- Email: [email protected]
Phase response of WC–Ni to cyclic compressive loading and its relation to toughness
Int. Journal of Refractory Metals & Hard Materials 27 (2009) 313–316
Contents lists available at ScienceDirect
Int. Journal of Refractory Metals & Hard Materials journal homepage: www.elsevier.com/locate/IJRMHM
Phase response of WC–Ni to cyclic compressive loading and its relation to toughness A.D. Krawitz a,*, A.M. Venter b, E.F. Drake c, S.B. Luyckx d, B. Clausen e a
Mechanical and Aerospace Engineering, Lafferre Hall, University of Missouri, Columbia, MO 65211, USA Necsa Limited, Pretoria, P.O. Box 582, Pretoria 0001, South Africa c ReedHycalog, Houston, TX 77252, USA d Chemical and Metallurgical Engineering, University of the Witwatersrand, Johannesburg, South Africa e Los Alamos National Laboratory, Los Alamos, NM 87545, USA b
a r t i c l e
i n f o
Article history: Received 15 August 2008 Accepted 22 November 2008
Keywords: Residual stress Neutron diffraction Cemented carbide composites Mechanical behavior
a b s t r a c t The interaction of uniaxial compressive load and thermal residual stress was measured in a WC–10 wt.% (16 vol.%) Ni cemented carbide composite using neutron diffraction. Loading was from 0 to 2500 MPa in increments of 250 MPa, and measurements were made in situ during load–unload cycles 1, 2, 3, 10, 25, 50 and 100. Plasticity is observed in the Ni from the lowest levels of applied load, leading to continuous curvature of the WC–Ni stress–strain curves, and is believed to be a significant contribution to the composite’s toughness. It is due to interaction between local extremes of the thermal residual microstress with the applied macrostress and leads to anisotropic relaxation of the thermal residual stress. Strain distribution and plasticity were observed through peak breadths. Although the initially strong hysteresis is reduced as the cycles increase, there are still changes taking place after 100 cycles. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Cemented carbide composites have interesting and important mechanical and physical properties. These include high toughness for materials of such high hardness [1], stress–strain curves that are continuously curved [2,3], and very large thermal residual stresses in the as-produced state [4]. In this paper, we have used neutron diffraction to follow the strain response in both phases of a WC–10 wt.% Ni cemented carbide under uniaxial compressive load. Because the absorption of neutrons is low in many engineering materials, including W, bulk average values of strain can be obtained – during loading and unloading – from both the WC and Ni phases through observation of changes in diffraction peak positions. These are changes in the elastic component of strain. Information concerning the distribution of these elastic strains, as well as plasticity, can be observed through diffraction peak breadth behavior. Previous work on WC–Ni, WC–Co and WC–(Co,Ni) has revealed three distinctive characteristics of thermal residual stresses in cemented carbides. First, they are ‘‘diffraction hydrostatic” in character, which means that, when averaged over many particles and binder regions, the same stress values are measured in any physical direction. This is due to the fine scale nature of the microstructure and the fact that, when the Ni binder surrounding the micron* Corresponding author. E-mail address: [email protected] (A.D. Krawitz). 0263-4368/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijrmhm.2008.11.010
sized WC particles tries to shrink during cooling from the sintering temperature, it is severely constrained by the many surrounding WC particles. The constraint leads to a largely hydrostatic stress state, when averaged over many particles. The degree of constraint is closely related to the WC particle size; the smaller the particles, the greater the constraint. In a study of WC–Co material, samples with 20 vol.% Co had thermal residual stress levels for the WC that ranged from 755 MPa (±65 MPa) for 0.6 lm WC to 383 MPa (±66 MPa) for 5.1 lm WC particles [5]. The counterbalancing Co values are +1712 MPa (±147 MPa) for the 0.6 lm WC and +869 MPa (±150 MPa) for the 5.1 lm WC. Thus, the magnitude of the thermal stress doubles as the WC particle size decreases from 5 to 0.6 lm. For the present composite, the values are 382 MPa for the WC and +2022 MPa for the Ni. Second, in spite of the scalar character of the stresses when averaged over many particles, they vary dramatically from point to point in both the particles and binder. This is revealed through diffraction peak breadths, which are significantly broadened by the presence of a range of strains in both particles and matrix regions. If, for example, the temperature is lowered, both the thermal stress magnitudes and breadth of distribution increase [4]. An FE study has addressed the issue of stress distribution, in two dimensions, and found, in agreement with diffraction data, that the range of stresses in angular WC particles is large [6]. Thus, even though the average stress is compressive, there are local regions with tensile stress and others with very high magnitudes of compressive stress.
