a positive, negative, and biofunctional resist all …...1. supplementary experiments and discussion...

SUPPLEMENTARY INFORMATIONDOI: 10.1038/NNANO.2014.47

NATURE NANOTECHNOLOGY | www.nature.com/naturenanotechnology 1

SUPPLEMENTARY INFORMATION

All water-based electron beam lithography using silk as

a positive, negative, and biofunctional resist

Sunghwan Kim1, §,*,, Benedetto Marelli,1, §

Mark A. Brenckle1, Alexander N. Mitropoulos1, Eun-Seok Gil,1 Konstantinos Tsioris1,**,

Hu Tao, 1 David L. Kaplan1, and Fiorenzo G. Omenetto1,2, ‡

All-water-based electron-beam lithography using silk as a resist

© 2014 Macmillan Publishers Limited. All rights reserved.

http://www.nature.com/doifinder/10.1038/nnano.2014.47

1. Supplementary Experiments and Discussion

1.1 Silk fibroin protein

In the gland of Bombyx mori caterpillar, the water-soluble silk fibroin and the

sericin, which is a water-soluble protein glue to coat the silk fiber, are spun together into

a fiber, leading to a new insoluble conformation due to a rapid change in structure of the

fibroin1,2. The sericin is added to the surface of silk fibroin during spinning, and the result

is the formation of the silk fiber. By chemical procedure, the silk fibroin solution can be

extracted from the cocoon, a bunch of the silk fibers. A silk film coming from the silk

solution has random coil complex, which is water-soluble. These random coils can be

modified to two structural forms by phase transition: silk I and silk II.3 Silk I refers to the

complex helix-dominated structure existing within the silkworm gland just before

spinning, which is different from the water-soluble random-coil conformation of

uncrystallized silk. Silk II is the water-insoluble antiparallel β-sheet crystal conformation,

which forms after the spinning of silk fibers from the spinneret of the silkworm.

Therefore the water-insoluble silk film can be achieved by increasing β-sheet content.

The most common method to induce β-sheet is based on the chemical change during

immersion of the silk film into organic solvents such as methanol or ethanol3. An

alternative method is water-vapor annealing3. Water molecules can plasticize protein

structures with hydrogen bonds and also promote solvent-induced crystallization. Since

β-sheet is composed of chains connected by hydrogen bond, the water-vapor treated silk

film become the water-insoluble (see Fig. S1).


Supplementary Figure S1. Mechanism of the formation of β-sheet. Schematic diagram of the formation of β-sheet secondary structure in silk fibroin protein.

To make water-insoluble silk films, any cross-linking methods widely studied in

polymer science would be good candidates4. In other polymer systems such as Poly(vinyl

alcohol), poly(ethylene glycol), and poly(acrylic acid), high energy irradiation using γ-

ray and electron beam generates free-radicals, thereby cross-linking water-soluble

polymers without any additional cross-linkers4. Recombination of the radical results in

the formation of covalent bonds and finally cross-links structure. This supports the

extension of irradiation based crosslinking to silk through a similar mechanism.


1.2 Characteristics of silk as a resist

Supplementary Figure S2. Sensitivity of silk resist. Sensitivity to electron dose characteristics for positive silk resist (a) and negative silk resist (b) at 100 keV acceleration. The y-axis illustrates the ratio of measured depth z to initial film thickness z0 after irradiation. (c) AFM images show the surface morphologies of a square pattern at different electron dose. Bottom graph is the cross section profile of the pattern. Red line is about 40 mC/cm2 and blue line is about 90 mC/cm2.


The sensitivity characteristics for silk resist are shown in Fig.S2 (a) and (b).

Electron beam is used to irradiate a 5-μm-square area in the crystallized and amorphous

silk film (270-nm-thickness) with different doses, using the 100 keV system. After

developing the film, atomic force microscopy (AFM) was used to obtain the depth

(height) of the etched (remaining) patterns. Since the surface profile of the pattern is not

even, the depth and height are averaged. In Fig. 2S(a) and (b), the threshold of sensitivity

is about 2.25 mC/cm2 (positive) and 25 mC/cm2 (negative). These threshold values will

scale based on electron acceleration and resist thickness considerations, but indicate that

silk resist requires ~11 times higher dose for the negative pattern than the positive pattern.

