The influence of methods of injection moulding and isostatic pressing on structural ceramics performance

Influência de métodos de moldagem por injeção e prensagem isostática no desempenho de cerâmicas estruturais

The influence of methods of injection moulding and isostatic pressing on structural ceramics performance

(Influência de métodos de moldagem por injeção e prensagem isostática no desempenho de cerâmicas estruturais)

Carlos Alberto Fortulan, Benedito de Moraes Purquerio

LAMAFE-EESC-USP - Av. Dr. Carlos Botelho, 1465, CP 359, S. Carlos, SP, Brazil

e-mail: [e-mail]

Abstract

Advanced ceramics for structural applications are present in engineering parts subjected to complex solicitations. Such unique combination of properties makes these ceramics the unique solution for material selection in special cases. However a material selection cannot be made independently of the choice of manufacturing processes, shape and cost. The determination of a methodology for the evaluation of the sintered ceramic parts reliability is as important as the selection and processing techniques. This work presents the recent trends in the structural advanced ceramics application, the cold forming isostatic pressing and injection processes molding, the elaboration of a methodology for the performance evaluation and choice of forming processes related with the mechanical properties, microstructure and unlubricated sliding wear.

Keywords: advanced ceramics, structural ceramics, forming, ceramic injection, isostatic pressing.

Resumo

Cerâmicas avançadas para aplicações estruturais estão presentes em componentes de engenharia submetidos a solicitações complexas. Tal combinação única de propriedades torna essas cerâmicas solução sem paralelo para seleção de materiais em casos especiais. Uma seleção de materiais não pode contudo ser feita independentemente da escolha de processos de fabricação, forma e custo. A determinação de uma metodologia para a avaliação do grau de confiança dos componentes cerâmicos sinterizados é tão importante quanto as técnicas de seleção e de processamento. Este trabalho apresenta as tendências recentes na aplicação de cerâmicas estruturais avançadas, na conformação a frio por prensagem isostática e moldagem por processos de injeção, a elaboração de uma metologia para avaliação de desempenho e escolha de processos de corformação relacionados com propriedades mecânicas, microestrutura e desgaste por atrito sem lubrificante.

Palavras-chave: cerâmicas avançadas, cerâmicas estruturais, conformação, injeção cerâmica, prensagem isostática.

(Versão do Editor)

INTRODUCTION

Structural ceramics are an attractive option for the solution of countless engineering problems, mainly when applications subjected to complex solicitation demand high performance materials. The combination of the inherent properties that such materials present in these situations has been the reason for the unique selection of these ceramics [1] .

The selection of the forming process is of fundamental importance to the product and it is related with the application, shape, production level and cost of the desired product. Among the several forming processes for structural advanced ceramics, two of them have been well applied: the ceramic injection and the cold isostatic pressing.

The products and components obtained by the isostatic pressing and ceramic injection processes present inherent characteristics to each of these processes; some are positives, other negatives. The application of the isostatic pressing process for the production of ceramic parts generally demands a machining stage to green. The machining to green is the main process responsible for the introduction of product defects and tensions. On the other hand, the injection process allows the production of a product in its final form; however it incorporates typical defects related to the process such as weld lines, emptiness, depressions and others.

Structural advanced ceramics

The structural advanced ceramics are materials that present high performance when subjected to high solicitations and they include components subjected to severe wear, such as pump elements and valves, components for automotive engines, machining tools and dies and biomedical components. Such ceramics are obtained from a careful chemical combination, a controlled microstructure and sophisticated production processes, generating a field of materials with a unique combination of properties [1] .

Injection of ceramic masses

The process of ceramic masses injection or ceramic injection consists basically on the injection of a melted binder loaded with ceramic powders inside a mold cavity. After the binder solidification a part formed to green is obtained. The organic vehicle of the formed part is removed (debinding) and it is afterward densified for sintering [1, 2].

This process presents as main characteristics the possibility for product forming with very complex shapes with a low relative cost and high performance [1, 3, 4].

Isostatic pressing

The isostatic pressing process, traditionally known by the production of ignition automotive sparking parts, stands out due to the uniformity and high resulting quality of the obtained products. Actually the process consists of a flexible elastomeric mould which cavity is filled with ceramic powders and sealed tightly. Placed afterward inside a pressure vase, the mold is pressed isostatically through the action of a pressurized fluid. The fluid, when pressurized, compresses the mould in all directions generating the compact to green.

