The Role of Powder Particle Size Distribution in the ...

This article was downloaded by: [Berenika Hausnerova] On: 20 July 2011, At: 00:11 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

International Journal of Polymer Analysis and Characterization Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/gpac20

The Role of Powder Particle Size Distribution in the Processability of Powder Injection Molding Compounds a

a

a

a

B. Hausnerova , T. Kitano , I. Kuritka , J. Prindis & L. Marcanikova

a

a

Polymer Centre, Faculty of Technology, Tomas Bata University in Zlín, Zlín, Czech Republic Available online: 19 Feb 2011

To cite this article: B. Hausnerova, T. Kitano, I. Kuritka, J. Prindis & L. Marcanikova (2011): The Role of Powder Particle Size Distribution in the Processability of Powder Injection Molding Compounds, International Journal of Polymer Analysis and Characterization, 16:2, 141-151 To link to this article: http://dx.doi.org/10.1080/1023666X.2011.547047

PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan, sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

International Journal of Polymer Anal. Charact., 16: 141–151, 2011 Copyright # Taylor & Francis Group, LLC ISSN: 1023-666X print=1563-5341 online DOI: 10.1080/1023666X.2011.547047

THE ROLE OF POWDER PARTICLE SIZE DISTRIBUTION IN THE PROCESSABILITY OF POWDER INJECTION MOLDING COMPOUNDS B. Hausnerova, T. Kitano, I. Kuritka, J. Prindis, and L. Marcanikova Downloaded by [Berenika Hausnerova] at 00:11 20 July 2011

Polymer Centre, Faculty of Technology, Tomas Bata University in Zlıń, Zlıń, Czech Republic The role of particle size distribution of hard-metal carbide powder compounds in structural changes caused by shear deformation is investigated via their response to dynamic viscoelastic strain. Materials employed in the study are intended for the production of sintered carbide components via powder injection molding. Four grades of hard-metal carbide powders differing in their particle size distribution were mixed thoroughly with a polymer binder containing low-density polyethylene, ethylene/vinyl acetate copolymer, and paraffin wax. From the results, it can be concluded that powder particle size distribution dominantly influences not only the magnitude of viscoelastic functions of highly filled compounds, but also their cause. As the demands on powder characteristics arising from the particular steps of powder injection molding are contradictory, thermogravimetric analysis was used as a complementary method for the investigation of the role of particle size distribution during debinding and sintering. Keywords: Carbide powder; Debinding; Filled polymers; Oscillatory flow; Particle size distribution; Polymeric binder; Thermogravimetric analysis

INTRODUCTION Powder injection molding (PIM) technology currently garners substantial interest in both fundamental and applied research as the number of successful applications (small-size, complex-shape parts) in terms of substituting traditional processing routes increases. During the PIM process, a powder (metallic or ceramic) is homogenously dispersed within a polymeric binder of a tailored composition. Such a compound is then processed in a conventional injection molding machine into a so-called green part. Submitted 4 November 2010; revised 23 November 2010; accepted 30 November 2010. The authors would like to acknowledge the Grant Agency of the Czech Republic (Project No. 103=08=1307) and The Ministry of Education, Youth and Sports of the Czech Republic for financial support through Project No. MSM 7088352101. The rheological results have been presented in part at the 3rd WSEAS International Conference on Engineering Mechanics, Structures, Engineering Geology, held in Corfu (2010). Correspondence: B. Hausnerova, Polymer Centre, Faculty of Technology, Tomas Bata University in Zlıń, TGM 5555, 760 01 Zlıń, Czech Republic. E-mail: [email protected] 141

