Vous êtes sur la page 1sur 8

Minerals Engineering 18 (2005) 481–488

This article is also available online at:


www.elsevier.com/locate/mineng

Corrosion protection of mild carbon steel media


in phosphate grinding mill using impressed current technology
a,*
D. Tao , G.L. Chen a, B.K. Parekh b

a
Department of Mining Engineering, University of Kentucky, 234E MMRB, Lexington, KY 40506-0107, USA
b
Center for Applied Energy Research, University of Kentucky, Lexington, KY 40514, USA

Received 20 May 2004; accepted 10 August 2004

Abstract

Corrosion plays a significant role in wear of mills used to grind acidic phosphate slurry at pH 2–4 in Florida fertilizer plants.
Approximately half of the grinding media wear results from corrosion or oxidation dissolution of metal surfaces in grinding. Catho-
dic protection using impressed current was investigated to reduce the wear rate. The effect of polarization potential on corrosion rate
and required current density were studied using a specially designed ball mill whose electrochemical potential can be controlled.
Experimental results indicate that the total wear rate was reduced by 42–46% for 1018 carbon steel when a potential of 1.0 V
was applied. The required current density to effectively reduce the wear rate was 210 mA/m2 in pH 3.1 solution, 180 mA/m2 in
pH 6.8 solution, and 160 mA/m2 in pH 9.2 solution. Scanning electron microscopy (SEM), and X-ray diffraction (XRD) methods
were used to investigate the surface morphology, corrosion products, and composition of metals during phosphate grinding with
and without cathodic protection. SEM results indicated that pitting corrosion was the main corrosion type. XRD images suggested
that Fe2O3 was the main corrosion product in pH 3.1 solution without cathodic protection, and the corrosion was significantly
reduced when potential of 1.0 V was applied.
Ó 2004 Elsevier Ltd. All rights reserved.

Keywords: Grinding; Industrial minerals; Process control

1. Introduction and potassium, is utilized as the major nutrient for


plants.
The United States is the largest phosphate rock pro- Grinding mills are commonly used in the Florida
ducer in the world. About 30% of the world production phosphate industry to reduce particle size. The corro-
was produced by the United States (Bartels and Gurr, sion of metallic grinding media and mill liner is a very
1994). The State of Florida accounts for approximately serious problem, particularly in acidic conditions as
80% of US phosphate production (Harben, 1980). The encountered in the Florida phosphate fertilizer industry
upgraded phosphate concentrate is widely used for (pH 2–4). At acidic pHs, protective passivation films of
tooth paste, detergents, food, fertilizers, etc. Approxi- metal oxides and/or hydroxides do not exist on the sur-
mately 95% of produced phosphate is consumed in fer- face of balls and liners, exposing fresh metal surface that
tilizer plants where phosphorus, together with nitrogen is more susceptible to corrosion. It is known that the
corrosive wear of grinding mill leads to an increase in
operating cost and plant downtime, a loss of process effi-
*
Corresponding author. Tel.: +1 859 257 2953; fax: +1 859 323
ciency, and product contamination. The direct operating
1962. costs in grinding are mainly the energy consumed and
E-mail address: dtao@engr.uky.edu (D. Tao). the metal lost through wear and corrosion (Allen,

0892-6875/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mineng.2004.08.003
482 D. Tao et al. / Minerals Engineering 18 (2005) 481–488

