Research Summary
Debinding Injection Molded Materials by Melt Wicking B.R. Patterson and C.S. Aria
For metal and ceramic injection molding procedures which use wax binders in the production of powderbased parts, melt wicking is commonly employed to debind the components prior to sintering. Because debinding is often a time-consuming procedure, the influence of such process variables as powder size, part height, green density and temperature have been investigated to reduce the amount of time required for debinding by melt wicking.
100 r--::::Z---=:::::====LJl 80 "0
.§ 60 '"
Cl
Powder Size (~m)
0-5
0-10
b,. -
25
0~--~5~0--~1~00~~15~0~~2~00~~2~50
Time (min.) Figure 1. Influence of part powder size on the percent debound with time.
INTRODUCTION Following the molding of a part by either metal or ceramic injection molding, but prior to sintering, a number of methods may be used to debind the component. 1- 3 Depending on the binder and powder employed, these techniques include thermal degradation, solvent extraction and melt wicking. While the debinding step is frequently time consuming, different process variables can be adjusted to minimize the duration of the procedure. Although theoretical models have been applied to some of these debinding mechanisms,4 few comparative tests have been performed, and little firsthand knowledge is available concerning the true effects of the different process variables on debinding rates. In debinding by melt wicking, which is commonly used with wax binders, newly formed parts are removed from the mold and embedded in a fine powder where, upon heating, the molten binder is absorbed with the aid of capillarity. A number of process variables-including powder size in the part and wick, part height, green density or compaction of the wick powder, and temperature-markedly affect the debinding rate. Controlled studies of these different variables have revealed how debinding rates can be influenced.5 This information can be used to optimize debinding conditions and minimize debinding time. EXPERIMENTAL PROCEDURES In these studies of debinding by melt wicking, fine stainless steel powder6 and paraffin wax were used as the part and binder materials, respectively. Activated alumina7 was used as the wicking powder. Debinding experiments were designed to examine the effect of variations in process variables under otherwise standard conditions. The standard powder used for injection molding was as-received spherical stainless steel powder, with an average particle size of 11 11m. This powder was air classified and sieved to obtain narrow cuts with mean sizes of 5, 10 and 25 11m, to examine the effect of part powder size on debinding rate. The wick powder was attritor milled for different times to vary the fineness; powder ground for 48 minutes was generally used as the standard wick material. Debinding was performed at 60°C except for comparative runs at 75°C. The standard wick powder green density after tapping was 30% offull density, with comparative runs performed at 40% compaction. Cylindrical specimens ofO. 724 cm diameter were prepared by extruding mixtures of 55 vol. % stainless steel powder and molten paraffin and sectioning the resulting rods into cylinders with lengths of 0.32 cm, 0.64 cm and 1.27 cm to examine the effect of part height on debinding rate. The 0.32 cm specimen was used as the standard part height. These different process variables and their range of variation are described in Table 1. The debinding procedure involved setting the wick powder in a stainless steel boat and placing the specimens vertically on the bed. The wick bed was made Table I. Experimental Parameters
80
Grind Time (min.) 0-12 0-24
<> -4B
b,. - 96
50
100
150
200
250
Time (min.) Figure 2. Effect of the wick powder fineness on the percent debound with time.
22
Wick Powder Sample Wick Powder Milling Powder Sample Green Time Size Temp. Height Density Experiment (min.) (/lm) ( OC) (cm) _ _--'-('*_0:....)_ _ 12,* 24* 11 60 0.32 30 Effect of Wick Powder Size 48,* 96* 60 5* 24 0.32 30 Effect of Sample Powder Size 10* 25* 0.32* 48 Effect of Sample Height 60 30 11 0.64* 1.27* Effect of Wick Powder Green Density 48 0.32 30* 11 60 40* Effect of Temperature
48
11
60* 75*
0.32
30
* Denotes the variables t est ed in each experiment.
JOM • August 1989
0.20..-------------,
100~----------------------~
80
0.15
Sample Height (em) 0-0.32 0-0.64 ~-1.27
:§
'" '" .3
E 0.10 Ol
~
Sample Height (em) 0-0.32 0-0.64 0-1.27
50
100 150 Time (min.)
200
250
Figure 3. Effect of part height on the percent debound with time.
50
100 150 Time (min.)
