Wear, 103
(1985)
253
CORROSIVE
253
- 261
AND ABRASIVE
WEAR IN ORE GRINDING*
I. IWASAKI, S. C. RIEMER and J. N. ORLICH Mineral (U.S.A.)
Resources
Research
Center,
University
of Minnesota,
Minneapolis,
MN
55455
K. A. NATARAJAN Indian
Institute
of Science,
Department
of Metallurgy,
Bangalore
560012
(India)
(Received March 26, 1985; accepted May 9, 1985)
Summary
The relative significance of corrosive and abrasive wear in ore grinding is discussed. Laboratory marked ball wear tests were carried out with magnetic taconite and quartzite under different conditions, namely dry, wet and in the presence of an organic liquid. The effect of different modes of aeration and of pyrrhotite addition on the ball wear using mild steel, high carbon low alloy steel and austenitic stainless steel balls was evaluated. Results indicate that abrasive wear plays a significant role in ore grinding in the absence of sulfides, and rheological properties of the ore slurry influenced such wear. The effect of oxygen on corrosive wear becomes increasingly felt in the presence of a sulfide mineral such as pyrrhotite. Wear characteristics of the three types of ball materials under different grinding conditions are illustrated.
1. Introduction Consumption of grinding media forms a significant part of the operating cost of a mineral processing plant. An estimated 0.25 X lo9 kg of grinding balls and rods in the U.S.A. and over 0.5 X lo9 kg in the world are consumed each year in wet grinding alone [ 11. The Minnesota taconite iron ore industry is considered to be the largest single regional consumer of grinding media and it has been estimated that about 55% of all rods and 32% of all balls consumed in the U.S.A. are used by the Minnesota taconite mills [2]. With the present trend towards lower grade ores requiring finer grind for liberation, the amount of media wear per tonne of concentrate recovered will be steadily on the increase. Hence studies leading to a better understanding of the principles underlying grinding ball wear must be advanced and thereby the means of minimizing the wear must be developed. *Paper presented at the International Conference on Wear of Materials, Vancouver, Canada, April 14 - l&1985. 0043-1648/85/$3.30
@ Elsevier Sequoia/Printed in The Netherlands
254
It has been known that the total wear of grinding media in ball mills is made up of abrasion, corrosion and impact, but it is difficult to isolate the relative contributions from each towards overall wear under the conditions in a mill. Nascent metals as well as mineral surfaces are continually being exposed in grinding and the electrochemical interactions between abraded and unabraded areas of the surface as well as between the grinding media and the minerals can occur in ground pulps. Media wear in dry grinding is reported by Bond [3] to average about one-seventh that of wet grinding of the same material, implying the seriousness of corrosion. The role played by corrosive wear was also demonstrated by using corrosion inhibitors in grinding [4,5] and by flushing nitrogen through a ball mill [6]. This work was undertaken in order to establish the relative significance of corrosive and abrasive wear in ore grinding. In laboratory ball mills, impact wear or spalling ,in macro-scale does not occur. Ball wear data obtained from laboratory dry and wet grinding tests with two types of ores, namely magnetic taconite and quartzite, are analyzed. The influence of different types of aeration, namely nitrogen, air and oxygen, on the wear of mild steel, high carbon low alloy (HCLA) steel and austenitic stainless steel balls was established. The results obtained from laboratory marked ball wear tests under dry and wet grinding conditions are compared with those of grinding in the presence of an organic liquid that does not promote electrochemical corrosion. The effect of addition of a sulfide mineral, pyrrhotite, to magnetic taconite and quartzite on the corrosive wear of the balls is also brought out under different aeration conditions, since small amounts of pyrrhotite are often present as impurities in many magnetite ores [ 71.
