Anda di halaman 1dari 16

Accepted Manuscript

Plasticity in polymeric honeycombs made by photo-polymerization and nozzle based


3D-printing

H. Goodarzi Hosseinabadi, R. Bagheri, L. Avila Gray, V. Altstädt, K. Drechsler

PII: S0142-9418(17)30818-8
DOI: 10.1016/j.polymertesting.2017.08.008
Reference: POTE 5119

To appear in: Polymer Testing

Received Date: 20 June 2017


Revised Date: 0142-9418 0142-9418
Accepted Date: 5 August 2017

Please cite this article as: H. Goodarzi Hosseinabadi, R. Bagheri, L. Avila Gray, V. Altstädt, K. Drechsler,
Plasticity in polymeric honeycombs made by photo-polymerization and nozzle based 3D-printing,
Polymer Testing (2017), doi: 10.1016/j.polymertesting.2017.08.008.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT

Test Method

Plasticity in polymeric honeycombs made by photo-polymerization and nozzle based 3D-


printing

H. Goodarzi Hosseinabadi a,b , R. Bagheri a,*, L. Avila Gray c, V. Altstädt b, K. Drechsler c

PT
a
Polymeric Materials Research Group, Department of Materials Science and Engineering, Sharif

University of Technology, Azadi ave., P.O. Box 11155-9466 Tehran, Iran

RI
b
Department of Polymer Engineering, University of Bayreuth, Universitätsstraße 30, D-95440

SC
Bayreuth, Germany
c
Institute for Carbon Composites, Faculty of Mechanical Engineering, Technische Universität

U
München, Boltzmannstraße 15, Garching, D-85748 Munich, Germany
AN
* Corresponding Author. Tel: +98-21-6616-5207; Fax: +98-21-6600-5717; Email address:
M

rezabagh@sharif.edu (Reza Bagheri)


D

Abstract
TE

Study of the plastic deformation in polymeric honeycombs can pave the way for understanding the
deformation localization in more complex cellular structures, which have received progressive
attention in the past few years. This study compares the strain localization in deforming honeycombs
EP

made by two cost-effective 3D-printing technologies. Hexagonal honeycombs and their unit cell
models were 3D-printed by both PolyJetTM, using a photo-crosslinkable polymer, and fused
deposition modelling (FDM) using a thermoplastic material. The state of the art digital image
C

correlation (DIC) technique was employed as the experimental route in order to calculate the strain
field during the deformation of manufactured parts. It was found that DIC is an effective tool to study
AC

the localization in 3D-printed honeycomb struts. Moreover, in comparison with FDM, PolyJet
technology provides more homogeneous strain distributions in struts. In addition, FDM decreases the
maximal strains generated on the side layers of the honeycomb struts. Accordingly, the ligament
damage under plastic deformation can be postponed and the energy absorption capability of the
product can be improved when PolyJet technology is utilized.

Keywords: PolyJet; Fused deposition modeling; Deformation localization; Digital image


correlation; Hexagonal honeycomb.

1
ACCEPTED MANUSCRIPT

Acknowledgements

The authors would like to acknowledge support of the New Materialian Bayreuth GmbH for

providing computational equipment to post process a part of digital image analysis presented in this

PT
research.

RI
SC
1. Introduction
Nowadays, the demand for 3D-printed polymeric parts is expanding in innovation-driven applied

U
fields such as robotics, aerospace and bio-engineering [1–4]. Among 3D printing technologies,
AN
PolyJet and fused deposition modeling (FDM) have received increasing attention for the production

of small, detailed prototypes due to their remarkable potential in commercialization with affordable
M

costs [5–7]. Among different structural models, 3D-printed honeycombs have received considerable
D

attention due to their space-filling capacity with the least consumed material, their structural stability
TE

and their high energy absorption potential. The energy absorption capability of honeycombs, which is

tightly related to the deformation localization mechanism, is of interest in numerous industrial


