Multi-scale additive manufacturing: an investigation on building objects with macro-, micro- and nano-scales features

Abstract

Additive manufacturing (AM) processes can fabricate objects with complex structures and different materials. By gradually accumulating material layer-by-layer, three-dimensional objects can be built from computer-aided design (CAD) models. The dimension of the processed CAD model can range from macro-scale to nano-scale. Currently different scales of features are accomplished in different platforms and are not available to achieve on a single object. However, for various applications such as optical and fluidic purposes, it is desired to fabricate macro-scale objects with micro- and nano-scale features. The aim of this research work is to investigate how to integrate the Two-Photon Polymerization (TPP) and Mask Image-Projection-based Stereolithography (MIP-SL) processes to address the multi-scale fabrication challenge, and to demonstrate some unique applications of such multi-scale objects. We developed macro-scale and micro-scale MIP-SL systems and applied a machine vision approach to integration the two processes. Process parameters such as cure depth, cure width and exposure time are optimized to improve the fabrication quality. Special fixtures are also added in the macro-scale component design in order to integrate with the nano-scale TPP process. Three test cases have been designed and tested. The experimental results validate the feasibility and efficiency of the proposed multi-scale fabrication approach. The multi-scale fabrication process would enable novel fluidic and optical applications.

Full text

1
11th International Conference on Micro Manufacturing
Orange County, California, USA, March 2016
Paper# 96
Multi-Scale Additive Manufacturing: An Investigation on Building Objects with Macro-,
Micro- and Nano-scales Features and Its Application
Xiangjia Li1 Tommaso Baldacchini 2 Xuan Song1 Yong Chen1
1 Department of Industrial and Systems Engineering, University of Southern California, Los Angeles, CA, USA
2 Technology and Applications Center, Newport Corporation, Irvine , Los Angeles, CA, USA
Abstract
Additive manufacturing (AM) processes can fabricate objects with complex structures and different materials.
By gradually accumulating material layer-by-layer, three-dimensional objects can be built from computer-aided
design (CAD) models. The dimension of the processed CAD model can range from macro-scale to nano-scale.
Currently different scales of features are accomplished in different platforms and are not available to achieve on
a single object. However, for various applications such as optical and fluidic purposes, it is desired to fabricate
macro-scale objects with micro- and nano-scale features. The aim of this research work is to investigate how to
integrate the Two-Photon Polymerization (TPP) and Mask Image-Projection-based Stereolithography (MIP-SL)
processes to address the multi-scale fabrication challenge, and to demonstrate some unique applications of such
multi-scale objects. We developed macro-scale and micro-scale MIP-SL systems and applied a machine vision
approach to integration the two processes. Process parameters such as cure depth, cure width and exposure time
are optimized to improve the fabrication quality. Special fixtures are also added in the macro-scale component
design in order to integrate with the nano-scale TPP process. Three test cases have been designed and tested. The
experimental results validate the feasibility and efficiency of the proposed multi-scale fabrication approach. The
multi-scale fabrication process would enable novel fluidic and optical applications.
Keywords: Multi-scale Fabrication; Additive Manufacturing; Stereolithography; Two-Photon Polymerization
1. Introduction
Additive manufacturing (AM) can be applied to
fabricate objects with free-form surfaces and
complex inner structures with high resolution. Due to
its unique capabilities, AM technology has been
widely used in a wide variety of applications.
Building macro-scale objects with micro- and nano-
scale features is a challenging but promising topic. It
would enable even wider applications such as
sensors, biomedical, structures, scaffolds, etc. Some
examples are shown in Figure 1.
Researchers have demonstrated the ability of
Stereolithography Apparatus (SLA) to print multi-
scale objects. Laser-based SLA is one candidate to
fabricate these complex multi-scales objects. It
achieves the high resolution for micro-scale by
controlling the beam size of laser injection. However,
the fabrication process is slow and costs hours to
fabricate a part [1]. Projection based SLA can
accomplish high-resolution fabrication within shorter
time since it can build one full layer with a single
exposure. Researchers investigated multi-scale
fabrication using the projection-based SLA. Fig.1 (b)
shows a needle with 100µm features; however, the
size of the whole object is subject to the fabrication
capability of a projection image [2]. Moreover,
Electrohydrodynamic (EHD) printing, which
integrate multi-materials for advanced micro- and
nano-scales devices, also sacrifices the efficiency to
satisfy the nano-scale fabrication requirements [3].