1993; Shishodia et al., 1993). An effective protection of per ring. The cathodic protection of the ball mill was
grinding mills from corrosion will improve the process accomplished using a 0.6 cm diameter and 30.5 cm
performance and economics, enhance product quality, length titanium-indium oxide, dimensionally stable
and increase the lifetime of mills and grinding media. anode (DSA) (acquired from Corrpro Companies Inc.,
Metal corrosion reactions are electrochemical in nat- Medina, OH) that was located in an anode compart-
ure and their reaction rates are controlled by the electr- ment and isolated from the slurry with a high-density
ochemical potential at the surface. A reduction in polyethylene (HDPE) membrane to prevent particulates
potential with excess electrons (e) will depress the ano- from contaminating the anode. Gases such as oxygen
dic dissolution reaction of metal (e.g., M = M2+ + 2e). and hydrogen produced at the anode were released from
This can be accomplished by supplying an impressed the anode chamber through a vent shown in Fig. 1. The
current to the object to create a negative potential potential and current applied to the coupons were con-
change called cathodic polarization which reduces the trolled using an EG & G PARC, Model 273 potentio-
rate of metal reaction. Fundamental corrosion analysis stat. An industrial standard calomel electrode (SCE)
indicated that a 0.1 V reduction in potential of the iron was used as the reference electrode.
or steel electrode can reduce the corrosion rate by more The coupons used to represent the grinding balls and
than 99% (Jones, 1996). mill shell were made of 1018 carbon steel acquired from
McMaster-Carr Supply Company, Cleveland, OH. The
compositions of 1018 carbon steel are shown in Table
2. Experimental 1. The coupon assembly consists of the metal coupons
in the center and the insulator around it. Total wear
Unlike earlier mill corrosion studies that used devices of metal was estimated based on the average weight loss
that merely simulated mill motion, the present study uti- of three coupons after grinding.
lized a specially designed ball mill whose electrochemical The phosphate samples and pond water used in this
potential can be controlled under operating conditions study were acquired from Florida CF Industries, Plant
(Kotlyar et al., 1987; Pozzo and Iwasaki, 1989; Pazhia- City, FL. The phosphate sample was thoroughly mixed
nur et al., 1997; Liu et al., 2001). The results from this and wet screened into seven different size fractions for
study more closely resemble those expected from the size distribution analysis and results are shown in Table
industrial ball mills. The mill was constructed of a stain- 2. It is clear that the particle is quite evenly distributed in
less steel pipe of 21.6 cm diameter and 30.5 cm length all size fractions except in 0.075 mm. The acidic phos-
with a wall thickness of 0.6 cm, as illustrated in Fig. 1. phogypsum pond water (pH 1.3) was used as part of
To characterize the weight loss of the ball mill shell or feed water, which is a standard industrial practice. The
its liner, three coupons made of the material represent- pond water was mainly comprised of water (90%), phos-
ing the mill shell or liner was installed flush with the phoric acid (<4%), and sodium potassium fluosilicate
interior surface of the mill. These coupons, 1.0 cm diam- (<1.5%). The contents of calcium sulfate, fluosilicic acid,
eter and 2.5 cm in height, were electrically isolated from
the mill. The potential or current of coupons was con-
trolled via copper brushes riding on the concentric cop-
Table 1
Compositions of 1018 carbon steel
Sample Fe (%) C (%) Mn (%) P (%) S (%)
1018 carbon 98.81–99.16 0.15–0.20 0.60–0.90 0.04 0.50
steel

Table 2
Phosphate size analysis results
P
Size fraction (mm) wt (%) wt (%)
+3.35 7.10 7.10
3.35–2.0 8.64 15.74
2.0–1.18 17.23 32.97
1.18–0.85 10.44 43.41
0.85–0.60 6.18 49.59
0.60–0.425 7.70 57.29
0.425–0.25 21.44 78.73
0.25–0.15 13.20 91.93
0.15–0.075 7.87 99.80
0.075 0.20 100.00
Fig. 1. Schematics of the specially designed mill for grinding tests.
D. Tao et al. / Minerals Engineering 18 (2005) 481–488 483

sulfuric acid, dissolved metallic impurities, fluorides (as -0.0


F), and monoammonium phosphate were less than 1%.
-0.2 pH 4.6

pH 6.8

3. Results and discussion -0.4


pH 9.2

In all grinding tests, the rotation speed of the grind- -0.6

Potential (V, SHE)


ing mill was fixed at 70 RPM which was 75% of the crit-
-0.8
ical speed. The solid percentage was chosen to be at 64%
which was close to the industrial value. The load per-
-1.0
centage was kept at 50% which was the middle level of
this factor, unless specified otherwise. The wear rate in -1.2
mills penetration per year (MPY) was calculated from
Eq. (1): -1.4
534W
MPY ¼ ; ð1Þ -1.6
DAT
where W is weight loss in milligrams, D is density in -1.8
grams per cubic centimeter, A is area in square inches, 0.01 0.1 1 10 100 1000 10000 100000
and T is time in hours (Jones, 1996). Current density (µ A/cm 2 )

