Subscriber access provided by UNIV OF SCIENCES PHILADELPHIA

General Research
Study on the foaming mechanism of polypropylene
in the gas-assisted microcellular injection molding
Junji Hou, Guoqun Zhao, Lei Zhang, Guiwei Dong, and Guilong Wang
Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.7b05389 • Publication Date (Web): 16 Mar 2018
Downloaded from http://pubs.acs.org on March 16, 2018

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W.,

Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 Study on the foaming mechanism of polypropylene in the gas-assisted
5
6 microcellular injection molding
7
8
Junji Hou, Guoqun Zhao*, Lei Zhang, Guiwei Dong, Guilong Wang
9
10
11 Key Laboratory for Liquid-Solid Structural Evolution and Processing of Materials (Ministry of Education),
12
13 Shandong University, Jinan, Shandong 250061, PR China
14
15
16
17
18 * Corresponding author
19
20
21 E-mail address: [email protected] (Guoqun Zhao)
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 1
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 2 of 31

1
2
3
4 Abstract
5
6 In our recent study, it's reported that the gas-assisted microcellular injection molding (GAMIM) is a promising
7
8
method to fabricate plastic foams with good surface quality, high weight-reduction and improved mechanical
9
10
11 performance. To clarify the foaming mechanism, the crystallization behaviors of polypropylene in both
12
13 conventional injection molding (CIM) and gas-assisted injection molding (GAIM) were studied. It's found that the
14
15
16 GAIM can significantly promote melt crystallization and refine crystals. To explore the improvement mechanism of
17
18 crystallization, a numerical model was developed to simulate the filling and cooling processes of CIM and GAIM.
19
20
21 The simulation results show that the much stronger shear field and the faster cooling rate lead to the refinement of
22
23 crystals in GAIM. Finally, a crystallization-driven cell nucleation model was proposed to explain the improved
24
25
26 foaming behavior of polypropylene in GAMIM. This paper provides a deep understanding of the foaming behavior
27
28 and hence benefits the foaming process design.
29
30
31
32
33 Keywords: microcellular injection molding; gas-assisted injection molding; gas-assisted microcellular injection
34
35 molding; crystallization; foaming; numerical simulation
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 2
59
60 ACS Paragon Plus Environment
Page 3 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 1 Introduction
5
6 Microcellular plastics are the foams with cell sizes of 0.1−10 µm, and cell densities of 109−1015 cells/cm3.1-3
7
8
Compared with the regular plastic foams with much larger cells, the microcellular plastics have many advantages,
9
10
11 such as high specific strength, excellent thermal and sound insulation, and good energy absorbing capacity.4 Thus,
12
13 they show a promising future in many industrial fields, for example, construction, automotive and packaging.4,5
14
15
16 Microcellular injection molding (MIM) is a superior technology to produce microcellular plastics due to its
17
18 high efficiency, outstanding design ability, and ability in directly producing parts with complex shape.6,7 In spite of
19
20
21 the advantages, the MIM suffers from several obvious defects. First, the foamed part has poor surface quality due to
22
23 many silver streaks and swirl marks on the surface.8-11 Second, the injection molded part's weight-reduction is
24
25
26 limited,9 even for the low-pressure MIM, during which the shot size is usually 65−95% of the cavity volume,12 the
27
28 molded part with uniform cells for weight-reduction higher than 20% is not easy to fabricate.13 In addition, the
29
30
foamed part has a seriously deteriorating mechanical performance,9 particularly for the ductile performance.7
31
32
33 For improving the injection molded foam's surface quality, the efficient method is to suppress foaming in the
34
35 filling stage or re-dissolve the gas in stretched bubbles on the part surface into the polymer melt. It has been
36
37
38 reported that adopting gas counter pressure,14,15 reducing the blowing agent content,11 and increasing the mold
39
40 cavity temperature10,16-18 can improve surface quality of the injection molded foam. However, these methods cannot
41
42
43 help to improve the mechanical performance or increase the weight-reduction of the molded part. The limited
44
45 weight-reduction of the part molded by the conventional MIM is due to the foaming occurs in a restricted space. To
46
47
48 increase the weight-reduction, the key issue is to increase the free space for cell expansion. For this purpose, the
49
50 foam injection molding with mold opening method was developed.19-23 In this process, a full-shot size of
51
52
polymer/gas solution is first used to fully fill the mold cavity, then the mold is opened for an expected distance to
53
54
55 provide free space for foaming. However, it is difficult to control the part's shape and dimension accuracy,
56
57
58 3
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 4 of 31

1
2
3
4 especially for the part with complex geometry. To improve the part's mechanical performance, the previous studies
5
6 were mainly focused on the modification of the polymer matrix by adding fillers.13,24-26 The additives themselves
7
8
can not only enhance polymer matrix but also refine cells,24,26 so the stress concentration caused by undesired
9
10
11 cellular morphology can be decreased.
12
13 In order to simultaneously endow foamed parts with good surface quality, high weight-reduction and improved
14
15
16 mechanical performance, the gas-assisted microcellular injection molding (GAMIM) method was introduced for
17
18 preparing the foamed polypropylene (PP) part in our previous work,27 and the foamed PP part meeting the
19
20
21 aforementioned requirements was successfully fabricated. Although the foaming process of PLA and TPU by using
22
23 the foam injection molding with gas-assisted processing have been studied in the researches of Mark et al.,28,29 the
24
25
26 surface quality and mechanical property of the part were not concerned. The GAMIM is actually the combination
27
28 of the gas-assisted injection molding (GAIM) and MIM. Its technical principle is illustrated in Figure 1. The
29
30
implementation process of GAMIM is briefly stated as follows. First, in order to reduce the part’s weight, a rather
31
32
33 small volume of the polymer/gas solution is used to fill mold cavity. In this process, the injected melt generates
34
35 phase separation due to a sudden pressure drop, and the numerous non-uniform cells shown in Figure 1(a) are
36
37
38 formed inside the melt. Second, the high-pressure assisted gas is injected into mold cavity, and the assisted gas
39
40 pushes the melt to fully fill the mold cavity. Then the assisted gas is held for building a high cavity pressure. The
41
42
43 non-uniform cells formed in the first foaming process can be re-dissolved into the melt by assisted gas's
44
45 compression, as shown in Figure 1(c), so the surface defects are reduced or eliminated. By the aid of the assisted
46
47
48 gas, the foaming and melt filing processes can be decoupled. Finally, a subsequent foaming is triggered by the
49
50 assisted gas’s rapid release, and the foamed part with uniform cells which is better for mechanical properties will be
51
52
obtained, as shown in Figure 1(d).
53
54
55
56
57
58 4
59
60 ACS Paragon Plus Environment
Page 5 of 31 Industrial & Engineering Chemistry Research