2. Experimental details Marked ball wear tests were carried out in a 203 mm Abbe procelain mill on a magnetic taconite sample of specific gravity 3.3 with 34.6% Fe and on a quartzite sample with 0.64% Fe and 0.02% S, both initially crushed to -10 mesh. Three different grinding ball materials were used, i.e. plain carbon steel (AISI 1020), HCLA steel (0.8% C, 1% Mn, 0.5% Cr) and austenitic stainless steel (type 304). Although mild steel balls are not used in commercial grinding operations, they were used in this study since they exhibit active behavior through a wide range of pH with relatively poor abrasion resistance. HCLA steels are harder (Rockwell A hardness, 85 HRA, as compared with 40 HRA for mild steel), having good abrasion resistance, and are used widely in industrial grinding mills. Austenitic stainless steel exhibits high corrosion resistance and thus a comparison of the ball wear data between the three types of steels could bring out the role of corrosive and abrasive wear. All the balls were in the as-forged condition, having a nominal diameter of 25.4 mm. The mill charge consisted of 1150 g of the crushed ore and 770 ml of distilled water (60% solids) unless specified otherwise together with 126
255
balls, 14 of which were marked. To control the atmosphere inside the mill, nitrogen, air or oxygen was flushed through a tubing inserted into the mill through a hole drilled at the closed end of the mill. To study the effect of an added sulfide mineral on the ball wear, pyrrhotite (crushed to -10 mesh, concentrated with a dry magnetic separator and demagnetized) containing 35.3% Fe and 23.3% S was added to the ore charge, maintaining the total mass of the mill feed at 1150 g. Ball preparation, cleaning and weighing procedures are given elsewhere [8]. Three to four grinding tests were performed for each experimental condition and the average mass loss reported as milligrams per ball. The size distribution as well as the Blaine (air permeability) surface area of the ground ore were also determined. For examination of ball surfaces under a scanning electron microscope, one unmarked ball was set aside prior to the standardized ball cleaning procedure for wear determinations [9]. A Hitachi S-450 scanning electron microscope was used. The manner in which ground pulp coats a grinding ball was investigated by subjecting a mild steel ball on a three-pronged wire stand to a stream of pulp discharging from a funnel through a tube of 6 mm diameter. The coated ball was weighed wet and dry, and the pulp layer thickness was calculated. Pulp viscosities were measured with a Brookfield RVT viscometer with a T-bar. Calibration and operating conditions are given elsewhere [lo].
3. Results 3.1. Dry versus wet grinding To study the difference in ball wear between dry and wet grinding, marked ball wear tests were carried out on a taconite sample with mild steel balls. The wear rate was markedly higher for wet grinding (37.1 mg h-’ ball’) than for dry grinding (5.4 mg h-’ ball-‘). In dry grinding, the balls were observed to be packed by ground ore particles, limiting continuous exposure of ball surfaces to abrasion. When the size distributions of the ground products after 60 min of grinding for the two cases are compared (Table l), it is evident that dry grinding was considerably less efficient than wet grinding. Hence the ball wear data for different periods of grinding are plotted against surface area increase due to grinding in Fig. 1. It is interesting to note that although the slopes of the lines, or the ball wear for a unit of new surface created, were approximately the same for dry and wet grinding, wet grinding produced a larger surface area change. From the dry grinding data for quartzite in Table 2, it is clear that the abrasion resistance of austenitic stainless steel balls is superior to that of mild steel balls but lower than that of HCLA steel balls. Also, the quartzite was found to be more abrasive than the taconite. The percentage increase in ball wear on changing from dry to wet grinding for the quartzite was 119%, 86% and 82% for mild steel, HCLA steel and austenitic stainless steel respectively .
256 TABLE 1 Size distribution of magnetic taconite balls, 60 min grind) Cumulative
Size (mesh) 65 100 150 200 270 400
under different grinding conditions
percentage
passing under the following
(mild steel
conditions
Dry
Organic liquid
Weta
99.0 98.7 96.7 86.9 76.6 67.3
100.0 100.0 99.7 96.2 86.9 74.0
100.0 100.0 100.0 98.9 93.2 80.8
aUnder nitrogen flushing condition.
