Corrosion Science, vol 39 Pergamon PI:s0010938X(96001424 THE EFFECT OF CORROSION AND EROSION ON CERAMIC MATERIALS Q. FANG, P S SIDKY and M. G. HOCKING Department of Materials, Imperial College, London, SW7 2BP, U. K. Abstract-The effects of corrosion/ erosion on the chemical / mechanical properties and microstructures of four ion bonded silicon carbide, sialon and Psz zirconia were investigated and characterised, using a chemical etch bed (hydrofluoric-hydrochloric acid mixtures, 1.5% HF+5%HCD)and a sand-water slurry erosion test rig. A series of corrosion and erosion tests were carried out and the morphology, structure and chemical analysis of the samples was investigated using optical microscopy, SEM, XRD and AAs(Atomic absorption spectroscopy). Under the test conditions, Sic shows the best chemical and erosion resistance of the four ceramics; and alumina has the highest corrosion and erosion rate. The erosion and rrosion response, the degeneration of mechanical properties and the micromechanisms of the surface damage for these ceramic materials are presented and discussed. C 1997 Elsevier Science Ltd. All rights reserved Keywords: A ceramic, B corrosion, B, erosion, C acid corrosion INTRODUCTION Engineering ceramic materials, such as zirconia-toughened ceramic, nitride and carbide ceramics,are finding ever increasing applications in the development of modern industry and technology and have been playing an active role in erosion resistant applications. One of the outstanding advantages which ceramics have over other classes of materials is their good combination of properties, such as high hardness, high melting temperature, good erosion and wear resistance and high level of chemical inertness in corrosive environments Slurry erosion by the impact of liquid-carrying solid particles reduces the life of mechanical components used in many industrial and research applications. It is generally accepted that ceramic materials are eroded in the form of radial and lateral cracking, ring cracking and ploughing on the surface. -6 A recent review describes many aspects of the erosion behaviour and erosion mechanisms of ceramics. In many cases, joint effects of erosion and corrosion in a slurry are inevitable; the surface of components are damaged, not only by the impact of the solid particles, but also by chemical or electrochemical corrosion. The term erosion-corrosion(E-C)", applies to the degeneration of a material due to the simultaneous action of mechanical and chemical forces Major advances in the study of the effect of erosion and corrosion have been made in recent years. Stephenson et al. studied the interaction between erosion and corrosion of C-Mn steel in dry and wet CO2 conditions. The use of wet co 2 increases the rate of metal loss by factor of 2-4, and loss from pure corrosion to erosion/corrosion increases by two orders of magnitude. Results from Stott revealed that in the dry case, there may be four Manuscript received 3 August 1995; in revised form 3 June 1996
Corrosion Science, Vol. 39, No. 3, pp. 51 I-527, 1997 0 1997 Elsevier Science Ltd Printed in Great Britain. All rights reserved 001&938X/97 $17.00+0.00 THE EFFECT PII: 80010-938x(%)001424 CORROSION AND EROSION MATERIALS CERAMIC Q. FANG, P. S. SIDKY and M. G. HOCKING Department of Materials, Imperial College, London, SW7 2BP, U.K. Abstract-The effects of corrosion/erosion on the chemical/mechanical properties and microstructures of four engineering ceramic materials, namely alumina, reaction bonded silicon carbide, sialon and PSZ xirconia, were investigated and characterised, using a chemical etch bed (hydrofluoric-hydrochloric acid mixtures, 1.5% HF + 5% HCl) and a sand-water slurry erosion test rig. A series of corrosion and erosion tests were carried out and the morphology, structure and chemical analysis of the samples was investigated using optical microscopy, SEM, XRD and AAS (Atomic absorption spectroscopy). Under the test conditions, Sic shows the best chemical and erosion resistance of the four ceramics; and alumina has the highest corrosion and erosion rate. The erosion and corrosion response, the degeneration of mechanical properties and the micromechanisms of the surface damage for these ceramic materials are presented and discussed. 