2.1.

Theory

Two coherent beams on a plane create an interference pattern. This pattern can be used to expose a photoresist layer to create permanent structures.¹² The Lloyd’s mirror interferometer consists of a mirror placed perpendicular to the sample plane [see Fig. 1(a)]. When a diverging laser beam from a spatial filter (lens and pinhole) illuminates a Lloyd’s interferometer, the sample is exposed both to the direct beam and to the reflected beam. The addition of two laser beams creates a light intensity interference pattern on the substrate. The periodicity of the pattern ($\mathrm{\Lambda }$) is given by Lloyd’s interferometer equation:¹²

Fig. 1

(a) Lloyd’s interferometer setup. (b) Intensity distribution of two individual exposures and the sum due to the double exposure.

A Gaussian beam intensity distribution on the substrate for Lloyd’s interferometer created by the direct and reflected beams can be represented by Eq. (2):²⁵

For larger values of magnification and larger distances between the pinhole and sample, the intensity distribution becomes relatively flat along the surface of the mirror geometry yielding uniform structures. This large distance geometry is used by most researchers implementing the Lloyd’s mirror interferometer. On the other hand, confining the Gaussian beam or bringing the Lloyd’s mirror setup closer leads to a higher variation of the intensity along the sample and mirror. A distinct intensity variation can create various photoresist structures on the substrate. Creating two dimensional structures can be achieved by rotating the photoresist coated substrate by 90 deg and then applying a second exposure. For instance, the intensity distributions for the first, the second, and the total exposure of an experiment with $d=23.5\mathrm{cm}$, ${\omega }_0=1.9\mathrm{mm}$, and $\theta =14°$ on a sample with a size of $4\times 4{\mathrm{cm}}^2$ are illustrated in Fig. 1(b).

2.2.

Experimental

A 442-nm wavelength HeCd Laser (Omnichrome Series 74, Model 4074-P-A03, CVI Melles Griot, New Mexico) with a coherence length $L_c$ of 30 cm was operated at 80 mW. The 1.5-mm diameter laser beam passed through a lens-pinhole system built with a $20\times$ objective lens (Zeiss LD Plan-Neofluar 20X/0.4 Corr Ph2), and a 5-μm pinhole mounted on a three axis stage (Newport Three-Axis Spatial Filter, Model M-900, Newport, California). A time-controlled shutter was situated between the laser and the objective lens to control the exposure time during each experiment. The pinhole was used to remove undesirable components of the laser beam such as donut mode contributions. An iris was used to prevent reflections of the expanded laser beam. The Lloyd’s mirror/sample was placed on a two axis stage ($X$ and $\theta$) in such a way that the common corner of the sample and the mirror was located on the optical axis. The distance between the sample and the pinhole along the direction of the optical axis was 23.5 cm.

2.3.

Fabrication of Nanostructures

Experiments were performed with $(2.5\times 2.5){\mathrm{cm}}^2$ Fisher Brand microscope slides (soda lime glass) or $(2.5\times 2.5){\mathrm{cm}}^2$ fused silica slides (Valley Design Corp., USA) immersed in Nano-Strip (Cyantec Inc., California) at 60°C for 20 min to remove possible organic and inorganic contamination. In order to increase the adhesion of the photoresist, substrates were silanized with hexamethyldisilazane (HMDS) in an oven (YES-3TA HDMS Oven, Yield Engineering, California). Then substrates were spin-coated at 4000 rpm for 45 s with $1:4$ Shipley S1805 photoresist (Shipley, Massachusetts) diluted with Microposit Thinner Type P (Shipley, Massachusetts). Samples were soft-baked for 5 min at 115°C on a hot plate. The resulting photoresist thickness was $320\pm 10\mathrm{nm}$ found by analysis of electron microscopy images of FIB cut samples. To achieve the photoresist structures which carry the same periodicity on both axes, samples were fabricated with a mirror-laser beam axis angle $\theta \sim 15\mathrm{deg}$, yielding a periodicity $\mathrm{\Lambda }$ of $\sim 800\mathrm{nm}$. The sample was exposed first for 18 s, rotated by 90 deg and exposed a second time for 18 s. A second sample was exposed first for 18 s, then 12 s after the 90 deg rotation.

For the biaxial periodic structures, with different periodicities in $x$- and $y$-direction, $\theta$ was set to $\sim 11\mathrm{deg}$, yielding a periodicity $\mathrm{\Lambda }$ of $\sim 1200\mathrm{nm}$. After the first exposure, (for various times) the sample was developed, then placed in the Lloyd’s setup rotated by 90 deg with respect to the first exposure and exposed a second time (for various times) at an angle $\theta$ of $\sim 15\mathrm{deg}$ to obtain a periodicity $\mathrm{\Lambda }$ of $\sim 800\mathrm{nm}$.

