Novel Route for Preparing Diamond-Enhanced Cemented Carbides via Reactive Sintering

Abstract

Hardmetals are cemented carbides consisting of the hard ceramic phase WC and the ductile metallic binder Co. They offer an outstanding combination of hardness and fracture toughness. Hence, they have a widespread use across the manufacturing industry. However, due to the increasing requirements for tool material, the combination of the beneficial properties of hardmetal and diamond is a long sought-after objective. In this work, a new approach was evaluated to reduce the formation of graphite due to the phase transformation of diamond during the sintering of compounds together with hardmetal. Earlier trials could not fully suppress the phase transformation despite using alternative Ni-instead of conventional Co-based binder systems and field-assisted sintering (FAST) to reduce required sintering temperatures and time. To lower the amount of graphite formed during sintering even further, a reactive sintering process was developed. The increased sinter activity due to the in situ synthesis of WC has the potential to decrease the needed temperature to achieve a pore-free compact. For the first time, a WC-Ni hardmetal produced from elemental powders was successfully used as a matrix in a diamond-enhanced cemented carbide (DECC). Different approaches regarding carbon sources and the extent of reactive material were pursued. The introduction of a carbon deficit by adding metallic W to a mixture of WC and Ni, which is essentially partial reactive sintering, leads to an increased relative density compared to the reference of 97.3%.

1. Introduction

Hardmetals are cemented carbides, which consist of the hard ceramic phase tungsten carbide (WC) and the tough metallic binder cobalt (Co). Other cemented carbides and carbonitrides are used, like TiCN with NiCo binder. However, they do not exhibit the same favorable combination of hardness and toughness that WC-Co offers. Therefore, hardmetals are used for tools to work a wide variety of materials especially metal and stone as well as wear parts of all kinds. Since their invention 100 years ago, hardmetals have been developed further. Their properties can be altered via the addition of secondary carbides, i.e., from chromium, molybdenum, and tantalum, as well as the alternative binder metals nickel (Ni) and iron (Fe), which are usually alloyed with Co. However, binder content and WC-grain size mainly determine their hardness and toughness, which can be tuned in a wide range [1,2]. This variability makes hardmetals very versatile in their application range. They are used in many industrial branches like automotive, aerospace, the mining and construction sector, chemistry, as well as the agri-food and pharmaceutical sector, and many more. The development of relatively cost-effective methods to synthesize super hard materials like diamond and cBN has led to new tools that compete with hardmetals in some applications, not only for milling high-strength materials like Ni-based superalloys and intermetallics (e.g., titanium aluminides) but also for abrasive materials like fiber-reinforced plastics. Among others, these are used to an increasing extent in the production of sustainable energy and to satisfy the requirements for efficiency in energy transformation, which is supported by lightweight design. For diamond and cBN cutting and milling tools, a mono- or polycrystalline insert is usually used, which must be supported by a carrier from hardmetal due to their low toughness. This drawback limits their use for tools with a defined edge. Diamond is especially used to a great extent in the form of small grains embedded in a matrix of resin or metal alloy, which provides toughness but exhibits comparatively low hardness and strength. This limits their application to processes like grinding and sawing. The use of hardmetal as a matrix material for diamonds bears the potential to combine the beneficial properties of both materials. The main challenge to overcome is the phase transformation of the metastable diamond structure into graphite at ambient pressure during the sintering process. The transformation temperature is further lowered in contact with Co, Ni, and Fe [3]. Sintering temperature and time thus have to be lowered in order to minimize graphitization. Conventional sintering processes for hardmetals like vacuum sintering or SinterHIP take several hours and require a liquid phase to form a pore-free body. Therefore, these methods are not suitable for the consolidation of this novel compound material. With high-pressure high-temperature (HPHT) techniques, stabilization of the diamond structure is possible [4], but these do not offer a cost-efficient production. The development of DECC at conditions diamond is not stable, experienced a breakthrough via the invention of field-assisted sintering techniques (FAST), among which direct current sintering (first invented as spark plasma sintering, SPS) is especially useful. For electrically conductive materials, the consolidation takes place via direct joule heating of the powder. Additionally, pressure is applied to support compaction. Sintering times and temperature can be reduced significantly to achieve a fully dense compact without the formation of a liquid phase. This technique was used in this study as well because it helps to minimize graphitization. However, the formation of a graphite layer on the diamond grains could never be fully suppressed without the use of a coating [5,6,7,8]. By avoiding the metallic binder in the hardmetal, the temperature stability of the diamond is greatly improved. Since the required temperatures to consolidate the pure WC are much higher, graphite forms inevitably [9,10]. The adaptation of alternative binder systems provided further improvements. Ni and NiCu alloys, especially, proved to be a promising way [11,12]. In this work, the aforementioned approaches were combined with reactive sintering. Due to the in situ formation of WC, a high sinter activity bears the potential of lowering the sintering temperature needed for full compaction. There have been several attempts at reactive sintering of hardmetals with conventional processes with good success [13,14,15]. Also, for FAST, there have been a few promising trials [16,17]. Mostly, these processes were based on carbothermal reduction. Since the tool used for FAST is quite gas-tight, the excessive formation of large amounts of gaseous reaction products could lead to a catastrophic rupture of the die when quickly raising the temperature. Thus, only elemental powders from W, C, and Ni were used in this work. Investigations conducted via conventional sintering techniques proved their feasibility and resulted in bimodal microstructures and platelet formation regarding the WC grain morphology [18,19]. In combination with FAST processes, this field is rather unexplored [20]. Therefore, preliminary trials on mixtures without diamond addition were carried out. Three different starting compositions for the DECC were investigated, which can be differentiated by the amount of reactive material and the carbon source. The density, microstructure, phases, and composition were analyzed.

