Journal of Materials Processing Technology 213 (2013) 101110

Contents lists available at SciVerse ScienceDirect

Journal of Materials Processing Technology

journal homepage: www.elsevier.com/locate/jmatprotec

Experimental and simulation studies of micro blanking and deep drawing compound process using copper sheet
M.W. Fu , B. Yang, W.L. Chan
Department of Mechanical Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong

a r t i c l e

i n f o

a b s t r a c t
Microforming process is a promising micromanufacturing technology for producing microparts due to its high efciency, low production cost and good product quality. The occurrence of size effect in the process, however, leads to the uncertainties in process determination, tooling design and product quality control and renders the design full of challenges. In-depth study of material deformation behavior in microforming process is thus crucial for development of quality microparts. In this research, the micro compound blanking and deep drawing of copper sheet is conducted. To investigate the size effect, the similarity theory is employed in die design and deformation process simulation. The grain size effect is studied by preparing the copper sheets with different grain sizes, while the feature size effect is also studied via using different punch radii. Based on the analysis of experimental results, it is found that the deformation load decreases with the increase of grain size, but this decrease is not signicant when there are only a few grains in the cross section of the sheet metal. The deformation becomes inhomogeneous with the decrease of formed part size and the increase of grain size. This further leads to the irregular geometry and rough surface nish of the formed part. Furthermore, the simulation is carried out to reveal the entire deep drawing process. The deformation loads predicted by simulation have the same trend as the experimental ones, but the difference exists in-between. This indicates that the simulation of microforming process needs to consider the inhomogeneous deformation of material in the process. 2012 Elsevier B.V. All rights reserved.

Article history: Received 12 April 2012 Received in revised form 20 July 2012 Accepted 25 August 2012 Available online 3 September 2012 Keywords: Microforming Micro compound blanking and deep drawing Deformation behavior Product quality analysis Finite element simulation

1. Introduction With the advance of microforming process in terms of the understanding of deformation behavior, process determination, and tooling design, microforming technology has become a more feasible, implementable, repeatable and controllable micromanufacturing process in recent years. The microparts with relatively small size, reliable structures, as well as the acceptable surface nish can be achieved nowadays. As a sub-set of micromanufacturing process, microforming is referred to the fabrication of parts or its feature with at least two dimensions in the sub-millimetre range via micro-scaled plastic deformation. It has a promising potential in mass production of micro components. Most typical microforming parts are leadframe, connector pin, IC-carriers, shaft of micro motor, etc. However, when the dimensions of parts are scaled down to micro/meso scale, the occurrence of size effect signicantly affects the microforming process. The traditional knowledge in macroscale forming cannot be leveraged into the microforming process. Therefore, in-depth understanding of material deformation behavior and

Corresponding author. Tel.: +852 27665527. E-mail address: mmmwfu@polyu.edu.hk (M.W. Fu). 0924-0136/$ see front matter 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jmatprotec.2012.08.007

process characteristics is critical in design and development of micro-formed parts. In tandem with the above-mentioned facts, many efforts have been made to the researches in microforming arena. Chan et al. (2010) have revealed that the ow stress decreases and inhomogeneous deformation occurs with the increase of grain size based on compression test using pure copper. Gau et al. (2007) conducted the tensile and bending tests of sheet metal to investigate the interactive effect of specimen size and grain size on material deformation behavior. It is found that the yield strength, tensile strength and stretchability (formability) decrease with the decreasing ratio of sheet metal thickness to grain size. Lai et al. (2008) established a constitutive material model for the change of ow stress with the specimen and grain sizes based on the surface layer model and HallPetch relation. In addition, Parasiz et al. (2007) found that the inhomogeneous deformation and irrational hardness distribution occur in the micro-extruded part when the grain size is in the order of micropart size. They also found the similar phenomena in micro-bent parts in the later research (Parasiz et al., 2010). To avoid the inhomogeneous deformation, Rosochowski et al. (2007) achieved the uniform material ow in the microextrusion experiment with ultrane-grained material. Eichenhueller et al. (2007) investigated the microforming with a heating unit. They found that additional slipping systems of grains are activated with the increase

102

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110

Fig. 1. Blanking and deep drawing compound tooling set. (a) Upper and lower die halves; (b) blanking punches and drawing dies; (c) drawing punches, punch radius = 3.75t; and (d) drawing punches, punch radius = 2t.