60’) 6-
0 1000
‘I’I’IIII’I”I’i
1500
Bloine
2000 Surface
Area
Change
2500 (c.m2/g)
Fig. 1. Mild steel ball wear as a function of Blaine surface area change of ground magnetic taconite [ 7 ] (the numerals in parentheses represent the grinding time in minutes). TABLE 2 Ball wear and surface area increase in quartzite grinding under different conditions min grind)
(60
Grinding condition
Mild steel
HCLA steel
Austenitic stainless steel
Ball wear (mg ball-‘) BY Wet (air) Organic liquid
21.7 f 0.8 47.6 f 0.7 57.2 f 0.8
14.7 f. 0.7 27.3 + 1.5 31.2 + 2.6
17.8 f 0.4 32.4 + 0.5 38.3 f 0.5
0.212 0.253 0.197
0.213 0.244 0.225
Blaine surface area increase (mZ g-l) Dry 0.158 Wet (air) 0.195 Organic liquid 0.166
257
3.2. Grinding with organic liquid Since the electrochemical corrosion of steels is expected to be negligible in organic liquids of low electrical conductance, it was thought that a comparison of the results of marked ball wear tests using water with those using an organic liquid might identify the contribution from corrosive wear towards overall wear. A mixture of ethylene glycol and methanol (53 vol.% and 47 vol.% respectively) with a density close to that of water was used. The organic liquid mixture had a density of 970 kg rne3, a viscosity of 3.0 mPa s and a surface tension of 38.5 mN m-l. The wear rate of mild steel balls at 60% solids with the organic liquid mixture (67.7 mg h-’ ball-‘) was notably higher than that in distilled water. However, the size distribution of the ground ore was coarser with the organic liquid than with distilled water (Table 1). The use of the organic liquid, therefore, led to higher ball wear for a unit of new surface created than when water was used (Fig. 1). In Table 2 the ball wear in the presence of the organic liquid mixture for quartzite grinding with the three types of balls is shown. Replacing the aqueous medium with the organic liquid increased the ball wear for all types of steel balls used in this study, although in varying degrees. The wear of mild steel, HCLA steel and austenitic stainless steel balls was found to increase by 30%, 15% and 19% respectively. The surface area increase for the ground product was smaller with the organic liquid than with distilled water. It appears, therefore, that the grinding efficiency was poorer with the organic liquid although the ball wear was higher. This was taken to mean that grinding with the organic liquid increased the abrasive wear of balls since electrochemical corrosion was insignificant under the conditions [ 81. The surface area increase for the three types of balls is again noted to be related to the hardness of the balls. 3.3. Wet grinding in nitrogen atmosphere For corrosion of steel in aqueous media, oxygen plays a major role and hence the marked ball wear tests under nitrogen flushing conditions would minimize the corrosive wear. Evidence that this is the case was given in the marked ball grinding test results in the presence of oxygen and pyrrhotite as will be discussed later, which was supported by the scanning electron microscopy (SEM) observations of grinding balls. Photomicrographs of typical surfaces of mild steel balls after wet grinding of taconite at 60% solids for 60 min under nitrogen flushing and oxygen flushing conditions are given in Fig. 2. The surface under nitrogen flushing conditions showed numerous indentations with sharp ridges, but very few gouges or scratch streaks, indicating that grinding with mild steel balls involves nipping rather than attrition. Under oxygen flushing conditions the ridges became rounded and the appearance of corrosion was particularly evident in the presence of pyrrhotite. It should be noted that the fine details of the indentation marks are also lost by corrosion. Perez and Moore [9] attributed such a surface texture to strain-induced corrosion. The absence of the strain-induced cor-
258
(b)
(a)
(cl Fig. 2. Scanning electron micrographs of mild steel ball surfaces after 60 min of grinding of magnetic taconite at 60% solids: (a) taconite in nitrogen; (b) taconite in oxygen; (c) taconite plus 10% pyrrhotite in oxygen.
rosion would suggest flushing conditions.