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17 Figure 1. Schematic principle of gas-assisted microcellular injection molding: (a) injecting polymer/gas solution; (b)
18
19
injecting high-pressure assisted gas; (c) holding the assisted gas; (d) releasing the assisted gas and foaming.
20
21
22 In our earlier work,27 it is mainly focused on the clarification of the GAMIM method and its implementation
23
24 procedure. By comparing the macroscopic properties of PP parts molded by MIM and GAMIM, such as the surface
25
26
27 quality, weight-reduction and mechanical performance, it is found that the GAMIM has great superiority and the
28
29 foaming behavior of PP can be noticeably improved. However, the mechanism behind the improved foaming
30
31
32 behavior is unclear. Therefore, this work mainly focuses on revealing the foaming mechanism of PP in GAMIM.
33
34 The differences melt crystallization behaviors of PP in both conventional injection molding (CIM) and GAIM are
35
36
37 firstly studied. Then, a numerical model was developed to simulate the filling and cooling processes of CIM and
38
39 GAIM. The influence of high-pressure assisted gas penetration on melt crystallization is further revealed by the
40
41
simulation results. Finally, the crystallization-driven cell nucleation model was proposed to explain the improved
42
43
44 foaming behavior of PP in GAMIM.
45
46 2 Experimental methods
47
48
49 2.1 Sample preparation
50
51 The isotactic PP resin (HJ730) used in this study is purchased from Hanwha Total Petrochemical Company in
52
53
54 Korea. It has a melt flow rate of 20 g/10 min, and a density of 0.91 g/cm3. The industrial grade compressed nitrogen
55
56 is used as the physical blowing agent and the assisted gas. The injection molding machine used is Borch BS800-III,
57
58 5
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 6 of 31

1
2
3
4 equipped with a supercritical fluid dosing unit purchased from Trexel Inc. and a high-pressure gas control device
5
6 purchased from Beijing Chn-top Machinery Co., Ltd. The CIM, GAIM, MIM and GAMIM samples are
7
8
respectively molded by using the mold shown in Figure 2. The mold temperature is 60 °C. The melt temperature is
9
10
11 220 °C. The injection rate is 85 cm3/s. The shot size of CIM is 100%, and for GAIM, MIM and GAMIM, the shot
12
13 sizes are all 55%. The pressure of assisted gas is 20 MPa and the physical blowing agent's content is 0.3 wt%. The
14
15
16 holding time of assisted gas is 5 s, 15 s, 25 s, 35 s and 45 s, respectively.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37 Figure 2. Structure of mold cavity and shape of injected parts.
38
39 2.2 Experimental characterizations
40
41
After injecting by using the above experimental parameters, the CIM and GAIM samples can be obtained.
42
43
44 Figure 3(a) gives the schematic drawing of specimen preparation. The cylinder sample with a diameter of 15 mm is
45
46 chosen as the characterizing object for CIM. The ring sample, of which the outer diameter is 15 mm and inner
47
48
49 diameter is 10.8 mm, is chosen as the characterizing object for GAIM. The length of all the samples is 135 mm.
50
51 For polarized optical microscopy (POM) observation, according to the marked positions shown in Figure 3(a),
52
53
54 the samples are first cut into two kinds of specimens, as shown in Figure 3(b). Then the approximately 15 µm thin
55
56 slices shown in Figure 3(c) are cut from the above two kinds of specimens using a manual rotating microtome
57
58 6
59
60 ACS Paragon Plus Environment
Page 7 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 (YD-2508B, YIDI). Finally, the prepared thin slices are placed on a POM (BX51, OLYMPUS) for observing.
5
6
7
8
9
10
11
12
13
14
15
16
17 Figure 3. Schematic drawing of specimen preparation: (a) CIM and GAIM sample; (b) specimens for cutting slices;
18
19
(c) thin slices for POM observation; (d) sampling method for differential scanning calorimetry measurement.
20
21
22 A scanning electron microscope (SU-70, HITACHI) is used to characterize the crystal morphology of the CIM
23
24 and GAIM samples and observe the cellular morphology of the MIM and GAMIM samples. Before characterizing
25
26
27 the crystal morphology, the fracture sections along the melt flow direction are etched for 18 h with an acid solution
28
29 containing 1.3 wt% potassium permanganate (KMnO4).30 All the specimens are then sprayed with platinum for
30
31
32 electric conduction.
33
34 A differential scanning calorimetry instrument (DSC 204 F1, NETZSCH) is employed to study the
35
36
37 crystallization behavior of the samples. According to the marked positions shown in Figure 3(b) of the CIM and
38
39 GAIM samples, the two kinds of slices with a thickness of 2 mm are obtained after cutting. Then, for the two kinds
40
41
of slices, three small specimens with a weight of about 10 mg are respectively cut off at the positions about 0 µm,
42
43
44 1000 µm, and 2000 µm away from the slice's skin layer, as shown in Figure 3(d), and they are marked as P0, P1000
45
46 and P2000, respectively. For these small specimens, each of them is heated from 40 °C to 220 °C at a rate of
47
48
49 10 °C/min under a nitrogen atmosphere using the DSC, and the melting process is recorded.
50
51 3 Modeling and simulation
52
53
54 3.1 Governing equations
55
56 Numerical simulation is an effective method to obtain the internal physical fields of the melt that is difficult to
57
58 7
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 8 of 31

1
2
3
4 measure in the injection molding process. For simulations of GAIM, the most used models are 2D Hele-Shaw31 or
5
6 surface models,32 but these models could not accurately describe some important physical phenomena between the
7
8
gas and liquid phases.33 In this paper, a 3D non-isothermal transient numerical model of two-phase fluid is
9
10
11 established to accurately describe the complicated mass and heat transfer phenomenon between gas and polymer
12
13 melt. The high-pressure gas (nitrogen) is treated as a fluid, and the process of nitrogen's penetration to the melt can
14
15
16 be simplified into the solution of the velocity, pressure and temperature fields of the transient flow between two
17
18 fluids.
19
20
21 Before simulation, in the polymer melt and gas filling process, the following assumptions are taken to simplify
22
23 the model and calculation: (a) since the filling is driven by the pressure, which is not very high in the filing
24
25
26 stage,34-36 the polymer melt and nitrogen can be all assumed as incompressible; (b) the nitrogen merely acts to
27
28 transfer pressure,35 and the Reynolds number is small in the polymer melt region,37 so the flow type can be
29
30
considered as laminar flow; (c) since the viscosity of polymer melt is high, the insignificant inertia force and
31
32
33 surface tension can be ignored37 and (d) no chemical occurs during the filling process. Based on the assumptions,
34
35 for simulation of the polymer melt and gas filling process, the relevant governing equations in the numerical model
36
37
38 are described as follows:
39
40 Continuity equation is written as
41
42
43 ∇ ⋅U = 0 (1)
44 where ∇ ⋅ denotes the divergence operator and U denotes velocity vector.
45
46
47 Momentum equation is written as
48
49 ∂ (ρU)
+ ∇ ⋅ ( ρ UU ) = −∇ P + ∇ ⋅ (2 µ ε ) (2)
50 ∂t
51 1
52 ε=  ∇ U + (∇ U ) T  (3)
53 2
54 where ρ means density, t means time, P means pressure field, µ means kinematic viscosity, and ε is the strain rate
55
tensor of the two-phase system.
56
57
58 8
59
60 ACS Paragon Plus Environment
Page 9 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 Energy equation used is expressed as
5
∂ ( ρ C PT )
6 + ∇ ⋅ ( ρ UC PT ) = ∇ ⋅ ( k ∇ T ) + µγ& 2 (4)
7 ∂t
8 γ& = ε ( U ) = 2ε ( U ) : ε ( U ) (5)
9
10 where C P means the isobaric heat capacity of the two-phase fluid, T means the temperature field of the mold
11
12 cavity domain, k is the heat conductivity of the two-phase fluid, and γ& is the modulus of the strain rate tensor.
13
14
15 When the filling stage is finished, the velocity field in the system is zero. During the simulation of cooling
16
17
stage, the convection heat transfer can be ignored, and the energy equation can be expressed as
18
19 ∂ ( ρ C PT )
20 = ∇ ⋅ ( k ∇T ) (6)
21 ∂t
22
3.2 Interface tracking method
23
24
25 In order to determine the interfaces of the two-phase flow, the phase fraction of each phase is calculated. The
26
27 phase fraction of polymer is defined as α, and the phase fraction of nitrogen is 1-α, correspondingly. The VOF
28
29
30 method38 is employed for solving the phase fraction in this paper. It satisfies the following equation:
31
32 ∂α
33 + ∇ ⋅ ( Uα ) = 0 (7)
34 ∂t
35
36 where, at any time, the phase fraction α of polymer in a mesh element satisfies the following formula:
37
38 1, point ( x, y, z, t ) is located in polymer
39 
α ( x, y, z, t ) = 0, point ( x, y, z, t ) is located in nitrogen (8)
40 
41 0 < α < 1, point ( x, y, z, t ) is located in interface area
42
43 Physical properties Ф in interface region are calculated by the following formula:
44
45
46
Φ = α Φ1 + (1 − α ) Φ2 , (Φ = C P , k , ρ , µ ) (9)
47 where, Φ1 and Φ2 are the physical properties of polymer and nitrogen, respectively. The physical properties of
48
49 polymer and nitrogen used are shown in Table 1.
50
51
52 Table 1. Physical properties of polymer and nitrogen.
53
54 Materials C P (J·kg−1·K −1) k (W·m−1·K−1) ρ (kg·m-3) µ (m2·s−1)
55
56
57
58 9
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 10 of 31