A.D. Krawitz et al. / Int. Journal of Refractory Metals & Hard Materials 27 (2009) 313–316
2. Experimental The sample was WC–10 wt.% (15.9 vol.%) Ni and was produced via liquid phase sintering and hot isostatic pressing. It had an average WC particle size of 1.17 lm, a binder mean free path of 0.84 lm, a contiguity of 0.48, a Rockwell A hardness of 88.6, and a density of 14.56 g/cm3. The in situ stress measurements were made at the Los Alamos Neutron Science Center (LANSCE) using the SMARTS spectrometer. LANSCE is a pulsed neutron source that measures the time-of-flight of neutrons from a white beam source for which the diffraction angles are fixed and there is a range of wavelengths from 0.05 to 0.5 nm. There are detector banks centered at ±90° so that data can be collected simultaneously in the axial and transverse directions of a cylindrical sample. The cylindrical sample was 8 mm in diameter by 19.2 mm in length. Count times were about a half hour per point. The samples were loaded in uniaxial compression from 0 to 2500 MPa to 0 in increments of 250 MPa. Diffraction measurements were made during load–unload cycles 1, 2, 3, 10, 25, 50 and 100. Strain was measured in both phases simultaneously, in directions parallel and transverse to the loading axis of cylindrical samples.
0
Ni -500
Applied stress (MPa)
Third, the pre-existing thermal residual stresses interact with applied stresses. This interaction is the primary subject of this paper. A previous FE study of a WC–10 wt.% Co sample has modeled the interaction of the thermal residual stresses with an applied uniaxial compressive load [7]. The FE model captured the general features of the measured response.
Transverse direction -1000
-1500
-2000
-2500 -5000
Cycle 1 Load Cycle 1 Unload Cycle 100 Load Cycle 100 Unload -4000
- 3000
-2000
- 1000
0
1000
Elastic microstrain (με)
(a) Ni response in the transverse direction
0
Applied stress (MPa)
314
-500
Cycle 1 Load Cycle 1 Unload Cycle 100 Load Cycle 100 Unload
-1000
-1500
Ni
-2000
Axial direction 3. Results and discussion
-2500
The macroscopic stress–strain curves for cycles 1 and 100 are shown in Fig. 1. They show curvature of the loading portion of each cycle, hysteresis, and a decreasing residual strain at the end of each cycle. The individual phase responses are shown in Fig. 2. In the transverse direction, Fig. 2a, there is an initial Poisson tensile response due to the compressive axial loading. However, the strain begins to reverse while the composite is still undergoing increasing tensile strain in the transverse direction. For cycle 1, the reversal begins at about 750 MPa and becomes more compressive than the initial value beginning at an applied load of about 1500 MPa. Upon unloading, there is a mild reverse curvature and a large net negative change in the strain from the starting value. In the axial direction, Fig. 2b, the strain is compressive
Applied stress (MPa)
0
cycle 1 load cycle 1 unload cycle 100 load cycle 100 unload
-500
-1000
-1500
-2000
-2500 -0.6
-0.5
-0.4
-0.3
- 0.2
-0.1
0.0
Macroscopic strain (%) Fig. 1. The macroscopic stress–strain curves for cycles 1 and 100.