2. Materials and Methods

2.1. Materials

In order to easily transfer results into industrial processes, only commercially available powders were used, which can be commonly found in the hardmetal and diamond industry (

Table 1. Details on the used raw materials.

Powders	Average Particle Size/µm	Manufacturer
WC	1.4	H.C. Stark Tungsten GmbH, Goslar, DE
W	0.7	Wolfram Bergbau und Hütten AG, St. Martin i.S., AT
C (Carbon Black)	0.005	Degussa GmbH, Düsseldorf, DE
Ni	2.5	Eurotungstene, Grenoble, FR
Diamond	14–20	Vollstädt Diamant GmbH, Seddiner See, DE

). The chosen particle size of 1.4 µm for the WC powders is considered fine to medium to obtain a hardmetal that has a high hardness while maintaining a sufficient toughness. The smaller particles of the W powder (0.7 µm) support sinter activity due to the high specific surface area available for reaction with C. Additionally, “Carbon Black” was used, which is an especially reactive, amorphous form of carbon. To achieve good mechanical retention of the diamond particles in the hardmetal matrix, the diamond grains have an irregular non-facetted morphology.

With the introduction of brittle diamond particles, a decrease in the toughness of the compound was expected. Since the content of metallic binder has a high impact on the toughness, an amount of 20 vol.% was chosen for the pure hardmetal. This was kept equal for all compositions. By adding 30 vol.% of diamond, this theoretically results in a composite with 14 vol.% metal binder when considering WC and diamond together as hard phases and no free tungsten or other unwanted phases being apparent. However, in one composition, the diamond should be the carbon source itself. Therefore, no Carbon Black was added. To maintain a content of 30 vol.% of diamond in the sintered samples, more diamond powder was added to compensate for the loss due to the reactive formation of WC.

Table 2. Composition of powders and theoretical densities (TD) of sintered compacts; the diamond content refers to the total share in the DECC.

Designation	Content/w.%					TD/g/cm3
	W	WC	C	Ni	Dia.	HM	HM + Dia.
(W+5.37C)-Ni	82.18	-	5.37	12.45	-	14.318	-
(W+5.55C)-Ni [+D]	82.05	-	5.55	12.40	9.74	14.318	11.079
W-Ni [+D]	86.85	-	-	13.15	14.39	-	11.079
WC-Ni+W [+D]	15.46	71.96	-	12.58	9.36	-	11.393 1
WC-Ni [+D] (Ref.)	-	87.55	-	12.45	9.53	14.318	11.079

¹ Higher TD because C consumption due to formation of WC was not compensated by increasing the share of diamond.

summarizes all compositions.