1
2
3
4 PP HJ730 2496 0.15 909.49 Eq. (12)
5
6 Nitrogen 1040 0.024 1.13 1.6×10-5
7
8
3.3 Viscosity model
9
10
11 The Cross-WLF model is used as viscosity model, and it is expressed as
12
13 η0
14 η=
1 + (η0γ& / τ ∗ )
1-n (10)
15
16
− A1 (T −T ∗ )
17 ∗

18 η 0 = D1 e A (T −T
2 )

19 T ∗ = D2 + D3 P (11)
20
21 A2 = A2 + D3 P
22
23 where η is dynamic viscosity, η0 is zero shear viscosity, and n, τ*, D1, D2, D3, A1 and A2 are the parameters
24
25 depending on materials. The dynamic viscosity of the momentum equation is
26
27 η
28 µ= (12)
29 ρ
30 For PP HJ730 material used in this study, according to the material database of Moldflow software, the
31
32 values of n, τ*, D1, D2, D3, A1 and A2 are 0.3301, 19385.1 Pa, 2.93023×1017 Pa·s, 263.15 K, 0 K·Pa-1, 40.995
33
34 and 51.6 K, respectively.
35
36
37 3.4 Boundary condition
38
39 The meshed geometric model of cylinder is shown in Figure 4. The polymer inlet, high-pressure gas (nitrogen)
40
41 inlet, wall and outlet boundaries are also given in the figure. Table 2 lists the setting methods of phase fraction,
42
velocity, temperature and pressure on boundaries, where Uin is the velocity vector of polymer melt on polymer inlet,
43
44 Tmelt is the temperature of polymer melt, and tgas is the starting time of filling gas or the end time of injecting
45
46 polymer melt and it is determined by the shot size and the velocity of polymer melt, and P(t) is the pressure variety
47
48 of high-pressure nitrogen with time at gas inlet, and Tgas is the high-pressure nitrogen's initial temperature , and
49
50 ΓU_filter and ΓP_filter are a mixed boundary condition of filtering gas37 at outlet.
51
52
53
54
55
56
57
58 10
59
60 ACS Paragon Plus Environment
Page 11 of 31 Industrial & Engineering Chemistry Research

1
2
3
4
5
6
7
8
9
10
11 Figure 4. Meshed geometric model of GAIM.
12
13 In order to improve convergence speed and calculation efficiency, the hexahedral grid and parallel operation
14
15
16 method are used in this paper. In solving the coupled problem of the velocity and pressure field in the momentum
17
18 equation, the PIMPLE algorithm,39 which is a combinational algorithm of PISO algorithm and SIMPLE algorithm
19
20
21 and more suitable for solving the transient problem whose flow field changed strongly, is used. The idea of
22
23 PIMPLE algorithm is that the sub-relaxation in every time step to obtain a bigger time step (same with the SIMPLE
24
25
26 algorithm), and the multiple pressure corrections of PISO is used to fully solve the coupled problem of pressure and
27
28 velocity in every iteration step. The specific calculation process is detailed described in the reference.40 According
29
30
to experimental design of injection process, the parameter values of numerical example on boundaries are shown in
31
32
33 Table 3. Based on above descriptions, this paper wrote the numerical simulation program by adopting C++
34
35 programming language on the development platform of Foam-extend 3.0.
36
37
38 Table 2. Setting methods of boundary conditions.
39
40 Variables Polymer inlet Gas inlet Wall Outlet
41
42
43  1, t ≤ tgas ∂α ∂α
α α =1 α = =0 =0
44  0, t > t gas ∂n ∂n
45
46  U in , t ≤ t gas U = U in , t ≤ t gas
47 U (m /s) U = U = ( 0, 0, 0 ) Γ U _ filter
48  (0, 0, 0), t > tgas ∇U = 0, t > t gas
49
50 ∇P = 0, t ≤ tgas
51 P(MPa) ∇P = 0 ∇P = 0 Γ P _ filter
P = P ( t ) , t > tgas
52
53
54 Tmelt , t ≤ t gas
T (° C ) T = Tmelt T = T = Tmold T = Tmold
55  Tgas , t > t gas
56
57
58 11
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 12 of 31

1
2
3
4 Table 3. Parameter values of numerical example on boundaries.
5
6 Uin (m/s) Tmelt (°C) tgas (s) P ( t ) (MPa) Tgas (°C) ΓU _ filter Γ P _ filter
7
8 ∇U = 0, α ≤ 0.3 P = 0.1, α ≤ 0.3
9 (0, 0, 0.3) 220 0.5 min [ 0.1 + 20(t − 0.5), 20] 25
U = (0, 0, 0), α > 0.3 ∇P = 0, α > 0.3
10
11
12 4 Results and discussion
13
14
4.1 Crystallization behavior
15
16
17 Figure 5 gives the crystal morphology of the CIM and GAIM samples under the polarized optical microscopy.
18
19 Figure 5(a)-(c) show the crystal morphology near the skin layer, in the middle layer, and away from the skin layer
20
21
22 of the CIM samples with the magnification of 200 times, respectively. Figure 5(d)-(f) show the crystal morphology
23
24 near the skin layer, in the middle layer and away from the skin layer (near gas channel) of the GAIM samples with
25
26
27 the magnification of 200 times, respectively. Figure 5(g) and (h) are the crystal morphology of the CIM and GAIM
28
29 samples with the magnification of 40 times, respectively. For the CIM sample, it can be seen from Figure 5(a) that,
30
31
32 near the skin layer, because the crystal size is small or there are numerous crystals in this region, it is difficulty to
33
34 observe a typical extinction phenomenon which the spherulitics have under this observation condition. But from the
35
36
Figure 5(b) and (c), the spherulites exhibit a typical extinction phenomenon, and the crystal boundaries are distinct
37
38
39 in the middle layer and away from the skin layer of the CIM sample. This indicates that the crystal size is large in
40
41 these areas. As shown in Figure 5(g), there is no obvious skin-core structure of the CIM sample, and this may be
42
43
44 related to the high mold temperature and large diameter of the sample. For the GAIM sample, similar with the
45
46 location near the skin layer of the CIM sample, there is also no obvious spherulitic crystal structure under the
47
48
49 observation condition, as shown in Figure 5(e)-(f). Besides, the crystal morphology of the GAIM sample shown in
50
51 Figure 5(h) is similar to that of the CIM sample shown in Figure 5 (g), and there is no obvious skin-core structure.
52
53
54
55
56
57
58 12
59
60 ACS Paragon Plus Environment
Page 13 of 31 Industrial & Engineering Chemistry Research