that
corrosive
wear
was minimal
under
nitrogen
3.4. Ball coating and pulp viscosity determinations In a recent study on abrasive wear in taconite grinding [ll], ball wear as well as product size distribution were shown to be strongly dependent on
259
per cent solids of pulps. To investigate the role that per cent solids plays, pulp viscosity and ball coating characteristics were determined as functions of per cent solids and mesh of grind, and attempts were made to relate these to ball wear and grinding rate. Figure 3 shows the pulp viscosity, ball wear and grinding rate expressed as the net mass of -325 mesh material produced after 60 min of grinding under nitrogen flushing conditions. In these tests the pulp volume was held constant at 840 ml (106% of the void volume of the ball charge) to simulate the overflow discharge mills. As the per cent solids of the pulp increased, the ball wear reached a maximum at 40% solids and then decreased, while the grinding rate increased steadily. Beyond 75% solids the pulp viscosity increased sharply which, in turn, slowed the grinding rate. After grinding, balls were observed to be coated heavily with a thick pulp. Hence the thickness of pulp layers that coat a mild steel ball was determined at each per cent solids, and the results are plotted against pulp viscosity in Fig. 4. The 40 -
6000
1
= 0 ,” 0 30 E
: B E
L g 20= P 10 -
OL
0
20
40 60 % Solids
60
Fig. 3. Effect of percentage solids on the wear of mild steel balls, net mass of -325 material produced and pulp viscosity after grinding magnetic taconite for 60 min.
Brookfield
Viscosity,
mesh
mPo.s
Fig. 4. Effect of pulp viscosity on pulp layer thickness around a mild steel ball with ground magnetic taconite pulps.
260
figure includes for 30 min. It with the pulp grinding action
3.5.
additional data points that were obtained with pulps ground is apparent that the pulp layer thickness increased directly viscosity, suggesting that too thick a coating cushions the between balls.
Wet grinding in the presence of oxygen and pyrrhotite
Grinding media wear for the three types of steel balls in the wet grinding of taconite and quartzite is illustrated in Table 3. The pH of the ore slurry in the grinding of taconite remained in the narrow range 8.1 - 8.4 in the presence and absence of pyrrhotite. In the grinding of quartzite, the pH of the slurry under oxygen aeration remained at 6.7 - 6.9, while a higher pH in the range 7.1 - 7.6 was observed under aeration with air and nitrogen. Ball wear, in general, was higher in the grinding of quartzite than taconite. HCLA steel balls exhibited the lowest wear in the absence of pyrrhotite, while austenitic stainless steel balls registered the lowest mass loss in the presence of the sulfide mineral under oxygen atmosphere. In the absence of pyrrhotite, aeration with oxygen increased the ball wear for mild steel by about 13% - 16% and for HCLA steel by about 9% lo%, in comparison with a deoxygenated mill for both taconite and quartzite grinding. The presence of pyrrhotite, in contrast, significantly increased the ball wear for both mild steel and HCLA steel, especially in the presence of oxygen. Ball wear for austenitic stainless steel balls remained virtually
TABLE
3
Grinding media wear in the wet grinding of magnetic taconite ence and absence of pyrrhotite (60 min grind, distilled water)
Ball material and aeration conditions
and quartzite
in the pres-
Ball wear a (mg ball-‘) Taconite alone
Taconite with 10% pyrrho tite
Quartzite alone
Quartzite with 10% pyrrhotite
33.6 37.3 38.1
f 0.4 + 0.7 f 0.5
34.8 38.6 55.5
* 0.4 + 0.5 + 0.3
44.0 47.6 51.3
+ 0.7 f 0.7 f 0.6
48.4 49.1 59.2
f 0.6 f 0.6 f 0.5
13.0 13.8 14.2
f 1.5 + 0.8 f 0.8
15.1 21.7 40.5
f 1.5 If: 2.3 f 3.5
27.3 27.3 29.9
f 1.6 + 1.5 f 1.4
27.3 29.4 48.4
f 1.0 t 1.3 f 4.9
19.9 18.5
f 0.3 f 1.1
32.1 + 0.6
33.0 f 0.5
34.2
-
18.7 f 0.2
32.1 + 0.4
Mild steel N2
Air 02
HCLA steel N2
Air 02
Austenitic stainless steel N2
19.5 f 0.1
Air
18.5
02
18.9 f 0.9
aMultiply
by 0.11
to convert
+ 0.1
the ball wear to kilograms
per tonne.