Fig. 2. Schematic illustration of the micro blanking and deep drawing compound process.

of forming temperature, which further results in more homogeneous deformation. Deng et al. (2011) carried out the researches into the inuence of specimen size, grain size and the asperity size of material on the surface deformation behavior via compression of pure copper cylinder. It is found that the toolingworkpiece interfacial friction increases with the decrease of specimen size due to the increase of the real contact area (RCA). It is also concluded that the open and closed lubricant pockets result in the nonuniform deformation on material surfaces. Furthermore, Chan et al. (2011)

conducted an investigation on different microextrusion processes. Their experimental results suggest that the grain size, part feature size and toolingworkpiece interfacial friction have a close relationship with the material ow behavior and deformation load. Therefore, the conventional friction evaluation methods might not be accurate in measurement of the friction coefcient in microforming systems. Among the worldwide efforts, micro deep drawing process is also one of the focuses. Hirt et al. (2003) raised the critical issues

Fig. 3. Different size-scaled tooling sets and the corresponding blank thicknesses.

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110 Table 1 Dimensions of the die set (mm).

103

D1, Diameter of drawing die; D2, diameter of drawing punch; D3, diameter of blanking punch; R1, radius of drawing die; R2, radius of drawing punch; C, drawing clearance.

in design and development of micro deep drawing system in terms of the size effect on material deformation behavior, tooling design, measurement of formed part geometry and simulation of micro deep drawing process. Justinger and Hirt (2007) formed the cups with the diameter ranging from 8.0 to 1.0 mm to investigate the inuence of punch velocity, microstructure and scaling factor on the formed cup geometry. They found that the decrease of deformation load with the increase of punch velocity in deep drawing process. They also found that the deformation load is independent of the scaling factor but dependent on the grain size and sheet thickness. Their nding of the effect of punch velocity can be explained based on the result of strip drawing test that the friction coefcient decreases with the increase of drawing velocity as reported by Vollertsen and Hu (2010). Furthermore, Justinger and Hirt (2009) further investigated the inuence of grain size and grain orientation in micro deep drawing via considering Taylor factor in their later research. Based on their experimental results, it is shown that the thinning of cup wall becomes signicant and the surface roughness increases with the decrease of the deformed part size. Manabe et al. (2008) studied the surface roughness of the deformed cups. They concluded that the increase of surface roughness is caused by the inhomogeneous deformation of polycrystalline metal. Saotome et al. (2001) conducted the investigation into the micro deep drawing, in which limit drawing ratio is experimentally determined for the thin steel sheets with the thickness ranging from 1 to 0.05 mm. Furthermore, a novel micro deep drawing process supported by laser beam was proposed by Erhardt et al. (1999). In this process, the deformation region is locally heated by laser beam to improve the formability. Vollertsen and Hu (2006) investigated the strip drawing process and developed a model to estimate the material size effect on tribology. Their model was further validated in analysis of the micro deep drawing of rectangular part (Hu, 2011). Based on the above-mentioned prior arts, it can be seen that many researches have been done in micro deep drawing process, but there is a lack of in-depth research on the micro blanking and deep drawing compound process. The objective of this research is to provide an in-depth study of the process considering the blank thickness, grain size and tooling feature size effects. The physics behind the change of deformation load and the quality of the deformed parts are also investigated.

Fig. 4. Microstructures of the pure copper sheets.

2. Experimental setup A blanking and deep drawing compound tooling set is developed, as shown in Fig. 1, to overcome the difculty in positioning the tiny workpiece in the die set. The schematic illustration of the forming operations is shown in Fig. 2. A circular workpiece is rstly blanked out and then drawn into the die cavity to form a cup by

Table 2 Grain sizes of the copper sheets.