Theoretical densities for comparison purposes were calculated from literature values of the starting materials by using the rule of mixtures assuming a complete reaction of W to WC. In contrast, the assumption that the total amount of W remains unreacted would result in lower TDs. The true TD will likely be somewhere in between. The partial solution of W in Ni, which is dependent on the share of C, complicates the calculation further. For compositions with diamonds, the first approach is, however, roughly correct for the hardmetal matrix. Here, the formation of graphite will lower the density and cannot be predicted in advance. Assuming a surplus of C and with no observation of pores for the compared specimen, the difference in graphitization can be roughly estimated from density measurements in the boundaries of measurement error.

2.2. Powder Preparation and Sintering Process

The powder mixtures for the hardmetal matrix were prepared via the conventional powder metallurgic route: milling for 24 h hours in a horizontal ball mill in a nitrogen atmosphere with n-heptane as solvent. Hardmetal balls (4 mm) were added with a ball-to-powder ratio of 1 to 10. The solvent was removed afterward using a vacuum-drying furnace. The powder was granulated and separated from the milling balls by sieving the mixture (315 µm mesh). Mixing with the diamond powder was subsequently performed by dry mixing them for 1 h with the addition of hardmetal balls in a nitrogen atmosphere. To prevent crushing and a change in the size distribution of diamond grains, a low rotational speed of 60 r min⁻¹ and a 1 to 5 powder-to-ball ratio was used. The powder was consolidated via FAST. It was manually filled into cylindrical graphite tools with graphite foil in-between punches and die (inner Ø: 20 mm) and pre-compacted with 10 kN. The amount of powder was chosen to result in a specimen height of 5 mm without pores or unwanted phases. This ensured there was enough distance from the edge-induced inhomogeneities caused by temperature and pressure gradients, as well as absorbed carbon from the graphite foil. Sintering was conducted with continuous, direct current. The time the diamond was exposed to temperatures at which graphitization takes place should be minimized. Therefore, a short dwell of 5 min and a high heating rate of 300 K/min were used. The small tools and powder mass would allow much higher heating rates. However, temperature gradients would increase accordingly. By wrapping the die into a graphite felt, excessive heat loss was prevented to maintain a more even temperature distribution. Temperatures during dwell varied between 1100 and 1180 °C. To ensure a full reaction of W and C, one trial with each of the pure hardmetal powders was performed at 1250 °C for 30 min. Sintering was conducted in a vacuum with a slight argon partial pressure (ca. 8 Pa) due to flushing of the tube for pyrometric measurement. Temperature was measured using a pyrometer at the bottom of the hollow top punch, which has a wall thickness of only 5 mm at this point. To allow enough time for degassing, until a temperature of 600 °C, only 100 K/min was used. At 50 °C before reaching the maximum temperature, the heating rate was reduced to 100 K/min to minimize overshooting. During those 30 s, the piston force was increased from the initial 10 kN to 25 kN (32 MPa to 80 MPa for Ø 20 mm die). Among other parameters, temperature, displacement, and chamber pressure were recorded.

2.3. Specimen Preparation and Characterization Methods

The sintered specimens were sandblasted and ultrasonically cleaned in ethanol prior to the density measurement according to ISO 3369 (Archimedes principle, repeatability: ±0.02 g/cm³). Pure hardmetal samples were cut, embedded into resin, and polished to a mirror finish using diamond-based suspensions. Diamond-containing samples were fractured, and a small area was polished by means of ion beam slope cutting (IBSC). This method made it possible to study the interface between diamond and hardmetal. Mechanical polishing did not yield a planar surface. The softer hardmetal around the diamond were removed in excess, and the diamond grains were pulled out of the surface and caused scratch marks. The microstructures were examined via field emission scanning electron microscopy (FE-SEM; ULTRA 55-Carl Zeiss Microscopy GmbH, Jena, DE). The FE-SEM was also used for elemental analysis in combination with a detector for energy-dispersive X-ray spectroscopy (EDS; X-MaxN 80-Oxford Instruments plc, Abington, GB). Phase analysis was carried out via X-ray diffraction (XRD; Bruker AXS). Measurements of carbon content were conducted via analysis of gases from the combustion of the powdered material in pure oxygen (WC600, Leco, St. Joseph, MI, USA). As a reference material, high-purity WC (ECRM No. 783-1, BAS Ltd., Middlesbrough, GB) with a defined C content of 6.188 ± 0.013 w.% was used. Theoretical calculations regarding phase diagrams and phase compositions were conducted via the CALPHAD method using FactSage 8.0 (GGT-Technologies, Herzogenrath, DE) in conjunction with the databases SGPS 2019 (v13.1) for pure substances and SGTE 2014 for intermetallic compounds and alloys (Scientific Group Thermodata Europe, Saint Sulpice, FR).

