Experimental Investigation of Key Parameters on the

Effects of Cavity Surface Roughness in

Microinjection Molding

H.L. Zhang, N.S. Ong, Y.C. Lam

School of Mechanical and Aerospace Engineering, Nanyang Technological University,
Nanyang Avenue, Singapore 639798

For microinjection molding, it is envisaged that cavity

surface roughness plays an important role in the cavity
lling of polymer melt. This article presents an experimental investigation into the surface roughness effects
on the ow area of a microthickness disk through
injection molding. Three core inserts, each of which
has different surface roughness on its two semicircular
halves but with the same roughness mean lines, were
machined and formed the mold cavity. The difference
in ow area (or volume) between these two semicircular halves of the molded part was investigated by varying the mold and melt temperatures. Regressive analysis of the signicance of mold and melt temperatures
and cavity thickness on the surface roughness effects
was carried out. Experimental results obtained indicated that the ow area on the smoother half is larger
than that on the rougher half during cavity lling. For
the same surface roughness, its effect on cavity lling
is a function of mold temperature, melt temperature,
and cavity thickness. An increase in mold temperature
or melt temperature will result in smaller surface
roughness effect on the ow area. When cavity thickness is reduced, the surface roughness effect will
become more signicant. Moreover, a larger difference
in the surface roughness between the two semicircular
halves of the insert will result in a larger difference in
the ow area between the two halves of the molded
part. POLYM. ENG. SCI., 48:490495, 2008. 2008 Society of
Plastics Engineers

INTRODUCTION
In recent years, plastic microparts have been used
extensively in various elds such as medical technology,
biotechnology, optics, telecommunication, and microuidics. Examples of products include pressure sensors, biological reactors, optical switches, ber connectors, micromotor and gears, and micropumps and valves [13].
Correspondence to: N.S. Ong; e-mail: [email protected]
Contract grant sponsors: Nanyang Technological University, Moldow
Pty. Ltd.
DOI 10.1002/pen.20981
Published online in Wiley InterScience (www.interscience.wiley.com).
C 2008 Society of Plastics Engineers
V

Injection molding is one of the most common processes for cost-effective mass production of plastic microparts. Compared with conventional injection molding (i.e.,
macroinjection molding), microinjection molding requires
higher injection pressure, injection rate, and mold temperature to achieve good quality molded parts [49]. Evacuation tools are usually needed to avoid incomplete lling
or degradation of polymer caused by entrapped air [2, 6].
Moreover, the factors such as viscous dissipation [7, 10,
11], wall slip [1215], and mold surface roughness [16
19], which are insignicant in conventional injection
molding, may play an important role in microinjection
molding.
For perfect polymer-mold interface (i.e., there is no
gap between polymer and mold contact interfaces), mold
surface roughness will enhance heat transfer, as heat
transfer rate is proportional to the contacted surface area
[17]. However, for imperfect polymer-melt interface,
mold surface roughness will increase thermal contact resistance, and thus resist heat transfer between polymer
and mold [20, 21].
In our previous work [22], the effects of surface roughness have been investigated experimentally through molding of a microthickness disk. However, only one core
insert was machined and used in the experiment. Moreover, the factors such as melt temperature and the cavity
thickness on the effects of cavity roughness were not
claried. In this work, three more core inserts, each of
which has different surface roughness on its two semicircular halves but with the same roughness mean lines,
were machined and formed the mold cavity. The ow
areas of the molded parts were measured more precisely
with the aid of graph digitalization software. The key parameters such as mold and melt temperatures and cavity
thickness on the effects of surface roughness were studied
through linear regression analysis.
EXPERIMENTAL PROCEDURE
Injection molding of microthickness disks using
POM (Ultraform W2320 003) was performed on a 5-ton
POLYMER ENGINEERING AND SCIENCE-2008

FIG. 2. Measurement locations of surface roughness of the core insert.