Fig. 2. Potentiodynamic cathodic polarization curve of 1018 carbon


3.1. Polarization study steel in oxygenated pond water solutions.

The polarization diagram gives a fundamental and reached a similar conclusion from the study of effect
quantitative assessment of the change in corrosion rate of solution pH on cathodic polarization of pure iron
caused by cathodic polarization. It was established using
a standard three-electrode electrochemical cell control- 2Hþ þ 2e ! H2 ð2Þ
led by a potentiostat which applies a sweeping potential
2H2 O þ 2e ! H2 þ 2OH ð3Þ
to the working electrode made of the same metal as the
mill or grinding medium (Tao et al., 1994). Additional potentiodynamic polarization studies of
Since phosphate grinding takes place in solution mix- 1018 steel were also carried out in nitrogenated pond
ture of gypsum pond water and fresh water, polarization water solution. Fig. 3 shows potentiodynamic cathodic
studies were carried out in simulation solution made of
gypsum pond water and distilled water. The pH of undi-
-0.0
luted gypsum pond water has been measured to be 1.31.
The pond water and NaOH were used to make solutions
-0.2 pH 4.6
of pH 4.6, 6.8, and 9.2. The potentiodynamic cathodic
polarization curves of 1018 steel specimen in oxygenated -0.4 pH 6.8
pond water solutions at these pHs are shown in Fig. 2.
The pond water solution was bubbled with oxygen for -0.6
pH 9.2
Potential (V, SHE)

10 min prior to each experiment. It is obvious that metal


corroded much faster in lower pH solutions than in -0.8
higher pH solutions, as indicated by higher current den-
sity. At pH 4.6, there was about one decade of Tafel -1.0
behavior for the hydrogen reduction reaction before
concentration polarization began with an eventual limit- -1.2
ing current density of about 20 mA/cm2. As pH in-
creased to 6.8, hydrogen ion activity became much -1.4
lower, concentration polarization of reaction Eq. (2) be-
gan at much lower current density and the limiting cur- -1.6
rent density decreased to about 1.6 mA/cm2. In the pH
9.2 solution, hydrogen evolution by direct reduction of -1.8
0.01 0.1 1 10 100 1000 10000 100000
water (reaction Eq. (3)) occurred above limiting current
Current density ( µ A/cm2 )
density for reaction Eq. (2). It is clear that the limiting
current density at pH 4.6 was about 12 times greater Fig. 3. Potentiodynamic cathodic polarization curve of 1018 carbon
than that at pH 6.8. Jones (1996) and Stern (1955) steel in nitrogenated pond water solutions.
484 D. Tao et al. / Minerals Engineering 18 (2005) 481–488

