Dynamic hardness of cemented tungsten carbides

Dynamic hardness of cemented tungsten carbides

International Journal of Refractory Metals & Hard Materials 75 (2018) 294–298 Contents lists available at ScienceDirect International Journal of Ref...

2MB Sizes 0 Downloads 34 Views

International Journal of Refractory Metals & Hard Materials 75 (2018) 294–298

Contents lists available at ScienceDirect

International Journal of Refractory Metals & Hard Materials journal homepage: www.elsevier.com/locate/IJRMHM

Dynamic hardness of cemented tungsten carbides a


Luke A. Hanner , John J. Pittari III , Jeffrey J. Swab a b c



Drexel University, Philadelphia, PA 19104, USA Oak Ridge Institute for Science and Education, Oak Ridge, TN 37831, USA US Army Research Laboratory, Aberdeen Proving Ground, MD 21005, USA



Keywords: Tungsten carbide-cobalt Tungsten carbide-chromium-nickel alloy “Binderless” tungsten carbide Knoop hardness Dynamic hardness testing Strain rate dependence

Cemented tungsten carbides are composed of a hard tungsten carbide (WC) phase held together by a soft, ductile binder phase, typically cobalt (Co). This study examines the role of the binder composition and content on the Knoop hardness at quasi-static and dynamic strain rates. Seven different tungsten carbide materials were tested: four WC-Co systems containing 10%, 15%, 20% and 25% cobalt, two with a chromium‑nickel alloy binder phase, and one “binderless” tungsten carbide. Quasi-static Knoop hardness testing was performed at strain rates of 10−3 s−1 and at indentation loads between 300 and 30,000 g using a Wilson Tukon 2100 unit. Dynamic Knoop hardness testing was conducted at strain rates of 103 s−1 using a Dynamic Indentation Hardness Tester over a range of indentation loads. All of the materials exhibited a rate-dependent Knoop hardness, with the hardness increasing by up to 60% with increasing strain rate.

1. Introduction Hardness is a measure of the resistance of a material to permanent deformation under an applied load. The problem with determining the hardness of ceramic materials is that ceramics are inherently brittle and exhibit minimal, if any, plastic deformation during the indentation process. This can lead to hardness indentations with excessive cracking and spalling, which complicates the hardness determination. The use of the Knoop hardness method has proven to be the most effective method of determining the hardness of many advanced ceramics over a broad range of indentation loads [1]. In the Knoop hardness method, a load is applied to an elongated, symmetric pyramidal diamond indenter, which is placed in contact with the material to be tested. This process causes deformation of the material leaving a permanent impression that is used to determine the hardness. The geometry of the Knoop indenter is such that the ratio of the long and short diagonals is approximately 1:7, while the ratio of the resulting long diagonal to the indentation depth is approximately 1:30, making the Knoop geometry useful for materials with a high likelihood of spallation and crack formation occurring during the indentation process. Dynamic testing at high strain rates (102–104 s−1) bridges the gap between the quasi-static laboratory testing and the very high strain

Corresponding author. E-mail address: jeff[email protected] (J.J. Swab).

https://doi.org/10.1016/j.ijrmhm.2018.05.007 Received 10 April 2018; Received in revised form 10 May 2018; Accepted 13 May 2018 Available online 14 May 2018 0263-4368/ © 2018 Published by Elsevier Ltd.

rates (105–106 s−1) experienced during high-speed machining [2] and ballistic impact events [3]. Rough bounds for these strain rate testing regimes are illustrated in Fig. 1. Many materials exhibit rate-dependent properties. Thus high strain rate testing is necessary to fully characterize and evaluate a material that may be exposed to such conditions. Previous work has established that many metals [4–8] and ceramics [4,6,9,10] exhibit a higher hardness during high-strain-rate indentation testing when compared to quasi-static hardness values. Cemented tungsten carbide materials have high strength, hardness, and fracture toughness; these properties are desirable for many engineering applications, including cutting tools for material machining and mining as well as military applications in armor-piercing projectiles. Binder content and tungsten carbide grain size can be used to tailor the hardness and fracture toughness for the intended application. Cobalt is the traditional binder material due to its ability to wet and dissolve the WC without forming a third phase [11]. However, the US Department of Health and Human Services' National Toxicology Program has listed Co as reasonably anticipated to be a human carcinogen [12]. These health concerns, combined with the fluctuation of global cobalt prices and availability due to economic and political instability in many of the countries [13–15] with cobalt deposits, have led to the exploration of alternative binder materials as replacements for cobalt.