0 1997 Elsevier Science Ltd. All rights reserved Keywords: A. ceramic, B. corrosion, B. erosion, C. acid corrosion. INTRODUCTION Engineering ceramic materials, such as zirconia-toughened ceramic, nitride and carbide ceramics, are finding ever increasing applications in the development of modern industry and technology and have been playing an active role in erosion resistant applications. One of the outstanding advantages which ceramics have over other classes of materials is their good combination of properties, such as high hardness, high melting temperature, good erosion and wear resistance, and high level of chemical inertness in corrosive environments. Slurry erosion by the impact of liquid-carrying solid particles reduces the life of mechanical components used in many industrial and research applications. It is generally accepted that ceramic materials are eroded in the form of radial and lateral cracking, ring cracking and ploughing on the surface.rm6 A recent review describes many aspects of the erosion behaviour and erosion mechanisms of ceramics.5 In many cases, joint effects of erosion and corrosion in a slurry are inevitable; the surface of components are damaged, not only by the impact of the solid particles, but also by chemical or electrochemical corrosion. The term erosion-corrosion (E-C)7,8 applies to the degeneration of a material due to the simultaneous action of mechanical and chemical forces. Major advances in the study of the effect of erosion and corrosion have been made in recent years.+i5 Stephenson et ~1.~ studied the interaction between erosion and corrosion of C-Mn steel in dry and wet CO* conditions. The use of wet CO2 increases the rate of metal loss by factor of 2-4, and loss from pure corrosion to erosion/corrosion increases by two orders of magnitude. Results from Stott” revealed that in the dry case, there may be four Manuscript received 3 August 1995; in revised form 3 June 1996. 511
512 erosion-corrosion regimes as a function of velocity at elevated temperatures. Newman" studied the transitions in erosion-corrosion regimes in slurries containing alumina sands in Na2CO3+ NaHCO3 solution at room temperatures, and tried to propose an erosion corrosion map for the more complex wet environments. a synergistic effect between erosion and corrosion is widely observed, namely the total material loss rate may be greater than that which would be observed from the erosion and corrosion processes operating separately. Some researchers reportedthat the synergistic effect between erosion and corrosion in an acidic slurry medium for either carbon steel or stainless steel was very (33-99%). Hutchingsstudied the slurry erosion-corrosion effect of aluminium in n: Na2CO3, and phosphate bufer, using electrochemical potential methods. Wood and 14 Stott o and Oka et al. I5 have reviewed the effect of n on various alloys under different conditions and gave some possible synergistic mechanisms. It was shown that predictions cannot be made of the relative erosion-corrosion resistance of alloys by extrapolation of data from one environment to another especially where velocities of impact are quite different Although much progress has been made in recent years in the understanding of the erosion-corrosion mechanism, there are some areas, especially slurry erosion and corrosion of ceramics, which are still not fully understood. Most of the research generally focuses on metals or alloys, actual research of the E C effect, is not as common in the field of ceramics as it is in metallurgy. Only during the past 25 years has a true understanding of the complexities of corrosion of ceramics begun to develop. Most of the research generally focuses on behaviours in neutral or mild medium such as sand-water slurry, saline wate slurry and wet CO2 slurry where the corrosion plays a less important role in the erosion corrosion process. On the other hand it is difficult to study the synergistic effects of E-Con ceramics simultaneously not or only because of the great number of parameters involved but also because some mechanisms of corrosion and erosion processes conceal each other. As ceramics are much more corrosion resistant than metals and alloys in general, corrosion effects are not instantaneous. This study aims to assess the corrosion and erosion resistance of ceramic components in solid-liquid environments with particular reference to offshore pumping applications. The effect of the cleaning solution(1.5% HF+5% HCi)used in the oil industry to clean downhole pipes is investigated. In this paper both simple effects of corrosion and erosion as well as the combined effect are shown in order to shed some light on the corrosion-erosion mechanism. Identifying the corrosion mechanism is useful as a first step in understanding corrosion-erosion effects EXPERIMENTAL METHOD Materials Four ceramic materials have been tested 85% alumina Kawenit EL, reaction bonded REFEL F SiC, Sialon 101, and partially stabilised zirconia(PSZ) Technox 1035. The dimensions of the first three ceramic specimens were 10 x 10 x 5 while the Sic was 5x 15 x 3 mm. Table I shows some of the physical properties of the four ceramic materials which a hydrofluoric-hydrochloric acid of 1. 5% HF+5%HCl was used to simulate the corroSion envI it during downhole cleaning in offshore pumping applications The erosion slurries were prepared from rounded silica sand (600-850 um)and water