Downloaded by [Berenika Hausnerova] at 00:11 20 July 2011

142

B. HAUSNEROVA ET AL.

This stage is followed by debinding, where the binder is extracted, and the resulting ‘‘brown part’’ is sintered to the final density. The combination of two different processing approaches—injection molding and sintering—necessarily involves contradictory requirements. Among the characteristic powder features, particle size distribution plays a dominant role, as has been demonstrated in our previous articles[1,2] investigating the capillary flow (high shear rates) of PIM compounds with regard to powder packing limits[1] and temperature as well as pressure sensitivities.[2] Powders suitable from the molding point of view should have a broad distribution of sizes, which provides a high loading level at a simultaneously suitable viscosity,[3,4] with a small amount of finely (below 0.5 mm) sized particles that prevent the agglomeration of the highly filled compounds. For the sintering stage, on the other hand, a wide distribution of particle sizes introduces microstructural inhomogeneity into the final sintered parts, but fine powders accelerate this step and improve the surface finish of the products. Therefore, a typical PIM powder with particle sizes varying between 0.1 and 20 mm[5] represents a compromise for various demands. To our best knowledge, detailed investigation of the role of particle size distribution on PIM compounds’ response to deformation has not been reported heretofore. Therefore, the aim of this work is to study particle size distribution influence on dynamic properties since this type of deformation mode has been successfully employed to provide information about the degree of dispersion of a filler in a polymer binder and the presence of yield indicating agglomerations in the system[6] as well as to demonstrate the effect of modification treatment.[7] Simultaneously, the compounds are subjected to thermogravimetric analysis in order to assess the suitability of the tested materials for the debinding and sintering steps of the PIM process, as these are often in contradiction to molding step. EXPERIMENTAL SECTION Viscoelastic properties of PIM compounds were measured with a plate-plate rotational rheometer (ARES, Rheometrics, USA) equipped with an RSI Orchestrator software package. The plate diameter was 25 mm, the gap between plates was set at 1 mm, and angular frequency ranged from 0.1 to 100 rad=s at 140 , 150 , and 160 C. The powders employed in the experiments were composites of tungsten carbide and cobalt (cemented carbides) supplied by Sylvania Tungsten (Czech Republic). They were compounded in a laboratory kneader (Brabender Plasticorder PL-2000-6) with a three-component binder (polyethylene, ethylene=vinyl acetate copolymer, and paraffin wax) in order to prepare 30, 40, 45, 50, and 55 vol.% feedstocks. Four grades were tested—BC10U, BC17S, BC37S, and BC75H—differing in their particle size distributions, as shown in Table I. Particle size distributions and the shape of the powders are depicted in Figures 1 and 2. Scanning electron microscopy (SEM) micrographs (Figure 2) were obtained with the scanning electron microscope Vega II (TESCAN, Czech Republic) operated at 10 kV; all samples were coated with a thin layer of gold using a polaron sputtering apparatus.

PSD OF POWDER INJECTION MOLDING COMPOUNDS

143

Table I. Particle size distribution of powders tested Powder grade

Particle distribution

BC10U:

Diameter at 10% Diameter at 50% Diameter at 90% Average diameter Diameter at 10% Diameter at 50% Diameter at 90% Average diameter Diameter at 10% Diameter at 50% Diameter at 90% Average diameter Diameter at 10% Diameter at 50% Diameter at 90% Average diameter

BC17S:


BC37S:

BC75H:

Size 0.45 mm 1.11 mm 3.75 mm 1.24 mm 0.43 mm 1.36 mm 2.71 mm 1.24 mm 1.04 mm 3.80 mm 7.14 mm 3.32 mm 2.19 mm 15.38 mm 28.06 mm 7.38 mm

The thermogravimetric (TG) analyses were carried out with a Setaram SETSYS Evolution 1200 thermogravimeter. The samples were examined under an inert atmosphere of He (5.5 purity), which should avoid Co evaporation at high temperatures; the constant gas flow rate was 30 sccm (i.e., 30 cm3 min 1 at pressure of 101.325 kPa for all experiments). Non-isothermal analyses were started from the ambient temperature and were continued until reaching 1200 C; the temperature growth rate was 20 C min 1. The samples were discs with a diameter of 2 mm and a thickness of 1 mm, thus the weight of samples varied approximately from 17 to 30 mg according to the density of each tested material.

Figure 1. Particle size distributions of carbide powders used.


144


Figure 2. SEM images depicting the shape and size differences among the powders: (a) BC10U, (b) BC17S, (c) BC37S, and (d) BC75H.