+ 0.5
33.7 * 0.8
261
the same, irrespective of the type of aeration in both the presence and the absence of pyrrhotite. The manner in which ball wear of the three types of balls is affected by the addition of pyrrhotite to taconite was investigated under different aeration conditions [12] and the results are illustrated in Fig. 5. For the mild steel balls, the ball wear was found to increase sharply with pyrrhotite addition up to about 196, beyond which no further significant increase in ball wear was observed in the presence of oxygen. Ball wear for HCLA steel increased with pyrrhotite addition and no steady state value for the wear was observed within the concentration ranges of pyrrhotite used in this study. For both types of steels, an oxygen atmosphere in the mill contributed to the increase in ball wear. Ball wear for austenitic stainless steel balls was influenced neither by changes in the type of aeration nor by pyrrhotite addition.
l-
o
0
’
’ ’ ’ ’ ’ 2 4 6 Percent Pyrrhofite
8
’ 8 Added
’
’ 10
Fig. 5. Effect of the level of pyrrhotite addition on the (denoted by MS), HCLA steel and austenitic stainless ASS) balls in magnetic taconite grinding (60 min) [ 111.
wear of mild steel steel (denoted by
4. Discussion 4.1. Relative significance of corrosion and abrasion Initially, the investigation was directed towards establishing the extent of corrosive wear in the total ball wear. In dry grinding, ball wear will be free of corrosive wear. Increased ball consumption observed on switching to wet grinding is often interpreted as due to a significant increase in corrosive wear. For example, conversion from wet to dry grinding at the Wabush mines pellet plant reduced ball wear from 3 to about 1 kg t-l [13]. This difference, however, may not be attributable solely to corrosion. As is evident from
262
Table 1 and Fig. 1, wet grinding resulted in a greater size reduction and ball wear than dry grinding. The ball coating tests demonstrated that, in wet grinding, the balls were coated with ground pulps to different thicknesses depending on per cent solids and mesh of grind, indicating a fundamental difference in the mode of grinding and wear from dry grinding. In the present investigation, it was originally thought that comparative grinding studies in the presence of water and an organic liquid would distinctly bring out the contributions from abrasion and corrosion towards overall ball wear but instead grinding of the taconite and quartzite in the presence of the organic instead of water increased the abrasive wear of the balls. However, there was no corresponding increase in size reduction of the ore. Dirmeik and Greig [14] compared the effect of water and an organic liquid (1: 10 mixture of carbon tetrachloride and a silicone fluid) and reported that the weight loss of the test rods in the organic liquid was about 3% lower than that in the presence of water although the corrosion rate for the steel specimens in water was almost two orders of magnitude higher than that in the organic liquid. The results reported by Dirmeik and Greig, therefore, are diametrically opposite to those of the present work. Apparently, organic liquids influence the slurry medium differently from water. In a study of the abrasive wear of mild steel and cast iron by mineral slurries, Bhattacharya and Bock [15] reported that water produced a 67% higher coefficient of friction than mineral oil. In the SEM photographs (Fig. 2) it was observed that mild steel ball surfaces were virtually free of corrosion when grinding was carried out under nitrogen flushing conditions. Such an observation may be taken to mean that the difference in metal wear under air or oxygen atmosphere and under nitrogen represents corrosive wear. In the present investigation, therefore, the ball wear under nitrogen atmosphere is assumed to represent essentially abrasive wear, and the electrochemistry of corrosive wear is discussed. In this manner the contribution from oxygen towards wear in the wet grinding of the taconite and quartzite with HCLA steel and mild steel balls was of the order of 9% - 16%. In the presence of pyrrhotite the ball wear in nitrogen atmosphere increased very little, whereas in the presence of oxygen the increase was roughly proportional to its partial pressure inside the grinding mill. Austenitic stainless steel balls, however, were found to be highly corrosion resistant and the wear remained essentially unchanged irrespective of the nature of aeration or the presence or absence of pyrrhotite in the wet grinding of taconite and quartzite. 4.2. Abrasive wear Grinding in a ball mill involves breakage via direct impact of falling balls as well as indirect impact propagated through other balls and abrasion between spinning balls. Ball wear in a mill operating at low (cascading) speeds is proportional to the square of its diameter [16 - 181, which suggests that abrasion by spinning balls may be the predominant mode of grinding. The pulp should be thick enough so that the ball surfaces are coated properly