512 Q. Fang er al erosion-corrosion regimes as a function of velocity at elevated temperatures. Newman” studied the transitions in erosion-corrosion regimes in slurries containing alumina sands in NazCO,+ NaHC03 solution at room temperatures, and tried to propose an erosioncorrosion map for the more complex wet environments. A synergistic effect between erosion and corrosion is widely observed, namely the total material loss rate may be greater than that which would be observed from the erosion and corrosion processes operating separately. Some researchers reported12 that the synergistic effect between erosion and corrosion in an acidic slurry medium for either carbon steel or stainless steel was very large (33-99%). Hutchings13 studied the slurry erosion-corrosion effect of aluminium in NaCl, Na2C03, and phosphate buffer, using electrochemical potential methods. Wood and HuttonI Stott” and Oka et ~1.‘~ have reviewed the effect of erosioncorrosion on various alloys under different conditions and gave some possible synergistic mechanisms. It was shown that predictions cannot be made of the relative erosion--corrosion resistance of alloys by extrapolation of data from one environment to another especially where velocities of impact are quite different. Although much progress has been made in recent years in the understanding of the erosion-corrosion mechanism, there are some areas, especially slurry erosion and corrosion of ceramics, which are still not fully understood. Most of the research generally focuses on metals or alloys, actual research of the E-C effect, is not as common in the field of ceramics as it is in metallurgy. Only during the past 25 years has a true understanding of the complexities of corrosion of ceramics begun to develop.16 Most of the research generally focuses on behaviours in neutral or mild medium such as sand-water slurry, saline water slurry and wet CO2 slurry where the corrosion plays a less important role in the erosioncorrosion process. On the other hand, it is difficult to study the synergistic effects of E-C on ceramics simultaneously not only because of the great number of parameters involved, but also because some mechanisms of corrosion and erosion processes conceal each other. As ceramics are much more corrosion resistant than metals and alloys in general, corrosion effects are not instantaneous. This study aims to assess the corrosion and erosion resistance of ceramic components in solid-liquid environments with particular reference to offshore pumping applications. The effect of the cleaning solution (1.5% HF + 5% HCl) used in the oil industry to clean downhole pipes is investigated. In this paper both simple effects of corrosion and erosion as well as the combined effect are shown in order to shed some light on the corrosionerosion mechanism. Identifying the corrosion mechanism is useful as a first step in understanding corrosion-erosion effects. EXPERIMENTAL METHOD Materials Four ceramic materials have been tested: 85% alumina Kawenit EL, reaction bonded REFEL F Sic, Sialon 101, and partially stabilised zirconia (PSZ) Technox 1035. The dimensions of the first three ceramic specimens were 10 x 10 x 5 mm3 while the SIC was 15 x 15 x 3 mm’. Table 1 shows some of the physical properties of the four ceramic materials which were used. A hydrofluoric-hydrochloric acid of 1.5% HF + 5% HCl was used to simulate the corrosion environment during downhole cleaning in offshore pumping applications. The erosion slurries were prepared from rounded silica sand (600-850 urn) and water
COl Table 1. Properties of the materials investigated Materials Size Hardness Fracture Toughness Brittlenesst (g/cm Kc( m°)(m PSZ zirconia 1489±50 rest SiO2) Sialon 101 324 1675+50 x=0.01 3.10 2255±50 2. 0.15 *The applied load of vickers hardness tests of 10 kg for alumina, 50 kg for the rest. tIndex of brittleness is defined as a quantity (H/ Kc).20 ±Ra: roughness of spe The erosion tests were carried out by impingement of a sand/water slurry jet onto the ceramic surface at an impact angle of 45. Figure 1 is the schematic design of the slurry erosion test rig. both the ejector and the exit nozzle(3.7 mm bore) are made of stainless steel The system gives a sand/ water loading(slurry flow)of 4.6 l/ min, feed pressure of 1. 4 bar, and the sand concentration in the slurry was 5 wt%. The jet velocity was calculated and found to be 7.3 m/s. The sample holder was situated on a stand in the tank and was so designed that parameters such as the nominal impact angle, the distance between the target and the eject suction valve valve by-pass line Fig. 1. Schematic diagram of the sand/ water slurry erosion test apparatus