200
250
Figure 4. Effect of part height on the absolute weight of wax removed with time.
sufficiently deep, and the specimens were spaced far enough apart to prevent interference with the flow of the wax into the wick. The specimen boat was heated in the furnace to the debinding temperature (above the melting point of the wax) and held at temperature for different times. The amount of wax lost was determined by weighing the specimen after removal from the furnace. EFFECT OF VARIABLES
Figures 1 through 4 illustrate the effects of the different process variables on debinding, expressed as percent or weight of wax removed versus time. These curves are generally parabolic in shape with the rate of weight loss decreasing with time. Figure 1 illustrates the effect of the part powder size on the debinding rate; coarser powder sizes debound most rapidly. The 25 J..lm powder sample required only one-third the time needed to debind the 5 J..lm powder sample. This increased debinding rate with increasing part powder size can be explained by two possible effects related to the resulting increase in pore sizes in the part: decreased capillarity holding the wax in the part or increased permeability of the part, either of which would cause the part to debind more rapidly. All variations in processing conditions examined herein affect both the permeability and capillarity of the part or wick, and both of these phenomena show strong, although sometimes opposite, influences on the resulting debinding rate. In Figure 2, it is apparent that the decreased pore size in the wick powder with longer grinding times markedly shortened the time for debinding. This result can only be attributed to increased capillary pressure drawing the wax into the more finely ground wick powder with finer pores, because the decreased permeability of this wick material would otherwise reduce the debinding rate. Figures 3 and 4 describe the effect of part height on debinding rate. Figure 3 shows that the shorter parts achieve a particular percentage of debinding in considerably less time than taller parts. Figure 4, showing cumulative weight loss versus time, illustrates that the taller parts actually have a greater absolute rate of wax removal due to their greater wax pressure head. This greater initial wax content in the tall parts causes the debound percentage to be less than for shorter parts. The implication of this result, with regard to a particular part size, is that the debinding rate should be greatest when the longest dimension of the part is vertical. This positioning may not always be practical due to part warpage during debinding. Embedding the parts in the wick powder would increase the overall debinding rate since debinding could occur from all surfaces. The surrounding powder would also support the part and prevent slumping. Experiments performed with varying wick green densities showed that increased wick green density significantly decreased the debinding rate, probably due to decreased pore volume fraction and permeability in the wick. Increasing the temperature from 60 to 75°C only slightly increased the debinding rate, through reduction of the wax viscosity. Figures 5 through 7 illustrate the above effects of part and wick powder size and part height on debinding rate, obtained from the slopes of the curves of Figures 1, 2 and 4, at different constant percentages or volumes of wax debound. Figure 5 clearly shows the increase in debinding rate with increased part pore radius, computed as one-half of the mean free pathS between particles, for the different powder sizes. The overall decrease in rate with increased percentage of debinding is due to the decrease in the combined (average) permeability of the sample and wick when more of the wax is in the low permeability wick. Figure 6 shows the decreased debinding rate, from Figure 2, with increase in wick pore size, due to decreased capillary pull into the wick. The wick pore size was obtained by mercury porosimetry.9 Figure 7 shows the increase in debinding rate with increased sample 1989 August • JOM
14
c 12 :§
ME g 'i
C>
x
10
Percent Debound ~-44
0-65 0-87
8
Q)
1ii
a:
6
c: '5 c: :0 Q)
4
Ol
0
2 0
2 4 3 5 Sample Pore Radius (~m)
6
7
Figure 5. Influence of sample pore size on debinding rate at constant levels of debinding.
36
c
32
:§
28 -
g
24
ME
'i
C>
x Q)
16
Ol
12
c: '5 c: :0 Q)
0
~-22
0-44 0-87
20
1ii
a:
Percent Debound
~
8 4 0
0.2
1.0
Figure 6. Influence of wick pore size on debinding rate at constant levels of deb in ding. 23
36 c:::
30
ME .2-
24
:§
height at constant amounts of sample weight loss, equivalent to constant volumes of wax in the wick. In Figures 5 and 6 specimens have a constant part height and comparisons of the debinding rate at constant percentages of debinding are also at equivalent amounts of wax in the wick. PERMEABILITY AND CAPILLARITY EFFECTS
"'1
0
x
18
Q)
«i
a:
C>
c:
'0 c:
:c Q)
0
12 6 0
1.6 Sample Height (em)
Figure 7. Influence of sample height on debinding rate at constant volumes of wax removed.