Along with these, the AFM images of the negative-tone patterns in which doses are

between 25 and 80 mC/cm2 show bumpy top-surfaces. At dose of over 90 mC/cm2, the

surfaces become flatter (Fig. S2(c)). We presume lower doses aren’t enough to induce

homogeneous crosslinks between silk molecules inside the silk crystalline matrix volume

exposed to the electron beam. Increasing depth of the patterns when dose is increasing

below the threshold would allow the formation of a three dimensional structure.

Fig.S3 shows the SEM and AFM image of an inverse pyramid-like geometry.

Each pyramid was exposed 4 times with different dose and size. We expect this technique

to be a building block for versatile nonmanufacturing.


Supplementary Figure S3. Three dimensional structure of silk resist at 100 keV acceleration. a, SEM image (top) and the distribution of dose (bottom). b, AFM image of the pattern (top) and the cross section profile (bottom).


1.3 Test structures in 1D, and 2D

Different structures were also tested to assess resist performance. One

dimensional grating patterns with a fixed lattice constant of 1 μm and a line-width

varying from 20, 60, 100, 200 to 400 nm were written using 100 keV acceleration. A

tilted SEM view of the silk fibroin grating exhibits a well-defined vertical profile,

showing promise towards the reduction of the minimum feature sizes obtainable by using

silk fibroin as an EBL resist, as shown in Fig. S4.

Supplementary Figure S4. 1D, 2D test structures SEM image of the 1D line pattern with the 100-nm width. (top) and tilted view with the 500-nm line width pattern showing 1:1 aspect ratio. 2D test structures (bottom) showing 50nm and 30nm pillars with 150nm and 250nm lattice constant at 1:1 aspect rato. Scale bars for the 1D patterns represents 1 μm, . For the 2D patterns, scales represent 500nm.


Resist resolution was investigated by writing 30 and 50nm diameter features in

negative silk resist. Here, the lattice constant of the features was varied depending on

diameter, with spacings of 150nm, and 250nm. The features were written at 125 keV

acceleration and 2nA current with a dose of 3750μC/cm2 (positive) and 75,000μC/cm2

(negative) on 40nm thick resist. SEM of the resulting features shows faithful replication

down to 30nm with good pattern fidelity. Any distortion is likely to due to slight swelling

of the features, investigated in a separate experiment.


1.4 Swelling of silk and aspect ratio limitations

Aspect ratio limitations due to swelling in the resist were tested by writing 2D

gratings of 200 nm diameter with 500 nm lattice constant and 50 nm diameter with 125

nm lattice constant in both positive and negative silk resist at a thickness of 200 nm. This

produces aspect ratios of 1:1 and 4:1. All features were written at 100 keV acceleration

and 2 nA current with doses of 3000μC/cm2 (positive) and 65,000 μC/cm2 . As shown in

Fig. S5, tilted SEM of the resulting structures shows good replication of both low and

high aspect ratios in the positive case, as well as good low aspect ratio features in the

negative case. The high aspect ratio negative features collapsed, which is likely due to

mechanical weakness in the swollen silk resist after development. This will ultimately

limit the aspect ratios of features producible in negative silk resist.

Supplementary Fig. S5 Swelling and aspect ratio SEM images of aspect ratio test structures in positive and negative resist at 1:1 (a) and 4:1 (b) aspect ratios.


1.5 Comparison with PMMA

A baseline comparison experiment was preformed in which the profile of silk

resist was compared with poly(methyl methacrylate) (PMMA), the most well-known and

widely-used EBL resist. The square-lattice pattern with 200-nm pitch size and 100-nm

diameter was generated on PMMA (130-nm-thick) and silk (80-nm-thick) resist films.

The dose value was 1 mC/cm2 (PMMA) and 3 mC/cm2 at 100 keV acceleration (silk).

Although the radius of holes in the PMMA film is slightly larger than silk due to the

proximity effect, silk favorably compares with PMMA with regard to hole shape and

cross section profile (Fig. S6).

Supplementary Figure S6. Comparison with PMMA. SEM and AFM image of the square lattice pattern on the PMMA film (a, b) and the silk film (c, d). All scale bars represent 400 nm.