As -received

Annealing 500 C 15 20 21
o

650 C 33 125 55

800 oC 75 150 100

Thickness (mm) 0.1 0.2 0.4 0.1 0.2 0.4 0.1 0.2 0.4 0.1 0.2 0.4 Grain size (m) 8 12 15

104

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110

260 350 300 240

Stress (MPa)

250 200 150 100 50 0 0.00 0.05 0.10 0.15 0.20 0.25 0.30

220 200 180 160 0.08 0.10 0.12 0.14 0.16 0.18

Thickness: 0.1mm As-received o 500 C o 650 C o 800 C Thickness: 0.2mm As-received o 500 C o 650 C o 800 C Thickness: 0.4mm As-received o 500 C o 650 C o 800 C

Stress (MPa)

Strain

Strain
Fig. 5. Material properties of the testing materials.

a single stroke. The bottom spring provides a force to the blank holder to prevent the material at ange region from wrinkling. To investigate the size effect, the similarity theory is employed in the experiment. A scaling factor ( ), which has the values of 1.0, 0.5 and 0.25 in this experiment design, is used to scale down the geometrical parameters of the workpiece and die set proportionally. Three size-scaled forming systems are thus designed and fabricated corresponding to the three scaling factors. Each forming system has a specic blank thickness, viz., 0.4, 0.2 and 0.1 mm, respectively, as shown in Fig. 3. The largest drawing punch diameter is 3.00 mm, while the smallest is 0.75 mm. The clearance between punch and die is 1.5 times the corresponding sheet thickness (1.5t). Two drawing punch radii, viz., 3.75t and 2t, for each size-scaled forming system are investigated. The drawing die radius remains to be 2 times the sheet thickness (2t) in all the studied cases. The detailed dimensions of different studied cases are presented in Table 1.

0.30 0.25 Fracture strain 0.20 0.15 0.10 0.05 As-received

Thickness 0.1mm 0.2mm 0.4mm 500 C

o

650 C Annealing

800 C

Fig. 7. The change of fracture strain with material state.

The pure copper sheets are used as the testing material in this study. They are annealed in a vacuum condition using different temperatures and holding times, viz., 500 C for 2 h, 650 C for 2 h, and 800 C for 3 h, to obtain different grain sizes. The heat-treated samples are then etched using a solution of 5 g of FeCl3 , 15 ml of HCl and 85 ml of H2 O for 1015 s. The grain sizes of the samples are examined with optical microscope. The microstructures of the copper sheets are shown in Fig. 4 and the corresponding grain sizes are presented in Table 2. It can be seen that the grain size increases with the annealing temperature. Furthermore, tensile test is conducted to obtain the ow stress curves of the testing materials and they are shown in Fig. 5. The micro deep drawing experiment is carried out in an MTS testing machine with load cell capacity of 25 KN for measuring the deformation load. The testing platform is shown in Fig. 6. The toolingworkpiece interface is lubricated with machine oil. Formability could be affected by forming speed and blank holder force as shown by Vollertsen and Hu (2010). In this study, the forming speeds and blank holder forces are determined based on the scale factors and the slow forming speeds are used in different studied cases. In such a way, the inuence of the punch speed is excluded. Table 3 shows the process parameters in the experiment.
Table 3 Process parameters in deep drawing experiment. Scale factor ( ) 0.25 0.5 1 Punch speed (mm/s) 0.01 0.02 0.04 Punch stroke (mm) 0.85 1.7 3.4 Blank holder force (N) 5 10 20

Fig. 6. Testing platform.

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110

105

Load-stroke curves of micro deep drawing with 0.1mm sheets (punch radius=3.75t)
1000

Load-stroke curves of micro deep drawing with 0.1mm sheets (punch radius=2t)

Scaled deformation load (N)

800

Scaled deformation load (N)

600

As received Grain size=15 m Grain size=33 m Grain size=75 m

800 600 400 200

400

200 0.0 0.2 0.4 0.6 0.8 Scaled punch stroke (mm) 1.0 1.2

As received Grain size=15 m Grain size=33 m Grain size=75 m 0.2 0.4 0.6 0.8 1.0 Scaled punch stroke (mm) 1.2

0.0

(a) =0.25 , punch radius=3.75t

800 Scaled deformation load (N)
Load-stroke curves of micro deep drawing with 0.2mm sheets (punch radius=3.75t)
Scaled deformation load (N)

