Vacuum-assisted self-contact (VASC) photolithography evacuates the sealed volume between a large-area photomask and a flexible polyimide membrane so that atmospheric pressure presses the membrane into near-contact with the mask, eliminating the tens-of-micrometer air gaps that otherwise ruin large-area contact lithography; it was demonstrated in 2020 with a φ400 mm two-level off-axis Fresnel zone lens on a 25 µm polyimide membrane (36.6% average diffraction efficiency, 4.2 µm minimum linewidth). The open problem is the method's practical extension: producing multilevel (four-level or more) phase profiles on large-aperture membranes requires sub-micron overlay registration between exposures on a flexible substrate; the predicted residual air gap (about 100 nm at 5 Pa, sub-nanometer below 0.05 Pa) has never been measured; the lens wavefront remains limited by the intrinsic optical uniformity of the membrane (0.576λ peak-to-valley for the bare film); and the proposed pattern transfer to curved substrates via membrane photomasks is undemonstrated. Later work has achieved multilevel profiles on rigid fused-silica substrates at 350 mm aperture (75.9% efficiency, 0.020λ RMS wavefront) and multilevel profiles on polyimide membranes only at about 80 mm (up to 16 levels, 91.6% focusing efficiency), leaving the combination — multilevel, large-aperture, on membrane — and all VASC-specific limits unresolved.

Background

A diffractive (Fresnel) lens focuses light with a pattern of concentric rings whose surface-relief steps impose a controlled phase delay, instead of with the curved surface of a refractive or reflective optic; because its focusing comes from a relief pattern only micrometers deep, its areal mass can be orders of magnitude below that of a mirror of the same aperture. A Fresnel zone lens (FZL; also called a Fresnel zone plate when the rings are absorptive) quantizes the ideal continuous phase blaze into LL discrete phase levels, and the fraction of incident power delivered to the design focus is ηL=[sin⁡(π/L)/(π/L)]2\eta_L=[\sin(\pi/L)/(\pi/L)]^2 — about 40.5% for a two-level lens, 81% for four levels, and 98.7% for sixteen — so multilevel profiles are essential wherever light throughput matters. Optical polyimide (PI), a tough and thermally stable polymer cast as a roughly 25 µm freestanding film, is the leading substrate for membrane diffractive lenses aimed at lightweight, deployable large-aperture space telescopes; reviews in this lineage motivate the work with 10–20 m class transmissive telescope concepts, and the source paper designs its element as one segment of a 1.5 m primary lens (eight φ400 mm off-axis FZL segments plus an 800 mm on-axis segment).

The fabrication bottleneck is photolithography over large areas. In contact lithography a rigid photomask is pressed against a resist-coated substrate and exposed with ultraviolet light; the printed linewidth is set by the light field at the mask plane, and that field is distorted by any air gap between mask and substrate. Commercial large-area photomasks have peak-to-valley (PV) flatness errors of roughly 25 µm, so over a 400 mm membrane the mask and film cannot conform, air gaps of many micrometers open up, the exposure field blurs, and the developed resist profile degrades. The 2020 source paper (Gao et al.) introduced vacuum-assisted self-contact (VASC) photolithography: a silicone-rubber sealing ring on the photomask, the PI membrane laid over it, and the enclosed volume evacuated to about 5 Pa, after which atmospheric pressure compresses the trapped volume; by the ideal-gas relation p1d1=p2d2p_1 d_1=p_2 d_2 (d the gap, p the pressure), a 2 mm ring at 5 Pa should leave only a ~100 nm residual gap, and pressures below 0.05 Pa are predicted to leave sub-nanometer gaps. With this method the authors patterned a φ400 mm two-level off-axis FZL on a 25 µm PI membrane with 4.2 µm minimum linewidth, uniformly across the aperture, measuring 36.6% average diffraction efficiency (over 90% of the two-level theoretical value) with a full-aperture efficiency map taken at a 10.5 m focal distance.

Two caveats frame the open problem. First, the lens wavefront error — 0.599λ PV — was limited not by the patterning but by the bare membrane's own transmission wavefront (0.576λ PV), that is, by the intrinsic optical uniformity of the PI film; the authors state that better imaging quality requires a membrane substrate with better uniformity. Second, the residual gap, and hence the ultimate resolution, is a theoretical ideal-gas estimate: no direct gap measurement was reported. The two-level demonstration also leaves the multilevel regime unattempted — reaching four or more levels requires overlay, the registration of successive exposures to each other, at sub-micron precision on a flexible membrane — as does the proposed transfer of the method to curved substrates, in which the pattern is first copied from the rigid mask onto a membrane that then serves as a conformal photomask for the curved surface.

Problem Statement

How far does vacuum-assisted self-contact photolithography extend as a fabrication method for large-aperture membrane diffractive lenses — the question the 2020 demonstration left open. Relative to the demonstrated φ400 mm two-level off-axis Fresnel zone lens (4.2 µm minimum linewidth, 36.6% average diffraction efficiency, 0.599λ peak-to-valley wavefront error, ~100 nm theoretically estimated residual gap at 5 Pa):

(i) Fabricate multilevel phase profiles — four levels or more — on large-aperture polyimide membranes by integrating sub-micron overlay registration between exposures into the vacuum self-contact process, and quantify the resulting diffraction-efficiency and wavefront improvements over the two-level result.