International Journal of Refractory Metals & Hard Materials 75 (2018) 294–298

L.A. Hanner et al.

Table 1 Properties of WC materials. Material a

Cercom WC Kennametal S105 MPI 15 MPI 20 MPI 25 Kennametal KFY Kennametal BFY

Fig. 1. Strain rate domains ranging from normal quasi-static laboratory testing to very high strain rates in the ballistic realm.


Binder wt%

Average WC Grain Size (μm)

N/A Co Co Co Co CrNi3 CrNi3

≈0% 10% 15% 20% 25% 10% 10%

0.4 0.6 1.7 1.1 1.5 0.6 0.9


This is considered binderless; however, it does contain trace amounts of Co and V.

Fig. 2. Scanning electron microscopy (SEM) images of material microstructures: A) Cercom binderless WC, B) Kennametal S105 10% Co, C) MPI 15% Co, D) MPI 20% Co, E) MPI 25% Co, F) Kennametal BFY 10% CreNi alloy, and G) Kennametal KFY 10% CreNi alloy.


International Journal of Refractory Metals & Hard Materials 75 (2018) 294–298

L.A. Hanner et al.

350 Ω strain gage mounted at the midpoint of the bar. The strain is measured using a quarter Wheatstone bridge in a MicroMeasurements 2310B Signal Conditioning Amplifier. The load is transformed into an electrical signal using a high-frequency Kistler Type 9213b piezoelectric force sensor connected to a Kistler Type 5010 Dual-Mode Amplifier. Both the strain and load signals are captured using a PicoScope 5000 series oscilloscope. The average strain rate can found as the ratio of the indenter velocity (v) over the long diagonal (d) of the indent as shown in Eq. (2).

2. Materials Seven tungsten carbides, varying in binder composition, binder content, and grain size, as described in Table 1 and Fig. 2, were tested. The Cercom (now part of CoorsTek, Golden, CO)1 WC is considered “binderless” since it only contains trace amounts of Co and vanadium (V), the latter acting as a grain pinner. The Kennametal (Latrobe, PA) S105 WC contains 10 wt% Co, while the Materials Processing Inc. (MPI) (Irving, TX) materials range from 15 to 25 wt% Co. The Kennametal KFY and BFY tungsten carbide materials contain 10 wt% chromium‑nickel alloy as the binder instead of the traditional Co. X-ray diffraction analysis of these two tungsten carbides indicates the binder phase in both is CrNi3. The difference between the latter two materials is the tungsten carbide grain size, with the KFY having a slightly finer grain size than the BFY.

ε ̇ (t ) =

All materials were ground and polished, sequentially, using 9, 6, 3, and 1 μm diamond suspensions to achieve a smooth, mirror-like finish. This finish was needed to facilitate observation and accurate measurement of the long diagonal of each indentation. These polished surfaces were also examined with a scanning electron microscope (SEM) to determine the tungsten carbide grain size and binder distribution in each material. A Wilson Tukon Model 2100 MicroHardness Tester was used to perform quasi-static Knoop indentations at strain rates of 10−4 s−1 with a 15-s dwell time using indentation loads of 4.9 N, 9.8 N, 19.6 N, 49.0 N, 98.0 N, 147.1 N, 294.2 N, and 490.3 N (0.5 kgf, 1 kgf, 2 kgf, 5 kgf, 10 kgf, 15 kgf, 30 kgf, and 50 kgf, respectively). A minimum of ten indentations were made at each load on each material in accordance with the procedures outlined in ASTM C1326-13 [16]. The Knoop hardness (HK) was calculated using Eq. (1), where C is the indenter geometry constant (0.014229), P is the indentation load (N), and d is the long diagonal (mm).