-5000
-4000
-3000
-2000
-1000
0
1000
Elastic microstrain (με) (b) Ni response in the axial direction Fig. 2. The applied composite stress vs. Ni phase strain in the axial and transverse directions for cycles 1 and 100. Microstrain is strain times 106.
throughout but the rate of decrease begins to slow at about 750 MPa. Upon unloading, there is again a mild reversed curvature and a net tensile increase in the strain relative to the starting value. The observed response is due to plasticity that occurs locally throughout the volume of the Ni phase, at sites in the Ni that experience shear yield due to the applied stress. The changes are greater in the transverse direction because the applied Poisson strain is tensile in this direction and adds to the pre-existing thermal residual strain, which is, on average, strongly tensile in the Ni. The reversal of curvature in the transverse direction indicates that most, if not all, of the changes are taking place during the loading portion of the cycle. The degree of net strain relaxation per cycle is shown in Fig. 3a. The changes slow dramatically after 10 cycles and by 50 cycles the system has entered a stable cyclic regime. After 100 cycles, the residual stress in the Ni has been lowered to +1750 MPa in the axial and +1363 MPa in the transverse direction compared to an initial value of +2022 MPa. The system behaves essentially elastically until the stress magnitude is raised or a crack develops. Evidence of the response when the stress magnitude is increased is shown in Fig. 4, which presents the Ni phase response in the transverse direction. This was obtained from another sample from the same batch that was loaded in three sequential steps: (1) three load–unload cycles to 500 MPa in steps of 100 MPa; (2) three load–unload cycles to 1000 MPa in steps of 200 MPa; and, (3) three load–unload cycles to 2000 MPa in steps of 200 MPa [8]. The general behavior
315
A.D. Krawitz et al. / Int. Journal of Refractory Metals & Hard Materials 27 (2009) 313–316
1500
-500 500
Applied stress (MPa)
Change in microstrain ( με)
0 1000
0 -500
Change in axial strain Change in transverse strain
-1000
-1000
Ni -1500
Axial
-2000
Cycle 1 Load Cycle 1 Unload Cycle 100 Load Cycle 100 Unload
-1500
-2500 -2000 0
20
40
60
80
100
60
90 100 110 120 130 Normalized breadth (a) Normalized average peak breadth for Ni in the axial direction
Cycles Fig. 3. The changes in elastic strain in the Ni phase from their initial values after cycles 1, 2, 3, 10, 25, 50 and 100 in the axial and transverse directions. Microstrain is strain times 106.
70
80
0
Ni -500
500 MPa
1000 MPa
-500
1st cycle at 1000 MPa
-1000
Transverse -1000
-1500
Cycle 1 Load Cycle 1 Unload Cycle 100 Load Cycle 100 Unload
-2000
-2500
-1500
2000 MPa
60 1st cycle at 2000 MPa
-2000
-1200
-800
-400
0
70
80
90
100
110
120
130
Normalized breadth (b) Normalized average peak breadth for Ni in the transverse direction
400
Microstrain (με) Fig. 4. Results due to step loading, in the transverse direction, for a WC–10 wt.% Ni sample. At each load level ( 500 MPa, 1000 MPa and 2000 MPa), three load– unload cycles were performed. The load increments were 100 MPa at the 500 MPa level, and 200 MPa at the other two. Microstrain is strain times 106.
