Wear of Anodized Aluminum under Three-Body Conditions - American

Department of Chemistry and Center for Surface and Coatings Research, Lehigh ... to the contact pressure and inversely proportional to the apparent de...
0 downloads 0 Views 626KB Size
Ind. Eng. Chem. Prod. Res. Dev. 1986, 25, 473-478

473

Wear of Anodized Aluminum under Three-Body Conditions Shem-Mong Chout and Henry Leldhelser, Jr." Department of Chemistry and Center for Surface and Coatings Research, Lehigh University, Bethlehem, Pennsylvania 180 15

The rates of wear of the anodic coating on an 1100 aluminum alloy were determined by gravimetric and optical methods. The coatings were formed at 10-30 "C and 10-30-V cell voltage. The wear rates were proportional to the contact pressure and inversely proportional to the apparent density. A wear theory was developed for porous anodized aluminum worn under three-body conditions.

Introduction This paper represents the third in a series dealing with the wear of anodized aluminum. The first paper (Noble and Leidheiser, 1981) described the experimental technique for simulating the wear process and applied the technique to anodized aluminum lithographic plates. The second paper (Hughes et al., 1983) outlined a sensitive optical technique for measuring the thickness of the anodized coating and demonstrated the application of the technique to studies of sealed anodized aluminum and to polymer coatings on anodized aluminum. The present study is focused on the wear process of flat anodized aluminum panels under conditions comparable to those that exist when another flat surface is in abrasive contact in a repetitious sliding motion. The conditions were selected as an approximation to the conditions that exist during lithographic printing with an abrasive ink. Experimental Section The aluminum used in this study was an 1100 alloy expressly anodized by the manufacturer for use as a lithographic plate. The anodic coating was removed by immersing the plate in a mixture of 2% Cr03 and 5% H3P04at 95 "C for 5-10 min according to ASTM Standard B137. The metal was then etched in an 8% NaOH electrolyte, desmutted in HN03, chemically brightened in a H3P04-HN03solution, and thoroughly washed with distilled water immediately before anodizing. The anodization was carried out in 15% H2S04under different conditions of electrolyte temperature and anodizing voltage. The thickness of the oxide coating was controlled by varying the anodizing time. Commercial electrolysis grade lead was used as the cathode. Electrolyte circulation was produced by compressed air, and the temperature of the anodizing solution was controlled by a water bath. After anodization, the panel was immersed in a 5% NaHC03 solution for 10 s to neutralize the residual acid, thoroughly washed with distilled water, and finally stored in a desiccator until testing. The coated plates were worn by the procedure outlined previously (Noble and Leidheiser, 1981). This procedure involves a back-and-forth motion of a weighted block against the sample plate immersed in an aqueous suspension of Sic abrasive. The average particle size of the abrasive was approximately 1 pm. The concentration of abrasive was 2 g of Sic dispersed in 40 mL of distilled water except as otherwise indicated. Contact pressure was adjusted by either changing the block or adding dead weight to the shaft. t Present address: Rockwell International, Graphic Systems Division, Chicago, IL.

0 196-432 1 f86/ 1225-0473$0l.50/0

Two techniques were used to evaluate the wear rate, weight loss and specular reflectance measurements. The weight of oxide film prior to and after wear was determined to 0.1 mg according to ASTM Standard B137. The optical technique for determining the film thickness has been described previously (Hughes et al., 1983).

