Fig. 1 Schematic of extrusion through a square die in a ram extruder (P is total extrusion pressure, Pe is die-entry pressure and P1 is die-land pressure of paste).

Rod Extrusion of Cordierite-Based Paste Containing Aluminum-Rich Anodizing Sludge The plasticities of various aluminum-rich anodizing sludge batch formulations that contained a plasticizer and a lubricant were tested using stress–deformation curves, and the effects of ram speed and pressure were evaluated using the Benbow–Bridgwater model. M.J. Ribeiro and J.A. Labrincha

Cordierite (2MgO·2Al2O3·5SiO2) ceramic materials show interesting thermomechanical properties, including high thermal shock resistance, because of their low thermal expansion coefficient.1,2 Accordingly, cordierite-based materials have found favor as honeycomb supports for catalytic converters in automobiles, furniture for self-cleaning ovens and industrial heat exchangers for gas turbines. Many processing routes, including slip casting and dry processing,2,3 result in devices based on these materials. Extrusion long has been used to shape ceramic objects, mostly for traditional applications, such as brick, tile and pipe.4 Extrusion also is commonly used in other industrial sectors, including food, agriculture, chemistry and pharmaceuticals.5 Benbow and Bridgwater5,6 have demonstrated that the extrusion of particulate pastes that are comprised of fine particles suspended in a liquid continuous phase through dies with circular cross section and having a square entry (see Fig. 1) can be described by P = Pe + P1 = 2(σ0 + αVn) ln(D0/D) + (τ0 + βVm)4(L/D)

(1)

where α is a velocity-dependent factor for the convergent flow, β the velocity-dependent factor for parallel flow, n and m exponents, σ0 the paste bulk yield value, τ0 the paste characteristic initial wall shear stress, D0 and D the diameters of the barrel and die, respectively, L the die-land length and V the extrudate velocity. In this equation, die-entry (Pe) and die-land (P1) pressures are separated. American Ceramic Society Bulletin, Vol. 86, No. 9

9101

Rod Extrusion A coefficient of static friction for the extrudate (µ) can be calculated by µ = τ0/σ0

(2)

and can be considered an important parameter for controlled extrusion.7 In this work, cordierite-based rods were extruded using a ceramic paste that contained aluminum sludge (waste from an aluminum anodizing process), diatomite and talc. To adjust the plasticity level to allow defect-free extrusion, various amounts of plasticizing and lubricating agents were added. Plastic behavior was characterized by stress–deformation curves and compared with those of standard industrially prepared pastes.8 A sludge-based formulation was used to test the applicability of the Benbow– Fig. 2 Detailed representation of the die used for ram extrusion of rods production Bridgwater model in the extrusion of rods. with the indication of partial pressure drops in the die-entry region and in die (Pe

Materials and Methods

is die-entry pressure, D0 = 34.14 mm, D1 = 25.88 mm, D2 = 35.00 mm, D3 = 7.75 mm, L2 = 13.50 mm, L3 = 8.60 mm, θ = 6° and Φ = 44°).

The cordierite paste (CP) was prepared from a premixed powder that contained 25 wt% precalcined (at 1400°C) aluminum anodizing sludge (Extrusal SA, Aveiro, Portugal), 43 wt% talc (Luzenac, France) and 32 wt% diatomite (Anglo-Portuguese Society of Diatomite, Óbidos, Portugal). An alternative paste (CP-S) was prepared using sand (Mibal-B, Barqueiros, Portugal) as a diatomite replacement. Details of preparation and characterization of the sludge have been given elsewhere.8,9 To adjust the plasticity level, commercial additives were incorporated in the test formulations: a plasticizer (Zusoplast PS1, Zschimmer and Schwarz, Germany) and a lubricant (Zusoplast O59, Zschimmer and Schwarz) (Table 1).

Table 1 Tested CP Batch Formulations Specimen

Zusoplast PS1

Zusoplast O59

Moisture

(wt%)

(wt%)

content (%)