is the same at each level. There is hysteresis and a decrease in the residual stress of the system at each load level. Three cycles is insufficient for the changes in residual stress to reach the plateaus shown in Fig. 3, but the response is the same – new plasticity at the higher stress magnitude. It is particularly noticeable for the transition from 1000 to 2000 MPa. The diffraction peak breadth response of Ni for cycles 1 and 100 is shown in Fig. 5. In the axial direction, Fig. 5a, the peak breadth broadens even though the applied compression acts with the opposite sign to the residual tensile stress in the Ni, indicating that the stress variance is increasing even though the mean stress is decreasing. This indicates that even though plastic flow is occurring at specific sites in the microstructure, other regions retain their elastic character and do not flow. The rate of increase decreases steadily with increasing applied stress magnitude. Upon relaxation, there is a linear decrease in the breadths, indicating that the elastic regions whose stress variance has increased due to the compressive loading are relaxing during unloading. The peak breadth at the end of the unloading portion of cycle 1 is much lower due to a net relaxation of the stress variance due to the local plastic flow. During cycle 100, a reversible process is seen, due to the stable cyclic
Fig. 5. Values of normalized average peak breadth for Ni during cycles 1 and 100 in the axial and transverse directions.
regime that has been established. In the transverse direction, for cycle 1, the peak breadth decreases with load up to about 1500 MPa, when it begins a slow, small increase. The strain variance decreases due to plastic flow at the extremities of the local strain distributions. Apparently, new constraint develops in the la-
100
Normalized Ni peak breadth
Applied stress (MPa)
0
1st cycle at 500 MPa
Applied stress (MPa)
cycles increasing
95 90 85
transverse axial
80 75 70 0
20
40
60
80
100
Cycles Change in normalized peak breadth of the Ni phase as a function of cycle. Fig. 6. The change in normalized peak breadths in the Ni phase from their initial values after cycles 1, 2, 3, 10, 25, 50 and 100 in the axial and transverse directions.
316
A.D. Krawitz et al. / Int. Journal of Refractory Metals & Hard Materials 27 (2009) 313–316
ter part of the loading. As for the axial direction, there is a linear decrease in peak breadth and a net decrease in peak breadth after unloading. The response during cycle 100 is basically elastic. Fig. 6 shows the normalized peak breadths at the ends of the measured cycles. There is net relaxation for both directions although the transverse direction shows somewhat greater relaxation, as expected.
stress relaxations can be expected to affect the shape of the macroscopic elastic stress field around notches. It has been shown that an increase in applied stress magnitude leads to a new flow response even if the region had become stable relative to a lower stress magnitude.
4. Conclusions
[1] Uphadaya GS. Cemented tungsten carbides: production, properties, and testing. Westwood, NJ: Noyes Publications; 1998. [2] Felgar RP, Lubahn JD. The behavior of cemented carbides. Proc Am Soc Test Mater 1957;57:770–90. [3] Drake EF, Krawitz AD. Fatigue damage in a WC–nickel cemented carbide composite. Metall Trans A 1981;12A:505–13. [4] Krawitz AD, Reichel DG, Hitterman R. Residual stress and stress distribution in a WC–Ni composite. Mater Sci Eng 1989;A119:127–34. [5] Coats DL, Krawitz AD. Effect of particle size on thermal residual stress in WC–Co composites. Mater Sci Eng 2003;A359:338–42. [6] Weisbrook CM, Krawitz AD. Thermal residual stress distribution in WC–Ni composites. Mater Sci Eng 1996;A209:318–28. [7] Livescu V, Clausen B, Paggett JW, Krawitz AD, Drake EF, Bourke MAM. Measurement and modeling of room temperature co-deformation in WC– 10 wt.% Co. Mater Sci Eng 2005;A399:134–40. [8] Paggett JW. Neutron diffraction study of load response and residual stresses in WC–(Ni/Co) composites. PhD dissertation, University of Missouri, Columbia, MO, 2005.