1.6 Gain enhancement at the band-edge and calculation of band-structure

The photons in the photonic crystal (PhC) can propagate through the medium with

reduced group velocity at the photonic band-edge, the point with zero slope in the

photonic band-structure.5 When the PhC contains quantum elements, such as metal

nanoparticles, quantum dots, and dye molecules, it provides a way to enhance their

emission or absorption characteristics at the photonic band-edge. To estimate the

enhancement of emission of green fluorescent protein (GFP) shown in Fig. 4 in the main

text, we calculated the photonic band-structure of the silk resist slab containing the

triangular lattice hole array. A 3D plane-wave expansion (PWE) calculation was

conducted using the MPB, open source PWE software distributed by MIT.6 The

refractive index used in the calculation was 1.54 (silk) and 1.45 (SiO2). The radius was

0.25a (a is the lattice constant.). The thickness in the calculation was set to 130 nm (silk)

and 100 μm (SiO2), - e.g. almost infinite thickness.

Supplementary Figure S7. Theoretical estimation of the gain enhancement. Photonic band-structure calculated by the PWE method. At the band-edges (red circle), the generated photons can be enhanced due to slow group velocity. The geometry of the super-cells that were used in the calculation is shown on the right. The adoption of a periodic slab geometry is a reason for the detection of more modes (that are irrelevant to the actual 2D PhC structure adopted in the experiment).


Fig. S7 shows the calculated band-structure. Since the calculation assumed that

the structure was periodically arranged in all three dimensional space, many unnecessary

modes (periodic slab modes) were shown in the band-structure. These modes could be

filtered out by multiple calculations changing the dimension of the super-cell with normal

direction to the slab. At around the normalized frequency of 0.8, there are several band-

edge modes at Γ-point that can enhance the gain of GFP. For the sake of convenient

comparison, the normalized frequencies were converted to the wavelength scale

considering the lattice constant in Fig. 4. The emission of GFP was enhanced at the band-

edge location in the photonic structure, which is consistent to what observed

experimentally.


1.7 Biological stabilization in vacuum

Supplementary Figure S8 Biological Stabilization in vacuum. Methodology and results of biological stabilization experiment, showing residual efficacy of HRP films after storage for 48 hours in reduced temperature (4 °C and -20 °C) and low-pressure conditions (4e-6 mbar).

Horseradish peroxidase (HRP) is widely used as an indicator enzyme for

immunoassays and was therefore used in our work to study enzyme stability of silk films

during and after the electron beam writing process (specifically the exposure of the doped

films to a high vacuum environment). HRP (Type VI, 250 unit/mg, Sigma) loaded silk

films were prepared by casting HRP dissolved silk solution (20 unit/mL, in 6 wt% silk

solution) onto glass slides and letting dry overnight at ambient conditions. The dried


films were peeled off from the substrate and cut into small squares (10 mm × 10 mm,

with a weight of ~ 10 mg), which were placed into Petri dishes (35 mm × 10 mm) and

then stored in a freezer (-20 °C), a refrigerator (4 °C), room temperature, or within the

chamber of a scanning electron microscope (Zeiss MA10, with a vacuum level of 4E-6

mbar) for 48 hours, respectively.

Three samples (N=3) were used for each analysis. Each sample was immersed in

200 μL of distilled water at room temperature for 10 minutes until it was fully dissolved.

200 μL of 3,3’,5,5’ Tetramethylbenzidine (Sigma) was added to the solution followed by

a gentle agitation for 60 seconds before being stopped by the addition of 400 μL of 0.1

mol/L sulfuric acid. Absorbance of solution was measured at 450 nm using a VersaMax

microplate reader (Molecular devices, Sunnyvale, CA). It was found that the activity

retention for silk films that were kept in SEM for 48 hours (which was significantly

longer than most e-beam writing processes which typically range from a few minutes to

several hours) was similar to the films kept in the fridge (86.87% vs. 86.77%), which had

slightly lower stability than the films kept in the freezer (normalized to 100%), but higher

stability than the films kept at room temperature (69.71%), implying no obvious loss of

enzyme activity due to exposure to high vacuum during ebeam writing, and as shown in

Fig. S8.