Results Rates of Formation of the Anodic Coating on Aluminum. Data obtained from gravimetric measurements over a 30-min anodization time have been given previously (Hughes et al., 1983). Data are presented in Figures 1and 2 for the formation rates over a 200-min time interval. The data in Figure 1indicate that under all the experimental conditions utilized the thickness of the coating increased linearly with time. Figure 2, however, shows that the mass of the coating increased at less than a linear rate with anodization time. These two sets of data permit one to calculate the apparent density of the anodic coating as a function of coating thickness. Figure 3 gives data at 10, 20, and 30 "C at a fixed anodizing voltage of 15 V, and Figure 4 gives data at 10, 15, and 20 V at a fixed temperature of 20 "C. In all cases it will be noted that the apparent density decreases with an increase in coating thickness with the effects being the most noteworthy at an anodizing voltage of 10 V. Rate of Wear of the Anodic Coating. Since the wear resistance of a material is not an absolute value, but is dependent upon the material itself and the wear environment, a series of experiments was carried out to determine the effect of contact pressure and abrasive concentration on the wear rate of anodized aluminum. Figures 5 and 6 show that the wear rate is directly proportional to the contact pressure for different anodizing times at both 20 and 30 "C. A transition was observed in all cases. The contact pressure at the transition point shifted to higher values for coatings exhibiting lower wear rates. Results for the wear rate as a function of abrasive concentration are shown in Figure 7 . The wear rates increased very rapidly at low concentrations and tended to level off at higher concentrations. The effect of anodizing conditions on the wear rate of the coatings on aluminum is summarized in Figures 8 and 9. The wear rate increased with the thickness of the anodic coating, the effect being more significant at higher electrolyte temperatures (Figure 8) or lower anodizing voltages (Figure 9). Discussion The wear results are well in accord with the known structure of the anodic oxide formed on aluminum in a sulfuric acid electrolyte. The simultaneous formation and 0 1986 American Chemical Society

474

Ind. Eng. Chem. Prod. Res. Dev., Vol. 25, No. 3, 1986 30

2.E

2.4

ANODIZATION TIME (min)

Figure 1. Rates of formation of the anodic coatings on aluminum determined by the interference method as a function of anodizing time. 20~C.20V

0

5

10

15

20

25

Thickness (pm)

Figure 4. Dependence of the apparent density of anodic coatings on the film thickness as a function of the anodizing voltage a t a constant temperature of 20 "C.

2O"C.lIV

t(min)

p'

100

50

0

45

200

150

ANODIZATION TIME (min)

Figure 2. Rates of formation of the anodic coatings on aluminum determined by a gravimetric method as a function of anodizing time.

Contact Pressure (g/cm2)

Figure 5. Effect of contact pressure on the wear rate of anodic coatings produced in 15% H2S04a t 30 "C and 15 V for different anodizing times.

2.0

c I 0

I 5

I 10

I

I

I

15

20

25

Thickness (rm)

Figure 3. Dependence of the apparent density of anodic coatings on the film thickness as a function of the electrolyte temperature at a constant voltage of 15 V.

dissolution of the oxide at selected points during anodization results in the formation of a porous oxide consisting of an approximately hexagonal arrangement of cells (Keller

et al., 1953). A t the bottom of the pores there is a thin layer of nonporous oxide, known as the barrier layer. This barrier layer is below the thickness of the oxide removed in the wear studies made herein and thus is not considered in this discussion. The cell size and barrier layer thickness are dependent upon the electrolyte and the applied voltage. The pore diameter and pore shape are determined t o an important extent by the electrolyte and the temperature of the electrolyte, but the pore diameter is also a function of the forming voltage (Wood and O'Sullivan, 1970). The chemical dissolution of the cell wall by anodizing electrolyte (Nagayama et al., 1970) results in an increase in pore diameter at the surface of the oxide film with anodizing time (Wood and O'Sullivan, 1970). The outermost oxide, formed first during the anodization process, undergoes the most severe chemical dissolution, resulting in the formation of conical rather than cylindrical pores. The rate of chemical dissolution increases gradually with

Ind. Eng. Chem. Prod. Res. Dev., Vol. 25,

No. 3, 1986 475

30d t h90 id

50

15

-

10

-

20 5 -

0

L 0

I

I

5

10

I 15

I

20

I 25

T h i c k n e s s (pm)

Figure 8. Rate of wear of the anodic coatings formed in 15% HzS04 at 15 V and different electrolyte temperatures as a function of film thickness. Contact pressure was 22.9 g/cm2. 20 I

20

40

80

60

Contact Pressure (g/cm2)

Figure 6. Effect of contact pressure on the wear rate of anodic coatings produced in 15% H2S04a t 20 OC at 15 V for different anodizing times.