polarization curves of 1018 steel specimen in nitrogen- ization potential in pH 3.1 solution, while it was 418
ated pond water solutions with pH values of 4.6, 6.8, MPY at 0.9 V potential in the same solution. Fig. 4
and 9.2, respectively. Again, metal corrodes much faster clearly indicates that further increasing the polarization
in lower pH solution than in higher solution, as indi- potential from 1.0 to 1.5 V did not improve the ball
cated by higher current density. The limiting current mill cathodic protection. This suggests that potential of
density was about 35 mA/cm2 at pH 4.6, which is about 1.0 V was sufficient to effectively reduce the wear rate
17 times greater than the limiting current density of 2.04 of 1018 carbon steel. Similar results were obtained with
mA/cm2 at pH 6.8. 1018 carbon steel in pH 6.8 and pH 9.2 solutions, except
Comparing Fig. 3 with Fig. 2 reveals that the limiting that the wear rate was higher in pH 3.1 than that in pH
current density is higher in nitrogenated pond water 6.8 and pH 9.2 solutions. These experimental data sug-
solution than in oxygenated pond water solution at gested that the total wear rate was reduced by 42–46%
the same pH value. This is believed to result from easier when a potential of 1.0 V was applied.
passivation by oxygen in oxygenated pond water The total wear rate of metal determined from the
solution. weight loss measurement has three components, which
are abrasive wear, corrosive wear and synergistic effect
3.2. Effect of polarization potential on corrosion rate of abrasive and corrosive wear. The polarization poten-
tial was mainly used to reduce the corrosive wear. The
Anodic polarization represents a driving force for corrosive wear rate was determined using the approach
corrosion by the anodic half cell reaction. When surface described elsewhere (Tao et al., 2004; Chen et al.,
potential measures more positive, the oxidizing (or cor- 2004) and the results are also shown in Fig. 4. This fig-
rosive) power of the solution increases. To investigate ure clearly indicates that the corrosive wear of 1018 car-
effects of polarized potential on corrosion rate of the bon steel decreased substantially with increasing
ball mill and grinding media, 3-h grinding tests were cathodic polarization potential. As shown in Fig. 4,
conducted using the experimental setup described earlier the corrosive wear rate was 291 MPY without polariza-
under controlled electrochemical conditions. The effects tion potential in pH 3.1 solution, which was reduced to
of polarization potential on wear rate of 1018 carbon 39 MPY at 1.0 V polarization potential in the same
steel in different solutions were carried out and the re- solution. The corrosive wear rate of 1018 carbon steel
sults at pH 3.1 are shown in Fig. 4. It is obvious that was reduced by 86% in pH 3.1 solution. Furthermore,
wear rate decreased with increasing cathodic polariza- cathodic protection was also effective for reducing the
tion potential or making the potential more negative. synergistic effect of abrasive wear and corrosive wear
For example, the wear rate was 760 MPY without polar- as indicated by 51% reduction in pH 3.1 solution by
applying 1.0 V polarization potential. But the effect
of polarization potential on the abrasive wear was neg-
800 ligible. Pazhianur et al. (1997) reported similar result
that cathodic protection reduced materials loss in tum-
Total Wear
700 Abrasive Wear
bling mills by 30–60% when grinding tests were con-
S ynergistic Effect
Corrosive Wear
ducted on quartz, pyrite and quartz–pyrite mixtures at
600
pH levels of 4.6, 7.1 and 9.2.

3.3. Required current density for effective protection


Wear Rate (MPY)

500

The ball mill can be protected from corrosive wear


400
using the impressed current technique which utilizes an
external power source to force current out of the
300 DSA, into the electrolyte, and eventually onto the ball
mill. The impressed current cathodically polarizes the
200 ball mill and reduces its corrosion rate. The magnitude
of impressed current determines the effectiveness of cor-
100 rosive wear rate reduction. Previous work (Emenike,
1995) indicated that the main requirement for effective
0
cathodic protection is that the protective system pro-
0.0 -0.2 -0.4 -0.6 -0.8 -1.0 -1.2 -1.4 -1.6 vides enough current to polarize the structure.
Potential (V, SHE) To investigate effects of impressed current density on
Fig. 4. Effect of polarization potential on corrosion rate of 1018 corrosion rate of the ball mill and grinding media, 3-h
carbon steel at pH 3.1, rotation speed 70 RPM, crop load 50% and grinding tests were conducted using the experimental
solid percentage 64%. setup described elsewhere (Tao et al., 2004; Chen
D. Tao et al. / Minerals Engineering 18 (2005) 481–488 485