The cold isostatic pressing is the most used forming method by the ceramic industry nowadays. Commercially pressures ranging between 70 to 200 MPa are usually applied.

METHODOLOGY AND EXPERIMENTS

In the present work an alumina sample formulation containing 99.5% of alumina was selected for the experiments (Table I). Ceramics based on high alumina content represent an appropriate material choice for high performance parts and their characteristics and properties are entitled to represent the general behavior of the structural advanced ceramics. They can successfully represent the general alumina performance in a lot of applications of sliding contacts where optimum mechanical properties are desirable.

Table I
- Formulation of the prime material employed.

Materials	Alumina 99,5%
Alumina S-5G (ALCAN)	99,90
MgO - P.A. (for analysis)	0,25

Alumina milling. The alumina milling is necessary as well as the optimization of its granulometric distribution. The appropriate granulometric distribution establishes the commitment to obtain the largest density, the lower porosity, the smallest growth of the grains (medium size of 3 mm) and the best grain uniformity. The mixture of ceramic powders obtained through both the vibratory and the ball milling was tested and the one that promotes the largest density with the smallest retraction rate was the chosen mixture for this work.

For the ceramic injection processing at low injection pressure (0.7 MPa), an Injectable Ceramic Mixture (ICM) was selected with 55% (in volume) of ceramic powder and 45% (in volume) of binder composition. The powder was dried in a kiln at 120 ºC and was crushed until passing through a #200 mesh strainer and then was mixed with the binder.

The binder composition was based on the low injection pressure as suggested by literature. This composition is presented in Table II. With this multiple component formulation was possible to settle down the characteristics of ICM and the easiness of the debinding from the injected part for instance. Each component presence direct the ICM characteristics and represents also the binder components, that is, for the main component, the LDPE; for the secondary component, the paraffin wax; for the plastificant, the carnauba wax, the bee wax and the pitch; and for the processing helper, the stearic acid.

Table II
- The employed binder-formulation

Materials	Volumetric Proportion (%)
LDPE	33
Paraffin wax	33
Carnauba wax	20
Bee wax	9
Pitch	3
Stearic acid	2

In the case of isostatic pressing, before the pressing, the material received the addition of 3% in weight of PVAL previously dissolved and the material was milled in a ball mill during 24 hours. After that the material was dried in a spray-drier and the granules were classified in a strainer that allowed #80 mesh sizes and retained all #200 mesh.

Experiments. The analysis of the test bodies performance was made in three main groups that were the microstructure characterization (microstructure, density and hardness), the mechanical properties and the wear resistance.

The mechanical properties were obtained through compression tests. The test methodology for structural ceramics was based on previous suggestions [5]. They comment that the Japanese Association of Fine Ceramics has been working to establish standards for the evaluation of the properties of the advanced ceramics in addition to the current JIS standards for compressive test methods for ceramics [(JIS R1608 (1990)]. They also anticipated the proposal for the future shape of the test bodies in addition to this standard, that is, the dumbbell shape. In the dumbbell shape test body when under compression loading the edge effect so common to the flexion tests is avoided. Fig. 1 presents the schematic drawing of the dumbbell test body shape used in the experiments.

Figure 1
- Schematic drawing of the test body for the compression tests [5].

Wear resistance. Wear tests were performed using a pin-on-disc equipment and the sliding distance of the dry contact of a ceramic pin on a ceramic disc was measured. For ceramics based on alumina, wear tests should be accomplished with sliding speed and loading below the transition limit between the soft and the severe wear. In such conditions the product life becomes more reliable [6].

Moulds. The mould project represents a stage of great commitment for the product quality. The concentration of strains and the defects must be minimized in this stage.

The ceramic injection moulds are very similar to those employed in the polymer injection. This aspect allows the polymer injection mould design techniques to be used after some basic adaptations in the design of ceramic injection moulds. The injection moulding allow to obtain the injected test body in its final shape avoiding the machining to green. The injected parts retractions foreseen for the injection process is about 2% for the retraction injection and 18% for the sintering retraction. Fig. 2 shows the cavity dimensions of the injection mould [7].