RESULTS AND DISCUSSION The particular powder grades were selected in such a way that BC10U and BC17S have the same average diameters along with similar particle size distributions in the range of small particles with a high tendency to form agglomerates (up to 0.5 mm). In contrast to BC17S, the powder named BC10U has bimodal distribution, with the second peak corresponding to particle sizes around 5 mm (Figure 1). As clearly seen on SEM micrographs (Figure 2), the powder grade BC37S is coarser with an average diameter more than two times higher than those of the BC10U and BC17S samples. Finally, a high portion of very large particles (28.06 mm diameter at 90%) is present in the BC75H powder. First, as filled systems are generally highly strain dependent even at low strain amplitudes, strain sweep tests at strains ranging from 0.1 to 100% at three selected angular frequencies of 0.37, 3.7, and 37 rad=s were performed. A strain of 1% was selected as a representative linear viscoelastic region for the following rheological testing. The viscoelastic properties obtained for the compounds based on the different powders correspond well with the powder characteristics. Complex viscosities as a



145

Figure 3. Complex viscosity vs. angular frequency of (a) 30 and (b) 50 vol.% carbide powder compounds at a temperature of 150 C.

function of the angular frequency data, depicted in Figure 3, obtained for BC10U are similar to those evaluated for BC17S feedstocks at all tested temperatures. In contrast, the lowest viscoelastic response to deformation shows the BC75H compound having complex viscosities about two and three orders of magnitude lower than those of BC10U (BC17S) feedstocks at 30 and 50 vol.% concentrations, respectively. Such low values of viscoelastic functions are attributed to particle size distribution of this powder represented by large particles, without a portion of small particles prone to agglomerate (the diameter at 10% for BC75H was 2.19 mm, while for BC10U it was 0.45 mm). Accordingly, the complex viscosities of BC37S compounds are found to occur between those of BC10U (BC17S) and BC75H, as its particle size distribution also fits in between. As can be seen from Figure 4, compounds based on BC10U reveal an apparent yield for 50 and 55 vol.% caused by the agglomerated particle network structure formed within a melt (e.g., see Metzner[8]). Since fine particles exhibit larger surface

Figure 4. Complex viscosity vs. angular frequency of (a) BC10U and (b) BC75H carbide powder compounds at a temperature of 140 C.


146


areas and higher interparticle friction, they have an enhanced tendency to form agglomerates. At higher shear rates, however, this structure is broken, and the viscosity is dominated by hydrodynamic interactions.[9] Interestingly, the particle size distribution affects not only the overall level of the viscoelastic properties but also their characteristic course (Figure 5). Generally, for viscoelastic polymer melts, upon lowering angular frequency, the storage modulus decreases to zero, while it becomes independent of the frequency for highly powder-filled melts. Concerning the PIM compounds studied, only the BC10U and BC17S samples retain the behavior typical for composites, exhibiting a plateau on their storage modulus versus angular frequency curves for powder loadings above 30% and a frequency ranging from 0.1 to 1 rad=s. On the other hand, the feedstocks based on BC75H show the modulus increasing with angular frequency for all loading levels, as was found for viscoelastic polymer melts. The influence of temperatures ranging from 140 to 160 C on viscoelastic response was not distinct (see Figure 6) for low (30%) and high (50%) powder concentration independent of the influence of particle size distribution. However, as demonstrated on the BC10U compound (Figure 7), the concentrations in between these two limits—at the lowest temperature considered (140 C)—approach the highest loading level, while they are uniformly distributed as the temperature is increased. Thermogravimetry (TG) was chosen for preliminary evaluation of the feedstock compositions’ suitability for PIM technology with respect to the debinding procedure. First, pure components, i.e., binder and cemented carbide powder, were analyzed. As demonstrated on the differential TG scan (Figure 8), three debinding phases might be discerned according to the typical degradation temperatures of used polymers; the first (275 C, peak A1) corresponding to a loss of acetyl groups (19 wt.%) from the ethylene=vinyl acetate copolymer, the second (360 C, peak A2), the chain scission of the rest of copolymer (6.5 wt.%), and the third (488 C, peak B), vaporization of paraffin and LDPE degradation products (74.4 wt.%). Obtained mass losses are in excellent agreement with the binder composition as well as the degradation process of vinyl acetate copolymers described in the literature.[10]

Figure 5. Storage modulus vs. angular frequency of (a) BC10U and (b) BC75H carbide powder compounds at a temperature of 150 C.