et al., 2004) under controlled electrochemical conditions. steel structures in the North Sea ranged from 50 to 215
The impressed current density of the ball mill was varied mA/m2.
at 10, 50, 100, 150, 160, 180, 210, and 250 mA/m2 during As described earlier, the impressed current density
the entire grinding period. The change in the total wear, was mainly used to reduce the corrosive wear. The cor-
corrosive wear, abrasive wear, and synergistic effect of rosive wear rate was determined using the same ap-
abrasive wear and corrosive wear were determined for proach and the results are also shown in Fig. 5. This
each test and correlated with applied current density. figure clearly indicates that the corrosive wear of 1018
The difference in the correlation at varying pHs indicates carbon steel decreased significantly with increasing im-
the effect of solution pH on corrosion rate, and possibly pressed current density. As shown in Fig. 5, the corro-
corrosion mechanism as well. Thus, the optimum sive wear rate was 293 MPY without impressed
magnitude of impressed current density was readily current density in pH 3.1 solution. It reduced to 39
identified. MPY at 210 mA/m2 impressed current density in the
The effects of impressed current density on wear rate same solution. The corrosive wear rate of 1018 carbon
of 1018 carbon steel in pH 3.1 solution are shown in Fig. steel was reduced by 85% in pH 3.1 solution. Cathodic
5. It is obvious that the total wear rate decreased with protection was also effective for reducing the synergistic
increasing impressed current density. For example, the effect of abrasive wear and corrosive wear. For example,
total wear rate was 760 MPY without impressed current the synergistic effect of abrasive wear and corrosive wear
density in pH 3.1 solution, while it was 412 MPY at 210 was reduced by 47% in pH 3.1 solution by applying 210
mA/m2 impressed current density in the same solution. mA/m2 current density. But the effect of impressed cur-
Fig. 5 clearly reveals that the wear rate maintained a rent density on the abrasive wear was negligible. Similar
constant value when the impressed current density was results were obtained with 1018 carbon steel in pH 6.8
increased from 210 to 250 mA/m2. This suggests that im- and pH 9.2 solutions, except that the corrosive wear rate
pressed current density of 210 mA/m2 was sufficient to was reduced by 83% in pH 6.8 solution and 85% in pH
effectively reduce the wear rate of 1018 carbon steel in 9.2 solution; the synergistic effect of abrasive wear and
pH 3.1 solution. Fig. 5 also suggests that the total wear corrosive wear was reduced by 40% in pH 6.8 solution
rate was reduced by 45% in pH 3.1 solution when cur- and 54% in pH 9.2 solution, respectively. These results
rent density of 210 mA/m2 was applied. The required are consistent with those obtained from potential con-
current densities for mill protection are in good agree- trol. As described earlier, the total wear rate was re-
ment with that obtained for cathodic protection of steel duced by 51% and corrosive wear rate was reduced by
structures by Gartland et al. (1983) who concluded that 86% in pH 3.1 by applying 1.0 V polarization
the required current densities for cathodic protection of potential.
Comparing these results reveals that the required cur-
rent density for effective protection was higher for acidic
800
solutions than for neutral or alkaline solutions. For
example, an impressed current density of 210 mA/m2
Total Wear
Abrasive Wear provided sufficient protection from corrosive wear for
700 Synergistic Effect
Corrosive Wear
pH 3.1 solution, while the required current density was
180 mA/m2 for pH 6.8 solution and 160 mA/m2 for
600
pH 9.2 solution.
Wear Rate (MPY)

500
3.4. Corrosion mechanism

400 3.4.1. SEM analysis


The microstructures of the corroded surfaces of the
300 coupons obtained at different pH values under different
impressed currents were investigated by SEM. Figs. 6
200 and 7 show the SEM images of 1018 carbon steel after
it was exposed to phosphate slurry at pH 3.1 for 3 h with
100 and without cathodic protection, respectively. In Fig. 6,
the gouges and scratches on the surface indicate wear
0 due to corrosion and abrasion. The darker spots might
0 40 80 120 160 200 240 280 be due to the formation of an oxide. This figure clearly
Current Density (mA/m 2) shows that there are a lot of deep and shallow pits on the
Fig. 5. Effect of impressed current density on corrosion rate of 1018 surface. This suggests that pitting corrosion was the
carbon steel at pH 3.1, rotation speed 70 RPM, crop load 50% and main corrosion mechanism. Pitting corrosion was
solid percentage 64%. caused by localized attack in an otherwise resistant
486 D. Tao et al. / Minerals Engineering 18 (2005) 481–488