Figure 2
- Schematic drawing of the injection mould.

Test body machining. The machining of the compressed test bodies and pins for the wear tests which were obtained through isostatic pressing were accomplished from the machining to green of a pressed part (Fig. 3). The test bodies after the sintering were ground in both base faces in order to guarantee the minimum parallelism of 10 mm.

Figure 3
- Machining of a test body for compression.

EXPERIMENTAL TESTS, RESULTS AND DISCUSSIONS

Performance of the granulometric distribution. The powder vibratory milling supplied particles with excellent equivalent medium diameter of 0.5 mm however; about 80% of these particles presented equivalent diameters ranging from 1.0 to 0.2 mm. Through the ball milling, alumina particles were obtained with an equivalent medium diameter about 1.5 mm and 80% of which were distributed between 3 and 0.5 mm.

The distribution composed by the mixture of 80% of the powders obtained by vibratory milling (VM) with 20% of the obtained by ball milling (BM) presented the best relationship for sintering kinetics with resulting packing in a larger apparent density with the smallest retraction rate.

Performance of the forming processes. The isostatic pressing test bodies followed four stages to be obtained: the pressing stage, the machining to green, the sintering and of the abrasive machining of the test body base faces. To obtain a lot of 10 pieces was spent a total time of 4.25 hours for each test body. So it was necessary 1.5 work/man hours for each test body. These numbers can be significantly reduced for larger lots and better process optimization. But if the product shape is complex these numbers can be raised considerably.

In the other hand, the injection process is relatively more complicated and more dependent on external variables than the isostatic pressing process. The time spent for the production of 10 test bodies was about 19.5 hours for each test body and consequently 0.3 of work/man hour were spent for each test body. It is observed that this later process is much less dependent on the man direct performance. However, it demanded a larger equipment readiness, mainly during the debinding stage, which should be optimized and time can therefore be reduced. These characteristics are in favor to the productivity and high production, in spite of the mould complexity and the larger cost that can be absorbed with a high production.

The employment of the relatively fine ceramic powder with an equivalent medium diameter of 0.6 mm and strained in #200 mesh, presented during the mixture and heating phases the gathering in small and rigid flakes of close dimensions around 1.5 mm in diameter. This gathering is certainly due to the increase of the surface energy, the Van der Waals forces and the capillary attraction forces. The planetary system of racket beaters used as the mixer of the injection machine was unable to break the adhesion force of these small agglomerates which, after the debinding, integrated themselves discretely into the ceramic mixture but became the test bodies crumbly after forming.

The debinding using the diffusion process followed by thermal degradation presented the largest difficulties and the largest damages to the processed test bodies. The damage was due to the small size of the ceramic particles that avoid enough diffusion at low pressure of the binder components in the liquid state. It was observed during the debinding that the external layer of the test body surface became impermeable with an thickness about 0.8 mm. Due to the superficial and capillary attraction forces in this external layer, after the initial diffusion of some melted components, there was an accentuated approach of the particles followed by retraction of the external layer which compressed the internal layer and blocked the exit of the remaining components inside the test body. The defects generated by this effect are presented in Fig. 4.

Figure 4
- Illustration of the main defects occurred during debinding. a) Due to the retract ion of the external layer, b) Due to evolution of gases inside the test body.

Microstructure Characterization. The test bodies isostaticaly pressed at 100 MPa presented the microstructure illustrated in Fig. 5. In this sample can be observed a great amount of pores, some located inside the grains and several others located in the triple points of the contours of grains and a relative larger amount due to current defects of compression. In this figure can also be observed some porous pockets, where occurred a considerable absence of grain growth. The average grain size remained around 5.5 mm; however, there was some heterogeneity in its morphology. Many angles of the diedros became very sharp and others around 180 degrees. This analysis confirms the occurrence of a lacking of compressive packing to promote the effective sintering of the test part and also the absence of conditions to promote the proper sintering and the grain growth for a better compactness. Nevertheless the medium apparent density was 3.78 g/cm³ and medium Vickers hardness was 8.3 Hv (98 N) GPa.

Figure 5
- Microstructure of the test bodies isostaticaly pressed. a) 5000x and in b) 1000x.

