The Erosion-corrosion Resistance of Tungstencarbide

ELSEVIER Pll: Int. J. of Refractory Metals & Hard Materials 15 (1997) 81-87 1997 Elsevier Science Limited Printed in Great Britain. All rights reserved 263-4368/97/$17. S263-4368 (96)16-9 The Erosion-corrosion Resistance of Tungstencarbide Hard Metals E. J. Wentzel Boart Longyear Research Centre, Krugersdorp, South Africa & C. Allen Department of Materials Engineering, University of Cape Town, South Africa (Received 22 November 1995; accepted 2 February 1996) Abstract: A series of cemented tungsten-carbides with different binder phases consisting of combinations of cobalt, nickel and chromium have been subjected to erosion-corrosion testing using a silica-water slurry. The polarisation characteristics of these cermets have been investigated using a potentiodynamic technique. The differences in binder composition influences the cermets' properties and corrosion behaviour, which in turn affects the synergistic action of erosion-corrosion. The inherent corrosion resistance of a pure nickel binder did not increase the slurry erosion resistance of the cermets, but the nickel-chromium-cobalt grades were found to improve the erosion-corrosion behaviour compared to the pure cobalt binder grade. Comparisons are made between the properties and behaviour of Ni-Cr-Co based cermets and pure metal grades with compositions similar to those found in the binder phase of the corresponding cermet grades. Explanations are advanced to explain the differences in behaviour linked to composition and mechanical properties. 1997 Elsevier Science Limited 1 INTRODUCTION It is apparent that materials subjected to slurry environments must not only resist the erosion of solid particles and liquids, but also be resistant to damage caused by corrosion. These two factors may interact synergistically to produce wear rates that are greater than the sum of their separate effects. ''2 The ability of the material to resist the synergistic action of slurry erosion and corrosion is thus of the utmost importance. The characteristic high hardness and fracture toughness of WC-Co cermets have in the past made them materials of choice for use as cutting tools, mining bits, sandblasting nozzles, and for a variety of other wear resistant applications. In most cases cobalt has been used as the tough metal binder phase, due to its excellent wetting, adhesion and adequate mechanical properties. Cobalt, however, has low corrosion resistance and thus to improve the wear resistance in aggressive environments, other metallic alloy compositions have been formulated) In this work, slurry erosion rates for cobalt, nickel, nickel-chromium and nickelchromium-cobalt based hard metals are compared with respect to their overall behaviour. In order to establish the influence of binder composition on the behaviour of cermets, the behaviour of pure metal alloys with compositions similar to those of the binder phase, in the corresponding cermets is examined. This will determine whether pure binder metal behaviour can be used to model cermet behaviour in slurry erosion or pure corrosive environments. 2 EXPERIMENTAL PROCEDURES A slurry jet erosion rig based on that of Zu 4 and Bester ~ was constructed from non-corrosive plastic materials. This rig permits variations in 81

82 E. J. Wentzel, C. Allen Fig. 1. st: A schematic representation of the slurry erosion rig. the impact angle, impact velocity, erodent size, erodent concentration and carrier fluid. A schematic representation of the test apparatus is shown in Fig. 1. The carrier fluid is circulated by means of a centrifugal pump from the holding tank, via a bypass valve, a rotameter, a pressure gauge and an ejector to the specimen surface. The ejector ensures a pressure drop in the fluid, resulting in a vacuum that is sufficiently strong to draw erodent particles up through a vertical suction tube. The bottom of the suction tube is placed in the submerged sand bed. The erodent and carrier fluid is mixed in the ejector, thus causing a slurry jet to be