FIG. 1. Schematic diagram of: (a) the mold cavity and (b) the molded
part.

injection molding machine. Three core inserts (numbered

I, II, and III) were machined using an electrical-discharge-machine (EDM) and used respectively to form the
core wall of the cavity. Figure 1 shows the schematic diagram of the mold cavity (see Fig. 1a) and the molded part
(see Fig. 1b). The diameters of the gate and the core
insert are 0.4 and 8 mm, respectively. The wall of the
cavity insert is smooth with comparably negligible surface
roughness. The wall of the core insert consists of two
semicircular halves of different surface roughness but
with the same roughness mean lines. The separating line
between the two halves was adjusted vertically in the
mold such that gravity has the same effect on both halves
during cavity lling. Therefore, it is reasonable to assume
that the two semicircular halves are lled under the same
processing conditions, and the difference in ow area is
predominantly caused by the difference in surface roughness between the two halves of the core insert.
A Talyscan 150 dual gauge system was used to measure the surface roughness at three different locations, as
illustrated in Fig. 2. The measurement length at each location was 6 mm. The scanning speed of the stylus was
60 lm/s. The average of the roughness values obtained at
the three locations was used to identify the surface of the
inserts. Table 1 gives the measured surface roughness of
the three core inserts as well as the cavity insert, where
Ra is the arithmetical average of the absolute values of
the roughness prole and Dd is the vertical distance from
the roughness mean line of the rougher half to that of the
smoother half. For the core insert, Dd 0 indicates that
the mean lines of the two halves are the same, Dd > 0
DOI 10.1002/pen

indicates that the mean line of the rougher half is higher

than that of the smoother half, and Dd < 0 indicates that
the mean line of the rougher half is lower than that of the
smoother half. Therefore, for insert I or III, the mean line
on the rougher half is lower than that on the smoother
half, which will result in underestimation of roughness
effects. Conversely, for insert II, the roughness effects
will be overestimated. To avoid signicant error caused
by Dd = 0, cavity thicknesses ranging from 250 to
300 lm were used, thus limiting Dd/H in the order of 1%.
The determination of ow area of the molded part is
not an easy task, considering the fact that the shape of the
molded part is usually not regular for short-shot molding.
To measure ow area precisely, a graph digitization software kit, GetData, together with AutoCAD was used. The
molded part was rst scanned using a scanner and saved
as an image. Subsequently, the image was imported into
the GetData, and the outline of the molded part was discretized by a set of points. These points were then
exported to a .dxf le. Finally, AutoCAD was used to
import the .dxf le and its area calculation function was

TABLE 1. Surface roughness of the inserts.

Core insert
Insert I (center)
Insert I (above)
Insert I (below)
Average
Insert II (center)
Insert II (above)
Insert II (below)
Average
Insert III (center)
Insert III (above)
Insert III (below)
Average

Rougher
half Ra (lm)

Smoother
half Ra (lm)

Dd
(lm)

3.23
4.55
4.59
4.12
6.78
5.80
5.60
6.06
3.69
3.40
3.80
3.63

0.24
0.24
0.22
0.23
0.27
0.25
0.25
0.26
0.25
0.27
0.28
0.27

4.00
211.00
1.50
21.83
0.50
0.50
0.00
0.33
6.00
25.00
29.00
22.67

Ra (lm)

Cavity insert
Cavity wall (center)
Cavity wall (above)
Cavity wall (below)
Average

0.032
0.038
0.037
0.076

POLYMER ENGINEERING AND SCIENCE-2008 491

line obtained through regressive tting of the experimental data using the linear function as follows:
Ar kAt

(1)

where Ar is the ow area of the rougher half and At is the

total ow area of the molded part. The value of k is
bounded between 0 and 1. There will be negligible roughness effect if k 0.5, indicating that the ow areas are
the same for both halves. The larger the difference in the
value of k from 0.5, the more signicant is the roughness
effect. As the ratio, k, is dened using the rougher half, it
is expected that the value of k is less than 0.5.
In the experimental investigations, injection rate, mold
and melt temperatures were selected as the controlled parameters. Four cartridge heaters were used to heat up the
mold, and the mold temperature was monitored using two
thermocouples. Partially lled parts (short-shot molding)
of different sizes were obtained by varying the injection
rate during molding. Mold and melt temperatures were
also varied to investigate their signicance for surface
roughness effects. For inserts I and II, the cavity thickness

FIG. 3. Flow area on both semicircular halves versus the total ow

area with Tmelt 453 K and Tmold 323 K: (a) insert I, (b) insert II,
and (c) insert III.

used to obtain the area of the region surrounded by the

set of points.
To investigate the signicance of melt and mold temperatures for the surface roughness effect, an impact factor k, dened as the ratio of ow area of the rougher half
with the ow area of the entire molded part, was used.
The value of k can be determined from the slope of the
492 POLYMER ENGINEERING AND SCIENCE-2008

FIG. 4. Fitted-lines against experimental data for insert I with: (a) Tmelt
453 K and (b) Tmelt 473 K.
DOI 10.1002/pen