Lo

2

15

x 0

. 0 r

E, "

?:/

0

40 -

10

01 I

m

a L

*

5

i 30

I 5

C

-

I 15

I

10

I

20

I 25

Thickness (rm)

20

-

i

0

Figure 9. Rate of wear of the anodic coatings formed in 15% H2SO4 at 20 OC and different anodizing voltages as a function of film thickness. Contact pressure was 22.9 g/cm2.

A-

Abrasive Concentration ( g / m L ) Figure 7. Effect of abrasive concentration on the wear rate of anodized aluminum. Anodic coatings were prepared (a) a t 20 OC and 15 V for 20 min and (b) a t 20 OC and 10 V for 140 min. The contact pressure was 22.9 g/cm2.

anodizing time, owing to the increased surface area of the oxide coating. When the pore mouths merge into each other at the surface of the oxide coating, the dissolution rate becomes equal to the formation rate and the oxide coating reaches its limiting thickness. This picture of a porous oxide coating with a density that increases with depth into the coating thus allows us to explain the data in Figures 1 and 2. The range of experimental conditions used to form the oxide are within the range where the pores do not merge into one another and

41

a a

g 0

2.3

2.4

2.5

2.6

2.7

Apparent Density ( g l c m 3 )

Figure 10. Linear relationship between the wear resistance and apparent density of anodized aluminum formed in 15% H2SOb Triangles and circles represent the data obtained from Figures 5 and 6, respectively.

the thickness, as determined by the optical method, is linear as a function of time of anodization, and thus the data in Figure 1result. However, there is an increase in porosity with film thickness as the anodizing temperature is increased at constant voltage and as the voltage is decreased at constant temperature. These effects are illustrated by the curvature of the lines in the data summarized in Figure 2.

476

Ind. Eng. Chem. Prod. Res. Dev., Vol. 25, No. 3, 1986 Microchip Ridge

Ridge

. Grain

t-dS-4 (a)

dtldS

=k

[qC/(l

- qC)] P /

H,

dtldS

=k

(d,/H,) (qCl(1

-

qC)] P / d,

Figure 12. Models of material removal by plastic deformation through (a) cutting and (b) plowing mechanisms.

0 0 37

0 39

0 41

0 43

0 45

I I A p p a r e n t Density (cm3/g)

Figure 11. Dependence of the wear rate on the apparent density of anodic coatings formed in a wide range of conditions.

The data in Figures 5 and 6 along with the calculated apparent density in Figures 3 and 4 permit the calculation of wear resistance (defined as the ratio of contact pressure to wear rate) as a function of apparent density. Such data are plotted in Figure 10, in which it will be noted that the wear resistance increases linearly with the apparent density. A similar relationship between abrasion resistance and apparent coating density was shown by Deal (1959), but Deal's relation was not unambiguous in the density range used in the present study. Figure 11shows that the wear rates are inversely proportional to the apparent density, as seen when the data in Figures 8 and 9 are plotted against the apparent density. Thus, it appears that the wear properties are solely a function of the porosity of the anodic coating formed in sulfuric acid: the denser the coating, the greater the resistance to wear. The data shown in Figure 7 indicate that the rate of wear is a function of the concentration of the abrasive. Valdma (1959) observed that there was an optimum composition of corundum powder in an oil which caused a maximum wear of the rubbing surface. The analysis of frictional torque and the condition of the worn surface led him to the conclusion that for mixtures richer in abrasive than optimum, the rolling of the abrasive particles dominated and the wear was decreased. An optimum concentration of abrasive slurry was not observed in this study. This discrepancy can be accounted for when the differences in the experimental conditions are considered. The device used in Valdma's experiment was a closed system in which all the abrasive particles were entrapped between a rotating disk and the specimen. As the abrasive concentration exceeds the optimum composition, a multilayer of abrasives is built up. Most abrasive particles make a rolling action, and the number of abrasive particles carried by the rotating disk was decreased there was a consequent decrease in both the frictional torque and the wear. The wear apparatus used in this study is different in that it is an open system. A large fraction of the abrasive particles is pushed out of the interaction zone (the interfacial region between the moving block and the specimen surface) by the moving block, and only a small fraction is entrapped. The moving block squeezes the abrasive particles present in the interaction zone into a compact layer, one abrasive particle thick, and it is only these particles that participate in the abrasive wear. The solid lines drawn