Interactions between pre-existing thermal residual stresses and applied load have been observed in a WC–10 wt.% Ni. The local plasticity induced by the application of external stress is an energy-absorbing mechanism. It will operate within the microstructure wherever the applied stress vector is sufficient relative to the local residual stress state. This means that during service, as the magnitude and direction of loading varies, plasticity will be induced in different regions in proportion to the applied stress, providing an energy-absorption mechanism operating to some degree throughout the component volume, but most strongly near geometric stress concentrators such as notches or defects. Moreover, the anisotropy of plastic response and of the resulting residual
References
Contents lists available at ScienceDirect
Int. Journal of Refractory Metals & Hard Materials journal homepage: www.elsevier.com/locate/IJRMHM
Phase response of WC–Ni to cyclic compressive loading and its relation to toughness A.D. Krawitz a,*, A.M. Venter b, E.F. Drake c, S.B. Luyckx d, B. Clausen e a
Mechanical and Aerospace Engineering, Lafferre Hall, University of Missouri, Columbia, MO 65211, USA Necsa Limited, Pretoria, P.O. Box 582, Pretoria 0001, South Africa c ReedHycalog, Houston, TX 77252, USA d Chemical and Metallurgical Engineering, University of the Witwatersrand, Johannesburg, South Africa e Los Alamos National Laboratory, Los Alamos, NM 87545, USA b
a r t i c l e
i n f o
Article history: Received 15 August 2008 Accepted 22 November 2008
Keywords: Residual stress Neutron diffraction Cemented carbide composites Mechanical behavior
a b s t r a c t The interaction of uniaxial compressive load and thermal residual stress was measured in a WC–10 wt.% (16 vol.%) Ni cemented carbide composite using neutron diffraction. Loading was from 0 to 2500 MPa in increments of 250 MPa, and measurements were made in situ during load–unload cycles 1, 2, 3, 10, 25, 50 and 100. Plasticity is observed in the Ni from the lowest levels of applied load, leading to continuous curvature of the WC–Ni stress–strain curves, and is believed to be a significant contribution to the composite’s toughness. It is due to interaction between local extremes of the thermal residual microstress with the applied macrostress and leads to anisotropic relaxation of the thermal residual stress. Strain distribution and plasticity were observed through peak breadths. Although the initially strong hysteresis is reduced as the cycles increase, there are still changes taking place after 100 cycles. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Cemented carbide composites have interesting and important mechanical and physical properties. These include high toughness for materials of such high hardness [1], stress–strain curves that are continuously curved [2,3], and very large thermal residual stresses in the as-produced state [4]. In this paper, we have used neutron diffraction to follow the strain response in both phases of a WC–10 wt.% Ni cemented carbide under uniaxial compressive load. Because the absorption of neutrons is low in many engineering materials, including W, bulk average values of strain can be obtained – during loading and unloading – from both the WC and Ni phases through observation of changes in diffraction peak positions. These are changes in the elastic component of strain. Information concerning the distribution of these elastic strains, as well as plasticity, can be observed through diffraction peak breadth behavior. Previous work on WC–Ni, WC–Co and WC–(Co,Ni) has revealed three distinctive characteristics of thermal residual stresses in cemented carbides. First, they are ‘‘diffraction hydrostatic” in character, which means that, when averaged over many particles and binder regions, the same stress values are measured in any physical direction. This is due to the fine scale nature of the microstructure and the fact that, when the Ni binder surrounding the micron* Corresponding author. E-mail address: [email protected] (A.D. Krawitz). 0263-4368/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijrmhm.2008.11.010
sized WC particles tries to shrink during cooling from the sintering temperature, it is severely constrained by the many surrounding WC particles. The constraint leads to a largely hydrostatic stress state, when averaged over many particles. The degree of constraint is closely related to the WC particle size; the smaller the particles, the greater the constraint. In a study of WC–Co material, samples with 20 vol.% Co had thermal residual stress levels for the WC that ranged from 755 MPa (±65 MPa) for 0.6 lm WC to 383 MPa (±66 MPa) for 5.1 lm WC particles [5]. The counterbalancing Co values are +1712 MPa (±147 MPa) for the 0.6 lm WC and +869 MPa (±150 MPa) for the 5.1 lm WC. Thus, the magnitude of the thermal stress doubles as the WC particle size decreases from 5 to 0.6 lm. For the present composite, the values are 382 MPa for the WC and +2022 MPa for the Ni. Second, in spite of the scalar character of the stresses when averaged over many particles, they vary dramatically from point to point in both the particles and binder. This is revealed through diffraction peak breadths, which are significantly broadened by the presence of a range of strains in both particles and matrix regions. If, for example, the temperature is lowered, both the thermal stress magnitudes and breadth of distribution increase [4]. An FE study has addressed the issue of stress distribution, in two dimensions, and found, in agreement with diffraction data, that the range of stresses in angular WC particles is large [6]. Thus, even though the average stress is compressive, there are local regions with tensile stress and others with very high magnitudes of compressive stress.