Effect of corrosion and erosion on ceramic materials Table 1. Properties of the materials investigated 513 Materials Density Grain Size (g/cm’) (pm) Hardness JL+ Fracture Toughness &(MPa rn’.‘) Brittlenesst (m-“.5Mpa-‘) RaS (urn) PSZ zirconia 85% alumina (rest SiO*) Sialon 101 SIC 6.05 1-2 1489 f 50 6 248 0.03 3.7 56 1169k50 4 292 0.7 3.24 2-3 1675+50 3 558 x=0.01 y = 0.04 3.10 4-6 2255 + 50 2.8 805 x=0.3 y=o.15 *The applied load of Vickers hardness tests of 10 kg for alumina, 50 kg for the rest. tIndex of brittleness is defined as a quantity (H/k&).2o $Ra: roughness of specimen surface. Equipment The erosion tests were carried out by impingement of a sand/water slurry jet onto the ceramic surface at an impact angle of 45”. Figure 1 is the schematic design of the slurry erosion test rig. Both the ejector and the exit nozzle (3.7 mm bore) are made of stainless steel. The system gives a sand/water loading (slurry flow) of 4.6 l/min, feed pressure of 1.4 bar, and the sand concentration in the slurry was 5 wt%. The jet velocity was calculated and found to be 7.3 m/s. The sample holder was situated on a stand in the tank and was so designed that parameters such as the nominal impact angle, the distance between the target and the ejector pressure Qaug Q elector II specimen / by-pass valve by-pass line Fig. 1. Schematic diagram of the sand/water slurry erosion test apparatus
nozzle as well as their relative position, can be easily adjusted. Currently a 15 mm distance is used between the target and the ejector nozzle. The construction details can be found elsewhere. 6.17 Experimental procedure Three samples per test were measured for each of the four ceramics under investigation Two main sets of experiments were carried out. The first involved corroding the sample for various lengths of time(up to 115 h)followed by erosion while the second was concerned with eroding the samples without any prior treatment. The average error in erosion tests was about 10%, while that in corrosion tests was less than 5%. In all cases monitoring by the following techniques was carried out (1) Weight loss: in all corrosion experiments, the sample was weighed periodicall eighing was also carried out before and after erosion (2)Hardness measurements: were carried out prior to and after corrosion as well as prior to and after erosion. The indentation marks and cracks which were made on the sample prior to corrosion were amplified as the corrosive solution found its way through the cracks This was very useful in identifying two modes of cracking which in essence form an integral part of the erosion mechanism. This is amplified later on in this paper ()Optical microscopy and sem were used to examine the surface morphology at each X-ray diffraction was used to investigate any phase changes, or preferential attack on phases due to corrosion →A203 1.5%HF=5%Hc Corrosion time(h) Fig. 2. Weight loss of alumina against corrosion time
514 Q. Fang et al nozzle as well as their relative position, can be easily adjusted. Currently a 15 mm distance is used between the target and the ejector nozzle. The construction details can be found elsewhere.6,‘7 Experimental procedure Three samples per test were measured for each of the four ceramics under investigation. Two main sets of experiments were carried out. The first involved corroding the sample for various lengths of time (up to 115 h) followed by erosion while the second was concerned with eroding the samples without any prior treatment. The average error in erosion tests was about lo%, while that in corrosion tests was less than 5%. In all cases monitoring by the following techniques was carried out. (1) Weight loss: in all corrosion experiments, the sample was weighed periodically. Weighing was also carried out before and after erosion. (2) Hardness measurements: were carried out prior to and after corrosion as well as prior to and after erosion. The indentation marks and cracks which were made on the sample prior to corrosion were amplified as the corrosive solution found its way through the cracks. This was very useful in identifying two modes of cracking which in essence form an integral part of the erosion mechanism. This is amplified later on in this paper. (3) Optical microscopy and SEM were used to examine the surface morphology at each stage. (4) X-ray diffraction was used to investigate any phase changes, or preferential attack on phases due to corrosion. (5) Atomic absorption spectroscopy (AAS) was used to determine dissolved elements from the ceramic in the corrosive liquid. 35 30 25 20 15 IO 5 0 0 20 40 60 60 100 120 Corrosion time (h) Fig. 2. Weight loss of alumina against corrosion time