263
for maximum exposure to grinding action, yet it should not be too thick to act as a cushion between balls [ 191. If the pulp is too dilute, the ball surfaces are insufficiently coated and the energy is wasted in wearing the balls. Although pulp viscosity is known to exert a major influence on ball wear, the mechanism has not been described in literature. Klimpel [20] has recently brought out the importance of rheological properties in ball mill grinding on the specific rates of breakages of minerals in the presence and absence of chemical additives. The effect of viscosity and surface tension of various liquids on grinding efficiency has also been studied by Meloy and Crabtree [21]. Conversely, rheological properties of the slurry in a grinding mill may have an influence on the abrasion of ball materials as well. The coefficient of friction, an important parameter influencing the wear of materials, is also affected by the type of metal (ball material), concentration of abrasive slurries and by solution properties such as viscosity and pH [ 151. The effect of pulp viscosity on size reduction, therefore, appears to be related to the manner in which pulps coat the balls and to the rheological behavior of the pulp at the ball surface. In Fig. 3 it is evident that the ball wear and the size reduction data as a function of per cent solids are closely related to pulp viscosity. The optimum condition appears to be at about 70% solids or a little higher where the grinding rate is near maximum and the ball wear is steadily decreasing with increasing per cent solids. At this condition the pulp layer thickness was about 150 pm. By comparing with the size distribution data, grinding rates were beginning to be adversely affected when the pulp layer thickness exceeded, say, a few times the maximum particle size in the ground product at these per cent solids. Excessive pulp layer thickness cushions the balls and decreases not only the grinding rates but also the ball wear. The absence of gouges and scratch streaks on mild steel ball surfaces in the SEM photomicrographs indicates that grinding involves nipping and breakage through compression as well as impact between balls rather than attrition through differential movement of balls. The abrasive wear, therefore, may be visualized to involve mechanical cutting action by hard mineral particles at the point of contact. HCLA balls, in contrast, improved not only the abrasive wear but also the grinding efficiency [ll]. SEM photomicrographs showed gouges and scratch streaks together with indentation marks by mineral particles indicating that both nipping and attrition may be involved. Apparently, the hardness of ball materials influenced not only the wear behavior but also the grinding efficiency. 4.3. Corrosive wear In the presence of electroactive minerals such as pyrrhotite and magnetite, galvanic coupling between the mineral and the grinding ball in the slurry would generate a corrosion current which would be further increased as the oxygen partial pressure is increased. The mineral surfaces serve as cathodes for oxygen reduction whereas steel balls, which exhibit active potentials with reference to most sulfides and oxide minerals, act as anodes,
264 TABLE
4
Electrochemical data for the magnetite-ball of surface abrasion [ 211
Type of electrode combination and aeration
Magnetite-mild NZ
material
No abrasion: steady state value after 1 h Combination potentiala (mv)
Galvanic current (PA)
-570 -450 -330
couples
in the presence
and absence
A brasion (against -48, +65 mesh of quartz bed): steady state value after 15 20 min Corn bination potentiala (mV
Galvanic current @A)
0.5 0.3 0.05
-660 -560 -500
1.8 1.5 0.5 - 1.5
-510 -400 -300
0.9 0.4 0.25
-600 -480 -420
0.9 - 1.0 0.9 - 1.2 0.9 - 1.2
-174 -110 -92
0.09 0.07 0.04
-700 -380 -350
1.8 - 2.0 1.5 - 2.0 1.0 - 2.0
steel
Air 02
Magnetite-HCLA steel N2
Air 02
Magnetite-austenitic stainless steel N2
Air 02 aWith respect
to a saturated
calomel
electrode.