P d2


In this investigation, indents were generated at average strain rates around 1500 s−1. These strain rates are similar to the 2200 s−1 reported by Subhash, [5] and are within the desired high-strain-rate regime of interest. Multiple indents were performed at a variety of loads in order to encompass as wide a characterization window as possible. Instances of double indentation and indents with significant spallation were discarded as these indents are considered unacceptable according to ASTM C1326 [16]. An optical microscope was used to measure the long diagonal of all valid indents generated in both the quasi-static and dynamic tests. The average quasi-static and dynamic hardness values for each tungsten carbide material were calculated by plotting the indentation load against the square of the long diagonal and then multiplying the slope of a linear fit trend line by the indenter constant (0.014229). This is a simple manipulation of Eq. (1), where the slope of the trend line is equal to P/d2, which can be used to calculate an average hardness value that is independent of the indentation load.

3. Experimental procedure

HK = C ×

v −c0 = εi (t ) d d

4. Results and discussion The quasi-static Knoop hardness values of the seven tungsten carbide materials are summarized in Fig. 3. It is not surprising that the binderless WC is the hardest of the seven materials, since it contains virtually no binder, while the hardness decreases as the Co content increases. The BFY and KFY both exhibit higher hardness values than all the WC materials containing Co, except for the one containing 10% Co. The larger grain size of the materials containing 15, 20, and 25 wt% Co


The high-strain-rate Knoop indentation tests were performed using a modified Dynamic Indentation Hardness Tester, similar to the device designed by Koeppel and Subhash [5,17]. The unit is based on a Kolsky (split-Hopkinson) Bar and uses a one-dimensional stress wave to generate the indentation. The incident bar is a 1220-mm long, 12.7-mm diameter 7075-T6 aluminum rod. A nitrogen gas gun fires a 100-mm long, 12.7-mm diameter aluminum striker bar, which makes contact with a 25.4-mm diameter flange at the front of the incident bar. The resulting compressive elastic pulse (known as the incident pulse), with a length of approximately 40 μs, travels down the length of the incident bar, forcing the sample mounted on the far end onto a stationary indenter that is attached to a load cell. A momentum trap assembly, consisting of a flange, sleeve, and rigid mass, captures and manipulates the wave to prevent multiple compressive pulses from reaching the sample, thus avoiding multiple indentation events. In this case, the larger diameter of the sleeve leads to a 4:1 ratio of impedance between the momentum trap sleeve and the incident bar. This increase in impedance causes a smaller portion of the original incident pulse to travel through the bar and a much larger portion of the pulse to travel into the momentum trap assembly. As a result, the reduced magnitude incident wave produces the indentation load while the much larger tensile wave from the momentum trap retracts the bar more efficiently to reduce the occurrence of multiple indentation events. The pulse in the incident bar is recorded using a Vishay Instruments 1 Mention of specific test equipment, materials, software, or test methodologies does not constitute an official endorsement by Drexel University, the Oak Ridge Institute for Science and Education or the US Army Research Laboratory.

Fig. 3. Quasi-static Knoop hardness over a range of indentation loads. 296

International Journal of Refractory Metals & Hard Materials 75 (2018) 294–298

L.A. Hanner et al.

materials, including the same binderless WC examined in this effort, but the ISE was most pronounced at indentation loads below 4.9 N. Since none of the WC materials exhibit a pronounced ISE for the quasi-static tests and because of the variability in the dynamic indentation load from test to test, it is best to compare the hardness by plotting the indentation load as a function of the diagonal squared. Fig. 4 shows data plotted for each of the tungsten carbide materials over both strain regimes. Each data set has been independently fit with a linear trend line. The plots show a clear rate-dependence for all seven tungsten carbides, as indenting at higher strain rates yielded significantly higher Knoop hardness values. The binderless WC only had a 12% increase in hardness, while the 10% Co material increased by more than 30%. When the Co content was 15% or higher, the increase was well in excess of 50%. The Knoop hardness of BFY and KFY increased by 39% and 20%, respectively. The finding of rate-dependent hardness for these materials is not surprising since other ceramics and metals also show a rate dependence. The increase shown by the binderless WC is comparable to the