Fig. 6. SEM image of 1018 carbon steel exposed to phosphate slurry at


Fig. 7. SEM image of 1018 carbon steel exposed to phosphate slurry at
pH 3.1 without cathodic protection.
pH 3.1 with polarization potential 1.0 V.

surface. The rate of corrosion was greater in some areas protection. This suggests that Fe2O3 was the main cor-
than others. Pitting corrosion was caused by the pres- rosion product in pH 3.1 solution (Wu and Kim,
ence of phosphoric acid, fluosilicic acid, and sulfuric 1999). Fig. 9 shows the XRD image when polarization
acid that manage to pass through the passive film and potential 1.0 V was applied. This figure reveals that
initiate corrosion, resulting in rupture of the passive corrosion was considerably reduced since Fe2O3 disap-
film. Fig. 7 shows the SEM image when polarization peared from the coupon surface. This conclusion is in
potential 1.0 V was applied. This image shows fewer good agreement with the Eh-pH diagram of Fe–O–H
dark spots, indicating a reduction of corrosion on the system, which indicates that Fe2+ and Fe2O3 were the
surface. Similar types of results were obtained with most stable phases in solution pH 3.1 without polariza-
1018 carbon steel in pH 6.8 and pH 9.2 solutions. tion potential, while Fe was the most stable phase when
polarization potential 1.0 V was applied due to the
3.5. XRD analysis polarized surface. Many investigators (Liu et al., 1997;
Zhang et al., 1998; Panda et al., 2001) studied the sur-
The structure of the corrosion product of the cou- face properties of Fe, Fe2O3, and Fe3O4 films formed
pons obtained after 3-h grinding at different pH value under different conditions. The characteristic peaks of
was investigated by X-ray spectrometer. Figs. 8 and 9 Fe and Fe2O3 obtained from this study are consistent
show the XRD images of 1018 carbon steel at pH 3.1 with these studies.
with and without cathodic protection, respectively. Similar results were obtained with 1018 carbon steel in
Fig. 8 indicates that there existed iron (Fe) and ferric pH 6.8 and pH 9.2 solutions, except that the corrosion
oxide (Fe2O3) on the coupon surface without cathodic products were ferric oxide (Fe2O3), magnetite (Fe3O4)

Fig. 8. XRD image of 1018 carbon steel exposed to phosphate slurry at pH 3.1 without cathodic protection.
D. Tao et al. / Minerals Engineering 18 (2005) 481–488 487

Fig. 9. XRD image of 1018 carbon steel exposed to phosphate slurry at pH 3.1 with polarization potential 1.0 V.

and ferric hydroxide (Fe(OH)3) in pH 6.8 solution and under Project Number 00-01-170 which made this work
pH 9.2 solution (Shan et al., 2002; Raymahashay and possible.
Khare, 2003).