Effect of corrosion and erosion on ceramic materials EXPERIMENTAL RESULTS AND DISCUSSION Corrosion resistance behaviour of ceramic specimens Figures 2 and 3 show the dependence of corrosion resistance of the ceramic samples after immersion in 1.5%HF+5% HCl solution in room temperature, on corrosion time Figure 2 shows the weight loss for Al2O3, while Fig. 3 is for Sialon, PS7-zirconia and SiC. It can be seen that there is a great difference in the corrosion resistance of the various cerami materials. The corrosion rates, which were plotted from the slope of the weight loss- corrosion time curves(Figs 2 and 3), for all specimens is also seen to decrease with corrosion time( Figs 4 and 5). Sic has the best corrosion resistance of the four ceramic specimens; the corrosion rate for Sic is only about 0.001 mg/cm*h after 115 h; while alumina( 85%Al O) has the least corrosion resistance with a corrosion rate about 0. 3 mg/em h after 115 h of immersion in the corrosive solution. There are many factors, which infuence the corrosion resistance of ceramic samples. The main factors are discussed below Acidic and basic character of the ceramic materials. According to the Lewis acid-base theory, in general, acidic oxides and compounds resist attack by acids and are prone to attack by bases. Pure Al2O3 and Zro2 are amphoteric, whereas SiC and Si3N4 are weakly acidic. The properties of mixed compound ceramics tend to be intermediate between those of their constituents; in the case of sialon, its acidic character is stronger than that of Al2O3 out weaker than that of Si3N4. The ranking obtained from the experiments shows the corrosion resistance of the various engineering ceramics to be Sic Sialon > ZrO2>Al2O3 This is in line with theoretical predictions Minor constituents. The criterion of acidity or basicity can only give a rough indication of the corrosion resistance of ceramics, because it can be greatly altered by the presence of minor constituents. These are commonly added to engineering ceramics in order to improve their properties and performance In the case of the ceramics used in the experiment 中zro2 Sialon - siC ig. 3. Weight loss of ZrO2, sialon and Sic against corrosion time
Effect of corrosion and erosion on ceramic materials 515 EXPERIMENTAL RESULTS AND DISCUSSION Corrosion resistance behaviour of ceramic specimens Figures 2 and 3 show the dependence of corrosion resistance of the ceramic samples, after immersion in 1.5% HF + 5% HCl solution in room temperature, on corrosion time. Figure 2 shows the weight loss for AlzOs, while Fig. 3 is for Sialon, PSZ-zirconia and Sic. It can be seen that there is a great difference in the corrosion resistance of the various ceramic materials. The corrosion rates, which were plotted from the slope of the weight losscorrosion time curves (Figs 2 and 3), for all specimens is also seen to decrease with corrosion time (Figs 4 and 5). Sic has the best corrosion resistance of the four ceramic specimens; the corrosion rate for Sic is only about 0.001 mg/cm2 h after 115 h; while alumina (85% Al,O,) has the least corrosion resistance with a corrosion rate about 0.3 mg/cm2 h after 115 h of immersion in the corrosive solution. There are many factors, which influence the corrosion resistance of ceramic samples. The main factors are discussed below: Acidic and basic character of the ceramic materials. According to the Lewis acid-base theory, in general, acidic oxides and compounds resist attack by acids and are prone to attack by bases. Pure A1203 and ZrOz are amphoteric, whereas Sic and S&N4 are weakly acidic. The properties of mixed compound ceramics tend to be intermediate between those of their constituents; in the case of sialon, its acidic character is stronger than that of A1203, but weaker than that of S&N+ The ranking obtained from the experiments shows the corrosion resistance of the various engineering ceramics to be SIC > Sialon > Zr02 > AlzOs. This is in line with theoretical predictions. Minor constituents. The criterion of acidity or basicity can only give a rough indication of the corrosion resistance of ceramics, because it can be greatly altered by the presence of minor constituents. These are commonly added to engineering ceramics in order to improve their properties and performance. In the case of the ceramics used in the experiment, 0.9 N- 5 0.8 2 0.76 2 9 0.5 . T 0 20 40 60 80 100 120 Corrosion time (h) Fig. 3. Weight loss of ZrO*, sialon and Sic against corrosion time
Q. Fang et al 08 ◆A2o3 HF + 5% HCI 05 room temperature coO 60 orrosion time(h) Fig. 4. Corrosion rate of alumina against corrosion time 0016 1.5% HF+5% HCI Zro2 0014 room temperatur ▲Sia 0012 iC 0004 0.002 60 Corrosion time(h) Fig. S. Corrosion rate of ZrO2, sialon and SiC against corrosion time. PSZ ZrO2(95%)is stabilised by the addition of MgO, Y2 O, and Sio. CI Silica is added to suppress grain growth during manufacture and to improve the mechanical properties, but it will confer a reduced corrosion to HF and HCl solution. Mgo surface complex ion was suggested as the surface reaction i& molecule. resistance towards as a basic ingredient will reduce resistance to acids, while SiO z has good acids but is easily attacked by HF; the substitution of HF n solution with a Al2O3 is mixed with 15% SiO2. As mentioned above, the presence of Sioz can be predicted to reduce the corrosion resistancc of Al2O3 in HF HCl solution and indeed this is corroborated by our results