(b) =0.25, punch radius=2t

Load-stroke curves of micro deep drawing with 0.2mm sheets (punch radius=2t) 800

600

As received Grain size=20 m Grain size=125 m Grain size=150 m

600

400

400

200 0.0 0.2 0.4 0.6 0.8 1.0 1.2

200 0.0 0.2 0.4 0.6

As received Grain size=20 m Grain size=125 m Grain size=150 m 0.8 1.0 1.2

Scaled punch stroke (mm)

Scaled punch stroke (mm)

(c) =0.5, punch radius=3.75t

Load-stroke curves of micro deep drawing with 0.4mm sheets (punch radius=3.75t)

(d) =0.5, punch radius=2t

Load-stroke curves of micro deep drawing with 0.4mm sheets (punch radius=2t)

Scaled deformation load (N)

Scaled deformation load (N)

800

600

As received Grain size=21 m Grain size=55 m Grain size=100 m

800

600

400

400

200 0.0 0.2 0.4 0.6 0.8 1.0 1.2

200 0.0 0.2 0.4

As received Grain size=21mm Grain size=55mm Grain size=100mm

0.6 0.8 1.0 1.2

Scaled punch stroke (mm)

Scaled punch stroke (mm)

(e) =1.0, punch radius=3.75t

Fig. 8. Loadstroke curves in micro deep drawing experiments. (a) = 0.25, punch radius = 3.75t; (b) punch radius = 2t; (e) = 1.0, punch radius = 3.75t; and (f) = 1.0, punch radius = 2t.

(f) =1.0, punch radius=2t

= 0.25, punch radius = 2t; (c) = 0.5, punch radius = 3.75t; (d) = 0.5,

3. Results and discussion 3.1. Deformation behavior The as-received copper sheets are cold-rolled, which usually have a strong (1 0 0)[0 0 1] cube texture. Based on the prior study (Chan et al., 2012) using EBSD, the texture of copper can be signicantly weakened with the increase of annealing temperature. Simons et al. (2006) also investigated the recrystallization texture of copper foil. It is shown that the annealing of cold-rolled copper foils at 300 C can signicantly weaken the texture and obtain equiaxed grains structure. In this study, higher annealing

temperatures are used, viz., 500, 650 and 800 C. It is thus assumed that the texture of the annealed copper does not have a signicant effect on the strength of testing samples. The change of fracture strain with material state is shown in Fig. 7. It can be seen that the as-received copper has a small fracture strain and the formability is increased when the copper is annealed at 500 C. However, the grains grow larger at a higher temperature and there are only a few grains over the thickness, as shown in Fig. 4. This makes the fracture strain decrease as the deformation could be localized at the grains which are favorable to deform in tensile direction. In order to have a better comparison with the experimental results of different size-scaled deep drawing processes, the

106

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110

Fig. 9. Microstructures of the formed cups. (a) Copper sheet with the thickness of 0.1 mm and annealed at 500 C and (b) copper sheet with the thickness of 0.1 mm and annealed at 650 C.

deformation load and punch stroke are normalized based on the following equations: measured force
2

states that the ow stress of material increases with the reciprocal square root of grain size:
y (d)

hp

+ Khp d1/2

(3)

scaled deformation load =

(1)

normalized punch stroke =

punch stroke punch diameter

(2)

Fig. 8 shows the normalized loadstroke curves of the six forming cases with different material states. Each curve in the gure is the averaged value of ve tests. In general, all the loadstroke curves show the same trend that the deformation load rstly uctuates to the minimum point, and then increases steadily to the maximum point and nally falls rapidly. From Fig. 8, it can be seen that the deformation load decreases with the increase of grain size, while the as-received copper sheet shows the highest deformation load. The difference of deformation load in the different studied cases can be explained based on both the change of surface grain fraction and the grain boundary strengthening behavior. The former can be described by surface layer model, while the latter can be explained by HallPetch relation. The surface layer model is developed based on the fact that the surface grains have less constraints and their ow stress is lower than that of the inner grains. The decrease of blank thickness and the increase of grain size in turn increase the fraction of surface grains. This further decreases the ow stress and the deformation load of material. For the HallPetch relation, which