CP/without additive 45.8H

45.8

CP/6P4L45.8H

6

4

45.8

CP/6P4L34.9H

6

4

34.9

CP/8P4L45.0H

8

4

45.0

CP-S/6P4L34.5H

6

2

34.5

The yield value and plasticity level the pastes were obtained from stress–deformation tests conducted using plastic compression (Model LR 30K, Lloyd Instruments) in special metal molds. A screw extruder was used to preextrud the formulations (Table 1) through a cylindrical die (diameter of ~33.0 mm) to improve mixing and homogeneity. These rods then were cut into test billets (diameter of ~33 mm and length of ~43.0 mm). A minimum of three specimens per composition were tested. Compressive tests were conducted at a constant loading rate of 2.0 mm/min until a maximum deformation of ~70% or the ultimate limit of the load cell (500 N) was reached.8 Extrusion tests were performed in a ram extruder using a rod die with various ram velocities: 1, 2, 5, 10, 20, 30, 60, 100 and 200 mm/min. Contributing pressure drops generated during the extrusion were recorded (Fig. 2). The apparatus used has been described in a previous work.4 Values of total pressure (P) applied through the ram and pressure at the die-entry (Pe) were measured using digital sensors.

Stress/Deformation Behavior Plastic deformations of CP and CP-S, with or without lubrication and plasticizer additions, were obtained using plastic compression (Fig. 3). Paste without additives exhibited low levels of plasticity, as defined in a previous work.8

9102

American Ceramic Society Bulletin, Vol. 86, No. 9

Rod Extrusion The plastic deformation region was narrow and yield stress values were high. Paste formulations that exhibited this curve form were predicted to fail in extrusion. This proved to be the case with several failed attempts to obtain simple rods, even when relatively high pressures were applied. An industrial earthenware paste also was tested. The determined plasticity level, assumed as normal to use in a common ceramic process (e.g., roller, plastic press and extrusion), was compared with the CP plastic behavior. This comparison showed that CP/6P4L45.8H paste had plasticity most like that of industrial paste. Therefore, it was selected for the extrusion studies (despite its higher level of moisture, which obliged a careful drying process).

Stress (MPa)

The CP-S tested only with plasticizer and lubricant addition (CP-S/6P4L34.5H) was more plastic when compared with the CP with the same moisture level. For the same plastic behavior, the CP-S needed ~10% less water content. This result confirmed that the water sequestration effect, induced by diatomite particles, was well documented4 and might explain the difference between the two pastes. A previous work8 showed water to be a major contributor to the level of plastic behavior. Where lubricant and/or plasticizer were added, the level of water required was reversed, such that the CP formulations required more water and lubricant when compared with CPS. This effect was attributed to the presence of diatomite. Because of the higher moisture content of the CP formulation (with additives), careful drying and firing operations were predictable. Yield stress values (Fig. 3) then were used as an input parameter for extrusion modeling.

Extrusion Characterization Extrusion was characterized by Deformation application of the Benbow– Bridgwater equations used for modelFig. 3 Stress–deformation test curves of CP containing aluminum sludge. ing the flow of pastes through dies with 4–6 complex geometry. The total pressure drop for the current die design comprised several definable contributions, p0 to p3 (Fig. 2): P = Pe + P1 = p0 + p1 + p2 + p3 = [2(σ0 + αVn + τ0 cotg θ) ln(D0/D1) + βVm cotg θ)] + [(σ0 + αVn) ln(A1/A2)] + [2(σ0 + αVn + τ0 cotg Φ) ln(D1/D3) + βVm cotg Φ] + [4(τ0 + βVm)(L3/D3)] where θ and Φ are die parameter angles (where θ is the angle of die-entry region) (Fig. 2), Ax the areas at locations x, Dx the diameters at locations x and Lx the die-land lengths at locations x. The plastic behavior of the CPs is similar to those of industrial pastes.8 Therefore, the extrudability of the CP was studied only with CP/6P4L45.8H and the five Benbow parameters (α, n, τ0, β and m) for this paste, which were obtained in a previous work4 and reported here (Table 2). The yield value (σ0) was assumed to equate to the yield stress determined from stress–deformation curves (Fig. 3). The measured values of the total pressure loss as a function of extrusion velocity were compared with fitting curves obtained

American Ceramic Society Bulletin, Vol. 86, No. 9

(3)

Table 2 Benbow Extrusion Parameters α (MPa(s∙m–1)n] Parameter

CP/6P4L45.8H 0.462

n

β [MPa(s∙m–1)m]

0.284

m

τ 0 (MPa)

0.433

σ0 (MPa)

0.00909

µ

0.152

0.126

0.06

9103

Rod Extrusion

Pressure (MPa)

using Eq. (3) for the CP/6P4L45.8H paste (Table 3 and Fig. 4). To investigate the differences (in percent) obtained between predicted and experimental work, values were estimated from the quotient (predicted – experimental)/experimental. Predicted values according to the model seemed to be correct, and differences to the measured values were