Substrates were developed in MF319 developer (Shipley, Massachusetts). Symmetric structures were developed once after the two exposures. Biaxial structures were developed twice, once after the first exposure and then again after the second exposure. The first development was 45 s; the second development was 20 s. After each development, substrates were rinsed with copious amounts of deionized water and dried under nitrogen. Finally, samples were hard-baked for 10 min at 115°C on a hot plate. All the procedures were performed in a clean room facility at $21.5\pm 0.5°\mathrm{C}$, relative humidity of $30\pm 10\%$, and under yellow light.

The photoresist patterns were also used as etching masks for the fabrication of metallic or silicon nanofeatures. A 300-nm photoresist layer was spun on a p-type silicon substrate (500 μm thickness) with a 30-nm chromium layer deposited by an electron-beam evaporation system. The photoresist acted as an etching mask for the chromium layer resulting in a chromium pattern identical to the photoresist pattern. The patterned chromium nanostructures acted as a masking layer for silicon ion etching. Therefore, the pattern and objects fabricated lithographically in photoresist were transferred to the chromium layer by wet etching as well as the silicon wafer by deep reaction ion etching (DRIE; Alcatel 601E Deep Silicon Etch, France).

Fused silica samples carrying line and dot structures at a periodicity of 1200 nm in photoresist were ion milled (Vacu Tec Plasma Systems: control unit CPU 500, matching Plasmamatch; ENI: RF-generator ACG-3XL; Witney, Oxfordshire, United Kingdom) with SF6 at a gas flow rate of 20 sccm, 200 W plasma power and a pressure of $\sim 8\mathrm{Pa}$. Under these conditions, a surface topography with a profile depth of $\sim 100\mathrm{nm}$ was achieved. The structures were analyzed with SEM after ion milling to acquire the profile depth. The fused silica nanotopographic samples were used in the cell studies.

2.4.

Osteoblast Culture

All studies involving rats were performed in compliance with the University Council on Animal Care at the University of Western Ontario under approved protocols. Rat calvarial osteoblasts (RCOs) were obtained from newborn rat calvariae and cultured as previously described.²⁶

2.5.

Immunocytochemistry

Osteoblasts were plated at a density of $39\text{cells}/{\mathrm{mm}}^2$ and were fixed in 4% paraformaldehyde at 24 h post-seeding. Samples were then stained with rhodamine-conjugated phalloidin (Sigma-Aldrich, Canada), 4,6-diamidino-2-phenylindole dihydrochloride (DAPI) (Sigma-Aldrich, Canada), and vinculin (Sigma-Aldrich, Canada) as previously described.^7,26 Images were captured from each surface on an AxioScope microscope (Zeiss, Germany) using an Axiocam digital camera and AxioImager software.

3.1.

2-D Multishaped Structures

To correlate nanostructure shape with corresponding exposure and light intensity conditions, nine regions on each sample were chosen and characterized. Figure 2 shows the exposure intensity map as well as the selected regions where SEM images were taken. The first region $R_{(0,0)}$ was chosen at the origin of the $xy$-coordinate system and corresponded to the location with the highest exposure intensity. The remaining eight regions were distributed 10 and 15 mm, respectively away from the origin in both $x$- and $y$-directions. For instance, $R_{(10,15)}$ represented the region that was displaced 10 mm in the $x$-direction and 15 mm in the $y$-direction relative to the origin. As Fig. 2 shows, the nanostructure shape correlated directly with the exposure time and the beam intensity at each region.

Fig. 2

Illumination intensity maps and SEM images of photoresist patterns of samples fabricated with LIL. (a) Cumulative illumination intensity map for $18+18\mathrm{s}$ of exposure. (b) SEM images of nine selected regions from samples receiving illumination corresponding to (a). Location of each image corresponds to the location in the accompanying illumination intensity map. Scale bars represent 1 μm. (c) Cumulative illumination intensity map for $18+12\mathrm{s}$ exposure. (d) SEM images of nine selected regions from samples receiving illumination corresponding to (b).

In the case of the equivalent double exposure of sample 1, it was expected from Eq. (1) that the exposure intensity dropped by 6% and 14% at a distance of 10 and 15 mm away from the highest value at the origin, respectively. The diagonal in the $xy$-coordinate system was an axis of symmetry for the substrate. At the origin, [$R_{(0,0)}$ in Fig. 2(b)], round pillars were observed. Moving from the origin to $R_{(10,10)}$ the dot diameter gradually increased and some were connected. After passing $R_{(10,10)}$, connected dots formed holes and at $R_{(15,15)}$ the exposure was not high enough for the developer to etch the photoresist through to the substrate.