in Figure 7 through experimental points fit quite well an equation of the Langmuir type. In the Langmuir adsorption equation, the extent of adsorption or the surface coverage increases with gas pressure and approaches unity at very high pressures when the absorbent surface is completely covered with a close-packed layer of the adsorbate. If we equate the fraction of maximum wear rate W / W,, with the concentration of abrasive particles C in the suspension, the equation then takes the form

W/W"

= KC/(l+ k C )

(1)

The constant k can be calculated to be 9.0 and 10.8 for the anodic coatings formed at 10 and 15 V, respectively. The similarity of these values for two different conditions is additional evidence that the number of abrasive particles per unit area present in the interaction zone is governed by a law similar to the Langmuir adsorption isotherm. The microhardness of anodized aluminum coatings has been shown by Sautter (1966) to be a function of thickness. The hardness is highest for the layer adjacent to the metal and decreases progressively from the metal to the surface, the effect becoming less as the electrolyte temperature is reduced. This behavior is very similar to that of the wear rate (Figures 5 and 6) and the apparent density (Figures 3 and 4). These findings are all consistent with a tapered pore structure. Figure 1 2 details the material removal by cutting and plowing mechanisms. When an abrasive particle with its sharp tip contacts the surface and the attack angle of the abrasive particle to the surface is greater than a critical value, it will cut a microchip from the surface and leave ridges on both sides of the groove (Figure 12a). If the attack angle is less than the critical value for a sharp tip or if the abrasive particle contacts the surface with its blunt side, it will plow up material and form ridges around the abrasive particle as demonstrated in Figure 12b. The front ridge will be removed as the abrasive particle reaches the edge of the specimen. Buttery and Archard (1970/1971) have observed in scratch tests that repeated transversals of an indentor on the same track caused further removal from the side ridges. The breakaway of the side ridges results from the repeated plastic deformation. The amount of surface material removed by the plow mechanism is much less than that removed by the cutting mechanism. Most abrasive particles merely roll over the surface and make elastic contact with the surface and thus do not produce any abrasive wear debris. The volume dvi of material removed by the abrasive particle i moving through a distance dS can be expressed as

Ind. Eng. Chem. Prod. Res. Dev., Vol. 25, No. 3, 1986 477 dUi

= fiA,i d S

(2)

where fi is the fraction of the groove volume removed as wear debris and A is the groove area swept by the abrasive particle i. The vaue of fi may vary from 0 to a maximum value that depends upon the shape and orientation of the abrasive particle in contact with the wearing surface and the relative hardness of the abrasive particles to that of the surface material. The vertical force may be expressed as Li = H,A,i

(3)

where Li is the load carried by the abrasive particle i, H, is the surface hardness of the anodic coating, and Aciis the horizontal projected area of contact. The surface hardness is an apparent property of the anodic coating because of its porous structure. For convenience, the anodic coating is assumed to be homogeneous in the further development of the wear equation. Since both A, and Aciare dependent upon the shape and hardness of the abrasive particle i, they can be related by A,i = f{Aci

(4)

(6)

where g is a geometric factor for the abrasive particles, d is the mean diameter of the particle, and P is the contact pressure. The mean volume removal per particle, du, is given by du = gfd2PdS/H,

(7)

where f is the mean value of the product of fi and f{. It has been shown previously that the wear is a function of the number of particles in the interaction by a law similar to the Langmuir adsorption isotherm (eq 1). We then equate W / W,, to N/No,where Nois the maximum number of particles per square centimeter that can be accommodated. Equation 1 becomes N = NokC/(l + kC) (8) The maximum number of abrasive particles per unit area in the abrasion region is dependent upon the size and shape of the abrasive particles and can be expressed as

No = l/g’d2

(9)

where g’ is a geometric constant. The total volume removal per unit area, dVA,i.e., the change in the apparent thickness of the specimen, dt, can be obtained by combining eq 7-9 to yield dt = dVA = N du = k’[kC/(1 kC)]P dS/H, (10)