A.D. Krawitz et al. / Int. Journal of Refractory Metals & Hard Materials 27 (2009) 313–316
2. Experimental The sample was WC–10 wt.% (15.9 vol.%) Ni and was produced via liquid phase sintering and hot isostatic pressing. It had an average WC particle size of 1.17 lm, a binder mean free path of 0.84 lm, a contiguity of 0.48, a Rockwell A hardness of 88.6, and a density of 14.56 g/cm3. The in situ stress measurements were made at the Los Alamos Neutron Science Center (LANSCE) using the SMARTS spectrometer. LANSCE is a pulsed neutron source that measures the time-of-flight of neutrons from a white beam source for which the diffraction angles are fixed and there is a range of wavelengths from 0.05 to 0.5 nm. There are detector banks centered at ±90° so that data can be collected simultaneously in the axial and transverse directions of a cylindrical sample. The cylindrical sample was 8 mm in diameter by 19.2 mm in length. Count times were about a half hour per point. The samples were loaded in uniaxial compression from 0 to 2500 MPa to 0 in increments of 250 MPa. Diffraction measurements were made during load–unload cycles 1, 2, 3, 10, 25, 50 and 100. Strain was measured in both phases simultaneously, in directions parallel and transverse to the loading axis of cylindrical samples.
0
Ni -500
Applied stress (MPa)
Third, the pre-existing thermal residual stresses interact with applied stresses. This interaction is the primary subject of this paper. A previous FE study of a WC–10 wt.% Co sample has modeled the interaction of the thermal residual stresses with an applied uniaxial compressive load [7]. The FE model captured the general features of the measured response.
Transverse direction -1000
-1500
-2000
-2500 -5000
Cycle 1 Load Cycle 1 Unload Cycle 100 Load Cycle 100 Unload -4000
- 3000
-2000
- 1000
0
1000
Elastic microstrain (με)
(a) Ni response in the transverse direction
0
Applied stress (MPa)
314
-500
Cycle 1 Load Cycle 1 Unload Cycle 100 Load Cycle 100 Unload
-1000
-1500
Ni
-2000
Axial direction 3. Results and discussion
-2500
The macroscopic stress–strain curves for cycles 1 and 100 are shown in Fig. 1. They show curvature of the loading portion of each cycle, hysteresis, and a decreasing residual strain at the end of each cycle. The individual phase responses are shown in Fig. 2. In the transverse direction, Fig. 2a, there is an initial Poisson tensile response due to the compressive axial loading. However, the strain begins to reverse while the composite is still undergoing increasing tensile strain in the transverse direction. For cycle 1, the reversal begins at about 750 MPa and becomes more compressive than the initial value beginning at an applied load of about 1500 MPa. Upon unloading, there is a mild reverse curvature and a large net negative change in the strain from the starting value. In the axial direction, Fig. 2b, the strain is compressive
Applied stress (MPa)
0
cycle 1 load cycle 1 unload cycle 100 load cycle 100 unload
-500
-1000
-1500
-2000
-2500 -0.6
-0.5
-0.4
-0.3
- 0.2
-0.1
0.0
Macroscopic strain (%) Fig. 1. The macroscopic stress–strain curves for cycles 1 and 100.