accelerated through the exit nozzle onto the specimen surface. Tap water is used as a carrier fluid to provide a base against which to compare the erosion rate of a substitute ocean water based slurry and thus determine the degree that a corrosion environment influences slurry erosion rate. The specimen is held in a specimen holder that can be rotated about the horizontal axis to any pre-set impact angle between 15 and 9. The complete specimen and holder are enclosed in a test chamber. The impact velocity was set at 7. m s- ' with an erodent concentration of 6.3 wt% (37.8 kg of erodent impacted the specimen surface per h). Although the 5/~m diameter silica sand erodent was found not to degrade during a four hour erosion test, new erodent was used for each specimen. The impact angle was kept at 75, where the maximum slurry erosion rate occurred. 6 The solutions used for the potentiodynamic polarisation experiments were a 1N HzSO4 (27"2 ml of 98% H2SO4 per litre distilled water) in accordance with the ASTM standard G5-827 and a substitute ocean water solution based on ASTM specification Dl141-9 but containing only the three main compounds. The salt water solution for potentiodynamic testing is the same as that used for the carrier fluid in slurry erosion testing. A standard potentiodynamic polarisation procedure was used. The specimen was polished to a 1 ~m diamond finish and sealed in a PTFE holder. This was the working electrode of a cell consisting of a saturated calomel reference electrode (SCE) and a carbon rod counter electrodes. The solution was deaerated with bubbling argon for 6 rain before the specimen was inserted and continued during the experiment. The specimen was allowed to reach equilibrium at E... (corrosion potential) for 3 min before polarisation began. The potential was then anodically ramped from -5mV to 15mV by an Amel potentiostat at 167gVs ~ 3 MATERIALS The materials used in this project were chosen to examine the effects of compositional changes within the binder on the slurry erosion-corrosion behaviour of WC-based hardmetals. The materials were all manufactured by the Boart Longyear Research Centre. The target materials were separated into two distinct groups, tungsten-carbid based cermets and binder metal alloys containing no tungstencarbide particles. The tungsten-carbide based cermets are again separated into two categories, those with 1 wt% binder and those with 6 wt% binder. The binder metal alloys have compositions similar to those of the binder phase of the 1 wt% cermets. These binder metal alloys were manufactured by powder metallurgical techniques and were used to examine the behaviour of the binders. In practice, however, liquid phase sintering leads to some solution of tungsten and carbon in the binder phase which did not occur here. The different grades of materials can be identified as follows: 1 wt% binder cermets e.g. 833; 6 wt% binder cermets e.g. P6; 1 wt% binder alloys e.g. Uct 1.

The erosion-corrosion resistance of tungsten-carbide hard metals 83 The chemical compositions of the different materials used in this work are tabulated in Table 1. Note that the 439 grade and the 833 grade are both pure cobalt binder grades. The mean WC grain size in all the alloys was 1-1.5/~m irrespective of the binder volume fraction used. Vickers hardness values were obtained, using a 3 kg load for both the WC-based cermets and the pure binder alloys, in order to facilitate comparisons between hardness and slurry erosion resistance. The results are the average of 1 readings. The values of hardness and density for all the grades tested are shown in Table 2. 