 35 22.2
Fig. 3. Surface elemental composition and surface roughness of the hard-alloy samples, as well as their surface micrographs, after (a) abrasive machining, (b) decobaltization, and (d) decobaltization with subsequent vacuum annealing at 1200°C. (c) Surface micrograph of the sample coated by the diamond film deposited at 700°C and annealed at 1200°C.
The surface roughness after decobaltization rises because the binder between crystallites is partially lost, which also improves the adhesion strength of the coating. To estimate the amount of cobalt has emerged on the surface of the covered and uncovered decobalted samples, they were annealed in the vacuum furnace at different temperatures. The thickness of the samples was the same, i.e., 7 μm. The results are presented in Fig. 4. The surface enrichment with cobalt on the coated and uncoated samples differs radically. After annealing at 1200°C, the cobalt content on the coated surface is no higher than 3.6 wt %, whereas that on the uncoated surface reaches 50.5 wt %. The surface micrographs, surface elemental composition, and surface roughness of these samples are presented in Figs. 3c and 3d. It follows from the Raman spectra taken of the lower sides of the coatings on the annealed samples that the amorphization of layer adjacent to the diamond surface is insignificant. Spectrum b in Fig. 1 was taken of the lower side of the film covering the sample grown at 700°C and annealed at 1200°C. It can be seen that the intensity of the diamond peak (1332 cm–1) slightly decreased and the fraction of the trans-polyacetylene peak (1450 cm–1) increased. In addition, the traces of vibrational states characteristic of nanodiamond (1140–1190 cm–1) [18] are observed. This variation in the film spectrum indicates the slow step cobalt-induced transformation of the diamond lattice. The adhesion of the diamond coatings on all annealed samples remained unchanged, and indentations had the same form as before annealing (Fig. 2b). This allows us to argue that the amorphization of the diamond coatings on the side adjacent to the substrate is insignificant upon annealing. TECHNICAL PHYSICS
Vol. 63
No. 2
2018
CONCLUSIONS Thus, the cobalt content on the hard-alloy surface considerably influences the phase composition of diamond coatings and this influence strongly depends on the deposition temperature. It was found that the catalytic amorphization of carbon is observed to the full extent only during the growth of the carbon coating. This is because the cobalt content in the gas phase over the substrate is high and prevents the normal growth of diamond crystallites. At the same time, the presence of cobalt on the already synthesized diamond layer does not radically change the phase composition of diamond even at 1200°C (no complete amorphization was observed). It was shown that the rate of cobalt emergence on the coated hard-alloy surface is an order of magnitude lower than for the uncoated surface. This Co content, wt % 60 Without diamond film With diamond film
50 40 30 20 10 0 500
600
700
800
900 1000 1100 1200 1300 Temperature, °C
Fig. 4. Cobalt content on the surface of the decobalted hard alloy with and without diamond coating vs. annealing temperature.
210
LINNIK et al.
indicates that the deposition of diamond coatings on hard-alloy substrates should first be carried out at temperatures no higher than 900°C; then, after a diamond coating is formed, the temperature may be raised without sacrificing adhesion to the substrate. It is necessary to decobalt the substrate prior to deposition, since direct contact with cobalt prevents the diamond growth, even at low deposition temperatures (