References
4. Conclusions
Allen, M.P., 1993. The vernal phosphate rock mill. In: El-Shall, H.,
Based on the above results, the following conclusions Moudgil, B.M., Wiegel, R. (Eds.), Beneficiation of Phosphate:
can be drawn: Theory and Practice. SME, Littleton, CO.
Bartels, J.J., Gurr, T.M., 1994. Phosphate rock. In: Carr, D.D. (Ed.),
Industrial Minerals and Rocks, sixth ed. SME, Littleton, CO, pp.
(1) Polarization diagrams indicated that the current
751–764.
density was higher in nitrogenated solution than in Chen, G.L., Tao, D., Parekh, B.K., 2004. Wear rate analysis for high
oxygenated solution at the same pH value. chromium alloy media in phosphate grinding mill, Wear, in press.
(2) Polarization potential of 1.0 V was sufficient to Emenike, C.O., 1995. Mathematical modeling of cathodic protection
effectively reduce the wear rate of 1018 carbon steel. systems for oil and gas facilities. Anti-corros. Meth. Mater. 42 (4),
6–8.
The total wear rate was reduced by 42–46% and cor-
Gartland, P.O., Strommen, R., Bardal, E., 1983. Current density
rosive wear rate was reduced by 84–86% for 1018 requirements for cathodic protection of steel structures in the
carbon steel when potential of 1.0 V was applied North Sea. Mater. Performance 22 (6), 40–48.
at pHs of 3.1–9.2. Harben, P., 1980. Ind. Miner. 148, 48–55.
(3) The required current density to effectively reduce the Jones, D.A., 1996. Principles and Prevention of Corrosion, second ed.,
pp. 3–31.
wear rate of 1018 carbon steel was 210 mA/m2 in pH
Kotlyar, D., Pitt, C.H., Wadsworth, M.E., 1987. Corrosion-NACE 44
3.1 solution, 180 mA/m2 in pH 6.8 solution, and 160 (4), 221–227.
mA/m2 in pH 9.2 solution. Liu, P.S., Liang, K.M., Gu, S.R., 2001. Corros. Sci. 43 (7), 1217–1226.
(4) SEM analysis indicated that pitting corrosion was Liu, Y., Zhu, W., Tan, O.K., Shen, Y., 1997. Structural and gas
the main corrosion mechanism. SEM photographs sensing properties of ultrafine Fe2O3 prepared by plasma enhanced
suggested that corrosion was significantly reduced chemical vapor deposition. Mater. Sci. Technol. B 47, 171–176.
Pazhianur, R., Adel, G.T., Yoon, R.-H., Richardson, P.E., 1997.
after polarization potential 1.0 V was applied. Cathodic protection to minimize corrosive wear in ball mills.
(5) XRD images suggested that Fe2O3 was the main Miner. Metall. Process. 11, 1–7.
corrosion product for 1018 carbon steel in pH 3.1 Panda, R.N., Gajbhiye, N.S., Balaji, G., 2001. Magnetic properties of
solution without cathodic protection, while Fe2O3, interacting single domain Fe3O4 particles. J. Alloys Compd. 326,
50–53.
Fe3O4 and Fe(OH)3 were the corrosion products in
Pozzo, R.L., Iwasaki, I., 1989. J. Electrochem. Soc. 136 (6), 1734–1739.
pH 6.8 and pH 9.2 solutions. Raymahashay, B.C., Khare, A.S., 2003. The arsenic cycle in late
quaternary fluvial sediments: mineralogical considerations. Curr.
Sci. 84 (8), 1102–1104.
Shan, G.B., Liu, H.Z., Xing, J.M., Zhang, G.D., Ma, Z.Y., Liu, X.Q.,
Liu, J.G., 2002. Surface modification of nano-magnetic Fe3O4
Acknowledgements
particles and its applications, EU Forum on Nanosized Technol-
ogy, Beijing, China, December, pp. 68–76.
The authors gratefully acknowledge the financial sup- Shishodia, V., Rastogi, S., Pitchumani, B., 1993. Effect of ball size
port of Florida Institute of Phosphate Research (FIPR) distribution on rock phosphate grinding in industrial ball mill. In:
488 D. Tao et al. / Minerals Engineering 18 (2005) 481–488

El-Shall, S.Q., Moudgil, B.M., Wiegel, R. (Eds.), 1998, Beneficiation Wu, X., Kim, S.B., 1999. Synthesis and electrode properties of a-Fe2O3
of Phosphate: Theory and Practice. SME, Littleton, CO, p. 1993. from iron phthalocyanine. Electrochem. Solid-State Lett. 2 (4),
Stern, M., 1955. J. Electrochem. Soc. 102, 609. 184–186.
Tao, D., Chen, G.L., Parekh, B.K., 2004. Statistical analysis of wear Zhang, H.Y., Chen, J., He, Y.Y., Xue, X.M., Peng, S.Q., 1998. The
rate of grinding mill in phosphate industry, Corrosion, in press. preparation of carbon-coated iron nanocrystals produced from
Tao, D.P., Li, Y.Q., Richardson, P.E., Yoon, R.-H., 1994. The Fe2O3-containing composite anode in arc discharge. Mater. Chem.
incipient oxidation of pyrite. Colloids Surfaces 93, 229–239. Phys. 55, 167–170.