22 mm2 were investigated. This is to avoid the effects of

back pressure at the mold edges.
Figure 4 shows the experimental data obtained for various mold and melt temperatures when insert I was used.
When the mold temperature was increased from 323 to
383 K, while the melt temperature was kept constant at
453 K, the value of k increased from 0.46 to 0.49, namely
an increase of 6.5% (see Fig. 4a). In Fig. 4b, when the
mold temperature was increased from 323 to 383 K with
a constant melt temperature of 473 K, the value of k
increased from 0.48 to 0.49, namely an increase of 2%.
Figure 5 shows the experimental data obtained for different mold and melt temperatures when insert II was
used. When the mold temperature was increased from 323
to 383 K with a constant melt temperature of 453 K, the
value of k increased from 0.38 to 0.41, namely an
increase of 7.9% (see Fig. 5a). When the mold temperature was increased from 323 to 383 K with a constant
melt temperature of 473 K, the value of k increased from
0.41 to 0.42, namely an increase of 2.5% (see Fig. 5b).
Similarly, Fig. 6 shows the experimental data obtained
for various mold and melt temperatures for insert III.
When the mold temperature was increased from 323 to

FIG. 5. Fitted-lines against experimental data for insert II with: (a)

Tmelt 453 K and (b) Tmelt 473 K.

was 300 lm. For insert III, it was 250 lm. For every
change in the parameter setting, specimens were collected
only after 10 cycles of the injection process. This was to
stabilize the parameter setting and to obtain consistent
molded parts.

RESULTS AND DISCUSSION

In this experiment, the maximum ow area of the
molded part is 50.2 mm2, with 25.1 mm2 on each half.
Figure 3 shows plots of the ow area on the two semicircular halves versus the total ow area for the three
inserts. The melt (Tmelt) and mold (Tmold) temperatures
were kept constant at 453 and 323 K, respectively. At the
early stage of cavity lling, polymer melt tends to ow
faster on the smoother half than on the rougher half.
However, once the smoother half is nearly lled with
polymer melt (e.g., the ow area of the smoother half is
greater than 24 mm2), the melt on the rougher half tends
to ow faster. This is due to the back pressure, and it
occurs when the melt in the smoother half reached the
enclosed edge of the mold cavity. Therefore, only molded
parts with a ow area in the smoother half of less than
DOI 10.1002/pen

FIG. 6. Fitted-lines against experimental data for insert III with: (a)
Tmelt 453 K and (b) Tmelt 473 K.
POLYMER ENGINEERING AND SCIENCE-2008 493

FIG. 7. The k values for various mold and melt temperautres.

383 K with a constant melt temperature of 453 K, the

value of k increased from 0.35 to 0.39, namely an
increase of 11.4% (see Fig. 6a). When the mold temperature was increased from 323 to 383 K with a constant
melt temperature of 473 K, the value of k increased from
0.42 to 0.43, namely an increase of 2.4% (see Fig. 6b).
Figure 7 summarizes the changes in k value for various
mold and melt temperatures. It can be observed that an
increase in mold temperature will result in a larger k
value, and thus smaller surface roughness effects on the
ow area. Furthermore, the k value is more sensitive to
the change in mold temperature when the melt temperature is kept constant at a lower level of 453 K. Similarly,
an increase in the melt temperature will result in a larger
k value, and thus smaller surface roughness effect on the
ow area, and the k value is more sensitive to the change
in melt temperature when the mold temperature is kept
constant at a lower level of 323 K.
Figure 8 shows the k values for various melt and mold
temperatures for inserts I and II. It can be observed that
the k value for insert I is signicantly smaller than that

FIG. 9. The k values for different melt and mold temperatures for
inserts I with the cavity thickness of 300 lm and III with the cavity
thickness of 250 lm.

for insert II, when the mold and melt temperatures and
cavity thickness are kept constant. This is because the
roughness on the rougher half of insert I (Ra 4.12 lm)
is lower than that of insert II (Ra 6.06 lm), and the
smoother halves of the two inserts have comparable surface roughness values (e.g., 0.23 and 0.26 lm, respectively). This demonstrates the signicance of surface
roughness effects, i.e., the higher the surface roughness,
the smaller is the k value.
Figure 9 shows the k values for various melt and mold
temperatures for insert I and III. It can be observed that
the k value for insert III is much smaller than that of
insert I when mold and melt temperatures are kept constant, although the two inserts have similar roughness
(i.e., 4.12 and 3.63 lm on the rougher halves, and 0.23
and 0.27 lm on the smoother halves). This is mainly
because of the difference in cavity thicknesses used for
the two inserts (300 and 250 lm, respectively), i.e., the
smaller the cavity thickness, the more signicant are the
roughness effects.

