Holographic interference lithography: static and scanning-line exposure
Conventional 3D interference lithography with a phase mask typically involves irradiating a statically-placed diffractive phase mask, such as a 2D periodic grating, with a laser beam to generate multiple transmitted beams of diffraction as shown by the 0th order and four 1st order beams in the schematic of Fig. 1a. The laser beam diameter defines the working distance for photoresist exposure over which all the diffraction beams will appreciably overlap and fully interfere on the opposite side of the phase mask. At an increasing exposure distance, the region of 5-beam interference (dark red) will shrink as depicted in Fig. 1b to reveal annular exposure zones of 4-beam (orange), 3-beam (yellow), 2-beam (cyan) and 1-beam (light blue) interference where lower exposure is expected together with undesirable light patterns due to missing diffraction orders as observed in ref. 45.
The present zones represent photoresist exposure of a 10 mm diameter top-hat beam through a 2D binary phase mask (570 nm period) at a 1 mm exposure distance, yielding a large uniform full interference zone of ~8.5 mm diameter surrounded by a ~2.5 mm annular exposure zone of incomplete beam interference. The annulus of incomplete interference shrinks to only ~27 μm for a closer exposure distance of 40 μm (refer to Supplementary Fig. S2a). Alternatively, reducing the beam diameter from 10 mm to 60 μm at 40 μm exposure distance leads to a significant decrease of the full five-beam interference zone from 99.0% areal overlap to 56% (refer to Supplementary Fig. S2c).
The various interference zones for the case of large diameter beam exposure (10 mm), reproduced in Fig. 2 (center), are assessed individually to determine the isointensity surface expected in the photoresist from each exposure zone. The calculated surfaces are shown in Fig. 2a, providing the predicted motif and crystal periodicities of a = Λx = Λy = 570 nm along the x and y axes and c = 1.845 μm along the z axis that follows the reciprocal relationship between wave vector differences, namely, . The peak value of interference intensity (0.46 to 3.99) expected in each zone is also given, normalized to that of the incident beam. The 5-beam zone (dark red) provides a symmetric TTR or BCT-like structure with a strong interference peak intensity of 3.99 (Fig. 2a-iii). A skewed BCT-like structure and unconnected column array structures are noted when one and two diffraction orders are missing, yielding weaker intensity peaks of 2.83 (Fig. 2a-ii) and 1.87 (Fig. 2a-iv), respectively. The isointensity surface was set at a threshold intensity exposure of Ith = 1.2 in Fig. 2a-ii through 2a-iv, while a lower threshold of 0.4 was required to reveal the pattern of 1D periodic plates (Fig. 2a-i) produced in the 2-beam zone (light blue) which had a low interference peak intensity of 0.46.
To meet the objective for extending the 3D PC structure over larger area and with better uniformity, a scanning exposure of the laser beam in a straight single line (y axis) along the phase mask was investigated to determine the influence of partial exposure from the incomplete interference zones in the beam periphery. These weak exposures may distort and overshadow the ideal 3D periodic interference pattern otherwise expected from the full 5-beam interference zone. The photoresist position is locked in the proximity zone of the phase mask to ensure the interference pattern does not shift and wash out the 3D patterning during the laser scan. The isointensity surface structure predicted for the center beam exposure position (x = 0) is presented in Fig. 2b-i, calculated for a threshold intensity of Ith_scan = 1.2 and compared with structures expected at the lateral offset positions of x = 2.2 (Fig. 2b-ii), x = 4.4 (Fig. 2b-iii) and x = 5.5 mm (Fig. 2b-iv). The isolated voxels generated for offsets of x = 0 to 4.4 mm share a similar BCT-like symmetry that attests to the dominant contribution of the 5-beam interference over the weaker but distorted symmetry patterns imposed by the incomplete interference annular exposure zones. However, this 3D structure is seen (Fig. 2b) to skew and dis-connect with increasing lateral offset (from x = 0 to 5.5 mm) due to lower net exposure as well as due to a larger relative intensity contribution from incomplete interference zones.
An experimental comparison of the predicted nanostructure shape for static (Fig. 2a-iii) and scanning (Fig. 2b) exposure was made in 40 μm thick photoresist exposed through a 1 mm thick binary phase mask with a 10 mm diameter laser beam as in the arrangement of Fig. 1a. (refer to Methods—Photoresist preparation and development and Laser scanning holography) The photoresist substrate was gently mounted against the phase mask surface and remained stationary during the scanning exposure. Static exposure yielded the ~8 mm diameter PC structure zone shown in Fig. 2c-ii that matches with the reduced size expected for the full 5-beam interference zone as shown to scale in Fig. 2c-i (reproduced from Fig. 1b). The formation of 3D BCT-like structure here has been previously reported in ref. 46 using static exposure, yielding a ≈ 570 nm transverse and c ≈ 1.54 μm axial periods.
Scanning of the laser across the same stationary phase mask arrangement at the same exposure power produced the elongated (7 mm × 18 mm) 3D PC structure in SU-8 photoresist as shown in the optical image of Fig. 2d-ii. The cross-sectional SEM views shown in Fig. 2e for two different scanning offset positions of x = −2 (Fig. 2e-i) and x = 3 mm (Fig. 2e-ii) reveal differing filling fractions, but similar formation of BCT-like structure. The photoresist structures, overlaid as insets in Fig. 2e, are well represented by the calculated isointensity surfaces after scaling of the structures in Fig. 2b to exposure threshold of Ith_scan = 0.5. The lateral fall off of laser exposure yields ~100% filling fraction at the center x = 0 mm position (solid zone in Fig. 2d-ii), opening into dense and low density 3D bicontinuous structure at 2 mm offset (Fig. 2e-i) and 3 mm offset (Fig. 2e-ii) positions, respectively, that transitions into underexposed and washed out photoresist beyond ~3.5 mm lateral offset. The observed lateral and axial periodicities of a = 570 nm and c = 1200 nm, respectively, indicate a strong ~35% c-axis shrinkage that is commensurate with previous reports18,27,47. Single line scanning is therefore robust against pattern distortion from exposure zones…