Conventional additive manufacturing processes
fabricate three-dimensional (3D) models with a
uniform resolution during the building process.
However, it is desired to fabricate multi-scale objects
with widely varying resolutions. Based on our
previous studies, the Mask Image-Projection-based

2
Stereolithography (MIP-SL) process is highly
efficient in building objects with features that are in
macro- and micro-scales; at the same time, the Two-
Photon Polymerization (TPP) process is an efficient
way of fabricating sub-microns and nano-scale
features with high resolution and accuracy. In this
research, we investigate a multi-scale fabrication
method by integrating the MIP-SL and TPP
processes and demonstrate some of its applications.
Fig. 1. An illustration of the multi-scale objects (a)
Bio-scaffold (b) Needle (c) Macro-sensor (d)
Optics (e) self-fold robot (f) Fluid channel
2. Macro- and Micro-scales Mask Image
Projection based Stereolithography
For MIP based Stereolithography, the light
illuminates the Digital Micro-mirror Device (DMD)
and is reflected by the DMD chip, where the light
intensity can be changed based on the angle of mirror
in the DMD. The focusing image irradiates at the
surface of the resin and the resin will be cured with
suitable exposure. After curing one layer, the z stage
will move to next layer position and refresh resin
surface and irradiates the next focusing image (refer
to Fig.2). Since the DMD has millions of mirrors and
one mirror represents one pixel in the focusing image,
the projected mask image can achieve high-
resolution fabrication. Using different optical image
designs, the mask images can have different
projection sizes, ranging from several millimeters to
several hundreds of millimeters. According to the
dimension of projection image size, the MIP based
Stereolithography can be divided into 3 fields:
macro-MIP process, micro-MIP process and multi-
scale MIP process, [4, 5].
Fig. 2. The principle of the MIP -SL
2.1. Set-up of the bottom-up based MIP-SL
The bottom-up layout is applied to our MIP-SL
machine. This bottom-up movement consumes
significantly less resin, since the whole fabricated
solid is not immersed inside the resin (shown in
Fig.3). Furthermore, the resin surface settling time
and particularly re-coater blade is not required in this
movement frame. Because each new layer is formed
at the bottom of the resin, in comparison bottom-up
movement, the layers are completely flat and it is
easier to control the thickness of each layer, which is
important for the fabrication precise of
microstructures in z direction. The Z stage of this set
up can reach 1 um layer thickness movement. The
projection image size of our macro- and micro-scale
MIP-SL systems are designed as 60×45mm and
3.6×2.75mm, respectively. The related resolutions of
the projection image are 0.06mm and 3.5µm.
Fig. 3. Bottom-up projection based MIP-SL
To improve the accuracy and resolution of the
MIP-SL systems, two procedures of calibrations
were applied. First, we conducted the physical
calibration to adjust the light intensity distribution.
Based on the result of the physical calibration, the
mathematical model of curing properties of the
material were generated; then the energy calibration
process was performed to further control the
accuracy and resolution of the MIP-SL systems.
2.2. Physical calibration of bottom-up based MIP-SL
The light distribution of the focusing image is
not uniformed distributed. To illustrate the non-
uniformed light distribution, the test image shown in
Fig.4a was projected and one test part was fabricated
by this projection image. Because of non-uniformed
light intensity distribution, some portions of the
fabricated model were over cured while other portion
with weaker light intensity cannot be fabricated
using same exposure time. Thus non-uniformed light
intensity has serious effect on the accuracy and
resolution of the MIP based Micro-Stereolithography.
In the physical calibration period, a set-up to capture
light intensity distribution in real time was conducted
and a gray scale value adjustment approach also was
developed to keep the light intensity uniform.