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34 Figure 5. Crystal morphology under the polarized optical microscopy: (a)-(c) the crystal morphology near the skin
35
36
37 layer, in the middle layer, and away from the skin layer of the CIM samples; (d)-(f) the crystal morphology near the
38
39 skin layer, in the middle layer, and away from the skin layer of the GAIM samples; (g) the crystal morphology of
40
41
the CIM samples with the magnification of 40 times; (h) the crystal morphology of GAIM samples with the
42
43
44 magnification of 40 times.
45
46 In order to accurately characterize the crystal size of the CIM and GAIM samples, Figure 6 gives SEM
47
48
49 micrographs of the samples etched by the etchant. Figure 6(a)-(c) correspond to the observed positions a, b, and c
50
51 of the CIM samples shown in Figure 5(g), and it can be seen that with the distance from the skin layer increasing,
52
53
54 the crystal size increases for the CIM sample. The main reason is that the crystals near the skin layer cannot fully
55
56 grow because of the low temperature in this area. On the contrary, the melt away from the skin layer has a high
57
58 13
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 14 of 31

1
2
3
4 temperature, and the crystals have a sufficient time for growth before the melt cooling down, so the crystal size is
5
6 relatively coarse.41 Figure 6(d)-(f) correspond to the observed positions d, e, and f of the GAIM samples shown in
7
8
Figure 5(h). For the GAIM sample, the crystal size away from the skin layer (near gas channel) is slightly larger, as
9
10
11 shown in Figure 6(f), the crystal size near the skin layer is secondly larger, as shown in Figure 6(d), and the crystal
12
13 size in the middle layer is the smallest, as shown in Figure 6(e). However, the crystal sizes in these areas have little
14
15
16 difference overall. Moreover, in the same observation positon, the crystal size of the CIM sample is much larger
17
18 than that of the GAIM sample. The CIM sample's average crystal size is 30 µm, and the GAIM sample's average
19
20
21 crystal size is 15 µm. The result indicates that the average number of nuclei generated in GAIM is four times as
22
23 many as that in CIM. It can be confirmed that the GAIM can refine the crystal size and stimulate nucleation of melt
24
25
26 crystallization.
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46 Figure 6. SEM micrographs of the samples etched by the etchant: (a)-(c) the crystal morphology near the skin layer,
47
48
49 in the middle layer, and away from the skin layer of the CIM samples; (d)-(f) the crystal morphology near the skin
50
51 layer, in the middle layer, and away from the skin layer of the GAIM samples.
52
53
54 Figure 7 shows the DSC endothermic thermograms of the CIM and GAIM samples, respectively. Among these
55
56 curves, P0, P1000, and P2000 represent the specimens that are sampled at the positions about 0 µm, 1000 µm, and
57
58 14
59
60 ACS Paragon Plus Environment
Page 15 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 2000 µm away from the skin layer (as shown in Figure 3(d)), respectively. It can be found that the melting curves in
5
6 Figure 7 are all single peak curves. For the CIM sample, the three melting curves almost have the same peak
7
8
temperature, that is, the melting temperature is essentially the same. But for the GAIM sample, there are some
9
10
11 differences among the peak temperature, and the temperature corresponding to the peak of curves P0, P1000, and
12
13 P2000 are 168.26 °C, 166.95 °C, and 164.97 °C, respectively. This demonstrates that the melting point is gradually
14
15
16 reduced from the GAIM sample's skin layer to the gas channel.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34 Figure 7. DSC endothermic thermograms of the different samples: (a) CIM samples; (b) GAIM samples.
35
36
37 In order to quantitatively compare the melting enthalpy of the two samples, based on the DSC curves shown in
38
39 Figure 7, the two samples' melting enthalpy is calculated and shown in Table 4. From the table, it can be seen that
40
41
with the distance from the skin layer increasing, the melting enthalpy increases for the CIM sample. That is, the
42
43
44 farther away from the skin layer, the greater the melt's crystallinity is. Because the melt away from the skin layer
45
46 has a relatively high temperature, and the crystals have a sufficient time for growth, the crystallinity in this area is
47
48
49 high. For the GAIM sample, the specimen (P1000) at the middle position has the lowest melting enthalpy, and the
50
51 other two specimens (P0 and P2000) almost have the same melting enthalpy. The main reason is that, compared with
52
53
54 the other two specimens, the specimen P1000 has the most nuclei, which can make the crystals grow imperfectly, and
55
56 the crystallinity is reduced. Besides, at the same sampling location, the melting enthalpy of the CIM specimens is
57
58 15
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 16 of 31

1
2
3
4 greater than that of the GAIM specimens, that is, the crystallinity of the CIM sample is larger than that of the
5
6 GAIM sample. This means that, although the GAIM can stimulate nucleation of melt crystallization, it cannot
7
8
increase the melt's crystallinity. In the next section, by combining the results of numerical simulation, this paper
9
10
11 will further explore the reason why the GAIM can stimulate nucleation of melt crystallization but not increase the
12
13 melt's crystallinity.
14
15
16 Table 4. Melting enthalpy (∆Hm (J·g-1)) of the specimens.
17
18 P0 P1000 P2000
19
20
21 CIM 103.35 103.41 109.52
22
23 GAMIM 98.18 80.83 98.35
24
25
26 4.2 Shear and temperature fields
27
28 Figure 8 shows the shapes of the GAIM samples obtained by the numerical simulation and experiment, where
29
30
31 Figure 8(a) is the simulation result and Figure 8(b) is the experiment result. The red color represents the polymer
32
33 melt, and the blue color represents the nitrogen. As shown in Figure 8(a), when the injection time is 0.5 s, the
34
35
polymer melt fills 55% of the whole sample's volume. Then the high-pressure nitrogen is injected to the polymer
36
37
38 melt through the gas inlet. After the time interval of 0.24 s, that is, when the time is 0.74 s, the high-pressure
39
40 nitrogen penetration to the melt makes a hollow cavity in the melt. Once the time reaches 0.75 s, the length of the
41
42
43 hollow cavity in the melt increases significantly. So the high-pressure nitrogen penetration to the melt is an
44
45 accelerating process. When the time is 0.751 s, the polymer melt arrives at the end of the sample, and this time is
46
47
48 the termination time of the simulation for GAIM. Figure 8(b) shows the axis section of the sample obtained in the
49
50 experiment of GAIM. Because the sample is translucent, it is difficult to distinguish the wall thickness and the
51
52
53 hollow cavity when taking picture. For this reason, the solid part of the sample's axis section was painted with the
54
55 red color, and a green paper was placed at the bottom of the sample as the background. Comparing the simulation
56
57
58 16
59
60 ACS Paragon Plus Environment
Page 17 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 result of 0.751 s in Figure 8(a) with the experiment result in Figure 8(b), it can be found the hollow cavity's shape
5
6 obtained by simulation is in good agreement with that obtained by experiment. The only difference is that the end
7
8
shape of the hollow cavity obtained by simulation is a semi-ellipsoid, but the end shape of the hollow cavity
9
10
11 obtained by experiment is a sharp cone. The reason is that the part shrinks due to the melt's cooling during the
12
13 high-pressure nitrogen's holding time in actual injection molding process, and for reducing shrinkage, the
14
15
16 high-pressure nitrogen continues to push the melt to fit the mold cavity. This phenomenon is termed as secondary
17
18 gas penetration,42,43 while the melt shrinkage and the secondary gas penetration are ignored in the simulation. In the
19
20
21 numerical simulation, the assumption of the incompressible melt was introduced. As a result, there is a little of
22
23 difference between the simulation and experiment results. In order to approach the experimental results more
24
25
26 closely, the PVT property of the material should be taken into account, that is, the compressibility of the melt can't
27
28 be ignored. In future work, we will try to establish the relevant PVT model of the material, and couple it to other
29
30
equations for further simulations. In fact, the residual wall thickness of the actual GAIM sample is 2.1 mm, and the
31
32
33 residual wall thickness obtained by simulation is 2.05 mm, so the simulation result is also in good agreement with
34
35 the experiment result for the residual wall thickness.
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 17
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 18 of 31