147

Figure 6. Storage modulus vs. angular frequency as a parameter of temperature for (a) BC10U and (b) BC75H carbide powder compounds. Open and filled symbols stand for 30 and 50 vol.% compounds, respectively.

As shown in Figure 9, raw powders were virtually inactive in the temperature region below 1000 C. Negligible water desorption is manifested by an almost undetectable (but reproducible) decrease in weight (0.02–0.07% of initial sample weight with the error estimation in the same order of magnitude) starting at temperatures above 100 C. A typical weight loss step of 0.2–0.3% followed by a linear weight decrease is observed at temperatures above 1000 C. As there is no substance available for vaporization after debinding in this temperature region (i.e., below 1200 C), we conclude that it is only an apparent mass loss due to a buoyancy effect caused by the change of sample volume. The step is manifested as a sharp peak (peak C) on a differential TG (dTG) curve and could be ascribed to a very steep shrinking rate at the beginning of the sintering of the particles, which is then followed by a

Figure 7. Storage modulus vs. angular frequency of BC10U carbide powder compounds at temperatures (a) 140 , (b) 150 , and (c) 160 C.


148


Figure 8. Weight loss (TG) and weight loss rate (dTG) as functions of temperature for the pure polymer binder (LDPE, paraffin, and EVA); peaks A1, A2, and B correspond to a loss of acetyl groups, the chain scission of the copolymer, and the vaporization of paraffin and LDPE degradation products, respectively.

linear decrease of sample volume. Such an explanation is supported in the literature by dilatometric observations of shrinking of similar powder systems.[11,12] The peak C was observed for BC10U powder at 1091 C, BC17S at 1018 C, BC37S at 1137 C, and BC75H at 1135 C. Hence, the start of the sintering process strongly depends on the type of powder. Thermogravimetric analysis of the compounds (as an example, see Figure 10) confirms the complete debinding of the binder system at 1200 C within the error limits of the weight loss estimation, except for the BC17S compound with 50 and 55% powder loading. In these cases, about 10% of the binder mass was retained in the sample after completion of the heating cycle. The A1 peak corresponding to acetyl moiety appeared at temperatures ranging from 250 to 275 C with corresponding weight losses from 1.2 to 3.3%. The second peak, A2, was detected only for BC75H feedstocks at 355.5 C as 0.96% weight loss, while for feedstocks with smaller particle sizes based on BC17S, BC10U, and BC37S, the remaining gasses did not have time to overcome the space hindrances due to smaller and longer pores, and thus the A2 and A1 peaks overlapped each other into one



149

Figure 9. Weight loss (TG) and weight loss rate (dTG) as functions of temperature for a 50% BC10U compound; peak C corresponds to the change in lifting force at the beginning of sintering stage of tungsten carbide particles.

peak, denoted A. A weight loss consistent with vaporization of paraffin and LDPE (peak B) occurred for all samples at temperatures ranging from 487 to 497.6 C. The last peak (C), appearing at temperatures from 891 C (for compounds made of small particles) to 1196 C (for large particles, BC75H), corresponds to the change in lifting force at the beginning of the sintering stage of tungsten carbide particles. TG analysis revealed that temperatures corresponding to the maximum debinding speed of the particular peaks (A and B) are only slightly influenced by the concentration and characteristics of the powder used, which makes the binder an ideal candidate for application in the compounding of the examined powders. The sintering onset, however, depends strongly on the type of powder used. This can be caused by particle size distribution, morphology, amount of carbon residues in the brown part, phase formation, and other factors that are yet not fully understood. Finally, after qualitative inspection of the samples, the binders that were not removed completely, as well as the samples with poor shape retention after debinding and=or the initial stage of sintering, were determined to be unsuitable for PIM.