where hp and Khp are material constants with the values of 25.5 MPa and 0.11 MN/m3/2 for pure copper, respectively. Actually, the characteristic of grain boundary has a close relationship with the grain size effect on material strengthening in deformation process. The grain boundary prohibits slip transfer, which further leads to dislocation pile-up and grain boundary strengthening effect. Fig. 9 shows the microstructures of two formed cups made of blanks with the thickness of 0.1 mm annealed at 500 and 650 C, respectively. It can be seen that the distribution of different grains becomes uneven when the grain size is larger or in the same order as the blank thickness. Some regions, such as the bottom region of the formed part shown in Fig. 9b, are constituted by only 12 grains over the thickness. Therefore, the blank with larger grain size and less number of grains across its thickness have a relatively low density of grain boundaries, resulting in a lower restriction to deformation, and thus have a low deformation load. The grains with the favorable orientation to the deformation direction could be considered as soft portions, while the others as hard portions. When there are a large number of grains evenly distributed in workpiece, soft grains are surrounded by hard grains. The soft grains do not deform unless the hard grains deform simultaneously. However, it is not the case when there are only a few grains constituting the workpiece. Deformation becomes inhomogeneous and the soft grains are not fully constrained by the hard grains, which leads to the large deformation localizing at the

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110

107

Fig. 10. Schematics of the interaction between grain size and feature size.

soft grains and the decrease of material strength, as reported by Henning and Vehoff (2007). In addition to grain size effect, the effect of punch radius is also investigated in this study. As illustrated in Fig. 10, considering the two copper sheets with the same grain size in micro deep drawing process, the inuence of punch radius on material ow is less signicant for the ne-grained copper sheet. In the case of using small

punch radius and coarse-grained copper sheet, the thickness at the bottom radius of the formed cup appears to be thinnest based on the experimental results. It is due to the severe deformation of sheet metal around the radius and a considerable strain gradient, which may cause material fracture. Therefore, large feature size and small grain size can facilitate even deformation on each grain and thus the better product quality can be obtained.

Copper sheet with thickness of 0.4mm ( =1)

Scaled deformation load (N)

800 As-received
Large radius (3.75t) Small radius (2t)

600

400
800 C annealed Large radius (3.75t) Small radius (2t)
o

200 0.0 0.2 0.4

0.6

0.8

1.0

1.2

Scaled punch stroke (mm)

(a) =1
Copper sheet with thickness of 0.2mm ( =0.5)

Scaled deformation load (N)

1000

Copper sheet with thickness of 0.1mm ( =0.25)

As-received Large radius (3.75t) Small radius (2t)

Scaled deformation load (N)

800

As-received Large radius (3.75t) Small radius (2t)

800

600

600

400
800 C annealed Large radius (3.75t) Small radius (2t)
o

400
800 C annealed Large radius (3.75t) Small radius (2t)
o

200

200

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Scaled punch stroke (mm)

Scaled punch stroke (mm)

(b) =0.5

(c) =0.25
= 1, (b) = 0.5 and (c) = 0.25.

Fig. 11. The inuence of punch radius on the deformation load in different size-scaled cases. (a)

108

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110

300

Flow stress curves of copper sheets o o annealed at 500 C and 650 C

250 200 150

100 50 0.00 0.05 0.10 0.15

Thickness: 0.1mm o 500 C (15 m) o 650 C (33 m) Thickness: 0.2mm o 500 C (20 m) o 650 C (125 m) Thickness: 0.4mm o 500 C (21 m) o 650 C (55 m)
0.20 0.25 0.30

Stress (MPa)

Strain
Fig. 14. The material properties used in simulation.