Platform
Z stage
Tank
Mirror
Lens
Fixture
Resin
Computer
(a)
(b)
(c)
(d)
(f)
(e)

3
Since the accumulation effect of light intensity
of pixels, it is hard to calculate the light intensity of
all the area in the same time. Then the area was
divided into several blocks and each block has a
square area of 25x25 pixels, where the gray scale
level was set at 255. The focusing image captured by
the CMOS (refer to Fig.4 c) contains three
information components-Red, Green, and Blue
values. To calculate the light intensity from these
three colors, we firstly converted this color image
into single gray scale G (i,j) by using weighted
average method (see Eq. 1). The light intensity of
each pixel was represented by calculated gray scale
level of the focusing image [6]. After identifying
effective block area, the light intensity L (i,j) of each
effective block was calculated. The original light
intensity distribution of the projection area is seen in
Fig.4 d.
𝐺𝑖,𝑗=0.30𝑅𝑖,𝑗+0.59𝐺𝑖,𝑗+0.11𝐵(𝑖,𝑗) [1]
𝐿𝑖,𝑗=
!!,!
!!!"×!!!!!!!"×!!!!
!!!"!;!!!"!
!"×!"
! [2]
Fig. 4. The light distribution of micro-MIP SL (a)
Test image with gray scale 255 (b) The focusing
image captured by CMOS (c) The fabricates
result by test image with exposure time 800ms (d)
The light intensity distribution model of focusing
image
After getting the light intensity distribution, the
calibration image was projected with different gray
scale values and the monitor setup was used to
capture the focusing images. For each image, the
gray scale of each pixel was calculated, which
respectively represents the light intensity, by using
image analysis method. Furthermore, for each block,
the mathematical model was fitted to figure out the
relationship between the respective light intensity
variations and different gray scale levels (refer to Fig.
5a).
Based on the light intensity variation
mathematical models of each block, the gray scale
level H (i,j) of each block was adjusted to reach the
same light intensity value Lmin. Finally, the light
intensity distribution of the whole projection image
area can reach the same level Lmin after several
times computations. Using the adjusted gray scale
level H (i,j) of each block derived from above
adjustments, the bilinear interpolation method can be
used to fit the gray scale distribution database for the
whole projection area (refer to Fig. 5.c-d).
Fig. 5. The gray scale database for uniform light
intensity distribution for MIP SL process (a) The
fitting model of the light intensity variation with
different gray scale levels in one block (b) The
focusing image captured by CMOS after first
time adjustment (c) The gray scale level database
after physical calibration (d) The gray scale level
distribution of the whole projection image
2.3. Energy calibration of bottom-up based MIP-SL
After the calibration process, the light intensity
distribution is close to uniform. For each MIP-SL
system, according to the material property and the
input light source of the setup in the energy
calibration process, a series of experiments were
conducted to establish the relationship between the
energy, cure depth using design of experiment (DOE)
method. The energy distribution of each pixel in the
mask projection image follows Gaussian function
(see Eq. 3) [6]. The energy of each position is
accumulative and the irradiation energy depends on
the light intensity and exposure time.
𝐺𝑥,𝑦=𝐴𝑒𝑥𝑝{−!
![(!!!!)!
!!
!+(!!!!)!
!!
!]} [3]
The thickness of each layer in MIP based
Stereolithography depends on the cure depth.
Especially for the fabrication process in micro-MIP
system, the thickness of one layer is only 50-10um,
the cure depth changes significantly with slight
modification of the exposure time. For the model
without overhanging feature, the influence of the
cure depth can be eliminated in the bottom-up
(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)

4
movement in the sense that the space between the
container and platform is fixed. However, for the
overhanging features, over-cured features are
generated if the exposure time exceeds the threshold
value δ (see Fig.6a). Based on the overhanging
model fabrication result, therefore a series of
experiments were designed to find the relationship
between the cure depths, gray scale level and light
absorber (see Fig.6b-c). With analysis the
experiment result of the cure depth, the adjustment of
the cure depth can be applied to reach the
requirement of fabrication.