EP

applications. As it is an important topic for the engineering community, the characterization of the

localization has received much attention. Although strain localization is commonly observed in struts
C

of foams, cellular solids and biomimetic 3D-printed designs, it is rarely calculated through
AC

experimental tests due to practical limitations. The majority of investigations with this aim have been

limited to numerical estimations based on simple material models [8–10]. Additionally, 3D-printed

polymeric parts are susceptible to showing high degrees of anisotropy and heterogeneity based on the

used 3D-printing route, the material used and orientation of the printed part [6,11]. Consequently, the

application of numerical approaches, such as finite elements for optimizing the plasticity and the
2
ACCEPTED MANUSCRIPT

analysis of localization in the design, is largely limited for 3D-printed polymeric parts where the

nonlinearity of the base material properties and the geometrical complexity of the structure add

together [12].

PT
In metallic bulk materials, the study of deformation localization by processing of digital image data

has advanced during past decades [13–18]. In recent years, Pierron et. al. [19] have investigated the

RI
contribution of localization effect on Poisson’s ratio of polymeric foams with micron size porosities

using digital image correlation (DIC). Bakhshaee et. al. [20] have successfully used DIC to quantify

SC
the elasticity of the vocal fold of excised porcine larynges during self-oscillation. Betts et. at. [21]

U
have developed a finite element model based on micro-tomography images for analyzing the energy
AN
absorption in a single strut of an open cell foam material. The current study is one of the first studies

that reveal the capabilities of DIC to characterize the localization in 3D-printed porous polymeric
M

parts with mesometric cell size. It is known that the deformation of full honeycombs can be

approximated by the analysis of deformation in their unit cells considering the number of deformed
D

unit cells and the mode of collapse in the cell walls [9]. Therefore, the focus of current research is
TE

placed on the evaluation of localization in unit cell models of honeycombs.


EP

2. Materials and methods


C

Honeycomb models were made with dimensions of 27 × 20 × 10 mm3 consisting of 6 × 7 hexagonal


AC

cells. Unit cell models were made with size 18 × 21 × 10 mm3 keeping the same symmetry elements

in honeycomb models. The thickness to length ratio in all samples was adjusted to unity in order to

facilitate strain calculation by the DIC instrument. The first series of samples were made by FDM

technology. RepetierTM software was used for slicing the sample and “concentric lines” printing

pattern was implemented. Commercial filaments (XYZprinting Co.) of acrylonitrile butadiene


3
ACCEPTED MANUSCRIPT

styrene (ABS) with 1.75 mm diameter were printed at 200 µm resolution and 20 mm/s linear speed.

Hexagons were printed layer by layer perpendicular to the z-axis of the printer.

The second series of samples were fabricated by PolyJetTM inkjet technology and VeroGrayTM-

PT
FullCure 850 commercial resins, where ultra-thin layer (40 µm resolution) of the resin was jet-

sprayed onto the bed and then the layer was cured by ultraviolet light through a photo-polymerization

RI
process. The typical quality of fabricated honeycomb and unit cell models by each technology and

the applied loading direction on the prototypes are shown in Figure 1-a,b.

SC
In order to approximate mechanical properties of the base material which was used in each

U
technology, standard compression samples were fabricated according to the size guidelines specified
AN
in ASTM D695. Typically, four replicate samples were printed in each case. Both printed materials

could tolerate large plastic deformations under compression loading and were well known for their
M

comparable ductility and strength in prototyping applications. An InstronTM universal machine was

used for compression tests with the cross head speed set to be constant at one mm/min. The surface
D

of samples was prepared by spraying white and black acrylic paint to generate a statistically random
TE

pattern according to the requirements of the DIC instrument. A typical prepared surface and the

obtained point size distribution are shown in Figure 1-c. Then, an advanced contactless Aramis-4MTM
EP

instrument (GOM mbH, Germany) was used to compute the full-field 3D strain fields from the
C

optical live images of the sample surface during compression testing according to the formulations

described in [22]. The optical sensors were calibrated for 35 × 25 mm2 measuring volume and
AC

resolution of 2352 × 1728 pixels [23]. The force-displacement data from the universal machine

were linked to the image acquisition controller. The image acquisition process was performed at one

frame per second. In order to compute the strain values, a 55% validity quote and a conventional 25

% overlap between neighboring facets were implemented.