In all ofthe aforementioned cases, relatively small variations in the processing variables resulted in manifold changes in debinding rate. All of the results from variation in the process variables can be explained on the basis of changes in the permeability or capillarity of the part or wick or both. From the standpoint of permeability, larger pore sizes and lower green densities of the part and wick should increase the debinding rate. Large part pore size and small wick pore size would maximize debinding from the standpoint of capillary pull out ofthe part and into the wick, since the attractive capillary pressure is inversely proportional to pore radius. Increased part height, with a greater wax pressure head, should increase flow by permeation, as was seen experimentally. These results clearly show that both permeability and capillarity considerations are important in melt wicking. The results of Figures 1 and 5, showing faster debinding with coarser part powder size, are explainable by either capillarity or permeability considerations. Figures 2 and 6, showing faster debinding with finer wick powder, indicate an overriding effect ofcapillarity since finer wick powder size at similar green density decreases permeability. Increasing debinding rate with increased part height, as in Figures 4 and 7, results from gravity driven permeability considerations alone. The decrease in debinding rate with increased wick green density, from compaction, was consistent with permeability expectations. CONCLUSIONS It is apparent that small changes in processing variables in melt wicking exert great influence on the resulting debinding rate. Understanding these effects through modeling and experimentation enables producers to optimize debinding conditions and reduce processing time-a significant limitation in metal and ceramic injection molding.
ABOUT THE AUTHORS • • • •_ B.R. Patterson received his Ph.D. in materials science and engineering from the University of Florida in 1978. He is currently an associate professor in the Department of Materials Engineering at the University of Alabama at Birmingham. Dr. Patterson is also a member of TMS. C.S. Aria is working on his M.S. in materials engineering at the University of Alabama at Birmingham. He is currently a metallurgist with Delco-Remy in Meridian, Mississippi. Mr. Aria is also a member of TMS. If you want more information on this subject, please circle reader service card number 53.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge Avesta Stainless, Inc., for donating the stainless steel powders, Alcoa for donating the alumina wick powder and Quantachrome Corporation for mercury porosimetry determination ofthe wick pore sizes. We also wish to acknowledge publication of portions of this study by the Metal Powder Industries Federation in Reference 5. References 1. RS. Libb, B.R Patterson and H.A. Heflin, "Production and Evaluation of PIM Injection Molding Feedstocks," Progress in Powder Metallurgy, vol. 42 (Princeton, NJ: Metal Powd. Ind. Fed., 1986), p. 95. 2. B.R Patterson, RJ. Waikar and M.T. Young, "Several Aspects ofPIM Injection Molding," op. cit. 1, p. 85. 3. A.R Erickson and RE. Wiech, Jr., "Injection Molding," Metals Handbook Volume 7: Powder Metallurgy, 9th ed. (Metals Park, OH: ASM, 1984), p. 495. 4. RM. German, "Theory of Thermal Debinding," Int. J. of Powd. Metall., 23 (1987), p. 243. 5. C.S. Aria and B.R. Patterson, "Influence of Process Variables on Debinding and Melt Wicking," Mod. Dev. Powd. Metall., vol. 18 (Princeton, NJ: Metal Powd. Ind. Fed., 1988), p. 403. 6. Avesta Stainless, Inc., Fairfield, NJ. 7. Grade F-l alumina, Aluminum Company of America, Chemicals Division, Pittsburgh, PA. 8. RL. Fullman, "Measurement of Particle Sizes in Opaque Bodies," Trans. AIME, 197 (1953), p. 447. 9. Courtesy of Quantachrome Corporation, Syosset, NY.
Free Placement Service for TMS Members ... As a member of TMS, you are entitled to all the benefits of a national job placement service, at no charge to you. TMS Careerline is a computerized network of placement services that is a joint effort ofTMS and Placement Services, Ltd.
If you are looking to advance your career through a job change, you will appreciate the value ofTMS Careerline. Free placement service is just another of the many valuable benefits that accompany of your membership in TMS.
For more information on TMS Careerline contact:
TMS
Minerals· Metals· Materials
Joanne Melder TMS • Membership Services 420 Commonwealth Drive Warrendale, PA 15086 412-776-9011