CONCLUSION

FIG. 8. The k values for various melt and mold temperatures for inserts
I (DRa 2.06 lm) and II (DRa 6.13 lm) with the same cavity thickness of 300 lm.
494 POLYMER ENGINEERING AND SCIENCE-2008

The mold surface roughness effects in cavity lling

were investigated experimentally by comparing the difference in ow area between the two semicircular halves of
the molded part. The difference is caused by the different
surface roughness of the two halves of the core insert. Experimental results showed that the ow area on the
smoother half is larger than that on the rougher half during short-shot lling. The surface roughness effect is a
function of mold temperature, melt temperature, the difference in the surface roughness between the two halves,
and the cavity thickness. An increase in mold temperature
or melt temperature will result in smaller surface roughness effects. The signicance of surface roughness effects
is more sensitive to the change in mold temperature,
when melt temperature is kept constant at a lower level.
Similarly, the signicance of surface roughness effects is
DOI 10.1002/pen

more sensitive to the change in melt temperature, when

mold temperature is kept constant at a lower level. The
cavity thickness is signicant for the roughness effects.
Moreover, cavity thickness plays an important role in the
roughness effects, i.e., the smaller the thickness, the more
signicant are the roughness effects.
REFERENCES
1. H. Lorenz, M. Despont, P. Vettiger, and P. Reenaud,
Microsyst. Technol., 4, 143 (1998).
2. M. Heckele and W.K. Schomburg, J. Micromech. Microeng., 14, 1 (2004).
3. H.P. Heim and H. Potente, Specialized Molding Techniques,
William Andrew Publishing, Norwich, NY (2001).
4. Y.K. Shen and W.Y. Wu, Int. Commun. Heat Mass Transfer, 29(3), 423 (2002).
5. L.Y. Yu, C.G. Koh, L.J. Lee, and D.W. Koelling, Int. Commun. Heat. Mass Transfer, 42(5), 871 (2002).
6. Y.C. Su, J. Shah, and L.W. Lin, J. Micromech. Microeng.,
4, 415 (2004).
7. B. Xu, K.T. Ooi, C. Mavriplis, and M.E. Zaghloul,
J. Micromech. Microeng., 13, 53 (2003).
8. D.G. Yao and B. Kim, J. Micromech. Microeng., 12, 604
(2002).

DOI 10.1002/pen

9. Y.H. Koh, N.S. Ong, X.Y. Chen, Y.C. Lam, and J.C. Chai,
Int. Commun. Heat Mass Transfer, 31(7), 1005 (2004).
10. J. Koo and C. Kleinstreuer, Int. J. Heat Mass Transfer, 47,
3159 (2004).
11. P.S.B. Zdanski and M. Vaz, Jr., Numer. Heat Transfer A:
Appl., 49, 159 (2006).
12. H. Hervet and L. Leger, C. R. Phys., 4(2), 241 (2003).
13. K.M. Awati, Y. Park, E. Weisser, and M.E. Mackay,
J. Non-Newton Fluid, 89, 117 (2000).
14. D.J. Henson and M.E. Mackay, J. Rheol., 39, 359 (1995).
15. R.D. Chien, W.R. Jong, and S.C. Chen, J. Micromech.
Microeng., 15, 1389 (2005).
16. C.D. Smialek and C.L. Simpson, ANTEC, 3, 3373 (1998).
17. U.R. Theilade, E.M. Kjaer, and H.N. Hansen, ANTEC, 1,
463 (2003).
18. L. Weber, W. Ehrfeld, H. Freimuch, M. Lacher, H. Lehr,
and B. Pech, SPIE, 2879, 156 (1996).
19. C.A. Grifths, S.S. Dimov, E.B. Brousseau, and R.T. Hoyle,
J. Mater. Process. Technol., 189, 418 (2007).
20. R.L. Goff, G. Poutot, D. Delaunay, R. Fulchiron, and E.
Koscher, Int. J. Heat Mass Transfer, 48, 5417 (2005).
21. H. Masse, E. Arquis, D. Delaunay, S. Quilliet, and P.H.L.
Bot, Int. J. Heat Mass Transfer, 47, 2015 (2004).
22. H. Zhang, N. Ong, and Y. Lam, Int. J. Adv. Manuf. Technol., in press.

POLYMER ENGINEERING AND SCIENCE-2008 495