1
2
3
4 Figure 8. Shapes of the GAIM samples obtained by the numerical simulation and experiment: (a) simulation result
5
6 (b) experiment result.
7
8
Figure 9 gives the simulation results of the shear rate along the radial direction of the samples. The abscissa
9
10
11 axis indicates the radial distance on the cross section of the samples, and ±7.5 mm represent the upper and lower
12
13 end on the cross section of the samples, respectively. Curve 1 and curve 2 shown in Figure 9(a) are the shear rate
14
15
16 distribution on section 1 (20 mm away from the gate) and section 2 (40 mm away from the gate) of the CIM
17
18 samples, respectively. From curve 1 and curve 2, it can be seen that near the upper and lower end of the sample, the
19
20
21 melt's shear rate reaches the maximum value, and at the center of the sample, the melt's shear rate reaches the
22
23 minimum value. By comparing curve 1 and curve 2, it is found that the maximum shear rate of curve 1 is slightly
24
25
26 larger than that of curve 2. This indicates that the maximum shear rate of the melt near the gate is greater than that
27
28 of the melt away from the gate in the melt filling process for CIM. Due to the filling pressure dropping along the
29
30
melt flow direction in melt filing process, the filling speed also drops, and this makes the shear rate decrease. Curve
31
32
33 3 and curve 4 shown in Figure 9(b) are the shear rate distribution on section 3 (40 mm away from the gate) and
34
35 section 4 (80 mm away from the gate) of the GAIM samples, respectively. As shown in the curves, the melt's
36
37
38 maximum shear rate also appears near the upper and lower end of the GAIM sample, but the maximum shear rate
39
40 of the melt away from the gate is greater than that of the melt near the gate. As the nitrogen continuously fills, the
41
42
43 pressure in the hollow cavity will continue to increase, and the nitrogen will accelerate the filling of the melt. As a
44
45 result, the melt's shear rate increases accordingly with the distance from the gate increasing. By comparing curve 2
46
47
48 in Figure 9(a) with curve 3 in Figure 9(b), it can be seen the melt's shear rate of the GAIM sample is much greater
49
50 than that of the CIM sample on the same cross section. This is mainly caused by the nitrogen rapidly pushing the
51
52
melt to fill the mold cavity. In a word, the melt's maximum shear rate of GAIM is about 45 times larger than that of
53
54
55 CIM.
56
57
58 18
59
60 ACS Paragon Plus Environment
Page 19 of 31 Industrial & Engineering Chemistry Research

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 Figure 9. Simulation results of the shear rate along the radial direction of the samples: (a) CIM samples; (b) GAIM
19
20
21 samples.
22
23 Figure 10 shows the variation of the melt temperature with the cooling time at the position 1.5 mm away from
24
25
26 the skin layer of the CIM and GAIM samples. As shown in the figure, the melt temperature of GAIM drops faster
27
28 than that of CIM. This indicates that the cooling rate of GAIM is faster than that of CIM. Because the high-pressure
29
30
nitrogen penetration to the melt makes the part have a thin residual wall thickness, the heat transfer between melt
31
32
33 and mold wall is fast for GAIM. However, for CIM, the heat is transferred from the sample's center to the mold
34
35 wall, and the thickness from the CIM sample's center to the mold wall is far larger than the residual wall thickness
36
37
38 of the GAIM sample, so the heat transfer between melt and mold wall is slow.
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54 Figure 10. Variation of the melt temperature with the cooling time at the position 1.5 mm away from the skin layer
55
56 of the CIM and GAIM samples.
57
58 19
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 20 of 31

1
2
3
4 In related literatures,44-46 it has been pointed out that the shear action helps to induce nucleation. From the
5
6 simulation result in Figure 9, the high-pressure nitrogen penetration to the melt in GAIM produces a stronger shear
7
8
effect on the melt compared with CIM. Under this strong shear, a large number of nuclei are induced in melt
9
10
11 crystallization process, and the number of crystals in a GAIM is increased. In addition, the residual wall thickness
12
13 of the GAIM sample is 2.1 mm, and the shear rate in the middle layer (about 1 mm away from the skin layer of the
14
15
16 GAIM sample) almost corresponds to the curve peak shown in Figure 9 (b), that is, the melt in the middle layer of
17
18 the GAIM sample is subjected to a strong shear action. As a result, the melt in the middle layer has the largest
19
20
21 number of nuclei and the smallest crystal size, as shown in Figure 6 (e). However, the GAIM has a rapid cooling
22
23 rate which is unfavorable to the melt crystallization, so the crystallinity of the final sample is reduced.
24
25
26 It should be noticed that the obviously oriented crystals are not found in this paper. Although the strong shear
27
28 can make the polymer chain stretched or oriented along the flow direction, the molecule chain relaxation in the
29
30
cooling process will disrupt the orientation. The comprehensive effect of these two factors will determine the
31
32
33 microstructure with oriented crystals or spherulites,44,47 so the strong shear is just one of the necessary conditions to
34
35 induce the oriented crystals.
36
37
38 4.3 Foaming mechanism
39
40 Figure 11 shows the samples' fracture sections and the cell morphology of the foamed area under SEM. Figure
41
42
43 11(a) represents the samples obtained by microcellular injection molding (MIM), and Figure 11(b)-(f) represent the
44
45 samples obtained by gas-assisted microcellular injection molding (GAMIM) under the gas holding time of 5 s, 15 s,
46
47
48 25 s, 35 s and 45 s, respectively. As shown in Figure 11(a) and (b), the number of cells is relatively small, and the
49
50 cell size distribution is very uneven for the MIM sample and the GAMIM sample under the gas holding time of 5 s.
51
52
With the gas holding time increasing, when it reaches 25 s, the number of cells increases, and the cell size
53
54
55 distribution is very uniform, as shown in Figure 11(d). However, with the further increase of gas holding time, the
56
57
58 20
59
60 ACS Paragon Plus Environment
Page 21 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 number of cells reduces again, and the cell size increases, as shown in Figure 11(e) and (f). From the fracture
5
6 sections, which can macroscopically assess the overall foaming quality, it also can be seen the foaming quality
7
8
under the gas holding time of 25s is the best. Because the melt is still in a high temperature, and cells are formed
9
10
11 mainly via homogeneous nucleation in MIM and in GAMIM under the short gas holding time, the final number of
12
13 cells (cell density) is small. When cells grow in the PP melt with a high temperature and a low melt strength, cells
14
15
16 are susceptible to coalesce and rupture. As a result, the final cell size distribution is non-uniform. With the increase
17
18 of gas holding time in GAMIM, the melt temperature can be reduced, and the melt begins to crystallize. When the
19
20
21 foaming occurs, there are a lot of imperfect crystals for the heterogeneous nucleation of cells,6,48 so the final
22
23 number of cells is increased. Moreover, the melt strength can be also increased due to the decrease of the melt
24
25
26 temperature and the melt crystallization. This increased melt strength can effectively suppress the cell coarsening
27
28 and rupture. Therefore, the cell density is the largest and the cell size distribution is the most uniform at the most
29
30
appropriate gas holding time. However, with the gas holding time further increasing, the melt temperature
31
32
33 decreases greatly, and the formed crystals tend to be perfect. As a result, the cell nucleation is inhibited, and the
34
35 number of cells begins to decrease in the final sample. At the same time, the cell size will also slightly increase
36
37
38 with a definite amount of foaming agent.
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 21
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 22 of 31