1.8 Biological Survivability and irradiation

Additionally, biological survivability during irradiation was assessed by writing

200 nm features into HRP doped positive silk resist a thickness of 200 nm (0.2 unit/mL, 6%

silk). The films were exposed at 3750 microC/cm2 using 125 keV acceleration and 2nA

current. Figure S9 shows darkfield and SEM images of the developed patterns written

into the HRP resist. Following development, the films were exposed to both water and

TMB solution (3,3′,5,5′-Tetramethylbenzidine Liquid Substrate System for ELISA,

Sigma) to probe for HRP activity. The resulting blue color on the TMB films along with

no color on the control confirmed activity of the enzyme following sample irradiation, as

shown in the main text.

Supplementary Figure S9. HRP-photonic lattices - SEM and darkfield images of test structures written into HRP-doped positive silk resist.


2. Supplementary Methods

2.1 Silk fibroin sample crosslinking

To crosslink the silk fibroin films, two equivalent techniques were adopted – the

samples were either (a) treated in methanol at room temperature for one minute and

water-vapor at 95 °C for two hours or (b) water-vapor annealing at room temperature also

formed highly crystallized silk films, at the expense of longer annealing time (over 12

hours). The results are equivalent – the advantage of one method versus the other in this

case is processing time. For reference, the samples presented in this work were processed

as follows: Figure 2a,b,c,d,e, Figure 4a (QD), Figure S2, Figure S3, Figure S4 (top),

Figure S5, Figure S6 all utilized the Methanol crosslinking method. Figure 2f,g,h, Figure

4a (GFP,HRP),d, Figure S4 (bottom), Figure S9 all utilized the water vapor annealing

method.

2.2 sfGFP expression and purification

The sfGFP plasmid vector was commercially obtained from Sandia

BioTech/Theranostech, Inc. and transformed into chemically competent BL21 E.coil

(Invitrogen) expression host. Transformed E. coli was plated on LB plates containing

Kanamycin (Kan) (Fisher) antibiotic and incubated at 37 ºC overnight. A single bacteria

colony was selected and cultured at 37 ºC overnight in 5 ml LB growth media (Fisher)

containing Kan. The next day, a 2 liter volume LB growth media containing Kan was

inoculated with the 5 ml starter culture and cultured shaking at 37 ºC. IPTG (Isopropyl β-

D-1-thiogalactopyranoside) (Fisher) was added at OD 0.8 and cultured for an additional 5

h. Subsequently an E. coil pellet was obtained by centrifugation, re-suspended in lysis


buffer and sonicated to aid cell lysis. sfGFP was purified by metal affinity column

chromatography (NiNTA, Qiagen) and dialyzed. Aliquots were subsequently flash frozen

and lyophilized until further use. For future use, sfGFP was re-suspended in DI water.


References

1 Hu, X., Kaplan, D. L. & Cebe, P., Determining beta-sheet crystallinity in fibrous

proteins by thermal analysis and infrared spectroscopy. Macromolecules 39,

6161-6170 (2006).

2 Jin, H. –J. & Kaplan, D. L. Mechanism of silk processing in insects and spiders.

Nature 424, 1057-1061 (2003).

3 Hu, X., Shmelev, K., Sun, L., Gil, E.-S., Park, S.-H., Cebe, P. & Kaplan. D. L.

Regulation of silk material structure by temperature-controlled water vapor

annealing. Biomacromolecules 12, 1686-1696 (2011).

4 Hennink, W. E. & van Nostrum, C. F. Novel crosslinking methods to design

hydrogels. Adv. Drug Del. Rev. 64, 223-236 (2012).

5 Dowling, J. P., Scalora, M., Bloemer, M. J. & Bowden, C. M. The photonic band

edge laser: A new approach to gain enhancement. J. Appl. Phys. 75, 1896 (1994).

6 Johnson, S. G. & Joannopoulos, J. D. Block-iterative frequency-domain methods

for Maxwell’s equations in a planewave basis. Opt. Express 8, 173-190 (2001).


a positive, negative, and biofunctional resist all …...1. supplementary experiments and discussion...

Documents