Fig. 6. Curing character with different gray scale
value and light absorber. (a) Principle of gray
scale and Cd (b) Fabricated parts: Cd vs Gray
scale (c) Sample with 0.05% light absorber at 255
gray scale
3. Two-Photon Polymerization for Fabrication of
Nano-structures
Two-photon polymerization (TPP) is unique for
the fabrication of three-dimensional Microstructures.
It is capable of producing geometries with no
topological constrains and with resolution smaller
than 100 nm. Because of its versatility, TPP has been
employed to create devices previously impossible to
manufacture with conventional Micro fabrication
procedures.
There are two key reasons that render TPP such
an attractive solution for the fabrication of
microstructures. The first reason is that TPP is
intrinsically a three-dimensional writing technique; it
does not require a layer-by-layer approach to create
complex objects. The second reason is that TPP can
create microstructures with sub-micron feature sizes
in a relatively straightforward manner. These
characteristics originate from the nonlinear optical
nature of light absorption in TPP and from the
chemistry of the polymerization that follows.
In a typical experiment (see Fig.7 a), NIR
emission is used to excite a photosensitive material
(resin) that upon light absorption undergoes a phase
change from liquid to solid through a polymerization
process. Since the resin is transparent in the NIR
region of the spectrum, high numerical aperture (NA)
lenses and ultra-short pulsed lasers are employed to
increase the probability of a multiphoton absorption
event to occur. That is two or more photons are
simultaneously absorbed by specialized molecules in
the resin (photoinitiators) to create the active species
that start polymerization. Under these conditions
multi-photon absorption occurs only in the region
where light intensity is the highest, thus confining
polymerization within the volume of the focused
laser beam (voxel).
Three-dimensional microstructures are created
by precisely overlapping voxels through the scanning
of either the laser beam or the sample around
predetermined geometries. Successively, the
microstructures are revealed by washing away the
unsolidified part of the resin using an organic solvent.
The exposure curves of resins employed in TPP are
quite nonlinear, thus creating a light intensity
threshold below which polymerization does not
occur [7]. It is the presence of this intensity threshold
that permits the formation of voxels with dimensions
that are considerably smaller than the wavelength of
light used for fabrication. For example, by adjusting
the intensity of the excitation laser barely above the
minimum intensity needed to start and sustain TPP,
voxels 100 nm wide can be obtained with a typical
exposure wavelength of 800 nm.
Fig. 7. Experimental setup and test result in TPP.
(a) set-up design (b) mecro-model made by TPP (c)
submicron model made by TPP
TPP process is an excellent method for
fabricating microstructures with submicron finesse.
However, It is not ideal for creating macro- and
meso-scale structures using TPP process. There are
several challenges in making large structures by TPP.
(b)
(c)
(a)
(a)
(b)
(c)

5
Firstly, since the width of the focused laser beam
(voxel) is limited in micro level, it is time-consuming
to fabricate macro- and meso-scale structures by TPP
process, which the beam should scan the entire
portion on each layer of the model. Furthermore,
during the fabrication, the flatness of the object is
critical factor in TPP process, because laser beam
should be focused at the same horizontal position.
While macro- and meso-scale fabrication increase
the adhesion difficulty for the perfect flatness and
material cracking brings out for specific material
used. Fig7.b-c showed an example where a hybrid
organic-inorganic material is used to create different
size structure by TPP process and it costs hours to
fabricate the whole part with whole dimension of
2mm and the material cracked during the fabrication.
To take advantages of high precision of TPP process
and expend future applications, it is necessary to
integrate TPP microstructures fabrication method
with other fast and high-resolution macro- and meso-
scale processes.
4. Integration of MIP-SL and TPP process
As mentioned above, the MIP-SL process is
highly efficient in building objects with features that
are in macro- and micro-scales, while sometimes the
micro-features are less in sharpness than the one
made by TPP process due to the light accumulation
effect. The efficient way is to integrate TPP and
MIP-SL processes for the purpose of achieving
fabrication of high resolution, accurate macro-
features together with sub-microns and nano-scale
features. In following chapters, we investigated the
feasibility of an integrated method of MIP-SL and
TPP processes.