4
ACCEPTED MANUSCRIPT

3. Results and discussions

Stress-strain curves obtained from compression tests are shown in Figure 2. Figure 2-a represents

PT
typical stress-strain curves based on performing compression tests on standard size samples. An

extreme post-yield softening response is observed in the PolyJet base material (VeroGrayTM) which

RI
shows high sensitivity of the base material to deformation localization under compression loading.

SC
However, FDM base material (ABS) shows homogenous plastic deformation analogous to a bi-linear

hardening material model. Figure 2-b compares the mechanical response of full honeycomb models

U
with their counterpart unit cell models. Comparing with FDM, it is clear that PolyJet samples have
AN
preserved a higher sensitivity to post-yield localization in their overall mechanical response, whether

in honeycomb or in unit cell models. Although unit cell models are highly successful to represent the
M

elasticity of full honeycombs, they do not directly reflect the post-yield behavior of full honeycomb

samples. This observation is mainly due to the fact that different collapse modes in different cells of a
D

full honeycomb model are activated by initiation of the post-yield deformation. Therefore, the
TE

number of collapsed cells and the mode of collapse in each cell should be distinguished for

estimating a full honeycomb stress strain behavior using unit cell models. For this aim, a precise
EP

study of localization in unit cell models is a requisite for understanding the behavior in full
C

honeycomb models. The current research has focused on studying the localization in those unit cells
AC

that follow freely compressive (FC) mode of collapse under deformation [9]. Figure 3, shows the

results of strain calculations by DIC on unit cell models.

At first glance, examination of the strain maps at different stages of deformation reveals a similar

trend for samples made by both technologies: formation of high strain shear bands at approximately

45o to the testing direction followed by localization at cell struts. Note that, in these structures, the

5
ACCEPTED MANUSCRIPT

plastic deformation is concentrated in the middle of cell struts rather than cell corners. Indeed,

formation of plastic joints inside cell struts instead of cell corners should be considered as a

characteristic of plastic deformation in high density cellular structures. In contrast, in low density

honeycombs (ρr<15%) previous numerical [9,24] and experimental [25,26] studies have reported the

PT
localization only at cell corners.

RI
Aside from general similarities, the contours obtained from PolyJet and FDM samples in Figure 3

also show significant differences. The contour lines computed by DIC are more homogenous and

SC
smooth in PolyJet samples compared with FDM samples. This observation can be attributed to the

U
higher typical resolution of 3D-printing by photo-polymerization using PolyJet technology- 40 µm, in
AN
comparison with that of the nozzle-based 3D-printing technology using FDM, 200 µm. In FDM

samples, the weld lines between 3D-printed layers (pale lines in schematic Figure 2-c) impose a
M

specific orientation to the microstructure which acts as obstacles to deformation propagation in the

direction of maximum shear stress –i.e. 45o to the compression direction. As a result, when FDM
D

technology is used, the lower resolution of printing and the existence of coarse weld lines between
TE

the printed layers have boosted the heterogeneity of computed strain contours by DIC.
EP

At the elastic region (Figure 3-a, d), the strain levels are almost evenly distributed throughout the

strut region in both printed systems. At a higher strain in the post - yield region (Figure 3-b, e), the
C

strain localization in the struts is remarkably different in the two systems. In PolyJet samples, the
AC

majority of the strut volume is exposed to a large major strain. In this case, the larger deformation

region facilitates the accommodation of the strut geometry with the deformed condition. As a result,

the maximum principal strain generated in the strut is limited to below 25 % (Figure 3-b). In the

FDM system, however, the strain is accumulated in a smaller volume of the strut region (Figure 3-e).