3. Results and Discussion

3.1. WC-Ni Hardmetal

Before the enhancement with diamond particles, the pure reactively sintered hardmetal was prepared to verify the feasibility of the suggested method. The two compositions, which vary in their carbon content, were sintered at 1100, 1140, and 1180 °C and 5 min dwell. Additionally, a set of sinter parameters (1250 °C, 30 min) was chosen to ensure the reaction of W and C to be finished during sintering.

The lower C content was chosen to be 5.37 w.%, the theoretical value for WC-12.45 w.% Ni (Figure 1). In the composition with more C, the addition should compensate for the higher C loss during sintering compared to the non-reactively sintered WC-Ni (

Table 3. C loss during sintering.

Specimen Designation	Sintering Parameters	C Loss/w.%
WC-Ni	1140 °C, 5 min	0.08
(W+5.37C)-Ni	1140 °C, 5 min	0.20
(W+5.37C)-Ni	1180 °C, 5 min	0.18

). The chosen amount of 5.55 w.% should avoid the formation of Eta phase while not being in the range of free graphite.

XRD analysis (Figure 2) reveals that no free, unreacted W is present even at the lowest sintering temperature. The composition with the lower C content sintered at 1250 °C shows the presence of a small amount of the Eta phase, which will occur at lower C content according to the calculated phase diagram (Figure 1). Other than that, no ternary phases besides WC and the Ni binder are visible. Notably, the peaks for the Ni binder phase are shifted toward smaller angles, i.e., higher lattice parameters for the sintered specimens in contrast to the powder mixture where the positions match nearly perfectly to the refined pattern of pure Ni (fcc). This implies dissolved W in the binder. Comparing the two compositions, it can be seen that the shift for the less C-containing specimens is slightly larger. This is in line with the calculated phase composition of the Ni binder, which has a high range of solubility for W depending on the available C of ca. 4.1–12.7 at.% within the boundaries of the two-phase area at 800 °C. EDS mappings are backing up the findings by XRD (Figure 3). Apart from different WC grain sizes and morphology, the two different compositions sintered with different parameters look quite similar in terms of elemental distribution. Even at the lowest sintering temperature and C content, no areas of higher W concentration than within the WC can be seen. It is assumed that W either reacted to form WC or is dissolved within the Ni binder. The same applies to the “Carbon Black”. Even at higher temperatures and dwell, no carbon was precipitated as graphite.

Looking at the SEM SE images (Figure 4), it becomes clear that in some specimens, residual porosity is still visible. Seemingly, fewer pores are apparent in the specimens sintered at 1100 °C compared to 1180 °C. However, when taking a closer look at (W+5.37C)-Ni, 1100 °C at a higher magnification, nanometer-sized pores become visible (Figure 5). At the highest sintering temperature and longest dwell, some porosity is still visible in (W+5.55C)-Ni in comparison to the less C-containing composition, which is fully densified.

Figure 6 shows the evolution of the microstructure of both compositions at different sintering parameters. From 1100 to 1180 °C, and among the two compositions, the grain size does not differ much. Mostly, the sub-micron portion has decreased, while bigger-sized grains did not grow much. In comparison, the specimen consolidated at 1250 °C for 30 min shows an overall increased grain size. Here, the difference between the two carbon contents becomes obvious. The WC grains are distinctively bigger with more C. This can be attributed to a higher sinter activity in the later stage of the sintering process and/or a higher volume of formed WC because of the available carbon. Inferring from the formation of the Eta phase (Figure 7) due to a carbon deficit in (W+5.37C)-Ni, the Ni binder should be saturated with W. Both phenomena will lower the amount of WC. The morphology of the WC grains changes as well with sintering parameters. While at lower temperatures, a lot of platelet-like grains can be observed, at higher temperatures, mostly rectangular and triangular shapes can be seen. The distribution of shapes indicates no major preferred orientation (the specimens are all cut radially through the middle). This is backed up by XRD (Figure 2): No significant change in relative diffraction peak intensities in relation to each other in comparison to the reference data can be observed.