4.1. Model design for integration
For the part with multi-scale features, we firstly
fabricate the macro object with micro features using
MIP-SL process and then add the submicron features
on the model by TPP process. The TPP process
requires the light go through from the bottom to
locate the fabrication position, the model, therefore,
should be fabricated on the top of transparent glass
and the potion, where submicron features to be added
by TPP process, should always be kept transparent.
The spacer with 200um thickness, which used for the
TPP process to determine the distance between the
fabrication position and focus point of laser beam,
can be fabricated during the fabrication of micro
features by MIP-SL process. The whole model for
our integrated process is designed with the features
shown in Fig.8.
When fabricate the micro features on the same
layer of the macro model using the micro MIP-SL
process, the attachment between each part can be a
critical factor. The micro model is susceptible to
damage and deformation due to self-gravity and lack
of attachment support. After the macro model was
built firstly, a certain amount of thin supporting
layers were coated successively (see Fig. 8.b) to
provide sufficient area for micro features to attach
with. Then we build the micro model on the surface
of transparent resin by micro-MIP-SL process
(shown in Fig.8.c), followed by the submicron
structures built by TPP method on the surface of
micro features (see Fig.8.d).
Fig. 8. Model design for integration of TPP
process.
4.2. Visual system for integration
To combine macro- and micro-scale MIP-SL
processes, it is necessary to determine the relative
position between the fabrication area of the set-up
and the location of processing object. To address this
issue, an optical system was designed to enable the
observation of the projection image and the macro
objects for position calibration. This system contains
a beam splitter, a convex lens and a CMOS camera
(see Fig. 9). The light of projection image, which
illuminates the macro object, goes through the
objective lens and is reflected by the beam-splitter.
To see what the macro object and focus image looks
like, a CMOS camera was set up behind a convex
lens, so the reflected light goes into the lens and the
final calibration image can be captured. This
calibration image is displayed on a PC’s monitor
with 50 times magnification. The relative position of
the focusing image and macro-scale object can be
determined through this calibration image by
following process.
Here we choose the 1024 X 768 center portion
as our building range. There are two movements that
were used to align these two portions. The movement
of X stage and Y stage can take the tank and the
macro-model together in X, Y direction (see Fig.10).
The target of the alignment of the two MIP-SL
processes is getting the relative position for macro
model made by the macro MIP-SL and projection
area in micro MIP-SL. After the homing of X and Y
stages, the position of the two objects are showed in
Fig. 10. P0~P3 are the four corners of the micro
(a)
(b)
(c)
(d)
(e)

6
projector image and M0~M3 are the four corners for
the processing area of macro mode where the micro-
feature locates. For each corner point there were two
related coordinates. One is the pixel position in the
projection image, defined as CI, and the other is the
actual position of the macro object on the platform,
defined as CM. As we find out the relative position
between the M0 point and the P0~P3 points, we can
address all the relationship between M0~M3 and
P0~P3. By adjusting the X and Y stage of the micro-
scale MIP-SL, the micro-scale projection image can
be located at relevant processing position of macro
model.
Fig. 9. The visual system in micro MIP-SL
machine for the integration of macro and micro
MIP-SL.
Fig. 10. Principle diagram of the relevant position
between the macro model and projection image in
micro MIP-SL process.
The alignment process has such several steps:
Step 1: Put a transparent check board on the top
surface of the tank, which will be used for marker the
positions of the focusing image and macro object.
Step 2: we first project a 1024 X 768 rectangular
white image by the micro- MIP-SL. And then marker
the four corners P0~P3 and get the relevant position
of the projection area in the micro MIP-SL system.
Step 3: Mount the macro model built by the
macro MIP-SL and move the X and y stages step by
step to make the point M0 closer to the projection
image_1’s corner P0. During the movement of X and
Y stages, record the steps for each axis. Because of
the size of the macro-model is big enough for us to
make markers at the corners M0~M4.
Step 4: When the two corners are close enough,
and try to capture the two corners both in the camera
view. Then move the X and Y by smaller steps and
observe from the calibration visual system at same
time. When they are merged together, we get the
relative position between the two corners from the
captures image。
Step 5: Use the same process for the other three
corners of the macro model, M1~M3. We can get all
the relationship between the projection image
position and processing area of macro-object.