that a longer stroke is needed to form the full cup without the ange with the increase of punch radius. 3.2. Quality of the formed cups Figs. 12 and 13 show the geometries of the formed cups. Based on the microstructures of the copper sheets shown in Fig. 4, the grains grow larger at a higher annealing temperature. The large grains in the thin sheets lead to the uneven deformation and the rough surface nish in the drawn cups. For the copper sheet annealed at 500 C, the grain size is in the range of 1521 m. The distribution of different grains is more even in the blank with the thickness of 0.4 mm than those in the cases with the thicknesses of 0.2 and 0.1 mm. Therefore, both the surface nish and the formed part geometry are better than those of the latter two cases. For the sheets annealed at 650 and 800 C, the grain sizes approach to the corresponding sheet thicknesses and only a few grains or single grain exist over the thickness in each case. Grain distribution and orientation highly affect the deformation behavior, which further affects the prole and accuracy of the formed cups, as shown in Figs. 9, 12 and 13. The formed cup geometry and surface nish tend to be irregular and rough with the increase of grain size and the

Fig. 12. Comparison of the different size-scaled cups using different annealed blanks. (a) Annealing temperature of 500 C; (b) annealing temperature of 650 C and (c) annealing temperature of 800 C.

The inuence of punch radius on the deformation load is shown in Fig. 11. It can be seen that the deformation load decreases with the increase of punch radius, due to the fact that a larger bending force is needed in the case with the small punch radius. Such a size effect becomes less obvious in the smallest size-scaled case with coarse grain size, as shown in Fig. 11. In addition, it can be seen that the maximum load occurs at the longer stroke with the increase of punch radius. This can be explained based on the fact that the plastic deformation satises the incompressibility condition, in such a way

Fig. 13. Comparison of the different formed cups with different material states for a given size-scaled case. (a) = 1.0, punch radius = 3.75t; (b) = 0.5, punch radius = 3.75t; (d) = 0.5, punch radius = 2t; (e) = 0.25, punch radius = 3.75t and (f) = 0.25, punch radius = 2t.

= 1.0, punch radius = 2t; (c)

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110

109

(a)
50

Load-stroke curve of deep drawing process obtained from epxeriment and simulation (thickness = 0.1mm)

(b)
180 160

Load-stroke curve of deep drawing prcess obtained from experiment and simulation (thickness = 0.2mm)

Deformation load (N)

40 30 20 10 0 0.0 Simulation (15 m) Simulation (33 m) Experiment (15 m) Experiment (33 m) 0.2 0.4 0.6 0.8

Deformation load (N)

140 120 100 80 60 40 20 0 0.0 0.2 0.4 0.6 0.8 1.0 Simulation (20 m) Simulation (125 m) Experiment (20 m) Experiment (125 m) 1.2 1.4 1.6 1.8

Punch Stroke (mm)

Punch Stroke (mm)

(c)
800

Load-stroke curve of deep drawing process obtained from experiment and simulation (thickness = 0.4mm)

Deformation load (N)

700 600 500 400 300 200 100 0 0.0 0.5 1.0 1.5 Simulation (21 m) Simulation (55 m) Experiment (21 m) Experiment (55 m) 2.0 2.5 3.0 3.5

Punch stroke (mm)

Fig. 15. Comparison of the deformation loads obtained from experiment and simulation.

decrease of the formed cup size. Similar nding is also reported by Justinger and Hirt (2009). Moreover, the anisotropic deformation appears based on the wave-shaped rim, i.e. earing, of the formed cups shown in Fig. 13. It is found that the earing in the group with a large punch radius is much obvious compared to that in the group with a small punch radius.

4.1. Comparison of the deformation loads Fig. 15 shows the comparison of the deformation loads obtained from experiment and simulation. The simulation results do not exactly match the experimental ones. However, there is a same trend of load variation in the experiment and simulation. The deformation load increases steadily to the maximum and drops quickly after it reaches the peak. The initial uctuation of the deformation load observed in experiment could be caused by the friction between the specimen and the side wall of the blanking die, which is not modeled in FE simulation. The deviation between simulation and experiment could attribute to the invalid material and frictional models used in simulation. The deformation in deep drawing process of this study could be inhomogeneous and the frictional behavior could be different from the one in macroforming. These inuencing factors are not considered in the conventional material and frictional models.