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 Figure 11. Samples' fracture sections and cell morphology of the foamed area under SEM: (a) MIM samples; (b)-(f)
23
24 GAMIM samples under the gas holding time of 5 s, 15 s, 25 s, 35 s and 45 s.
25
26
27 Figure 12 gives the schematic illustration of foaming mechanisms of the two kinds of foam injection molding.
28
29 In MIM, the high temperature polymer/gas solution shown in Figure 12(a)-I is formed after mixing the polymer and
30
31
32 physical foaming agent by the screw. When the high temperature melt is injected into the mold cavity, the melt
33
34 pressure drops, and cells begin to nucleate. Since the melt crystallization hardly occurs at this time,20 the cell
35
36
37 nucleation mainly belongs to homogeneous nucleation, and the number of cells is small, as shown in Figure
38
39 12(a)-II. After cell nucleation, the melt temperature is still high and the melt strength is low, so cells are easy to
40
41
coalesce and rupture in the growing process. As the melt cools down, the melt strength increases, but this limited
42
43
44 increase of melt strength still does not inhibit the coarsening of cells. And at this moment, the crystals shown in
45
46 Figure 12(a)-III begin to appear in the melt. Subsequently, the cells and crystals continue to grow, as shown in
47
48
49 Figure 12(a)-IV. In short, for MIM, cell nucleation occurs and cells grow in the environment where the melt
50
51 temperature is high, no crystallization occurs, and the melt strength is low. This foaming environment is harmful for
52
53
54 the tailoring of cell morphology. In GAMIM, after a proper gas holding time, the nuclei begin to form in the melt,
55
56 as shown in Figure 12(b)-II. Since the nitrogen content (0.3 wt%) in the melt is very small, the nitrogen dissolved
57
58 22
59
60 ACS Paragon Plus Environment
Page 23 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 in the PP melt has a diminutive plasticization effect. Thereby, this plasticization can only counterbalance the
5
6 nitrogen's negative hydraulic pressure effect on crystallization,49,50 and the nitrogen's effect on melt crystallization
7
8
can be ignored.51 Therefore, in the presence of the nitrogen, a large number of nuclei can also be generated and the
9
10
11 crystals are refined in GAMIM. When the high-pressure nitrogen is released, the foaming gas in the melt begins to
12
13 nucleate. Here, besides the homogeneous nucleation of cells, the heterogeneous nucleation of cells also occurs
14
15
16 around these numerous crystals. This nucleation mode can be described as the crystallization-driven cell nucleation.
17
18 So the nucleation density is increased, as shown in Figure 12(b)-III. After a proper gas holding time, the melt
19
20
21 temperature can be reduced, and the melt strength can be increased. So the cell coarsening can be effectively
22
23 prevented, as shown in Figure 12(b)-IV. In a word, the foaming process occurs after the melt crystallization in
24
25
26 GAMIM, and the crystals can drive nucleation of cells. The strong shear in GAMIM generates a large number of
27
28 crystals, of which the boundaries can provide a lot of nucleation sites. In addition, these crystals can also cause a
29
30
local stress on the melt around them.20,48 These two factors are very helpful for the crystallization-driven nucleation
31
32
33 of cells, which plays an important role on the improvement of the PP's foaming behavior in GAMIM. Moreover,
34
35 the decrease of the melt temperature and the melt crystallization make the melt strength increase. As a result, the
36
37
38 cell morphology of GAMIM can be greatly improved.
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 23
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 24 of 31

1
2
3
4 Figure 12. Schematic illustration of foaming mechanisms of the two kinds of foam injection molding processes: (a)
5
6 MIM; (b) GAMIM.
7
8
5 Conclusions
9
10
11 In this paper, the crystallization behaviors of PP in conventional injection molding (CIM) and gas-assisted
12
13 injection molding (GAIM) were studied. The three-dimensional numerical model was established to simulate the
14
15
16 filling and cooling processes of CIM and GAIM. The effect of the high-pressure gas penetration on melt
17
18 crystallization was also revealed. The cell morphology of microcellular injection molding (MIM) and gas-assisted
19
20
21 microcellular injection molding (GAMIM) was investigated. The foaming mechanism of PP in GAMIM was finally
22
23 proposed. The following conclusions were drawn:
24
25
26 (1) The GAIM can stimulate nucleation of melt crystallization and refine crystals, so the average crystal size
27
28 of the GAIM sample is smaller than that of the CIM sample. The crystal sizes in the positions with different
29
30
distances away from the GAIM sample's skin layer have little difference, while the crystal size increases with the
31
32
33 increase of the distance from the skin layer of the CIM sample.
34
35 (2) In GAIM, the gas penetration to the melt brings a strong shear action, and the thin residual wall thickness
36
37
38 helps to accelerate the heat transfer between the melt and mold cavity. This strong shear can induce nucleation of
39
40 the melt, and help to reduce the crystal size. However, the fast heat transfer and the large existing nuclei make the
41
42
43 crystal grow inadequately, so the crystallinity of the GAIM sample is smaller than that of the CIM sample.
44
45 (3) For MIM, cell nucleation occurs and cell grows in the environment where the melt temperature is high,
46
47
48 no crystallization occurs, and the melt strength is low. This foaming environment is very harmful for the tailoring of
49
50 cell morphology. But for GAMIM, when the foaming process begins, a large number of crystals has formed in the
51
52
melt. The mode of crystallization-driven cell nucleation can help to greatly enhance the nucleation of cells. In
53
54
55 addition, after a proper gas holding time, the melt temperature can be reduced, and the melt strength can be
56
57
58 24
59
60 ACS Paragon Plus Environment
Page 25 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 increased. Therefore, the foaming behavior of PP in the GAMIM is significantly improved.
5
6 Author Information
7
8
Corresponding author
9
10
11 * E-mail address: [email protected] (Guoqun Zhao).
12
13 ORCID
14
15
16 Guoqun Zhao: 0000-0002-9387-801X
17
18 Notes
19
20
21 The authors declare no competing financial interest.
22
23 Acknowledgements
24
25
26 The authors are grateful to the financial support from the Climbing Program for Taishan Scholars of Shandong
27
28 Province of China (NO. 20110804) and National Natural Science Foundation of China (NO. 51405267). The work
29
30
was also financially supported by State Key Laboratory of Materials Processing and Die & Mould Technology,
31
32
33 Huazhong University of Science and Technology (No. P2018-002).
34
35 References
36
37
38 (1) Park, C. B.; Baldwin, D. F.; Suh, N. P. Effect of the pressure drop rate on cell nucleation in continuous
39
40 processing of microcellular polymers. Polym. Eng. Sci. 1995, 35, 432−440.
41
42
43 (2) Doroudiani, S.; Park, C. B.; Kortschot, M. T. Effect of the crystallinity and morphology on the microcellular
44
45 foam structure of semicrystalline polymers. Polym. Eng. Sci. 1996, 36, 2645−2662.
46
47
48 (3) Matuana, L. M.; Park, C. B.; Balatinecz, J. J. Cell morphology and property relationships of microcellular
49
50 foamed PVC/wood-fiber composites Polymer. Polym. Eng. Sci. 1998, 38, 1862−1872.
51
52
(4) Wang, L.; Hikima, Y.; Ishihara, S.; Ohshima, M. Fabrication of lightweight microcellular foams in
53
54
55 injection-molded polypropylene using the synergy of long-chain branches and crystal nucleating agents.
56
57
58 25
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 26 of 31