The diagram below (Figure 8) shows the relative densities (RD) for the different sintering conditions and compositions, with a non-reactively sintered WC-Ni hardmetal for comparison. For the reference hardmetal form WC and Ni powders, it was possible to achieve full density at 1140 °C. The reactively sintered compositions achieve lower densities at comparable sintering conditions, which increases with sintering temperature and dwell time. Since no phases that would lower the density were detected, the lower values can be attributed to residual porosity. At 1100 °C at high magnification, nano-porosity is visible (Figure 5). Considering the measured density, the size of the pores initially increases with higher temperature while porosity in total decreases. Upon rising the temperature to 1250 °C and dwell time to 30 min, the pores are smaller or not apparent, which results in the highest densities. At this sintering condition, only the composition with less C shows an RD above 100% due to the formation of the Eta phase, which has a higher density than WC-Ni. The addition of more Carbon Black did not lead to an overall increased densification.

3.2. DECC

Three reactively sintered DECC with WC-Ni hardmetal matrix were prepared via FAST and compared to a similar non-reactively sintered system which was investigated in an earlier study [11]. The powder mixtures were composed to mimic the reference when consolidated. The first one was based on the findings from the prior studies of pure hardmetal, and thus (W+5.55C)-Ni was used for the DECC. The second did not contain Carbon Black as a C source. Instead, more diamonds were added as a C source for WC formation. Only W and Ni were mixed with diamonds. This was carried out to gain a better understanding of the graphitization of diamonds related to the formation kinetics of WC. The third approach can be described as partial reactive sintering. W was added to a conventional WC-Ni mixture to create a carbon deficit of 1 w.%. However, no additional diamond over the 30 vol.% was added to compensate for the loss due to WC formation because the exact amount of C from the diamond that will be part of the reaction was not possible to predict in advance and to verify after consolidation.

3.2.1. (W+5.55C)-Ni+D

The powder was sintered at 1180 °C for 5 min. Via SEM examination, no residual porosity is visible (Figure 9). WC grain size and morphology are similar to the pure in situ prepared hardmetal. Angular shapes are visible, and some platelets can be seen. Notably, there are Ni binder “lakes” where no WC grains can be found, which did not occur in the hardmetal without diamond addition. These WC-free volumes seem to appear especially within a cluster of diamond grains in their proximity (Figure 9, red circles). When taking a closer look at the different forms of carbon in the specimen, it is obvious that next to the diamond, two other morphologies are apparent: a compact and a very porous form. The first one can be identified as graphite formed upon degradation from diamond during sintering. Most likely, the second is residual Carbon Black, which did not react with W. The significant amount is probably caused by the competitive participation of the graphite in the formation of WC.

3.2.2. W-Ni+D

The results from the Carbon Black-containing specimen showed that a considerable amount of the C from the diamond must have been part of the WC formation. To study this effect separately, this trial without the addition of Carbon Black was performed. SEM images show a slightly greater formation of graphite compared to the reference material (Figure 10). More graphite can be found on the surface of the diamond as well as within the hardmetal. Additionally, the edges of the diamond seem to be more rounded. This could be attributed to the, at least initially, much larger diamond surface exposed to the matrix material due to its higher share. However, the exact amount was not possible to quantify.