4 RESULTS Under steady-state conditions slurry erosion follows a linear relationship and is thus quoted as volume loss per mass of impacting erodent. Linear regression analysis forcing the y-intercept through the origin allows the slurry erosion rates to be determined. For all the grades tested the slurry erosion rates in the salt (substitute ocean water) solution were greater than those in the tap water solution. Figure 2 shows the linearity of slurry erosion and the increase in erosion rate caused by the more corrosive salt water carrier fluid. The slurry erosion rates were established for all the materials tested in both carrier fluids and the results are plotted in Figs 3-5. in all these graphs the composition of the binders varies from the high cobalt grades on the left to the high nickel grades on the right. The intermediate grades have chromium additions of 5 wt% of the binder phase. Note the pure nickel (Uct 7) does not fit the trend. The slurry erosion rates shown in Figs 3 and 4 exhibit a trend with the lowest slurry erosion rates occurring when the composition of the binder phase has a high wt% cobalt and small additions of chromium or chromium and nickel. The influence of target material hardness on the slurry erosion resistance of the hardmetals and the binder metal grades is shown in Fig. 6(a) and (b). The general trend of decreasing erosion rate with increasing hardness is followed in the cermets' grades. The influence of binder composition can be seen in Fig. 7 where hardness values of the cermets and the pure metal binder grades are compared. Again the left side of the graph is the high cobalt and Table 1. Chemical compositions of the materials Grade Composition (wt%) WC Ni Cr Co Total C 1 wt% binder cermets 833 9 9.86 5.64 842 + 9.13.46 9.9 5.55 839 9 2.13.47 7.4 5.58 84 9 4.57.57 5.71 5.5 841 + 9 5.49.45 3.89 5.65 864 + 9 7.42.51 2.28 5.5 834 + 9 9.15 < -5 5.54 6 wt% binder cermets C6 94 + 6. -- P6 + 94 + 3. _ "6 + 2.4 -- V7 + 94 5.67.35.35 5.79 V6 + 94 + 6. -- 1 wt% binder alloys Uct 1.1.1 99.96 -- Uct 2.1 5.8 94.2 -- Uct 3 33' 4.1 72.9 -- Uct 4 6-5.5 34-3 -- Uct 5 7-3.5 26.8 -- Uct 6 95.2 4.1.1 -- Uct 7 99.91.1.1 --

84 E. J. Wentzel, C. Allen Table 2. tested Hardness and density values for the materials Grade Hardness Density HV3 (g/cm~) 1 wt%binder cermets 833 1473 14-53 842 1546 14.53 839 1498 14.51 84 1499 14.58 841 1424 14.48 864 14 14.58 834 149 14.55 6 wt%binder cermets C6 155 14.97 P6 1687 14.95 V7 1674 14.95 V6 1448 14-93 1 wt% binder alloys Uct 1 258 8.52 Uct 2 227 8.25 Uct 3 146 8. Uct 4 91 7"71 Uct 5 87 7.64 Uct 6 8 7.65 Uct 7 95 8.54 the right side the high nickel. The trends for the cermets and the pure metal binder grades are superimposed and it must be noted that the nickel rich binders are softer than the cobalt rich binders influencing the cermets hardness in the same manner, although to a lesser degree, as a result of the low volume fraction of binder in the cermets. The influence of the more corrosive salt water carrier fluid can be seen by plotting the slurry erosion rate ratio of tests in tap water solution to tests in salt water solution, as a percentage. Figure 8 shows this change in erosion rate for the 1 wt%, 6 wt% and pure binder metal grades. The average change is 52.2, 51. and 75.4% for the 1 wt%, 6 wt% and pure binder and metal grades respectively..7.6 ~'E.5.oooo4.~.3 m 2 w ~.1111112 (~.1 833 842 839 84 841 864 834 Fig. 3. Slurry erosion rates and trends in erosion rate for 1 wt% binder grade cermets in salt water solution (SWS) and tap water solution (TWS). Anodic polarisation curves for 1 wt% grades and pure binder metal grades are shown in Figs 9(a) and (b). The expected increase in passivation behaviour, decreases in corrosion potential and drop in current density occur as the nickel volume fraction within the binder phase is E g 25 ~.15 re" -~.1 e w.5 C6 P6 V7 V6 Fig. 4. Slurry erosion rates and trends in erosion rate for 6 wt% binder grade cermets in salt water solution (SWS) and tap water solution (TWS)..6 Fig. 2..9.8 e~ I::.7.6- ~.5 O ~.4- E O.3 ~>.2.1 -" i j i i J ~ i i.5 1 1.5 2 2.5 3 3.5 Exposure Time (hours) Linear nature of slurry erosion with substitute ocean water and tap water carrier fluids..5 1 [ ~ Increasing Nickel 1 ~.1111114 re".3._g P o.oo2 ul o. uctl uct2 uct3 uct4 uct5 uct6 uct7 Fig. 5. Slurry erosion rates and trends in erosion rate for pure metal binder grades in salt water solution (SWS) and tap water solution (TWS).