The bodies injected at 0.7 MPa presented the microstructure illustrated by Fig. 6 where a larger amount of pores is found, but with a better distribution than the compressed parts, characterizing a better uniformity of the grain packing. It can be observed in some grains a preferential growth in some directions in a number greater than the observed in the isostatic pressing. Also, many angles of the diedros are very sharp and many around 180 degrees. This analysis illustrates the lack of packing and the poor sintering. If a better sintering conditions is applied (temperature, landing) an increase the medium size of the grains is possibly obtained, but this effect should be undesirable. It can be in this figure a medium size of grains around 10 mm and in some grains the preferential growth in some directions in a great number like the observed in the isostatic pressing parts. The macro Vickers hardness obtained in this case was of 8.3 Hv (98 N) GPa.

Figure 6
- Microstructure of the injected alumina . a) 5.000x and in b) 1000x.

It was observed that the retraction values for the injected test bodies were higher than those for the pressed were. In fact, this retraction is near 30% larger than those pressed at 100 MPa, but the densities remained lower (3.33 g/cm³). The amplification of the retraction commits the precision of a structural component design and increases the distortion probability and current dimensional gradients of the sintering. The smallest apparent density evidences the existence of emptiness and internal defects.

Analysis of the Compression Resistance. Bodies pressed at 100 MPa and injected at 0.7 MPa where tested at a loading rate of 0.5mm/min up to the occurrence of the rupture.

Nine pressed test bodies were tested and the Weibull distribution was obtained as well the determination of the Weibull module. The fracture medium tension of 3.17 GPa and the value m=13 were obtained, which satisfy the expectations for structural ceramics based on high alumina.

The injected test bodies presented very low values for the rupture tensions with an average of 0.05 GPa. The analysis of fracture of these bodies indicates that occurred a rupture of the external layer initially and soon after there was a tension fall and subsequent loading of the internal layer that fractured initially with a smaller tension than the initial. This very low value is due to the defects related with the debinding. The formation of an external layer seemingly homogeneous did hide a great amount of internal defects. These defects mainly the great internal holes and the long cracks promoted the tension concentration in certain areas of the test body resulting in previous fracture.

These values cannot represent a class of injected bodies. The evaluation of the resistance to the compression of these bodies will only be possible after the optimization of the binder composition and of its removal.

Analysis of the Wear Resistance. The experimental conditions for the wear tests followed the literature recommendations as follows: load of 20 N; disk texture of 0.60 mm Ra; pin texture of 0.80 mm Ra; ambient temperature of 23 to 25 ºC; sliding speed of 0.314 m/s; U_rel 53 to 55%. The sample preparation consisted in the washing up with acetone, drying at 70 ºC during 30 minutes, cooling in a dryer during 30 minutes before the weighting and permanent sample maintenance in a dryer when not under test operation. The distance traveled in each test run was of 5000 m. For these test conditions the regime of soft wear was adopted.

The wear behavior of the pressed test bodies is illustrated in Fig. 7 where an intense wear is observed in the first 1000 m of sliding and a wear rate accumulated for the first 3000 m is around 28x10^-6 mm³/Nm. The wear rate occurred during the analyzed stages showed smaller numbers in the stage from 2000 to 3000 m with an average of 7.7x10^-6 mm³/Nm as observed. These numbers belong to the order of 10^-6 mm³/Nm, which can be considered as inside a transition regime between the soft and severe wear [8].

Figure 7
- Wear rate for the pressed pin.

The worn surface after 3000 m of sliding is presented by Fig. 8. It can be observed that in this stage the pin surface worn away and that the debris presenting spherical forms smaller than 1mm went repacked in layers in the form of scales with dimensions varying from 10 to 100 mm.

Figure 8
- Worm surface of the pressed test body n^o 2 after 3000m of sliding em a) 100x, em b) 1000x e em c) 5000x.

The Injected Pins Wear Resistance. The injected test pins presented the wear behavior as illustrated by Fig. 9 where can also be observed an intense wear in the first 500 m of sliding and a wear rate accumulated for the first 2000 m about 26x10^-6 mm³/Nm. In the stage from 2000 to 3000 m an average of 7.7x10^-6 mm³/Nm is observed. These wear rate values can be considered inside of the transition regime between the soft and severe wear and they are lightly inferior to the observed for the pressed pins.

Figure 9
- Wear rate of injected pins.