20 40 60 80 Corrosion time (h) 0 100 120 Fig. 4. Corrosion rate of alumina against corrosion time. 0.016 1 1.5% HF + 5% HCI 0.014 -~ room temperature 0.012 -- M Urn I I I 0.01 -- I A 0.006 .- A A IZro2 A Sialon 1 l SiC j 1 I 0.006 -- A 0.004 .~ A A 0.002 A A 04. 0. : 0 0 0 0 0 0 20 40 60 80 Corrosion time (h) Y 100 120 Fig. 5. Corrosion rate of ZrOz, sialon and SC against corrosion time - RB SIC contained 2-10% non-reacted Si, which is insoluble in HF and HCl. - PSZ ZrOz (95%) is stabilised by the addition of MgO, Yz03, and Si02. - Silica is added to suppress grain growth during manufacture and to improve the mechanical properties, but it will confer a reduced corrosion to HF and HCl solution. MgO as a basic ingredient will reduce resistance to acids, while SiOz has good resistance towards acids but is easily attacked by HF; the substitution of HF molecules in solution with a surface complex ion was suggested as the surface reaction.‘* - A1203 is mixed with 15% SiOz. As mentioned above,” the presence of SiO;, can be predicted to reduce the corrosion resistance of A1203 in HF + HCl solution and indeed this is corroborated by our results
Effect of corrosion and erosion on ceramic materials Porosity in ceramics and roughness on surface. Reaction bonded SiC(RB SiC) is produced by the reaction of either liquid or gaseous silicon or Sio with carbon in a silicon carbide/carbon compact. This results in a porous body with a continuous silicon carbide phase, however these pores can be filled with non-reacted Si(2-10%)yielding a dense product that results in excellent mechanical properties and corrosion resistance. This is confirmed by the results which show good corrosion resistance of SiC. In the case of Al2O3 it is possible that the sintering agent, 15% Sio2(which is included during manufacture)is preferentially attacked by the corrosive solution which contains hydrofluoric acid Etching would increase the penetration of solution and result in the formation of a surface layer which is depleted in SiO2 and hence mechanically weaker. This is confirmed by XRD and AAS (see below). It is worth noticing that the corrosion rate of Al2O3 decreases with time at a higher rate than for the other three ceramic materials. As the corrosion time increases the depleted layer thickness increases. Access of the corrosive solution would thus be via diffusion through this layer and thus slow down with time. Another factor for the decrease in corrosion rate with time for Al203 compared to the other ceramics could be due to the nitial surface finish which was rough a rougher surface would increase the exposure area to the corrosive solution at the start of immersion In this case Al,O3 had a start surface finish of Ra=0.7 um, while the other materials had a smoother surface of 0.3-0.01 um Strictly speaking, using the rate of ceramic weight loss as a criterion for corrosion rate is not accurate because there are two different corrosion processes: one is that the outer surfac can dissolve into solution, while the other is due to corrosion of one phase leaving a porous and fragile outer layer which is then easily undermined by erosion. The weight loss in the latter case may not be as high as in the first case but the top layer material strength is lost all the same Figure 6 shows the undermined surface layer on the Al2O3 surface after corrosion for 115h. ZrO and sialon also behave in a similar manner while such a top surface degradation is not obvious in Sic Xrd analysis X-ray diffraction peaks of the tested ceramic specimens are shown in Figs 7-9 Figure 7 shows the X-ray difiraction peaks for alumina(85% Al2O3, 15% SiO2)before nd after immersion in 1.5% HF+5% HCl solution for 35h. Prior to immersion of the mple in the corrosive solution, other peaks, apart from Al2O3, which are mainly due to SiO2, exist. These decrease remarkably, as the immersion time increases indicating preferential attack by the corrodent on the Sioz. This means that silica is particularly Fig. 6. Corroded layer on the Alzo3 surface after corrosion for 1 15 h