may also contribute to this difference. The KFY was consistently harder than the BFY even though they both contain the same binder composition and binder content. This difference may be due to the slightly smaller grain size of the KFY material, as finer-grained materials typically have higher hardness values. The Knoop hardness at 490 N could only be measured on three of the tungsten carbides. Extensive cracking and spalling around the indents in the binderless WC made it impossible to measure the indent size. In the case of the BFY, KFY, and WC with 25% Co, the indentation load was so high that it pushed the entire diamond indenter, as well as some of the mounting material that holds the diamond indenter in place, into the WC. This resulted in impressions that bulged out on the sides making these indents invalid for hardness determination. Further analysis of the quasi-static hardness shows that none of the tungsten carbide materials examined exhibit a pronounced indentation size effect (ISE) (a decrease in hardness with increasing indentation load) that is commonly observed in many ceramics. A previous investigation [1] showed a pronounced ISE in several tungsten carbide

Fig. 4. Indentation load versus indent diagonal squared for both quasi-static and dynamic strain rates: A) Cercom binderless WC, B) Kennametal S105 10% Co, C) MPI 15% Co, D) MPI 20% Co, E) MPI 25% Co, F) Kennametal BFY 10% CreNi alloy, and G) Kennametal KFY 10% CreNi alloy. 297

International Journal of Refractory Metals & Hard Materials 75 (2018) 294–298

L.A. Hanner et al.

four materials with varying Co content and grain size, and two with a CreNi alloy as the binder material. All materials tested showed ratedependent Knoop hardness values where hardness increased with increasing strain rate when compared to hardness values determined at a quasi-static strain rate. Higher binder content and larger grain size resulted in a greater increase in dynamic over quasi-static hardness. In the four WC-Co materials tested, a linear relationship between quasi-static hardness and an increase in dynamic hardness was established. Increases in dynamic hardness over quasi-static hardness values of up to 60% were recorded. Acknowledgements The efforts of Luke Hanner were supported by the Army Research Laboratory (W911-SR-15-2-001) College Qualified Leaders program administered by The Academy of Applied Sciences. The efforts of Dr. John J. Pittari, III were supported in part by an appointment to the Postgraduate Research Participation Program at the US Army Research Laboratory (ARL) administered by the Oak Ridge Institute for Science and Education (120-1120-99) through an interagency agreement between the US Department of Energy and ARL.

Fig. 5. Relationship between cobalt content, quasi-static Knoop Hardness, and hardness change. BFY and KFY materials do not contain any cobalt but are included for comparison.

increase observed in other monolithic ceramics [7,9], and the increase exhibited by the 10% Co material and the BFY and KFY are slightly higher than, but comparable to, some of the metals examined. [5–7]. However, the increase of 50% or more when the Co content was 15% or more is significantly higher than any value previously reported. This significant increase may imply that Co is an extremely rate-dependent metal. Similar to the quasi-static results, the binder content and tungsten carbide grain size also influence the dynamic hardness. As the binder content and tungsten carbide grain size increased, the dynamic hardness decreased. The reduction in hardness with an increase in tungsten carbide grain size shows that these cemented carbides follow a HallPetch type of relationship. Fig. 5 shows that as there is a linear relationship between quasistatic Knoop hardness and the change in Knoop hardness from the quasi-static to dynamic strain rates. This relationship implies that cobalt has a higher strain rate sensitivity than tungsten carbide. Additionally as the Co content increased, the quasi-static Knoop hardness decreased and the difference between the quasi-static and dynamic hardness values increased. There was no clear evidence of any mechanistic change occurring in these materials as a function of strain rate or indentation load. Since many metals with a hexagonal close packed (HCP) crystal structure exhibit mechanical twinning at high strain rates [18] it was anticipated that since Co also has a HCP structure and has been shown to twin [19] that twins would be observed, especially in the material containing 25% Co. A preliminary examination inside and underneath the indentations using scanning and transmission electron microscopy revealed some subtle differences, but nothing definitive and no clear evidence that the Co had twinned. The only observed change in any of these materials occurred in the binderless material where catastrophic cracking, i.e., complete fracture of the specimen, occurred at the higher strain rates and indentation loads. None of the materials containing a binder phase exhibited a significant change in the amount of cracking as a function of indentation load or strain rate. This is not unexpected as the binderless WC is the most brittle, i.e., has the lowest fracture toughness, of the all of the materials examined in this study. [20]