-5000
-4000
-3000
-2000
-1000
0
1000
Elastic microstrain (με) (b) Ni response in the axial direction Fig. 2. The applied composite stress vs. Ni phase strain in the axial and transverse directions for cycles 1 and 100. Microstrain is strain times 106.
throughout but the rate of decrease begins to slow at about 750 MPa. Upon unloading, there is again a mild reversed curvature and a net tensile increase in the strain relative to the starting value. The observed response is due to plasticity that occurs locally throughout the volume of the Ni phase, at sites in the Ni that experience shear yield due to the applied stress. The changes are greater in the transverse direction because the applied Poisson strain is tensile in this direction and adds to the pre-existing thermal residual strain, which is, on average, strongly tensile in the Ni. The reversal of curvature in the transverse direction indicates that most, if not all, of the changes are taking place during the loading portion of the cycle. The degree of net strain relaxation per cycle is shown in Fig. 3a. The changes slow dramatically after 10 cycles and by 50 cycles the system has entered a stable cyclic regime. After 100 cycles, the residual stress in the Ni has been lowered to +1750 MPa in the axial and +1363 MPa in the transverse direction compared to an initial value of +2022 MPa. The system behaves essentially elastically until the stress magnitude is raised or a crack develops. Evidence of the response when the stress magnitude is increased is shown in Fig. 4, which presents the Ni phase response in the transverse direction. This was obtained from another sample from the same batch that was loaded in three sequential steps: (1) three load–unload cycles to 500 MPa in steps of 100 MPa; (2) three load–unload cycles to 1000 MPa in steps of 200 MPa; and, (3) three load–unload cycles to 2000 MPa in steps of 200 MPa [8]. The general behavior
315
A.D. Krawitz et al. / Int. Journal of Refractory Metals & Hard Materials 27 (2009) 313–316
1500
-500 500
Applied stress (MPa)
Change in microstrain ( με)
0 1000
0 -500
Change in axial strain Change in transverse strain
-1000
-1000
Ni -1500
Axial
-2000
Cycle 1 Load Cycle 1 Unload Cycle 100 Load Cycle 100 Unload
-1500
-2500 -2000 0
20
40
60
80
100
60
90 100 110 120 130 Normalized breadth (a) Normalized average peak breadth for Ni in the axial direction
Cycles Fig. 3. The changes in elastic strain in the Ni phase from their initial values after cycles 1, 2, 3, 10, 25, 50 and 100 in the axial and transverse directions. Microstrain is strain times 106.
70
80
0
Ni -500
500 MPa
1000 MPa
-500
1st cycle at 1000 MPa
-1000
Transverse -1000
-1500
Cycle 1 Load Cycle 1 Unload Cycle 100 Load Cycle 100 Unload
-2000
-2500
-1500
2000 MPa
60 1st cycle at 2000 MPa
-2000
-1200
-800
-400
0
70
80
90
100
110
120
130
Normalized breadth (b) Normalized average peak breadth for Ni in the transverse direction
400
Microstrain (με) Fig. 4. Results due to step loading, in the transverse direction, for a WC–10 wt.% Ni sample. At each load level ( 500 MPa, 1000 MPa and 2000 MPa), three load– unload cycles were performed. The load increments were 100 MPa at the 500 MPa level, and 200 MPa at the other two. Microstrain is strain times 106.
is the same at each level. There is hysteresis and a decrease in the residual stress of the system at each load level. Three cycles is insufficient for the changes in residual stress to reach the plateaus shown in Fig. 3, but the response is the same – new plasticity at the higher stress magnitude. It is particularly noticeable for the transition from 1000 to 2000 MPa. The diffraction peak breadth response of Ni for cycles 1 and 100 is shown in Fig. 5. In the axial direction, Fig. 5a, the peak breadth broadens even though the applied compression acts with the opposite sign to the residual tensile stress in the Ni, indicating that the stress variance is increasing even though the mean stress is decreasing. This indicates that even though plastic flow is occurring at specific sites in the microstructure, other regions retain their elastic character and do not flow. The rate of increase decreases steadily with increasing applied stress magnitude. Upon relaxation, there is a linear decrease in the breadths, indicating that the elastic regions whose stress variance has increased due to the compressive loading are relaxing during unloading. The peak breadth at the end of the unloading portion of cycle 1 is much lower due to a net relaxation of the stress variance due to the local plastic flow. During cycle 100, a reversible process is seen, due to the stable cyclic
Fig. 5. Values of normalized average peak breadth for Ni during cycles 1 and 100 in the axial and transverse directions.