4. FE-simulation A commercial FE system, ABAQUS/Explict, is used to simulate the deep drawing process in this study. The toolings are modeled as 3D discrete rigid bodies, while the blank as a 3D deformable body. The drawing clearance between punch and die is 1.5 times the corresponding blank thickness. The radii of both the drawing punch and drawing die are modeled to be 2 times the blank thickness. Actually, the experimental results could be interactively affected by many factors, such as the amount of lubricant used, tooling and workpiece surface nishes, surface roughening effect, dimensional deviation of the tooling, and material ow stress. Therefore, it is not easy to accurately determine the friction coefcient in microforming process. In this research, friction coefcient ( ) is set to 0.1, which is widely accepted value in simulation of metal forming process. The blank is meshed with 8-node linear brick elements (C3D8R). The mesh sizes are determined based on having ve elements over the thickness direction in different size-scaled cases, which can obtain the satisfactory accuracy and stability in simulation of deep drawing process, as reported by Witulski et al. (2004). The blank holder forces in the simulation are determined based on the actual experiment. The material properties (ow stress) used are shown in Fig. 14. The constitutive equations are based on von Mises yield criterion and the normality hypothesis of plasticity.

4.2. Thickness distribution The thickness variation of the formed cups is investigated in this study. Fig. 16 shows the experimental and simulation results of the thickness distributions of the formed cups using the blanks annealed at 600 C. The three size-scaled cases have a similar thickness distribution based on the simulation results. The side wall thickness increases gradually from the bottom radius to the rim of the cup. However, the thickness remains unchanged at the bottom centre, and then suddenly decreases to its thinnest at the bottom radius. Table 4 presents the comparison between the experimental and simulation results of the maximum thickness reduction. There is a large deviation in the case with the original blank thickness of 0.2 mm. This could be explained based on the microstructures

110

M.W. Fu et al. / Journal of Materials Processing Technology 213 (2013) 101110

the predicted maximum thickness reduction deviates from experimental result when there is only one grain over the blank thickness. The deviation of the simulation results from the experiment ones indicates that the simulation of microforming process should consider the inhomogeneous deformation behavior of material and the change of interfacial friction behavior. Acknowledgments The authors would like to thank the funding support to this research from Innovation and Technology Commission of Hong Kong government under the project of ITS/313/09 and the projects of A-PJ29 and G-U923 from The Hong Kong Polytechnic University.
Fig. 16. Cup thickness distributions in different size-scaled cases using the blanks annealed at 600 C.