thereby promoting anodic dissolution [ 22,231. Repeated galvanic contacts between the mineral particles and grinding balls could readily be established inside the mill due to (a) agitation of the slurry inside the mill with the balls in constant motion, thereby exposing nascent mineral surfaces to freshly abraded ball surfaces, and (b) the possibility of attachment of ore particles to the ball due to magnetic and/or physicochemical interactions. Galvanic currents generated between magnetite-ball material electrode couples under stationary conditions and under conditions of continuous surface abrasion are given in Table 4. The results show that the galvanic currents of magnetite-metal couples depended strongly on the corrosion resistance of the metal in the absence of surface abrasion, while in the presence of surface abrasion the galvanic currents became less sensitive to the presence of oxygen. An attempt was made to establish a correlation between the electrochemical data and the grinding test results. Assuming that the difference in metal wear under air or oxygen atmosphere and under nitrogen represents corrosive wear, the difference in their weight losses may be converted to corrosion currents using Faraday’s law. Then the average current density
265
at the magnetite surface is estimated by taking the arithmetic average of the initial and the final surface areas and by assuming that the magnetite content of the taconite sample is l/3. The corrosion current thereby calculated for mild steel under oxygen flushing conditions is 0.8 PA cm-‘. This value is in good agreement with the galvanic current values with abrasion (Table 4), which suggests that the effect of oxygen on wear is due mainly to galvanic interaction involving the following reactions: ‘0, 2
+ Hz0 + Ze- -
Fe0 --+
Fe2+ + 2e-
20H-
on magnetite surfaces (cathode)
on ball surfaces (anode)
A closer examination of Table 3 shows that the presence of magnetite appears to provide added protection to the austenitic stainless steel balls. The corrosion rate of austenitic stainless steels was found to be drastically reduced in contact with a cathodic protector such as magnetite even in sulfuric acid [24]. Such a cathodic protection due to the presence of magnetite would be effective only for readily passivating metals or alloys (stainless steel, for example). The presence of a sulfide mineral such as pyrrhotite in either taconite or quartzite significantly increased the ball wear for mild steel and HCLA steel balls, especially in the presence of oxygen and such an increase has been attributed to corrosive wear. Electrochemical measurements indicated that the galvanic currents generated in a mild steel-pyrrhotite or HCLA steelpyrrhotite combination were several times higher than those observed in a magnetite-ball material couple [ 121. The corrosion currents, estimated from corrosive wear data in a laboratory mill, and the galvanic currents measured between metal-pyrrhotite electrodes under non-abrasive and abrasive conditions were shown to be in reasonably good agreement [23] as in the case of metal-magnetite interaction discussed earlier. It is also noted in Table 3 that, with mild steel and HCLA steel balls, the corrosive effect of the sulfide mineral was greater in the presence of magnetite as compared with that of quartzite. For example, the per cent wear increase for mild steel balls was 59.5% and 22.3%, and for HCLA steel balls was 167.4% and 76.9%, in the wet grinding of taconite and quartzite respectively in the presence of 10 wt.% of pyrrhotite and oxygen. Thus the presence of magnetite appears to be playing two different roles: decreasing the corrosive wear of stainless steel balls as mentioned earlier, while increasing the corrosion of mild and HCLA steels. Galvanic currents in a magnetite-pyrrhotite-ball material combination were higher than those observed with pyrrhotite-ball material couples and increased further in the presence of oxygen [12]. The corrosive wear of grinding balls thus also depends on the nature and type of possible mineral and mineral-ball material combinations inside the mill. The presence of more than one electroactive mineral in an ore slurry would have an effect on ball wear different from that of a single active mineral or an inert mineral system.
266
Acknowledgments The financial support provided for this project by the National Science Foundation under Grant CPE 8018395 and the Department of Energy under Contract DE-FC07-83ID12441 is gratefully acknowledged.