1
2
3
4 Polymer 2017, 128, 119−127.
5
6 (5) Suh, K. W.; Park, C. P.; Maurer, M. J.; Tusim, M. H.; Genova, R. D.; Broos, R.; Sophiea, D. P. Lightweight
7
8
Cellular Plastics Advanced Materials. Adv. Mater. 2000, 12, 1779−1789.
9
10
11 (6) Wang, L.; Ando, M.; Kubota, M.; Ishihara, S.; Hikima, Y.; Ohshima, M.; Sekiguchi, T.; Sato, A.; Yano, H.
12
13 Effects of hydrophobic-modified cellulose nanofibers (CNFs) on cell morphology and mechanical properties
14
15
16 of high void fraction polypropylene nanocomposite foams. Compos. Part A: Appl. Sci. Manuf. 2017, 98,
17
18 166−173.
19
20
21 (7) Sun, X.; Kharbas, H.; Peng, J.; Turng, L. S. A novel method of producing lightweight microcellular injection
22
23 molded parts with improved ductility and toughness. Polymer 2015, 56, 102−110.
24
25
26 (8) Yoon, J. D.; Hong, S. K.; Kim, J. H.; Cha, S. W. A mold surface treatment for improving surface finish of
27
28 injection molded microcellular parts. Cell. Polym. 2004, 23, 39−47.
29
30
(9) Spörrer, A. N. J.; Altstädt, V. Controlling morphology of injection molded structural foams by mold design and
31
32
33 processing parameters. J. Cell. Plast. 2007, 43, 313−330.
34
35 (10) Chen, S. C.; Li, H.-M.; Hwang, S. S.; Wang, H. H. Passive mold temperature control by a hybrid
36
37
38 filming-microcellular injection molding processing. Int. Commun. Heat Mass Transfer 2008, 35, 822−827.
39
40 (11) Lee, J.; Turng, L. S.; Dougherty, E.; Gorton, P. A novel method for improving the surface quality of
41
42
43 microcellular injection molded parts. Polymer 2011, 52, 1436−1446.
44
45 (12) Shaayegan, V.; Wang, G.; Park, C. B. Effect of foam processing parameters on bubble nucleation and growth
46
47
48 dynamics in high-pressure foam injection molding. Chem. Eng. Sci. 2016, 155, 27−37.
49
50 (13) Ameli, A.; Jahani, D.; Nofar, M.; Jung, P. U.; Park, C. B. Development of high void fraction polylactide
51
52
composite foams using injection molding: mechanical and thermal insulation properties. Compos. Sci. Technol.
53
54
55 2014, 90, 88−95.
56
57
58 26
59
60 ACS Paragon Plus Environment
Page 27 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 (14) Bledzki, A. K.; Kirschling, H.; Steinbichler, G.; Egger, P. Polycarbonate microfoams with a smooth surface
5
6 and higher notched impact strength. J. Cell. Plast. 2004, 40, 489−496.
7
8
(15) Li, S.; Zhao, G.; Wang, G.; Guan, Y.; Wang, X. Influence of relative low gas counter pressure on melt foaming
9
10
11 behavior and surface quality of molded parts in microcellular injection molding process. J. Cell. Plast. 2014,
12
13 50, 415−435.
14
15
16 (16) Lee, J.; Turng, L. S. Improving surface quality of microcellular injection molded parts through mold surface
17
18 temperature manipulation with thin film insulation. Polym. Eng. Sci. 2010, 50, 1281−1289.
19
20
21 (17) Chen, S. C.; Hsu, P. S.; Hwang, S. S. The effects of gas counter pressure and mold temperature variation on
22
23 the surface quality and morphology of the microcellular polystyrene foams. J. Appl. Polym. Sci. 2013, 127,
24
25
26 4769−4776.
27
28 (18) Dong, G.; Zhao, G.; Zhang, L.; Hou, J.; Li, B.; Wang, G. Morphology evolution and elimination mechanism of
29
30
bubble marks on surface of microcellular injection-molded parts with dynamic mold temperature control. Ind.
31
32
33 Eng. Chem. Res. 2018, 57, 1089−1101.
34
35 (19) Spoerrer, A. N. J.; Bangarusampath, D. S.; Altstaedt, V. The challenge of foam injection-moulding-possibilities
36
37
38 to improve surface appearance, foam morphology and mechanical properties. Cell. Polym. 2008, 27, 101−121.
39
40 (20) Ameli, A.; Nofar, M.; Jahani, D.; Rizvi, G.; Park, C. B. Development of high void fraction polylactide
41
42
43 composite foams using injection molding: crystallization and foaming behaviors. Chem. Eng. J. 2015, 262,
44
45 78−87.
46
47
48 (21) Wang, L.; Ishihara, S.; Ando, M.; Minato, A.; Hikima, Y.; Ohshima, M. Fabrication of high expansion
49
50 microcellular injection-molded polypropylene foams by adding long-chain branches. Ind. Eng. Chem. Res.
51
52
2016, 55, 11970−11982.
53
54
55 (22) Zhao, J.; Zhao, Q.; Wang, C.; Guo, B.; Park, C. B.; Wang, G. High thermal insulation and compressive
56
57
58 27
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 28 of 31