References
Chan, W.L., Fu, M.W., Lu, J., Liu, J.G., 2010. Modeling of grain size effect on micro deformation behavior in micro-forming of pure copper. Materials Science and Engineering A 527, 66386648. Chan, W.L., Fu, M.W., Yang, B., 2011. Study of size effect in micro-extrusion process of pure copper. Materials & Design 32, 37723782. Chan, W.L., Fu, M.W., Yang, B., 2012. Experimental studies of the size effect affected microscale plastic deformation in micro upsetting process. Materials Science and Engineering A 534, 374383. Deng, J.H., Fu, M.W., Chan, W.L., 2011. Size effect on material surface deformation behavior in micro-forming process. Materials Science and Engineering A 528, 47994806. Eichenhueller, B., Egerer, E., Engel, U., 2007. Microforming at elevated temperature forming and material behaviour. International Journal of Advanced Manufacturing Technology 33, 119124. Erhardt, R., Schepp, F., Schmoeckel, D., 1999. Micro forming with local part heating by laser irradiation in transparent tools. In: The 7th International Conference on Sheet Metal, pp. 497504. Gau, J.T., Principe, C., Wang, J.W., 2007. An experimental study on size effects on ow stress and formability of aluminum and brass for microforming. Journal of Materials Processing Technology 184, 4246. Henning, M., Vehoff, H., 2007. Statistical size effects based on grain size and texture in thin sheets. Materials Science and Engineering A 452, 602613. Hirt, G., Justinger, H., Witulski, N., 2003. Analysis of size effects in micro sheet forming. In: The 1st Colloquium Processscaling, Bremen. Hu, Z., 2011. Realisation and application of size dependent FEM-simulation for deep drawing of rectangular work pieces. CIRP Journal of Manufacturing Science and Technology 4, 9095. Justinger, H., Hirt, G., 2007. Analysis of size-effects in the miniaturized deep drawing process. Key Engineering Materials 344, 791798. Justinger, H., Hirt, G., 2009. Estimation of grain size and grain orientation inuence in microforming processes by Taylor factor considerations. Journal of Materials Processing Technology 209, 21112121. Lai, X.M., Peng, L.F., Hu, P., Lan, S.H., Ni, J., 2008. Material behavior modelling in micro/meso-scale forming process with considering size/scale effects. Computational Materials Science 43, 10031009. Manabe, K., Shimizu, T., Koyama, H., Yang, M., Ito, K., 2008. Validation of FE simulation based on surface roughness model in micro-deep drawing. Journal of Materials Processing Technology 204, 8993. Parasiz, S.A., Kinsey, B., Krishnan, N., Cao, J., Li, M., 2007. Investigation of deformation size effects during microextrusion. Journal of Manufacturing Science and Engineering Transactions of the ASME 129, 690697. Parasiz, S.A., VanBenthysen, R., Kinsey, B.L., 2010. Deformation size effects due to specimen and grain size in microbending. Journal of Manufacturing Science and Engineering Transactions of the ASME 132, 011018. Rosochowski, A., Presz, W., Olejnik, L., Richert, M., 2007. Micro-extrusion of ultra-ne grained aluminium. International Journal of Advanced Manufacturing Technology 33, 137146. Saotome, Y., Yasuda, K., Kaga, H., 2001. Microdeep drawability of very thin sheet steels. Journal of Materials Processing Technology 113, 641647. Simons, G., Weippert, C., Dual, J., Villain, J., 2006. Size effects in tensile testing of thin cold rolled and annealed Cu foils. Materials Science and Engineering A 416, 290299. Vollertsen, F., Hu, Z., 2006. Tribological size effects in sheet metal forming measured by a strip drawing test. CIRP Annals Manufacturing Technology 55, 291294. Vollertsen, F., Hu, Z., 2010. Analysis of punch velocity dependent process window in micro deep drawing. Production Engineering 4, 553559. Witulski, N., Justinger, H., Hirt, G., 2004. Validation of FEM-simulation for micro deep drawing process modeling. In: The 8th International Conference on Numerical Methods in Industrial Forming Processes, Columbus, Ohio, USA. pp. 952957.

Table 4 Maximum thickness reduction in three size-scaled cases. Scale factor ( ) 0.25 0.5 1 Original thickness (mm) 0.1 0.2 0.4 Maximum thickness reduction (experiment) 21% 29% 23% Maximum thickness reduction (simulation) 19% 19% 19%

shown in Fig. 4. The number of grains in the cross section of the blank annealed at 600 C with the thickness of 0.2 mm is less than those in the cases with the thicknesses of 0.1 and 0.4 mm. Much signicant thinning thus occurs in the case with the thickness of 0.2 mm. Such a kind of size effect on the localized thinning behavior has not yet been considered in the conventional isotropic material model. 5. Conclusions In this paper, different size-scaled compound micro blanking and deep drawing dies are designed and fabricated. The experiment of micro blanking and deep drawing is carried out using the pure copper sheets with different grain sizes. The material size effect on the deformation load, deformation behavior and nal formed part quality is further examined. When the grain size is in the order of the blank thickness, grain orientation and distribution signicantly affect deep drawing process. It is veried by comparing the deformation loads and the geometries of the formed cups in different size-scaled cases with different grain sizes. Furthermore, there is an interactive effect of grain size and formed part feature size on the deformation load. The fraction of grain boundary increases with the decrease of grain size, which leads to the increase of ow stress and deformation load. The large formed part feature size can facilitate even distribution of strain on different grains to accomplish deformation, resulting in the decrease of deformation load. The thickness of the formed cup increases gradually along the side wall from the bottom radius to the rim of the formed cup. The thinnest region appears at the bottom radius, while the bottom center region remains almost unchanged. There is no qualitative difference in the thickness distribution of the cups between micro and macro deep drawing. In addition, simulation is carried out to verify the deformation load and the material thinning behavior. The load predicted by simulation qualitatively agrees with the experimental one and