References [1] J.J. Swab, Recommendations for determining the hardness of armor ceramics, Int. J. Appl. Ceram. Technol. 1 (3) (2004) 219–225. [2] J.P. Davim, C. Maranhão, A study of plastic strain and plastic strain rate in machining steel AISI 1045 using FEM analysis, Mater. Design. 3 (2009) 160–165. [3] S.M. Walley, Historical review of high strain rate and shock properties of ceramics relevant to their application in armour, Adv. Appl. Ceram. 109 (8) (2010) 446–466. [4] C.J. Fairbanks, R.S. Polvani, S.M. Weiderhorn, B.J. Hockey, B.R. Lawn, Rate effects in hardness, J. Mater. Sci. Lett. 1 (1982) 391–393. [5] G. Subhash, B.J. Koeppel, A. Chandra, Dynamic indentation hardness and rate sensitivity in metals, J. Eng. Mater. Technol. 121 (1999) 257–263. [6] B.J. Koeppel, G. Subhash, Characteristics of residual plastic zone under static and dynamic Vickers indentations, Wear 224 (1999) 56–67. [7] G. Subhash, Dynamic indentation testing, in: H. Kuhn, D. Medlin (Eds.), ASM Handbook, Mechanical Testing and Evaluation, vol. 8, ASM International, Materials Park, OH, 2000, pp. 519–529. [8] A.H. Almasri, G.Z. Voyiadjis, Effect of strain rate on the dynamic hardness of metals, J. Eng. Mater. Technol. 129 (2007) 505–512. [9] R.J. Anton, G. Subhash, Dynamic Vickers indentation of brittle materials, Wear 239 (2000) 27–35. [10] E.J. Haney, G. Subhash, Static and dynamic indentation response of basal and prism plane sapphire, J. Eur. Ceram. Soc. 31 (2011) 1713–1721. [11] P.A. Kuhn, R. Jenkins, G. Robinson Jr., Investigation of Substitutes for Cobalt in WC-Co Alloys and Effect of Grain Size on the Properties of WC-Co Alloys, MA: Manufacturing Laboratories, Inc, Boston, June, 1955 (Report WAL 330/13). [12] National Toxicology Program, Report on Carcinogens, Background Document for Cobalt-Tungsten Carbide: Powders and Hard Metals, Research Triangle Park (NC): US Department of Health and Human Services, March, 2009. [13] M. Seddon, The cobalt market—current volatility versus future stability? Appl. Earth Sci. 110 (2) (2001) 71–74. [14] N. Tsurukawa, S. Prakash, A. Manhart, Social Impacts of Artisanal Cobalt Mining in Katanga, Democratic Republic of Congo, Freiburg (Germany): Öko-Institut eV, 2011. [15] K.B. Shedd, E.A. McCullough, D.I. Blelwas, Global trends affecting the supply security of cobalt, Min. Eng. (December, 2017) 37–42. [16] ASTM C1326-13, Standard Test Method for Determining the Knoop Hardness of Advanced Ceramics, vol. 15, West Conshohocken (PA): ASTM International, 2017, p. 01. [17] Subhash G, Chandra A, Koeppel BJ, Inventors; Michigan Technological University, Assignee. Apparatus and Method for Determining the Dynamic Indentation Hardness of Materials. United States Patent US6343502B1. October 7, 2002. [18] Y.T. Zhu, X.Z. Liao, X.L. Wu, Deformation twinning in nanocyrstalline materials, Prog. Mater. Sci. 57 (2012) 1–62. [19] Zhang XY, Tu J and Liu Q, High-Resolution Electron Microscopy Study of the {10−11} Twin Boundary and Twinning Dislocation Analysis in Deformed Polycrystalline Cobalt. Scrip. Mat. [67] 991–994 (2012). [20] J.J. Swab, J.C. Wright, Application of ASTM C1421 to WC-Co fracture toughness measurements, Int. J. Refract. Met. Hard Mater. (58) (2016) 8–13.

5. Conclusions Quasi-static and dynamic Knoop hardness testing was performed on seven cemented tungsten carbides, including a binderless” material,