Fig. 11: Flow chart of position calibration process
in visual system for integration
In addition, the initial position is also important
in the fabrication of the micro features since the
deformation of PDMS under pressure is around
100um, which is significantly bigger than the
fabrication thickness 20-50um in micro MIP-SL
process. So for the several layers fabrication in the
beginning, there is no resin filling in the projection
area and the first several layers of feature is missing.
The model starts building by the following projection
images. To avoid such situation, it is necessary to
figure out the initial position of the platform in z
direction, where was mounted with macro object
built by macro MIP-SL process. To address this
issue, taking advantage of the integrated optical
monitoring system, an imaging analysis method can
be applied. As the macro object moves down, the
distance L between the PDMS and macro object
decreases that the volume of the resin in between
loss as well. In the visual system of micro MIP-SL
set-up, the color of resin in the captured image
becomes lighter with the decrease of distance L.
When the outer surface of macro object just touches
with the top surface of PDMS, where is no pressure
loaded on the top of PDMS, the surface of macro
model replaced the resin and was appeared in the
captured images. By image-analysis, the relationship
between the initial distance Li and the color level R
of captured image can be fitted. Based on the above
model, the distance Li can be calculated during the
movement of platform. Finally, the plateform should
be moved to initial position until the value of Li

7
equal to the fabrication layer thickness of micro
MIP-SL process. In the case of integrating submicron
TPP process, the calibration process of relative
position is similar as the above integration process of
macro and micro MIP-SL systems. After whole
integration process, the MIP-SLA and TPP set-ups
can be utilized together to develop a more versatile
solution of novel multi- scale fabrication process.
5. Experiment and Test case
A mask planning testbed has been developed
using the C++ language with Microsoft Visual C++
compiler. The testbed integrates modules of
geometry slicing, mask image generation, mask
image projection and motion controlling. A Matlab
program is used to achieve the physical calibration
module in MIP-SL process. To achieve multi-scale
fabrication, the testbed also synchronizes the
integration processing with the X, Y and Z
movements. The hardware of the macro-, micro- and
nano-scales AM system are shown in Fig. 12.a-d,
respectively. The flow chart of the multi-scale
fabrication process is shown in Fig. 12 f.
Fig. 12. Hardware setup and flow chart of multi-
scale process. (a) Macro-scale MIP based SLA (b)
Micro-scale MIP based SLA(c) the GUI of the
software used in MIP-SL (d) Two-photon
polymerization set-up (f) the flow chart of the
multi-scale fabrication process
Fig. 13 and 14 show the two fabricated results of the
test case for the integration on macro and micro
MIP-SL processes, using resins SI500, MakerJuice
G+ and E-Glass 3SP. Fig.13a shows the CAD model
of one component. One fabricated physical model of
the device is shown in Fig. 13b. Fig. 13d and Fig.
13e show the microscope image of the
microstructures fabricated by the micro MIP-SL
processes. Fig. 14.a-b shows the CAD model of the
Optics with 100µm holes. One fabricated physical
model of Optics is shown in Fig. 14c. Fig. 14.d-e are
the projection image and microscopic image of the
fabricated micro-scale features.