Effect of corrosion and erosion on ceramic materials 517 Porosity in ceramics and roughness on surface. Reaction bonded Sic (RB Sic) is produced by the reaction of either liquid or gaseous silicon or SiO with carbon in a silicon carbide/carbon compact. This results in a porous body with a continuous silicon carbide phase, however these pores can be filled with non-reacted Si (2-10%) yielding a dense product that results in excellent mechanical properties and corrosion resistance. This is confirmed by the results which show good corrosion resistance of Sic. In the case of A1203, it is possible that the sintering agent, 15% SiOz (which is included during manufacture) is preferentially attacked by the corrosive solution which contains hydrofluoric acid. Etching would increase the penetration of solution and result in the formation of a surface layer which is depleted in Si02 and hence mechanically weaker. This is confirmed by XRD and AAS (see below). It is worth noticing that the corrosion rate of A1203 decreases with time at a higher rate than for the other three ceramic materials. As the corrosion time increases the depleted layer thickness increases. Access of the corrosive solution would thus be via diffusion through this layer and thus slow down with time. Another factor for the decrease in corrosion rate with time for AlzOs compared to the other ceramics could be due to the initial surface finish which was rough. A rougher surface would increase the exposure area to the corrosive solution at the start of immersion. In this case Al203 had a start surface finish of Ra =0.7 urn, while the other materials had a smoother surface of 0.3401 urn. Strictly speaking, using the rate of ceramic weight loss as a criterion for corrosion rate is not accurate because there are two different corrosion processes: one is that the outer surface can dissolve into solution, while the other is due to corrosion of one phase leaving a porous and fragile outer layer which is then easily undermined by erosion. The weight loss in the latter case may not be as high as in the first case but the top layer material strength is lost all the same. Figure 6 shows the undermined surface layer on the A120s surface after corrosion for 115 h. ZrOz and sialon also behave in a similar manner while such a top surface degradation is not obvious in Sic. XRD analysis X-ray diffraction peaks of the tested ceramic specimens are shown in Figs 7-9. Figure 7 shows the X-ray diffraction peaks for alumina (85% AlzOs, 15% SiO$ before and after immersion in 1.5% HF + 5% HCl solution for 35 h. Prior to immersion of the sample in the corrosive solution, other peaks, apart from A120s, which are mainly due to SiOz, exist. These decrease remarkably, as the immersion time increases indicating preferential attack by the corrodent on the SiOz. This means that silica is particularly Fig. 6
Q. Fang et al [counts] 7 (d) 500 20 (a) Fig. 7.X-ray diffraction peaks for alumina (a) Before corrosion; (b)after corrosion of I h; (c) after corrosion of 6h; (d )after corrosion of 30 h 1000 800 (b) 20 (a) Fig. 8. X-ray diffraction peaks for sialon. (a) Before corrosion; (b)after corrosion of 30h prone to hydrochloric-hydrofluoric acid mixtures solution attack. In engineering cerami additives are added to aid the sintering process of the ceramic and improve various properties, such as increased fracture toughness, phase stability or wear resistance. This
518 Q. Fang et al. 1000 [countal 900 I--- -._!___....._..s 700 1 Cd) ‘1 .-..J (cl 600 500 I--+ _ J’ ‘= 4.J Fig. 7. X-ray diffraction peaks for alumina. (a) Before corrosion; (b) after corrosion of I h; (c)after corrosion of 6 h; (d) after corrosion of 30 h. [countal 1400 1200 1000 800 600 400 200 0.0 - - u 0 Fig. 8. X-ray diffraction peaks for sialon. (a) Before corrosion; (b) after corrosion of 30 h prone to hydrochloric-hydrofluoric acid mixtures solution attack. In engineering ceramics, additives are added to aid the sintering process of the ceramic and improve various properties, such as increased fracture toughness, phase stability or wear resistance. This
Effect of corrosion and erosion on ceramic materials 519 [counter 1日D 1400 1400 1200 1000 m 600 m: ZrO, monoclinic phas t: ZrO, tetragonal phase. Fig9. X-ray diffraction peaks for PSZ zirconia (a) Reforecorrosion; (b )after corrosion of 30 h; (e) increasing diffraction peak of ZrOz-monoclinic phase in 20=28.23
Effect of corrosion and erosion on ceramic materials 519 kountml iaoo 1600 1000 600 I-- 600 400 1 200 0.0 [count6l- 1600. w m 600. 400- 200- m: ZrO, monoclinic phase. t: ZrO, tetragonal phase. t Fig. 9. X-ray diffraction peaks for PSZ zirconia. (a) Before corrosion; (b) after corrosion of 30 h; (c) increasing diffraction peak of ZrOz-monoclinic phase in 28 = 28.23