For areas off the axis of symmetry, either the $x$- or the $y$- component, the structures were elongated with respect to the other direction: elongated meaning that when lines form, the lines were along the $x$-direction on the right hand side of the symmetry line and along the $y$-direction on the left hand side of the symmetry line. The directions of the long axes of ellipses were analogous. The result was that $R_{(10,0)}$ contained objects elongated in the $x$-direction and $R_{(0,10)}$ contained structures elongated in the $y$-direction.

The elongation trend continued into wavy lines in either of the directions. The off-diagonal areas $R_{(10,15)}$ and $R_{(15,10)}$ contained broad wavy lines that were well-separated from one another. The ensemble of the wavy lines can be interpreted as a prestructure for the hole array located on the axis of symmetry at $R_{(15,15)}$ obtained with an exposure of lower intensity.

Another observation from Fig. 2 was the fact that in the high intensity areas, separated islands or nanodots were formed and with decreasing illumination intensity, the structures became connected with sharp linear features. Further decreases in the intensity led to a shape inversion, i.e., the formation of nanoholes.

The use of different exposure times when the substrate was in the $x$- and $y$-orientations in sample 2 resulted in samples with unique nanostructures within each region. Figure 2(d) clearly shows that the overall structure was not symmetric with respect to the diagonal. For example, the nanostructures in $R_{(10,0)}$ had a linear shape, while the nanostructures in $R_{(0,10)}$ were nominally oval-shaped. Furthermore, the nanostructures formed at the $R_{(0,0)}$ in Fig. 2(d) were slightly elongated in the $x$-direction, making them elliptical, while the nanostructures observed at $R_{(0,0)}$ in Fig. 2(b) were nearly circular.

3.2.

Nanostructures with Biaxial Periodicity

The combination of different exposure time, different exposure angles for the $x$- and $y$-axes, and post processing resulted in a wide variety of structures with controllable biaxial periodicity and morphology, including dots, holes, ellipses, elliptical holes, lines, and wavy lines. For instance, Fig. 3(a) shows elliptical nanodots in photoresist on a glass substrate with 1200 nm periodicity on the $x$-axis and 800-nm periodicity on the $y$-axis that resulted from exposure sets of $20+20\mathrm{s}$. Figure 3(b) shows elliptical nanoholes in photoresist on a glass substrate with similar biaxial periodicity to the structure in Fig. 3(a) that resulted from an exposure set of $22+18\mathrm{s}$. Figure 3(c) and 3(d) shows silicon nanopillar and nanohole structures with a height equal to 300 nm. In order to achieve 800 nm spacing on the $x$-axis and 1200 nm on the $y$-axis, the exposures sets of $15+25\mathrm{s}$ and $11+22\mathrm{s}$ were used to form nanodots and nanoholes, respectively.

Fig. 3

SEM images of nanostructures with biaxial periodicity (a) ellipses (in photoresist), (b) elliptical holes (in photoresist), (c) round pillars (in silicon), and (d) round holes (in silicon). Scale bars represent 1 μm.

3.3.

Cell Response to Nanotopographies

We tested the influence of the nanotopographies, explicitly on a dot pattern [Fig. 4(a)] and on a line pattern [Fig. 4(b)]. Cell adhesion and spreading of primary rat calvarial osteoblasts (Fig. 5) was tested. The vinculin stain (green) depicts the adhesions of the cell onto its substratum, whereas the f-actin (red) depicts actin microfilaments and hence relays information on cell morphology. The cell nuclei were stained with DAPI (blue). In the control experiments on smooth surfaces (no pattern) outside the structure of the substrates (Fig. 5, top row), adhesion sites (vinculin) were randomly oriented, as was cell spreading (f-actin). On the dot patterns (Fig. 5, middle row), adhesion sites were formed on the peaks of the topographies, which was evident from the dot patterning apparent in the vinculin-stained image (see inset). The cell morphology did not conform to the shape of the topography. On the 1200 nm spaced lines (Fig. 5, bottom row), osteoblasts formed densely adhesions in parallel with the long axis of the topography (vinculin), and the overall morphology (f-actin) was also aligned with the groove direction.

Fig. 4

SEM images of a dot (a) and a line (b) pattern on ion-milled glass substrates coated with 30-nm thick gold for SEM imaging. These example samples were cut with a focused ion beam to estimate the depth. The tilt corrected cursor height is $\sim 100\mathrm{nm}$ excluding the gold coating. Scale bars represent 100 nm.

Fig. 5

Rat calvarial osteoblasts cultured on smooth surfaces (top row), 1200 nm spaced nanodots (middle row) and 1200 nm spaced nanolines (bottom row). Adhesion sites were stained with vinculin (left column). Actin microfilaments were stained with f-actin (middle column). Overlay of the vinculin, f-actin and DAPI (nuclei) are depicted as green, red, and blue, respectively (right column). Insets show high resolution images of marked areas.