References 1 J. F. Remark and 0. J. Wick, Corrosion control in ball and rod mills, Corrosion-76, Houston, TX, 1976, Paper 121. 2 D. E. Nass, Steel grinding media used in the United States and Canada, Materials for the Mining Industry, Climax Molybdenum Company, Ann Arbor, MI, 1974, pp. 173 183. F. C. Bond, Metal wear in crushing and grinding, Chem. Eng. Prog., 60 (1964) 90 100. G. R. Hoey, W. Dingley and C. Freeman, Corrosion inhibitors reduce ball wear in grinding sulphide ore, CZM Bull., 68 (1975) 120 - 123. G. R. Hoey, W. Dingley and C. Freeman, Corrosion behavior of various steels in ore grinding, CIM Bull., 70 (1977) 105 - 109. D. A. Rice, Pilot plant evaluation of corrosion in wet grinding mills, Proc. 44th Annu. Mining Symp., Center for Continuation Study, University of Minnesota, Minneapolis, MN, 1983, Paper 20. J. J. Pavlica and I. Iwasaki, Electrochemical and magnetic interactions in pyrrhotite flotation, Trans. Sot. Min. Eng. AIME, 272 (1982) 1885 - 1890. K. A. Natarajan, S. C. Riemer and I. Iwasaki, Corrosive and erosive wear in magnetic taconite grinding, Miner. Metall. Process., 1 (1984) 10 - 14. R. Perez and J. J. Moore, The application of the scanning electron microscope in determining the interaction of abrasion and corrosion in mineral grinding media, Microstruct. Sci., 11 (1982) 207 - 234. 10 J. N. Orlich, The effect of slurry rheology on the erosive wear of grinding media, MS. Thesis, University of Minnesota, 1984. 11 I. Iwasaki, S. C. Riemer and J. N. Orlich, Erosive wear in taconite grinding, SMEAZME Preprint 84-316, 1984 (Society of Mining Engineers of AIME). 12 K. A. Natarajan, S. C. Riemer and I. Iwasaki, Influence of pyrrhotite on the corrosive wear of grinding balls in magnetite ore grinding, Znt. J. Miner. Process., I3 (1984) 73 - 81.
13 A. Sobering and D. N. Carlson, Dry grinding at the Wabush mines pellet plant, CZM Bull., 64 (1971) 254 - 258. 14 I. D. Dirmeik and J. D. Greig, Corrosion as a factor in the wear of tube-mill liners, J. S. Afr. Inst. Min. Metall., 68 (July 1967) 684 - 686. 15 S. Bhattacharya and F. C. Bock, Abrasive wear of engineering materials by mineral and industrial wastes, Wear, 46 (1978) 1 - 18. 16 T. K. Prentice, Ball wear in cylindrical mills, Trans. AIME, 169 (1946) 147 - 154. 17 T. E. Norman and C. M. Loeb, Jr., Wear tests on grinding balls, Trans. AZME, 176 (1948) 490 - 520. 18 R. T. Hukki, Correlation between principal parameters affecting mechanical ball wear, Trans. AZME, 190 (1954) 642 - 644. 19 E. J. Pryor, Mineral Processing, Elsevier, New York, 1965, 844 pp. 20 R. Klimpel, Slurry rheology influence on the performance of mineral/coal grinding circuits, Min. Eng., 34 (1982) 1665 - 1672; 35 (1983) 21 - 26.
267 21 T. P. Meloy and D. Crabtree, Surface tension and viscosity in wet grinding, Proc. 2nd Eur. Symp. on Comminution, Amsterdam, in DECHEMA-Monogr., 57 (1967) 405 426. 22 K. A. Natarajan and I. Iwasaki, Electrochemical aspects of grinding media-mineral interactions in magnetite ore grinding, Int. J. Miner. Process., 13 (1984) 53 - 71. 23 K. Adam, K. A. Natarajan, S. C. Riemer and I. Iwasaki, Electrochemical aspects of grinding media-mineral interaction in sulfide ore grinding, Corrosion-85, Boston, MA, 1985, Paper 362. 24 N. D. Tomashov and G. P. Chernova, Passivity and Protection of Metals Against Corrosion, Plenum, New York, 1967, pp. 151 - 158.