0
Research Papers

# Direct-Write Photolithography for Cylindrical Tooling Fabrication in Roll-to-Roll Microcontact Printing

[+] Author and Article Information
Larissa F. Nietner

Laboratory for Manufacturing and Productivity,
Massachusetts Institute of Technology,
Room 35-131,
Cambridge, MA 02139
e-mail: nietner@mit.edu

David E. Hardt

Fellow ASME
Laboratory for Manufacturing and Productivity,
Massachusetts Institute of Technology,
Room 35-231,
Cambridge, MA 02139
e-mail: hardt@mit.edu

The ratio of surface energy $γs$ (about 20 mJ/m2) to elastic modulus $E0$ (about 2 MPa) for conventional Sylgard 184 PDMS is quite large. The material radius of curvature $ρm=γs/E0$ is thus on the order of 10 nm. The surface energy has a significant effect at this length scale and permits conformal contact over a substrate with a surface roughness on the order of $ρm$.

It should be noted that the system was focused manually at a location near the outlet (for visual access) by minimizing the spot size on the surface of the resist.

1Corresponding author.

Contributed by the Manufacturing Engineering Division of ASME for publication in the JOURNAL OF MICRO- AND NANO-MANUFACTURING. Manuscript received February 10, 2015; final manuscript received June 1, 2015; published online June 25, 2015. Assoc. Editor: Sangkee Min.

J. Micro Nano-Manuf 3(3), 031006 (Aug 01, 2015) (10 pages) Paper No: JMNM-15-1011; doi: 10.1115/1.4030766 History: Received February 10, 2015; Revised June 01, 2015; Online June 25, 2015

## Abstract

The scale-up of microcontact printing (μCP) to a roll-to-roll technique for large-scale surface patterning requires scalable tooling for continuous pattern printing with μm-scale features (e.g., 1–50 μm). Here, we examine the process of creating such a tool using an optical direct-write or “maskless” method working on a rotating cylindrical substrate. A predictive model of pattern formation is presented along with experimental results to examine the key control factors for this process. It is shown that factors can be modulated to vary the cross-sectional shape in addition to feature height and width. This feature can then be exploited to improve the robustness of the final printing process.

<>

## Figures

Fig. 1

A centrifugally cast tool with circumferential features of 25 μm width at 100 μm pitch arranged within bands of 20 lines each

Fig. 2

Schematic of the process sequence. (a) The laser-resist patterning step determines the shape of the stamp feature in the positive resist. (b) The centrifugal casting of the PDMS stamp. (c) The resulting stamp feature, which in the printing process has to withstand normal pressure with the substrate required to transfer ink, capillary forces, and van der Waals interactions with adjacent feature surfaces.

Fig. 3

Stamp feature deformation modes: (a) sidewall collapse, (b) roof collapse, and (c) buckling

Fig. 4

Model layout showing cross section of resist with the photoresist state M(x,y,t) and input intensity I0(x)

Fig. 5

Result of the simulation for SPR 220. Exposure durations t=t0+1,t0+5,t0+10 , and t0+30 in 0.1 ms intervals. Concentration of remaining PAC is shown by color. Note the highly sloped sidewall prediction.

Fig. 6

Result of the simulation for AZ 9260. Exposure durations t=t0+1,t0+5,t0+10 , and t0+30 in 1 ms intervals. The simulation showed nearly straight sidewalls, as the low absorption attenuates light less significantly when passing through the material.

Fig. 7

The system comprises a laser with focusing optics, mounted on a linear axis, which can write on the inside of a drum that serves as a centrifuge for casting the tools

Fig. 8

Sequencing of features on the cylinder surface for experiments. For any resist film thickness, the experiment comprised multiple line arrays (sections) with different substrate speeds. Each speed was applied for three rotations.

Fig. 9

Table of experiments for different speeds and film thicknesses for both photoresists AZ 9260 and SPR 220

Fig. 10

Feature cross section dimensions. Note that the top width will be the corresponding printing surface for the stamp.

Fig. 11

Cross section of stamp features produced on AZ 9260 master with 12 μm feature height and 12 μm feature width at the top. (a) Microscope picture of cross section and (b) ESEM picture of same stamp.

Fig. 12

Width and height of features formed in SPR 220 (thickness 5 μm). Burning or dual-tone effects were visible at speeds lower than 10 rev/s, creating a narrow process window.

Fig. 13

Typical stamp created with a SPR 220 pattern. With SPR 220, straight but highly angled sidewalls were formed. The image above shows the features formed for 8 rev/s. At the printing side, the feature of the stamp has a width of 11 μm at a total height of 5.1 μm.

Fig. 14

Feature height and width comparison for different film thicknesses in AZ 9260. For an increase in speed (a decrease in dose), the width of the feature decreases as expected. The height data indicate that both low and high dose can lead to a decrease in feature height.

Fig. 15

Effect of speed on feature cross section for AZ 9260. Increasing writing speed reduces local energy dose and thus feature width. Note that for low speeds, the sidewalls of the features are formed by the edge of the Gaussian laser beam, resulting in a quality loss, and that for high speeds the resist threshold is not met to produce a full feature down to the substrate layer. These images were taken in a sample with 13 μm pattern thickness.

Fig. 16

Feature cross sections at different axial locations for a fixed substrate speed of 2 rev/s (i.e., a fixed dose). As the laser moves from the inner most to outermost location, the effect of variable focus produces a variable morphology in a nominal 10 μm thick AZ 9260 layer.

Fig. 17

Cross-sectional geometry optimized for μCP. With a feature height of 20 μm, a root width of 25 μm, and a top width of 9 μm, a sidewall angle of 25 deg is achieved.

Fig. 18

Simulation versus experimental results for SPR 220 and AZ 9260

Fig. 19

Simulated effect of beam overlap for AZ 9260. Both passes were with a 5 μm beam, and the second pass was offset by 10 μm. The result was a total width of 19 μm with fully developed root geometry on both sides.

## Discussions

Some tools below are only available to our subscribers or users with an online account.

### Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related Proceedings Articles
Related eBook Content
Topic Collections