(ii) Experimentally establish the attainable residual mask–membrane air gap and the corresponding patterning-resolution limit as functions of sealing-ring thickness and vacuum pressure, testing — rather than only assuming via the ideal-gas estimate — the prediction that sub-nanometer gaps are reached below 0.05 Pa.

(iii) Determine whether the optical uniformity of large-area polyimide membrane substrates can be improved enough that the wavefront error of a patterned large-aperture membrane lens is set by the patterning rather than by the bare substrate (which contributed 0.576λ PV in the demonstration).

(iv) Demonstrate, or experimentally bound the feasibility of, the proposed extension to curved substrates, in which patterns are transferred from the rigid mask to a membrane and that membrane then serves as a conformal photomask on the curved surface.

(i)–(iii) jointly answer whether this contact-lithography route can produce the multilevel, optically uniform, large-aperture membrane lenses needed for lightweight space imaging systems, while (iv) probes the same conformal-contact mechanism outside the flat-substrate geometry. Answers may use any process derived from vacuum-assisted membrane-to-mask self-contact; the question concerns the method's extension limits, not one fixed recipe.

References

  1. Guohan Gao, Lihua Wang, Heng Shi, Dun Liu, Bin Fan, Chunlin Guan, "Facile large-area uniform photolithography of membrane diffractive lens based on vacuum assisted self contact method", Scientific Reports 10:9005 (2020), DOI 10.1038/s41598-020-65990-2, https://doi.org/10.1038/s41598-020-65990-2
  2. Guohan Gao, Heng Shi, Lihua Wang, Dun Liu, Jiong Wang, Junfeng Du, Jiang Bian, Bin Fan, Hu Yang, "Large aperture high diffraction efficiency off-axis Fresnel lens fabrication and analysis", Optics Express 30(16):28932–28940 (2022), DOI 10.1364/OE.464494, https://doi.org/10.1364/OE.464494
  3. Qian Luo, Guohan Gao, Dun Liu, Junfeng Du, Bin Fan, "Wavefront measurement of a large aperture high image quality off-axis Fresnel lens", Optics Express 31(2):1249–1257 (2023), DOI 10.1364/OE.479611, https://doi.org/10.1364/OE.479611
  4. Xin Liu, Min Li, Bincheng Li, Bin Fan, "Membrane–Fresnel Diffractive Lenses with High-Optical Quality and High-Thermal Stability", Polymers 14(15):3056 (2022), DOI 10.3390/polym14153056, https://doi.org/10.3390/polym14153056
  5. Leslie Howe, Tharindu D. Rajapaksha, Kalani H. Ellepola, Vinh X. Ho, Zachary Aycock, Minh L. P. Nguyen, John P. Leckey, Dave G. Macdonnell, Hyun Jung Kim, Nguyen Q. Vinh, "High-Efficiency Multilevel Phase Lenses with Nanostructures on Polyimide Membranes", Advanced Optical Materials 12(25):2400847 (2024), DOI 10.1002/adom.202400847, https://doi.org/10.1002/adom.202400847 (see also arXiv:2502.17889, https://arxiv.org/abs/2502.17889)
  6. Xin Liu, Changchang Gong, Bin Fan, Jiang Bian, "High Precision Fabrication Method of Fresnel Diffractive Lenses on Ultrathin Membrane for Imaging Application", IEEE Photonics Journal 12(4):1–10 (2020), DOI 10.1109/JPHOT.2020.2998917, https://doi.org/10.1109/JPHOT.2020.2998917
  7. Guohan Gao, Jiong Wang, Xin Liu, Junfeng Du, Jiang Bian, Hu Yang, "Improving Laser Direct Writing Overlay Precision Based on a Deep Learning Method", Micromachines 16(8):871 (2025), DOI 10.3390/mi16080871, https://doi.org/10.3390/mi16080871
  8. Jiajia Yin, Mengxia Hou, Bin Fan, Jiang Bian, Junfeng Du, "Laser Radar System Based on Lightweight Diffractive Lens Receiver System", Photonics 12(1):86 (2025), DOI 10.3390/photonics12010086, https://doi.org/10.3390/photonics12010086
  9. Heng Shi, Junfeng Du, Lihua Wang, Jiang Bian, Guohan Gao, Dun Liu, Bin Fan, Hu Yang, "A High-Precision Real-Time Pose Measurement Method for the Primary Lens of Large Aperture Space Telescope Based on Laser Ranging", Sensors 23(10):4833 (2023), DOI 10.3390/s23104833, https://doi.org/10.3390/s23104833
  10. Junyi Lu, Cuifang Kuang, "Laser Direct Writing of CGH for Aspherical Surface Testing", 2024 3rd International Conference on Advanced Sensing, Intelligent Manufacturing (ASIM), pp. 131–136 (2024), DOI 10.1109/ASIM66270.2024.00031, https://doi.org/10.1109/ASIM66270.2024.00031