+

and the linear rate becomes d t / d S = k’[kC/(l

+ kC)]P/H,

P1 = Poco/c,

Pz = Poco/cz

(13)

The total load on the specimen, L,, is divided between the two phases

L, = PoAo = 7-431 + (1 - Y)A$z

(14)

and

+ (1 - r)Pz

=

Y(CO/Cl)PO

+ (1 - Y)(CO/CZ)PO (15)

(5)

Assuming that the applied load is equally distributed to all abrasive particles entrapped in the interaction zone, the mean normal load per particle, L, that contacts the surface is

L = gd2P

where c is a proportionality constant and 0, 1,and 2 represent the anodic oxide film, the aluminum oxide phase, and the air phase, respectively. The following relationship then exists:

Po = YP,

where f{ is a constant. From eq 2-4 one can obtain dui = fji’Li dS/H,

takes into account the concentration of the abrasive particles. The anodic oxide coating can also be considered as a composite material containing air cylinders dispersed in an alumina matrix. Krushchov and Babichev (1958) have shown that for a composite material the applied pressure is distributed between phases. Since the wear rate is proportional to the applied pressure, the rate of wear of a composite material is given by d t / d S = ~ $ 0 = C l P l = cZPZ (12)

(11)

where k’ = fg/g‘ and this constant is determined by the nature of the abrasive, i.e., its shape and hardness. Equation 11 is similar to the generalized wear equation, dV/dS = (k/3)L/H, and predicts that the wear rate is directly proportional to the contact pressure and is inversely proportional to the surface hardness, but it also

and co = l/[Y/Cl + (1 - Y ) / C Z l

(16)

where y is the volume fraction of oxide in the anodic coating and (1 - y) is the porous fraction of the oxide coating. Since the wear rate of the air phase may be considered to be infinite, c2 is also infinite. The wear rate of the anodic oxide is thus simplified to dt/dS = ~ $ 0 = (ci/y)Po

(17)

The volume fraction of oxide in the anodic coating can be obtained by Z = d,yV = d,V (18) and

Y = d,/dr

(19)

where Z is the weight of the anodic coating; V is the apparent volume of the anodic coating, which is equal to the product of coating thickness and specimen size; d, is the real density of the aluminum oxide; and d, is the apparent density of the aluminum oxide. The combination of eq 11, 17, and 19 gives dt/dS = k’(d,/H,)[kC/(l + kC)(P/d,)] (20) and

Hr = (dr/da)Hs

(21)

where H,is the real hardness of aluminum oxide. If the real density and hardness of the aluminum oxide formed on a particular aluminum alloy are material properties determined primarily by the nature of the anodizing electrolyte, then eq 20 predicts that the wear rate of anodic films produced under various conditions should be inversely proportional to the apparent density when the wear tests are carried out under identical conditions. The satisfactory agreement between the theory and experimental results suggests that the real density and hardness of the anodic films are indeed material properties. They are determined primarily by the nature of the electrolyte employed in producing oxide films, while the anodizing conditions such as anodizing voltage, electrolyte temperature, and anodizing time have considerable influence on

Ind. Eng. Chem. Prod. Res. Dev. 1986, 2 5 , 478-480

478

the apparent density and hardness of the coating.

Acknowledgment We are grateful to M. C. Hughes and W. Bilder for advice and assistance during the course of this research. Registry No. AA 1100, 11146-12-6. Literature Cited Buttery, T. C.; Archard, J. F. Proc.-Inst. Mech. Eng. 1970/1971, 185. 537-5 1. Deal, 6. E. Plating 1959, 46, 823; known through Wernick, S.;Pinner, R. The Surface Treatment and Finishlng of Aluminum and Its Alloys ; Robert Draper: Teddington, England, 1972; Vol. 2, p 775.