phase however is preferentially attacked by the corrosive medium leading to a depleted and weak surface layer, Figure 8 shows a similar tendency for sialon Figure 9 shows the X-ray diffraction peaks for PSZ zirconia. Before corrosion both the monoclinic and tetragonal structures are evident. After corrosion the peaks of monoclinic structures increase dramatically. ZrCl, or ZrF4 was not detected Increasing the length of corrosion time, the monoclinic structure is markedly increased. Obviously, the corroded layer is main source of contribution for increasing monoclinic phase. The doped tetragonal structure transformed into monoclinic phase; this may be due to the pure monoclinic phase being more stable than the tetragonal phase in PSZ zirconia at low temperatures. The possible reaction steps in HF HCl solution are as follows Psz-ZrO, HftHum-zro Y,O3+6HCI-2YCl3+3H,O SiO2+4HCl→→SiCl4+2H2C m-ZrO2+ 4HCI-ZrCl4 +2H2O m一ZrO2+HF→ZrF4+2H2O According to the view point of thermodynamics, a phase structure leans to the ansformation of a metastable phase to the more stable under the attack of a corrodent, just like that under the impact stress or tensile stress. However, the difference of the two cases is that induced by impacting stress, part of I-ZrO2 transforms into m-ZrO2 and microcracks develop, which consume crack propagation energy and release stress concentration, whilst t-m transformation in corrosion process forms a porous m-zirconia layer. From above experimental results, the surface structure of PSZ ZrO2 changes and bccomes morc porous during corrosion process, not only due to the t-m phase transformation, in which cracks and void are anticipated owing to 3-5% volume change but due to the effect of higher corrosion rate of additives in PSZ Zro2 as well. the hydrothermal effect of acidic solutions upon the corrosion of yttria(14 mol %)stabilised zirconia(YSZ) single crystals was investigated and a similar result was given by Yoshimura et al. 9 They found that rapid dissolution of yttria occurred forming an interface of polycrystalline monoclinic ZrO2 in acidic solutions(those containing HCl or H2 SO4)at 600C and 100 MPa for 24 h For SiC, the XRD peak has not significantly changed aftcr corrosion Analysis by atomic absorption spectrometry (AAS Table 2 shows the results of AAS for the corrosion solution after immersion of the ceramic samples for 115h Table 2. Chemical elemental analysis of the corrosive solution after corrosion of the ceramics for 115 h(ppn) A Mg Alumina 170±4 482+5 283+4 0±0.2 72+0.2 220±4 243+4
520 Q. Fang er al phase however is preferentially attacked by the corrosive medium leading to a depIeted and weak surface layer. Figure 8 shows a similar tendency for sialon. Figure 9 shows the X-ray diffraction peaks for PSZ zirconia. Before corrosion both the monoclinic and tetragonal structures are evident. After corrosion the peaks of monoclinic structures increase dramatically. ZrCl, or ZrF4 was not detected. Increasing the length of corrosion time, the monoclinic structure is markedly increased. Obviously, the corroded layer is main source of contribution for increasing monoclinic phase. The doped tetragonal structure transformed into monoclinic phase; this may be due to the pure monoclinic phase being more stable than the tetragonal phase in PSZ zirconia at low temperatures. The possible reaction steps in HF + HCl solution are as follows: PSZ - ZrOz - m - ZrOz YzOj + 6HCl R7, 2YCls + 3H20 SiO2 + 4HCl R’ SiC14 + 2H20 m - ZrOz f 4HCI z ZrC& + 2H20 m - ZrOz + HF R’, ZrF4 + 2H20 According to the view point of thermodynamics, a phase structure leans to the transformation of a metastable phase to the more stable under the attack of a corrodent, just like that under the impact stress or tensile stress. However, the difference of the two cases is that induced by impacting stress, part of f-ZrO, transforms into m-ZrOz and microcracks develop, which consume crack propagation energy and release stress concentration, whilst t-m transformation in corrosion process forms a porous m-zirconia layer. From above experimental results, the surface structure of PSZ ZrOz changes and becomes more porous during corrosion process, not only due to the t-m phase transformation, in which cracks and void are anticipated owing to 3-5% volume change; but due to the effect of higher corrosion rate of additives in PSZ ZrOz as well. The hydrothermal effect of acidic solutions upon the corrosion of yttria (14 mol%) stabilised zirconia (YSZ) single crystals was investigated and a similar result was given by Yoshimura et al.‘9 They found that rapid dissolution of yttria occurred forming an interface of polycrystalline monoclinic ZrOz in acidic solutions (those containing HCl or H2S04) at 600°C and 100 MPa for 24 h. For Sic, the XRD peak has not significantly changed after corrosion. Analysis by atomic absorption spectrometry (AAS) Table 2 shows the results of AAS for the corrosion solution after immersion of the ceramic samples for 115 h. Table 2. Chemical elemental analysis of the corrosive solution after corrosion of the ceramics for 1 I5 h (ppm) Sample Al Zr Si Mg Y Alumina 170+4 482k5 Zr02 1345+10 283k4 IO&O.2 <5 Sialon 7.2kO.2 220+4 Sic 2431-4