1
2
3
4 strength polypropylene foams fabricated by high-pressure foam injection molding and mold opening of
5
6 nano-fibrillar composites. Mater. Des. 2017, 131, 1−11.
7
8
(23) Zhao, J.; Zhao, Q.; Wang, L.; Wang, C.; Guo, B.; Park, C. B.; Wang, G. Development of high thermal
9
10
11 insulation and compressive strength BPP foams using mold-opening foam injection molding with in-situ
12
13 fibrillated PTFE fibers. Eur. Polym. J. 2018, 98, 1−10.
14
15
16 (24) Hwang, S. S.; Hsu, P. P. Effects of silica particle size on the structure and properties of polypropylene/silica
17
18 composites foams. J. Ind. Eng. Chem. 2013, 19, 1377−1383.
19
20
21 (25) Peng, J.; Walsh, P. J.; Sabo, R. C.; Turng, L. S.; Clemons, C. M. Water-assisted compounding of cellulose
22
23 nanocrystals into polyamide 6 for use as a nucleating agent for microcellular foaming. Polymer 2016, 84,
24
25
26 158−166.
27
28 (26) Yousefian, H.; Rodrigue, D. Morphological, physical and mechanical properties of nanocrystalline cellulose
29
30
filled Nylon 6 foams. J. Cell. Plast. 2016, 53, 253−271.
31
32
33 (27) Hou, J.; Zhao, G.; Wang, G.; Dong, G.; Xu, J. A novel gas-assisted microcellular injection molding method for
34
35 preparing lightweight foams with superior surface appearance and enhanced mechanical performance. Mater.
36
37
38 Des. 2017, 127, 115−125.
39
40 (28) Mark, L. H.; Chu, R. K.; Ameli, A.; Jahani, D.; Jung, P.; Park, C. B. Gas-assist injection molding of PLA for
41
42
43 foaming applications. SPE ANTEC 2014, paper # 1886669.
44
45 (29) Mark, L. H.; Wang, G.; Park, C. B.; Jung, U. P. TPU foam injection molding with gas-assisted processing. SPE
46
47
48 ANTEC 2015, paper # 2138090.
49
50 (30) Olley, R. H.; Bassett, D. C. An improved permanganic etchant for polyolefines. Polymer 1982, 23,
51
52
1707−1710.
53
54
55 (31) Chen, S. C.; Cheng, N. T.; Hu, S. Y. Simulations of primary and secondary gas penetration for a gas-assisted
56
57
58 28
59
60 ACS Paragon Plus Environment
Page 29 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 injection-molded thin part with gas channel. J. Appl. Polym. Sci. 1998, 67, 1553−1564.
5
6 (32) Chen, L.; Li, J.; Zhou, H.; Li, D.; He, Z.; Tang, Q. A study on gas-assisted injection molding filling simulation
7
8
based on surface model of a contained circle channel part. J. Mater. Process. Technol. 2008, 208, 90−98.
9
10
11 (33) Li, Q.; Ouyang, J.; Yang, B.; Li, X. Numerical simulation of gas-assisted injection molding using CLSVOF
12
13 method. Appl. Math. Model. 2012, 36, 2262−2274.
14
15
16 (34) Zhou, H. M.; Li, D. Q. Computer simulation of the filling process in gas-assisted injection molding based on
17
18 gas-penetration modeling. J. Appl. Polym. Sci. 2003, 90, 2377−2384.
19
20
21 (35) Polynkin, A.; Pittman, J. F. T.; Sienz, J. Gas assisted injection molding of a handle: three-dimensional
22
23 simulation and experimental verification. Polym. Eng. Sci. 2005, 45, 1049−1058.
24
25
26 (36) Li, Q.; Ouyang, J.; Yang, B.; Jiang, T. Modelling and simulation of moving interfaces in gas-assisted injection
27
28 moulding process. Appl. Math. Model. 2011, 35, 257−275.
29
30
(37) Zhang, L.; Zhao, G.; Wang, G. Formation mechanism of porous structure in plastic parts injected by
31
32
33 microcellular injection molding technology with variable mold temperature. Appl. Therm. Eng. 2017, 114,
34
35 484−497.
36
37
38 (38) Hirt, C. W.; Nichols, B. D. Volume of fluid (VOF) method for the dynamics of free boundaries. J. Comput.
39
40 Phys. 1981, 39, 201−225.
41
42
43 (39) Aguerre, H. J.; Damian, S. M.; Gimenez, J. M.; Nigro, N. M. Modeling of compressible fluid problems with
44
45 OpenFOAM using dynamic mesh technology. Mecán. Comput. 2013, 32, 955−1011.
46
47
48 (40) Zhang, L.; Zhao, G.; Dong, G.; Li, S.; Wang, G. Bubble morphological evolution and surface defect formation
49
50 mechanism in the microcellular foam injection molding process. RSC Adv. 2015, 5, 70032−70050.
51
52
(41) Zhang, J.; Shen, K.; Na, S.; Fu, Q. Vibration-induced change of crystal structure in isotactic polypropylene
53
54
55 and its improved mechanical properties. J. Polym. Sci. Part B: Polym. Phys. 2004, 42, 2385−2390.
56
57
58 29
59
60 ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 30 of 31

1
2
3
4 (42) Chen, S. C.; Cheng, N. T.; Hsu, K. S. Simulation and verification of the secondary gas penetration in a
5
6 gas-assisted-injection molded spiral tube. Int. Commun. Heat Mass Transfer 1995, 22, 319−328.
7
8
(43) Turng, L. S. Development and application of CAE technology for the gas-assisted injection-molding process.
9
10
11 Adv. Polym. Tech. 1995, 14, 1−13.
12
13 (44) Tribout, C.; Monasse, B.; Haudin, J. M. Experimental study of shear-induced crystallization of an impact
14
15
16 polypropylene copolymer. Colloid. Polym. Sci. 1996, 274, 197−208.
17
18 (45) Lee, O.; Kamal, M. R. Experimental study of post-shear crystallization of polypropylene melts. Polym. Eng.
19
20
21 Sci. 1999, 39, 236−248.
22
23 (46) Elmoumni, A.; Winter, H. H.; Waddon, A. J.; Fruitwala, H. Correlation of material and processing time scales
24
25
26 with structure development in isotactic polypropylene crystallization. Macromolecules 2003, 36, 6453−6461.
27
28 (47) Zhang, Q.; Wang, L.; Xia, X.; Feng, J.; Fu, X.; Yang, M. Tailoring the crystalline morphologies and
29
30
mechanical properties of high-density polyethylene parts by a change in the fluid flow pattern under
31
32
33 gas-assisted injection molding. J. Appl. Polym. Sci. 2014, 40349.
34
35 (48) Nofar, M. Effects of nano-/micro-sized additives and the corresponding induced crystallinity on the extrusion
36
37
38 foaming behavior of PLA using supercritical CO2. Mater. Des. 2016, 101, 24−34.
39
40 (49) Li, G.; Wang, J.; Park, C. B.; Simha, R. Measurement of gas solubility in linear/branched PP melts. J. Polym.
41
42
43 Sci. Part B: Polym. Phys. 2007, 45, 2497−2508.
44
45 (50) Naguib, H. E.; Park, C. B.; Song, S. W. Effect of supercritical gas on crystallization of linear and branched
46
47
48 polypropylene resins with foaming additives. Ind. Eng. Chem. Res. 2005, 44, 6685−6691.
49
50 (51) Nofar, M.; Tabatabaei, A.; Ameli, A.; Park, C. B. Comparison of melting and crystallization behaviors of
51
52
polylactide under high-pressure CO2, N2, and He. Polymer 2013, 54, 6471−6478.
53
54
55
56
57
58 30
59
60 ACS Paragon Plus Environment
Page 31 of 31 Industrial & Engineering Chemistry Research

1
2
3
4 Abstract Graphic
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 31
59
60 ACS Paragon Plus Environment