regime that has been established. In the transverse direction, for cycle 1, the peak breadth decreases with load up to about 1500 MPa, when it begins a slow, small increase. The strain variance decreases due to plastic flow at the extremities of the local strain distributions. Apparently, new constraint develops in the la-
100
Normalized Ni peak breadth
Applied stress (MPa)
0
1st cycle at 500 MPa
Applied stress (MPa)
cycles increasing
95 90 85
transverse axial
80 75 70 0
20
40
60
80
100
Cycles Change in normalized peak breadth of the Ni phase as a function of cycle. Fig. 6. The change in normalized peak breadths in the Ni phase from their initial values after cycles 1, 2, 3, 10, 25, 50 and 100 in the axial and transverse directions.
316
A.D. Krawitz et al. / Int. Journal of Refractory Metals & Hard Materials 27 (2009) 313–316
ter part of the loading. As for the axial direction, there is a linear decrease in peak breadth and a net decrease in peak breadth after unloading. The response during cycle 100 is basically elastic. Fig. 6 shows the normalized peak breadths at the ends of the measured cycles. There is net relaxation for both directions although the transverse direction shows somewhat greater relaxation, as expected.
stress relaxations can be expected to affect the shape of the macroscopic elastic stress field around notches. It has been shown that an increase in applied stress magnitude leads to a new flow response even if the region had become stable relative to a lower stress magnitude.
4. Conclusions
[1] Uphadaya GS. Cemented tungsten carbides: production, properties, and testing. Westwood, NJ: Noyes Publications; 1998. [2] Felgar RP, Lubahn JD. The behavior of cemented carbides. Proc Am Soc Test Mater 1957;57:770–90. [3] Drake EF, Krawitz AD. Fatigue damage in a WC–nickel cemented carbide composite. Metall Trans A 1981;12A:505–13. [4] Krawitz AD, Reichel DG, Hitterman R. Residual stress and stress distribution in a WC–Ni composite. Mater Sci Eng 1989;A119:127–34. [5] Coats DL, Krawitz AD. Effect of particle size on thermal residual stress in WC–Co composites. Mater Sci Eng 2003;A359:338–42. [6] Weisbrook CM, Krawitz AD. Thermal residual stress distribution in WC–Ni composites. Mater Sci Eng 1996;A209:318–28. [7] Livescu V, Clausen B, Paggett JW, Krawitz AD, Drake EF, Bourke MAM. Measurement and modeling of room temperature co-deformation in WC– 10 wt.% Co. Mater Sci Eng 2005;A399:134–40. [8] Paggett JW. Neutron diffraction study of load response and residual stresses in WC–(Ni/Co) composites. PhD dissertation, University of Missouri, Columbia, MO, 2005.
Interactions between pre-existing thermal residual stresses and applied load have been observed in a WC–10 wt.% Ni. The local plasticity induced by the application of external stress is an energy-absorbing mechanism. It will operate within the microstructure wherever the applied stress vector is sufficient relative to the local residual stress state. This means that during service, as the magnitude and direction of loading varies, plasticity will be induced in different regions in proportion to the applied stress, providing an energy-absorption mechanism operating to some degree throughout the component volume, but most strongly near geometric stress concentrators such as notches or defects. Moreover, the anisotropy of plastic response and of the resulting residual
References