As a result of the deformation localization in a smaller region, the maximum principal strain value

6
ACCEPTED MANUSCRIPT

has increased to 28.5%. The heterogeneity of the microstructure and the presence of coarse weld lines

can be responsible for the higher deformation localization in 3D printed samples by FDM. At larger

deformations up to 11% global strain, the same mechanisms govern (Figure 3-c, f). Indeed, in the

PolyJet system, 50% of the strut volume has experienced utmost 20-40% strain, while in the FDM

PT
system 30% of the strut volume has experienced around 30-50% strain. As a result, in the PolyJet

RI
system, lower local maximal strains in a larger volume fraction of honeycomb struts are generated.

However, in the FDM system, higher local maximal strains in a smaller volume fraction of the cell

SC
struts are generated.

U
Since failure in a deforming body initiates in the regions of the highest strain, PolyJet samples are
AN
somehow insured against failure and can absorb more energy under deformation before failure.

Owing to the high resolution of printing and employing a flexible base material which shows post-
M

yield softening sensitivity, the 3D-printed honeycombs by PolyJet can tolerate larger global strains

by maintaining more homogenous load distribution and having flexibility of the ligament geometry
D

under large deformations and rotations.


TE

Figure 4 provides more detailed analysis of the generated strain field by comparing variations of the
EP

local principal strain along the strut length in the central (Center path line) and the side layers (Out-1

and Out-2 path lines) of the strut. The orientation of these path lines is schematically shown in Figure
C

2-c. As shown, in the PolyJet system, the maximum local strain is always generated at the central
AC

layer of struts and it is significantly lower in the lateral layers. In other words, the localization of

deformation in the central layer of cell struts has secured the outer layers from experiencing higher

strains. More uniform distribution of the strain field due to the higher resolution of printing together

with the imposed flexibility due to the post yield softening behavior of the base material in the

PolyJet system are responsible for this observation. On the contrary, in the FDM system the maximal

7
ACCEPTED MANUSCRIPT

local strain is observed in the lateral layers of struts. In addition, the magnitude of the maximum local

strain is higher comparing with PolyJet samples. Therefore, the risk of tearing and failure initiation is

increased in the nozzle-based 3D-printing FDM system.

PT
4. Concluding remarks

RI
In summary, the deformation localization was investigated by DIC technique in struts of honeycomb

SC
unit cell models fabricated by the nozzle-based fused deposition modeling as well as photo-

polymerization based PolyJet technology. It was shown that DIC is an effective tool to study the

U
strain localization in 3D-printed layers which form honeycomb struts. The homogeneity of strain
AN
distribution was improved using the photo-polymerization technology, which used higher resolution
M

for 3D-printing and exploited a flexible base material susceptible to post yield strain softening. In

addition, the maximum local strains on the outer layers of struts were remarkably decreased, which
D

prevented the printed ligaments failing when the photo-polymerization based PolyJet technology was
TE

used for rapid prototyping. On the contrary, larger local strains at a smaller volume fraction of the

strut region were generated when the nozzle-based FDM technology was used, mainly due to the
EP

lower flexibility of the oriented material and the presence of coarse weld lines in microstructure.

Moreover, higher local strains at a smaller volume fraction of the strut were generated which could
C

accelerate the failure of the strut ligaments under plastic deformation in the nozzle-based printing
AC

technology.

References

[1] R. Pandey, Photopolymers in 3D printing applications, Arcada- Finland, 2014.

[2] S. Rahmati, F. Abbaszadeh, F. Farahmand, An improved methodology for design of custom-


8
ACCEPTED MANUSCRIPT

made hip prostheses to be fabricated using additive manufacturing technologies, Rapid


Prototyp. J. 18 (2012) 389–400. doi:10.1108/13552541211250382.

[3] M.T. Arafat, I. Gibson, X. Li, State of the art and future direction of additive manufactured
scaffolds-based bone tissue engineering, Rapid Prototyp. J. 20 (2013) 13–26. doi:10.1108/RPJ-

PT
03-2012-0023.