After 3000 m of sliding the worn surface of the pin nº 5 was observed and is illustrated in Fig. 10. In this figure a surface with a larger damaged relief and with several superficial cracks can be observed. The formation of scales which were generated from the debris packing is also observed, but with less uniform sizes and consequently with less contact area than observed in the surface of the pressed pins. This fact should inhibit the waste lightly.

Figure 10
- Worm surface of pin n^o 5 after 3000m of sliding. a) 100x, b) 1000x, c) 5000x.

The behavior of the test bodies pressed at 100 MPa showed a small superiority related do the test bodies injected at 0.7 MPa concerning the wear rate as observed.

CONCLUSIONS

The selection of the forming method is related to the shape, cost, dimensional precision, material, flexibility and level of production of a specific product. The injection process aroused in the second half of last decade and is since them backed up by the great evolution obtained with the injection of polymers. The isostatic pressing process on the other hand has been consecrated for many years and has its innovations related with the automation levels and in the simplification related to the press design.

Through the microstructure analysis of the injected pressed test bodies, a similarity is observed concerning the injected at 0.7 MPa and the pressed at 100 MPa. The main similar characteristics are related to the amount of pores, morphology of the grain and surface hardness. The apparent density is a main dissonant characteristic that should be interpreted distinctly. Taking into account that the porosity observed in the microstructure is much smaller than the estimated for the measure of the density, it can be admitted that this disparity is related to the great current internal emptiness due to the OV (organic vehicle). However, if this injection stage were optimized, the injected bodies would present an increment in the apparent density that would establish the levels observed for pressed at 100 MPa. The approach of the observations verified with the test bodies pressed isostaticaly at 100 MPa allows to consider that with an optimization of the OV removal technique for the injection process the bodies injected at 0.7 MPa would present compatible performance in relation to the bodies pressed at 100 MPa concerning their commercial applications as structural ceramics.

The measurements of the mechanical properties obtained by the compression tests certified the tendency in establishing the dumbbell shape for the test bodies and it was not verified in the experiments great disparity between the largest and the smallest values obtained. The bodies pressed isostaticaly at 100 MPa presented acceptable values of the rupture tension and a reasonable Weibull module (m=13). So, it could be concluded through the experimental results that products obtained through the isostatic pressing present acceptable spreading values and they can be used in structural applications with a good reliability level. The values obtained for the injected test bodies as they are do not represent the class of the injected ceramic itself. For achieve better performance, it is necessary to optimize the ceramic formulation and the debinding.

ACKNOWLEDGEMENTS

This work was supported by FAPESP.

(Rec. 04/98, Ac. 06/98)

• [1] M. R. Pascucci, R. N. Kats, Modern day applications of advanced ceramics, Interceram 42, 2 (1993).
• [2] C. A. Fortulan, B. de M. Purquerio, O processo de injeçăo de cerâmicas, In: Proc. XXXVII Brazilian Congress on Ceramics, Curitiba, PR, Brazil (1993) 77-80.
• [3] M. J. Edirisinghe, J. R. G. Evans, Review: fabrication of engineering ceramics by injection molding, I: Materials Selection, Int. J. High Techn. Ceram., 2 (1986a) 1-31.
• [4] R. M. German, K. F. Hens, S. T. P. Lin, Key issues in powder injection molding, Am. Ceram. Soc. Bull.70, (1991) 1294-1302.
• [5] H. Awaji, Y. Nagno, Compressive testing of ceramics, J. Ceram. Soc. Jap., Int. Edition, 101 (1993) 920-924.
• [6] P. Anderson, A. Blomberg, Alumina in unlubricated sliding point, line and plane contacts, Wear 170 (1993) 191-198.
• [7] P. A. S. Wrege, B. de M. Purquerio, C. A. Fortulan, Projeto de moldes para massas cerâmicas. In Proc. XXXVIII Brazilian Congress on Ceramics, Blumenau, SC, Brazil, (1994) 1191-1196.
• [8] J. Mangels, The low-pressure injection molding, Am. Ceram. Soc. Bull. 73, (1994) 37-41.

Publication Dates

• Publication in this collection
  12 June 2000
• Date of issue
  Oct 1998

History

• Received
  Apr 1998
• Accepted
  June 1998