The erosion-corrosion resistance of tungsten-carbide hard metals 85.2 &V6 go -- Increasing Nickel.15 Ig i 439 ~ ~ 8 3 3 &C61&6wt% Binder.~.1 u.l 2 ~.5 135.55.5 E.45 14 145 15 155 16 165 17 Hardness (HV 3) m ~ UCT6 ~.4._~.111135 UCT5 t CT4 LU 2 o.3 _~_._*UCT3 ~ct1 ~.25 _= e~ct7 UCT2.5O2 I I - I I --- I 5 1 15 2 25 3 Hardness (HV3) Fig. 6. (a) Salt water slurry erosion rates for 1 and 6 wt% grades related to hardness; (b) salt water slurry erosion rates for pure metal binder grades related to hardness. O F- 5 4O [:-.e-1wt%.-b-6wt%,-a.- Binders l 3 --I I I I ~- Fig. 8. Ratio as a percentage of tap water slurry erosion rate over salt water slurry erosion rate for all grades. 5 DISCUSSION The loss of material during a wear process is a complex relationship between many interacting variables. Thus the changing of one variable can 15 increased. A more distinct passivation region can be seen in the pure metal binder grades than in the cermet grades, although the passivation current density of the cermets is lower than that of the pure binder metals. This is not unexpected as the cermets have high volume fractions of the more corrosion resistant WC phase. The passivation behaviour of the 6 wt% grades follows the same trends highlighted by the two categories shown. g 1 lu k~ c Q. O 5O._o c 2 i_ O - 5 ~ : l 1 I 15O UC t S" 16 3 Increasing Ni v 155 -~ 145 '1-135 \ eerme,; 1 Binder Grades I 13 33 842 839 84 841 834 Uctl Uct2 Uct3 Uct4 Uct5 Uct6 Uct7 Fig. 7. The influence of binder hardness on the hardness of the cermet grades. Note the left side of the graph is high cobalt and the right high nickel. 2511 2 > 1511 c 1.o~ I11 511 i -5oo ~Z.1 I 1 1 1 1 1 1 Current Density (pa/cm 2) Fig. 9. (a) Anodic polarisation curves for the 1 wt% binder grades in 1N H2SO4; (b) Anodic polarisation curves for the pure metal binder grades in 1N H2SO4.

86 E. J. Wentzel, C. Allen have important consequences on the behaviour pattern of the total system. In this work, cemented carbides have been subjected to slurry erosion conditions where impact velocity, impact angle, erodent concentration and erodent characteristics remain constant. Thus only the effect of changing the binder composition on the erosive-corrosive wear process has been investigated. The relationship between corrosion resistance and/or erosion-corrosion resistance of the binder and that of the cermet is not simple, as related factors such as interface strength and the sinterability of the composite also influence the behaviour of the cermet. In this work however all the related factors are included and the performance of the cermet is considered in its entirety. When discussing the effects of binder composition it is important to bear certain fundamentals of cermet erosion-corrosion in mind. The mechanism of material removal is one of binder loss followed by WC grain pullout, ''8 thus the ability of a binder to resist removal by either improved mechanical properties or improved corrosion resistance"'"' is fundamental to the success of that material as a cermet binder. Slurry erosion is a balance between the erosion by solid particles and the corrosion of the target. The importance of the corrosion resistance of the binder is only of significance when the corrosion aspect of the process plays the dominant role or is the rate controlling factor. When the erosive component of the slurry process is dominant, the target material mechanical properties will be of more importance than the corrosion resistance of the binder material. The severity of the erosive component relative to the corrosive component is critical and it is thus very difficult to generalise as to the behaviour of cermets in slurry erosion conditions. Intuitively it might be expected that a cermet's slurry erosion resistance would improve with the addition of more corrosion resistant WC phase. In this work, however this is not the case as can be seen from Fig. 6(a) where the 1 wt% cermets outperform the 6 wt% cermets. This could be due to a dynamic situation that exists where the rate controlling factor for material removal is the corrosion of the binder phase, although the mechanical properties of the cermet must not be ignored, and thus the more binder phase present in the material the longer it will take to be removed to the extent where WC grain pullout is possible. This could however also be due to a dynamic situation where the erosion component is dominant and the small amount of binder in the 6 wt% grades causes more severe binder extrusion between WC grains and thus increased wear rates. A balance of these two conditions is also possible. Electron microscopy of the wear surfaces of the target materials in this work indicates that the first option is the most probable, with corrosion of the binder phase being the rate determining or dominating factor. In both tap water and the salt water carrier fluids, the slurry erosion resistance of the 1 wt% grades decreased when additions of chromium and chromium-nickel were made to the predominantly