[4] S.P. Soe, P. Martin, M. Jones, M. Robinson, P. Theobald, Feasibility of optimising bicycle

RI
helmet design safety through the use of additive manufactured TPE cellular structures, Int. J.
Adv. Manuf. Technol. 79 (2015) 1975–1982. doi:10.1007/s00170-015-6972-y.

SC
[5] C.W. Ziemian, M.M. Sharma, S.N. Ziemian, Anisotropic Mechanical Properties of ABS Parts
Fabricated by Fused Deposition Modelling, in: M. Gokcek (Ed.), Mech. Eng., InTech,
Shangha, 2012: pp. 159–180. doi:10.5772/34233.

[6]
U
J. Lee, A. Huang, Fatigue analysis of FDM materials, Rapid Prototyp. J. 19 (2013) 291–299.
AN
doi:10.1108/13552541311323290.

[7] S. Amirkhani, Prediction of mechanical behavior of bone scafolds produced by rapid


M

prototyping based on regular cells architecture, Sharif University of Technology, PhD


Dissertation, 2013.
D

[8] M. Alkhader, M. Vural, Mechanical response of cellular solids: Role of cellular topology and
TE

microstructural irregularity, Int. J. Eng. Sci. 46 (2008) 1035–1051.


doi:10.1016/j.ijengsci.2008.03.012.
EP

[9] H. Nakamoto, T. Adachi, M. Higuchi, Approximate analysis of progressive deformation in


honeycomb structures subjected to in-plane loading, Arch. Appl. Mech. 83 (2012) 379–396.
C

doi:10.1007/s00419-012-0685-6.

[10] J.M. Albuquerque, M.F. Vaz, M.A. Fortes, Effect of missing walls on the compression
AC

behaviour of honeycombs, Scr. Mater. 41 (1999) 167–174. doi:10.1016/S1359-


6462(99)00117-7.

[11] J.F. Rodríguez, J.P. Thomas, J.E. Renaud, Mechanical behavior of acrylonitrile butadiene
styrene fused deposition materials modeling, Rapid Prototyp. J. 9 (2003) 219–230.
doi:10.1108/13552540310489604.

[12] D. Asprone, F. Auricchio, C. Menna, S. Morganti, A. Prota, A. Reali, Statistical finite element
9
ACCEPTED MANUSCRIPT

analysis of the buckling behavior of honeycomb structures, Compos. Struct. 105 (2013) 240–
255. doi:10.1016/j.compstruct.2013.05.014.

[13] R.B. Joshi, A.E. Bayoumi, H.M. Zbib, Evaluation of macroscopic shear banding using a
digital image processing technique, Scr. Metall. Mater. 24 (1990) 1747–1752.

PT
[14] J. Kang, Y. Ososkov, J.D. Embury, D.S. Wilkinson, Digital image correlation studies for
microscopic strain distribution and damage in dual phase steels, Scr. Mater. 56 (2007) 999–

RI
1002. doi:10.1016/j.scriptamat.2007.01.031.

[15] A. Kurtovic, T. Niendorf, T. Hausoel, H.W. Hoeppel, M. Goeken, H.J. Maier, Surface strain

SC
evolution of ultrafine-grained aluminum alloy laminates under tension-Microscale plastic
instabilities and the Portevin-Le Chatelier effect, Scr. Mater. 68 (2013) 809–812.
doi:10.1016/j.scriptamat.2013.01.036.

[16]
U
C.C. Tasan, J.P.M. Hoefnagels, M.G.D. Geers, Microstructural banding effects clarified
AN
through micrographic digital image correlation, Scr. Mater. 62 (2010) 835–838.
doi:10.1016/j.scriptamat.2010.02.014.
M

[17] Q. Han, Y. Kang, P.D. Hodgson, N. Stanford, Quantitative measurement of strain partitioning
and slip systems in a dual-phase steel, Scr. Mater. 69 (2013) 13–16.
D

doi:10.1016/j.scriptamat.2013.03.021.
TE

[18] G. Li, F. Xu, G. Sun, Q. Li, Identification of mechanical properties of the weld line by
combining 3D digital image correlation with inverse modeling procedure, Int. J. Adv. Manuf.
EP

Technol. 74 (2014) 893–905. doi:10.1007/s00170-014-6034-x.