150


Figure 10. Weight loss (TG) and weight loss rate (dTG) as functions of temperature for a 40% compound based on BC17S; peaks A, B, and C correspond to the copolymer release, the vaporization of paraffin and LDPE, and the beginning of sintering, respectively.

This holds first and foremost for feedstocks containing the BC75H powder, where the samples’ surfaces were accompanied with cracks, blisters, and bubbles as the binder was removed from large interstices between particles. From the comparison of the resulting products it is clear that even if the mean size diameter is the same, the high portion of small particles also excludes BC10U feedstocks, while BC17S feedstocks prove suitable for PIM. Nevertheless, the results, combining both rheological and thermogravimetric analyses, demonstrated that the compounds based on the BC37S powder are the most likely candidates for usage as PIM feedstock, having a smooth surface after binder removal and good shape retention after pre-sintering. CONCLUSION Processability of powder injection molding compounds is considerably influenced by the characteristic of powder used: size and particle size distribution, shape, and tendency to form aggregates. These parameters can be hardly be distinguished from each other, which complicates the reading of the results and the formulation


151

of general rules. Nevertheless, it is clear that particle size distribution is a parameter of key importance. The compounds based on powders having a high portion of small particles tend to have higher viscoelastic functions and the character of a filled material, while the behavior of material containing large particles remains the course of a viscoelastic polymer melt. The contradictory demands on the powders suitable for production of sintered parts by powder injection molding are even more stressed if debinding=sintering is considered. The study demonstrates that, next to rheometry, thermogravimetry might be used as a complementary and fast method for the preliminary assessment of compound suitability for powder injection molding applications.


REFERENCES 1. Hausnerova, B., T. Honek, and P. Saha. 2005. Relative viscosity models and their application to capillary flow data of highly filled hard-metal carbide powder compounds. Polym. Compos. 26: 29–36. 2. Hausnerova, B., T. Sedlacek, P. Filip, and P. Saha. 2011. The effect of powder characteristics on pressure sensitivity of powder injection moulding compounds. Powder Technol. 206 (3): 209–213. 3. Chang, C., and R. L. Powell. 1994. Effect of particle size distribution on the rheology of concentrated bimodal suspensions. J. Rheol. 38: 85–98. 4. Zhang, J. G., M. J. Edirisinghe, and J. R. G. Evans. 1989. A catalogue of ceramic injection moulding defects and their causes. Ind. Ceram. 9: 72–82. 5. German, R. M., and A. Bose. 1997. Injection Molding of Metals and Ceramics. Princeton, N.J.: Metal Powder Industries Federation. 6. Hausnerova, B., T. Honek, P. Saha, and T. Kitano. 2004. Viscoelastic properties of powder injection moulding compounds. J. Polym. Mater. 21: 1–10. 7. Novak, S., S. M. H. Olhero, J. M. F. Ferreira, and A. Zupancic. 2004. Rheological properties of paraffin suspensions of surface-modified alumina powder for low-pressure injection moulding. Rheol. Acta 43: 559–566. 8. Metzner, A. B. 1985. Rheology of suspensions in polymeric liquids. J. Rheol. 29: 739–745. 9. Husband, D. M., and N. Aksel. 1993. The existence of static yield stresses in suspensions containing noncolloidal particles. J. Rheol. 37: 215–235. 10. Hull, T. R., D. Price, Y. Liu, C. L. Wills, and J. Brady. 2003. An investigation into the decomposition and burning behaviour of ethylene-vinyl acetate copolymer nanocomposite materials. Polym. Degradation Stabil. 82: 365–371. 11. Reichel, B., K. Wagner, D. S. Janisch, and W. Lengauer. 2010. Alloyed W-(Co,Ni,Fe)-C phases for reaction sintering of hardmetals. Int. J. Refract. Metals Hard Mater. 28: 638–645. 12. Laptev, A. V. 2007. Theory and technology of sintering, thermal and chemicothermal treatment. Densification of WC-Co alloys in solid-phase sintering (review). Powder Metall. Met. Ceram. 46: 317–324.