cobalt binder. As the nickel content in the binder increased, the erosion rate also increased thus having little positive effect on retarding the erosive-corrosion process. High nickel content in the binder adversely affected the hardness of the cermet and that of the pure metal binder grades. The trend seen in the 1 wt% binder cermets is replicated in the 6 wt% binder grades and the pure metal binder grades. The exception to this rule is the pure nickel binder metal that shows decreased slurry erosion rates in both carrier fluids. Figure 8 indicates that the influence of the more corrosive carrier fluid on the slurry erosion rate is lessened as the binder wt% increases, and as the binder composition tends toward the high nickel binder grades. As nickel is the more corrosion resistant metal this influence is expected. The polarisation behaviour of the different grades indicates that the corrosion resistance of the grades improves with the addition of the more corrosion resistant nickel to the binder. Not only do the pure binder grades show this improvement, but the cermets also show this behaviour. The importance of this similarity in behaviour between the pure binder grades and the related cermet grades is the possibility of modelling the corrosion resistance of the cermets on the corrosion resistance of the pure binder grades. It is interesting that this improvement in the corrosion resistance of the cermets with the increase in nickel content does not necessarily give rise to an improvement in the slurry erosion resistance of the cermets. In all cases the high nickel grades have better passivation

The erosion-corrosion resistance of tungsten-carbide hard metals 87 behaviour than the Co-Cr-Ni grades, but the slurry erosion behaviour deteriorates. It should be noted that the passivation behaviour of the nickel binder may well negate its superior corrosion resistance under conditions when any passive film formation is continuously being removed, such as in slurry erosion. In such conditions the hardness, deformation characteristics and possible phase transformations of the binder may be more important in determining the wear rates. These factors will be investigated in future work. Nevertheless, the nickel-chromium-cobalt and in particular the chromium-cobalt, although not always having improved pure corrosion properties, do show a distinct improvement over a cobalt binder cermet in slurry erosion-corrosion conditions. The influence of the binder composition is an important factor in the design of a long life cermet for erosioncorrosion conditions as the importance of synergistic wear must not be neglected. 6 CONCLUSIONS Based on the results of this work on slurry erosion-corrosion of hard metals and the influence of binder composition, we conclude the following: (1) The slurry erosion resistance of WC based cermets with pure metal binders of either Co or Ni can be improved by alloying. (2) Of the grades tested the Co-Cr grade with 1 wt% showed the lowest slurry erosion rates in both a salt water and a tap water solution. (3) No simple relationship exists between any one tested property and the slurry erosion resistance of the hard metals. (4) An improvement in the passivation behaviour of a binder metal does not show any relationship to an improvement in the slurry erosion-corrosion resistance of the grades tested. ACKNOWLEDGEMENTS The financial support of Boart Longyear Research Centre, Eskom and the FRD are gratefully acknowledged. REFERENCES 1. N6el, R. E. & Ball, A., On the synergistic effects of abrasion and corrosion during wear. Wear, 87 (1983) 351-61. 2. Madsen, B. W., Measurement of erosion-corrosion synergism with a slurry wear test apparatus. Wear, 123 (1988) 127-42. 3. Human, A. M., Northrop, I. T., Luyckx, S. B. & James, M. N., A comparison between cemented carbides containing cobalt- and nickel-based binders. J. Hard Mater., 2 (3-4) (1991) 245-55. 4. Zu, J. B., Hutchings, I. M. & Bursting, G. T., Design of a slurry erosion test rig. Wear (199). 5. Bester, J. A., The slurry erosive-corrosive wear of a selection of AI alloys, AI MMC's and steels. MSc Thesis, University of Cape Town (1992). 6. Wentzel, E. J. & Allen, C., Erosion-corrosion resistance of tungsten carbide hard metals with different binder compositions. Wear, 181-183 (1995) 63-9. 7. ASTM, Standard reference method for making potentiostatic and potentiodynamic anodic polarisation measurements. In Annual Book of Standards, Section 3, G5-82, No. 1, 1985. 8. Wright, I. G., Shetty, D. K. & Clauer, A. H., Slurry erosion of WC-Co cermets and its relationship to material properties. Proc. 6th Intern. Conf. on Erosion by Liquid and Solid Impact, (1983) pp. 1-8. 9. Tomlinson, W. J. & Linzell, C. R., Anodic polarisation and corrosion of cemented carbides with cobalt and nickel binders. J. Mater. Sci., 23 (1988) 914-18. 1. Tomlinson, W. J. & Molyneux, I. D., Corrosion, erosion-corrosion and the flexural strength of WC-Co hardmetals. J. of Mater. Sci., 26 (1991) 165-8.