Looking at the matrix material, it is obvious that the formation of WC could be achieved with the carbon from the diamonds. This is supported by XRD examination (Figure 11). Only peaks for WC, Ni, and diamond could be identified. Additionally, there is a peak visible which aligns well with the diffraction angle connected to the strongest intensity from the reference data for graphite. Most notable is the very different morphology of the WC grains compared to the specimen with Carbon Black addition (Figure 10). They have rounded edges and Ni inclusions, as well as small pores within the grains. This gives a hint toward the kinetics: The carbon source (diamond) is not as finely distributed and not as reactive as the mainly amorphous Carbon Black. Thus, the formation of WC might take place via the intermediate phases like an intermetallic compound and the mixed carbide Eta phase (M₆C). With the ongoing sintering process, more carbon will be available, leading to further reaction into Ni and WC. Since sinter activity has already decreased at this stage, driving forces are not high enough to form rectangular shapes. This could additionally be suppressed by the simultaneous dissolution of the intermediate phase: Not only does C react with W but also Ni has to be separated, restraining the rearrangement of W. Since C diffuses quicker in a solid state than W due to its smaller atoms C will travel from the diamond toward the areas with C deficit rather than the other way around. Thus, the intermediate phases will react from the outside to the inside, which can trap the released Ni within WC grains.

The formation of hollow tubular structures protruding from the surface into the diamond illustrates how aggressive the W-Ni mixture was on the diamond (Figure 12 and Figure 13). A similar phenomenon was recently described for a Ni-diamond compound on a smaller scale [21]. These tubes were only present in specimens from this composition. The walls of these tubes are partially covered with matrix material, which is not present in every structure and therefore most likely is a redeposit from IBSC. However, at the bottom of each tube, there is a lens containing W and Ni (Figure 13, right).

3.2.3. WC-Ni+W+D

From the two trials before the third composition was developed. The addition of W to a conventional WC-Ni mixture should help to increase the sintering activity and absorb the inevitably forming graphite by reaction to WC. This can be described as partial reactive sintering. In comparison to the reference samples the microstructure looks similar. Both SEM micrographs show a graphite layer around the diamond grains which is a few hundred nanometers thin as well as graphite precipitations within the binder area (Figure 14). The WC grains are slightly more angular for the composition with W addition, which gives an indication that some WC precipitated and rearranged during sintering. Comparing the specimen sintered at 1140 °C with the reference DECC at 1160 °C, less graphite can be found within the matrix and on the surface of the diamond grains of the reactively sintered sample.

Figure 15 illustrates the densities of all three reactively sintered samples and the reference for comparison. The sample with Carbon Black addition had the lowest density at 1180 °C despite having close to no visible porosity because of the large amount of free carbon in the form of graphite and remaining Carbon Black. Therefore, these trials were not continued. The composition with W addition and thus a slight carbon deficit achieved nearly the same maximum RD of the reference (97.4 vs. 97.5 %). However, at 1140 °C, WC-Ni+W+D had a higher RD. Together with the results from SEM micrographs, the potential to reduce the amount of graphite is feasible. W-Ni+D achieved lower RDs at all temperatures than WC-Ni+D and WC-Ni+W+D. This is likely caused by the hollow tubes, which developed due to the attack of W-Ni on the diamond grains, in addition to a higher amount of free graphite.

4. Conclusions

Previous trials on the diamond-enhanced cemented carbides mainly focused on conventional WC-Co hardmetal as a binder matrix. In this work, reactively sintered diamond-enhanced cemented carbides with an alternative Ni binder have been successfully prepared via FAST and characterized in terms of densification behavior and microstructural properties. The preliminary trials without the addition of diamond showed that WC-Ni hardmetal could be prepared from elemental powders via FAST in the solid state and with a very short dwell compared to conventional sintering techniques. Even with the lower reactivity of diamond compared to Carbon Black, W reacted to the carbide in W-Ni+D to form a hardmetal matrix during sintering. This can be attributed to the properties of the Ni binder. The good solubility and the quick diffusion of C in Ni enable the distribution of C throughout the matrix material during sintering. The different morphology of WC grains in comparison to the other compositions can be a sign of the different kinetics regarding the availability of carbon and the formation of intermediate phases, but it is still a subject of discussion. It could be shown that with partial reactive sintering with a diamond as the carbon source, the temperature needed for full densification could be lowered down to 1140 °C with less visible graphitization compared to the reference WC-Ni hardmetal matrix in dense samples. In future investigations, the promising approach of partial reactive sintering will be followed. The aim is the optimization of sintering parameters while using coated diamond grains to suppress graphitization to a minimum.