[19] F. Pierron, Identification of Poisson’s ratios of standard and auxetic low density polymeric
C

foams from full-field measurements, J. Strain Anal. Eng. Des. 45 (2010) 233–253.
doi:10.1243/03093247JSA613.
AC

[20] H. Bakhshaee, J. Young, J.C.W. Yang, L. Mongeau, A.K. Miri, Determination of strain field
on the superior surface of excised larynx vocal folds using DIC, J. Voice. 27 (2013) 659–667.
doi:10.1016/j.jvoice.2013.05.009.

[21] C. Betts, D. Balint, J. Lee, J. Lin, P. Lee, In situ microtensile testing and X-ray
microtomography-based finite element modelling of open-cell metal foam struts and sandwich
panels, J. Strain Anal. Eng. Des. 49 (2014) 592–606. doi:10.1177/0309324714539307.

10
ACCEPTED MANUSCRIPT

[22] Aramis, User Manual - Software, v6-3rd ed., GOM mbH, Braunschweig, Germany, 2011.

[23] Aramis, User Manual - Hardware, v6-3rd ed., GOM mbH, Braunschweig, Germany, 2011.

[24] D.-H. Chen, K. Ushijima, Deformation of Honeycomb with Finite Boundary Subjected to
Uniaxial Compression, Metals (Basel). 3 (2013) 343–360. doi:10.3390/met3040343.

PT
[25] S.D. Papka, S. Kyriakides, In-plane compressive response and crushing of honeycomb, J.
Mech. Phys. Solids. 42 (1994) 1499–1532. doi:10.1016/0022-5096(94)90085-X.

RI
[26] M.K. Khan, T. Baig, S. Mirza, Experimental investigation of in-plane and out-of-plane
crushing of aluminum honeycomb, Mater. Sci. Eng. A. 539 (2012) 135–142.

SC
doi:10.1016/j.msea.2012.01.070.

U
AN
M
D
TE
C EP
AC

11
ACCEPTED MANUSCRIPT

Figure Captions

Fig. 1- sample preparation: a) Quality of fabricated honeycomb models by PolyJet, and FDM

PT
technologies, b) Quality of fabricated unit cell models by PolyJet, and FDM (insert: 15 mm), c)
Quality of surface preparation for DIC analysis and its point size distribution analysis

RI
Fig. 2- Compression tests: a) stress-strain response of standard compression samples made by
PolyJet/ FDM, b) stress-strain response of full honeycomb & unit cell models, c) schematic

SC
representation of a typical honeycomb strut under deformation (F and M denote reaction force and
moment vector applied, respectively; pale lines represent the printing rout for FDM samples using
“concentric-lines” option in Repetiter software; “Out-1” and “Out-2” and “Center” demonstrate paths

U
where strain variation profile is illustrated)
AN
Fig. 3- Full-field strain contours of unit cells: contour of the maximum local principal strain
calculated by DIC for samples made by a-c) PolyJet, d-f) Fused deposition modeling, under different
M

global strains in accordance with Figure 2-b


D

Fig. 4- Variations of the maximum principal strain along the central and lateral printed layers
of unit cell struts: for a-c) PolyJet, d-f) FDM samples
TE
C EP
AC

1
ACCEPTED MANUSCRIPT

PolyJet Fused deposition modeling


a) Honeycomb model

PT
RI
SC
b) Unit cell model

U
AN
M
D

80
c) Surface preparation

60
TE

Frequency

40
20
0
EP

0 2000 4000
Point size (μm2)
C
AC

Fig. 1-

2
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Fig. 2-
M
D
TE
EP
C
AC

Fig. 3-

3
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Fig. 4-
M
D
TE
EP
C
AC

Anda mungkin juga menyukai