Fig. 13 A Test case of Macro-structure with Micro
features. (a) CAD model (b) Physical model
fabricated by Macro & Micro MIP-SLA (c) CAD
model of micro-features (d) One-layer projection
image (e) Microscope image of micro features
build by micro MIP-SL process
Fig. 14 Test case of Optical component (a) CAD
model (b) CAD model of micro-features (c)
Physical model fabricated by macro & micro MIP
based SLA (d) Microscope image of features by
macro and micro MIP-SL processes (e)
Microscope image of micro features made by
micro MIP- SLA
(b)
(c)
(d)
(a)
(f)
(a)
(b)
(c)
(d)
(e)
(d)
(e)
(a)
(b)
(c)

8
Fig. 15. A test case of Macro-structures with
Micro and Nano features (a) Macro-scale
structure built by micro MIP-SLA (b) Macro and
micro scale structures built by macro and micro
MIP-SL processes (c) the microscope image of the
micro grid built by micro MIP-SL in central
rectangular (d) the microscope image of the micro
grid built by micro MIP-SL in heart-shaped hole
(e) the microscope image of the hexagon model
fabricated by TPP (f) the microscope image of the
hexagon model fabricated by TPP with bigger
magnifications (j) the microscope image of A log-
pile microstructure built by TPP on the top
surface of the micro grid (h) the microscope
image of A log-pile microstructure built by TPP (i)
the SEM image of A log-pile microstructure built
by TPP (g) the SEM image of A log-pile
microstructure built by TPP
As shown in Fig. 15, a macro-scale model
with micro- and nano-scales features was fabricated
to illustrate the resolution capability. The layer based
macro- and micro- MIP-SL processes were used to
build the part firstly. The layer thicknesses were set
as 100µm and 20µm, respectively. As shown in
Fig.15a, in the center of the model build by macro
MIP-SL process, there are five holes of different
shapes with diameter 2mm. In the micro MIP-SL
process, micro-scale grids with side length of 100µm
were fabricated in these five holes (see Fig.15b).
Fig.15.c-d showed the microscopic images of the
micro-scale structures in the rectangle and heart-
shaped holes fabricated by macro MIP-SL. The
macro- and micro-scales features are built using
resins SI500 and E-Glass 3SP respectively.
The SEM images of the micro- and nano-scales
features fabricated by the TPP process are shown in
Fig. 15.e-g. Fig. 15.e-f showed the hexagon model
fabricated by TPP on the surface of the micro-grid
built by MIP-SL. A log-pile microstructure with 2.5
µm period consisting of eight layers are built by TPP
process, the overall dimensions of which are 50µm x
50µm x 15µm are shown in the Fig. 15.j-g. The
nano-scale features are built using home-made
LabView software capable of importing 2D and 3D
CAD files and transforming them in polyline
coordinates that are read by the XYZ stages stack
used for TPP.
4. Conclusions
A novel multi-scale fabrication approach by
integrating macro- and micro-scale MIP-SL process
and nano-scale TPP process has been presented. The
integration effort has been made for the fabrication
of 3D objects with multi-scale features. Taking
advantage of the visual monitoring systems, macro-
and micro-scales MIP-SL systems and nano- scale
TPP system are integrated with satisfactory accuracy
and resolution. Both hardware and software systems
have been constructed to verify the developed multi-
scale fabrication approach. The experimental results
based on the test cases demonstrated its effectiveness.
Acknowledgements
We acknowledge National Science Foundation
(NSF ) for funding this research, and Center for
Electron Microscopy and Microanalysis (CEMMA)
at USC for providing microscopic measuring
equipments.
References
[1] Yi Lu et al., "Macro and nano-fabrication of
biodegradable polymers for drug delivery,"
Advanced Drug Delivery Reviews, 2004; 56(11):
1621–1633
(b)
(c)
(d)
(e)
(f)
(i)
(j)
(h)
(g)
(a)

9
[2] Jae-Won Choi et al., "Combined Macro and
macro additive manufacturing of a swirling flow
coaxial phacoemulsifier sleeve with internal Macro-
vanes", Biomed Macrodevices, 2010; 12:875–886
[3] Chuang Wei et al., "Direct Fabrication of High
Resolution Three-Dimensional Polymeric Scaffolds
using Electrohydrodynamic Hot Jet Plotting,"
Journal of Macromechanics and Macroengineering,
2013; 23(2), 025017.
[4] Y.M. Ha et al., "Mass production of 3-D
Macrostructures using projection",
Macrostereolithography, J. Mech. Sci. Technol. 2008;
22: 514–521.
[5] Yayue Pan et al., “Rapid Manufacturing in
Minutes: the Development of a Mask Projection
Stereolithography Process for High Speed
Fabrication”, ASME Journal of Manufacturing
Science and Engineering, 2013,;34(5).
[6] Zhou C et al., "Calibrating large-area mask
projection stereolithography for its accuracy and
resolution improvements.' In: International solid
freeform fabrication symposium. 2009.