Hughes, M. C . ; Leidheiser, H., Jr.; Chou, S.-M.; Bilder, W. Ind. Eng. Chem. Prod. Res. Dev. 1983, 22, 132-38. Keller, F.; Hunter, M. S. Robinson, D. L. J . Electrochem. SOC. 1953, 100, 411-19. Krushchov, M. M.; Babichev, M. A. Frict. Wear Mach. (USSR)1958, 12, 5-23. Nagayama, M.; Tamura. K.; Takahashi, H. Corros. Sci. 1970, 10, 617-27. Noble, J. W., 111; Leidheiser, H., Jr. Ind. Eng. Chem. Prod. Res. Dev. 1981, 2 0 , 344-50. Sautter, W. Aluminum 1968, 42, 636-42. Valdma, L. E. Frict. Wear Mach. (USSR) 1959, 13, 17-30. Wood, G. C.; O'Sullivan, J. P. Nectrochim. Acta 1970, 15, 1865-76.

Received for review February 20, 1986 Accepted March 12, 1986

Flash Pulse Method for the Determination of Thermal Conductivities of Ceramic Abrasivest Laura L. Smith Sterimatics Company, Bedford, Massachusetts 0 1730

John W. Walkinshaw" Chemical Engineering Department, University of Lowell, Lo well, Massachusetts 0 1854

This work was conducted to determine the applicability of the flash pulse method in determining thermal diffusivities of ceramic abrasives. Experimentation was performed at 25 OC for both thermal diffusivity and heat capacity; the value of the heat capacity was verified by use of a differential scanning calorimeter (DSC). After irradiation of the front surface of prepared samples and recording of back surface temperature histories, heat-transfer correlations were used to determine the thermal diffusivities. The thermal conductivity was found by use of the heat capacity and density, which were calculated by a volume percent method. The results obtained showed that MgO, which gave a thermal conductivity of 35.41 W/(m.C), was within 3.49% of the published data. Additional ceramics which have not been previously evaluated, due to their single crystal size, were evaluated in the same manner as the MgO standard. Among these materials were (by Norton Co. designation) NZ, E-344, E-314, and dark alundum.

Introduction Abrasives have a large application in grinding wheels. They consist of abrasive grains bonded with a suitable agent into a three-dimensional network having a controlled porous structure. Most grinding wheels are used at high operating speeds, thereby raising wheel operating stress levels. Consequently, much attention has been given to the properties and characteristics that govern the mechanical strength and safety of the product along with the thermal properties of the grit of the wheel which becomes important as heat passes through the grit to the adhesive binder, effecting the mechanical integrity of the wheel. The abrasive materials most widely used in the manufacture of grinding wheels are fused aluminum oxide (A1,0,) and silicon carbide (Sic). Both materials are produced in tonnage size ingots (Haywood, 1984) and are then crushed and graded into sized abrasive grains ranging from 6 down to 1200 mesh. The choice of a particular type of variety and particle size of abrasive depends on the nature of the grinding application and workpiece material. 'This work was performed as part of a Master of Science thesis a t the University of Lowell, Chemical Engineering Department, Lowell, MA 01854.

The flash pulse method was chosen to determine the thermal conductivities of the abrasive samples due to its simple nature and its ability to be utilized on very small samples. The method was developed by Parker, Jenkins, Butler, and Abbott at the Naval Radiological Defense Laboratory in San Francisco, CA, in 1961. Their studies included work on Armco iron, molybdenum and titanium. The flash pulse method consists of the absorption of a very short pulse of radiant energy, from a source such as a xenon flash tube, on the front surface of a specimen and recording the rear surface temperature history by means of a thermocouple assembly attached to the rear of the specimen. The time it takes for the rear surface to reach one-half of its maximum temperature is then used in calculations to determine the thermal diffusivity. Once the densities and heat capacities are known, the thermal conductivity of the specimen can be found. Parker et al. based their development work on the 1959 work of Carslaw and Jaeger, who carried out mathematical modeling on various situations involving unsteady-state heat transfer with various boundary conditions. The particular case considered was a sample that is perfectly insulated with a constant emissivity across its surface. The sample is irradiated with a pulse of thermal energy of very short duration in comparison to the time necessary for the

0196-4321/86/1225-0478$01.50/00 1986 American Chemical Society