Grayscale Voxel Tuning in 3D Nanofabrication

• Grayscale voxel tuning is a technique that maps continuous grayscale intensities to voxel-specific parameters in nanoscale 3D fabrication.
• It employs nanopillar geometry manipulation and digital micromirror device exposure control to achieve precise optical properties such as brightness and scattering.
• This method enables high-dynamic-range, sub-diffraction patterning for applications in meta-optics and photonic devices while contending with hardware and material limitations.

Grayscale voxel tuning is a scheme in nanoscale 3D fabrication that enables precise, voxel-resolved control of the perceived brightness (gray level) in a printed structure. It involves mapping a continuous range of grayscale intensity values to tunable physical parameters of the materials or exposure conditions at the voxel scale, most commonly in two-photon polymerization (TPL) lithography. Key approaches utilize either direct geometric manipulation of nanostructure dimensions—such as pillar height, diameter, and array periodicity—or exposure dose control via spatial light modulation, to define locally resolved transmission, reflectivity, or scattering strength, thus reproducing arbitrary grayscale patterns or continuous 3D gray ramps within fabricated devices (Wang et al., 2020, Zhong et al., 28 Dec 2025).

1. Fundamental Mechanisms of Grayscale Voxel Tuning

In nanoscale additive manufacturing, grayscale voxel tuning is realized through two principal physical mechanisms: geometry-induced scattering regulation and exposure dose modulation. The first approach, exemplified by nanopillar-based metasurfaces, maps the desired brightness to a specific set of nanostructure geometric parameters—pillar height ($h$), diameter ($d$), and lattice periodicity ($p$)—leveraging Mie-like optical resonances and sub-resonant scattering to control the portion of incident light transmitted or reflected in the visible range (Wang et al., 2020). The second framework, prominent in advanced TPL systems, employs spatially programmable exposure via high-speed digital micromirror devices (DMDs) to modulate the delivered dose at each voxel, exploiting the nonlinear threshold response of two-photon polymerization to define the final polymerized volume and hence the local optical density or morphology (Zhong et al., 28 Dec 2025).

Both mechanisms share a reliance on deterministic mapping between the input image (or CAD model) grayscale and measurable physical properties of the fabricated voxel, underpinned by calibration tables or mathematical models relating geometry or dose to optically perceived brightness.

2. Geometric Control in Hybrid Nanopillar Arrays

The approach of Wang et al. (Wang et al., 2020) implements grayscale voxel tuning by direct manipulation of single nanopillar geometry. The unit cell, composed of a low-index polymer pillar of variable $h$, $d$, and $p$:

• For $h \lesssim 0.7\,µ$m, pillar-induced scattering is spectrally flat, and brightness is set primarily via $h$. Moderate adjustments to $d$ and $p$ refine the scattering cross section and thus the effective light transmission.
• At $h \gtrsim 0.7\,µ$m, resonant features introduce hue shifts, transitioning from gray to colored states.
• Periodicity $p$ controls inter-pillar coupling and serves as a secondary knob for brightness/saturation manipulation.

The mapping process uses a pre-fabricated "3D palette" with densely sampled ($h_i$, $d_i$, $p_i$) points, each associated with a measured brightness $B_i$ (in CIE L*a*b* or HSB space). For a given normalized gray value $g \in [0,1]$, the system selects the set $(h_i,d_i,p_i)$ from the palette with $B_i$ nearest to $g$. In the sub-resonant regime, the relation is approximately linear:

$$B \simeq \alpha h + \beta, \qquad h(g) \simeq \frac{g-\beta}{\alpha}$$

with empirical fits $\alpha \approx 0.30\,µ\text{m}^{-1}$ and $\beta \approx 0.05$.

This procedure enables continuous, single-pillar-level brightness tuning across at least 18 distinct gray steps in typical implementations, limited by fabrication precision and instrumentation (Wang et al., 2020).

3. Exposure Dose Modulation in Line-TF TPL

Line-illumination temporal-focusing TPL (Line-TF TPL) employs real-time grayscale voxel tuning through spatial control of laser exposure. Central to this system (Zhong et al., 28 Dec 2025) are:

• The dose-threshold model: A voxel forms only if the local cumulative time-integrated exposure $E(x,y,z)$ exceeds a polymerization threshold $E_{th}$. The relationship between exposure and voxel size follows:

$$R_{x}(E) = w_{x} \sqrt{\frac12 \ln \frac{E}{E_{th}}},\quad R_{z}(E) = w_{z} \sqrt{\frac12 \ln \frac{E}{E_{th}}}$$

where $w_{x}$, $w_{z}$ are beam radii.

• Grayscale levels are defined spatially: A DMD pattern line comprises $N_{total}$ mirrors per column; to encode a gray value $G \in [0,G_{max}]$, $$N_{on}=\mathrm{round}\left(\frac{G}{G_{max}} N_{total}\right)$$ mirrors are activated, with total intensity proportional to $N_{on}$. Effective per-line gray resolution is up to $\sim$1,600 steps, dictated by the number of mirrors.

Calibration is performed with a pre-determined look-up table $g_\mathrm{map}(i)$ that compensates for the femtosecond beam’s Gaussian intensity envelope, yielding uniform grayscale response along the illumination line. This methodology achieves sub-100 nm voxel resolution with hardware-limited rates ($\sim 3.3 \times 10^7$ voxels/s) and seamless continuous-field fabrication (Zhong et al., 28 Dec 2025).

4. Lookup Procedures and Mapping Algorithms

Both geometrically tuned nanopillar techniques and exposure-controlled TPL modalities depend on pre-calibrated mapping between target gray levels and controllable parameters. For nanopillars, this involves spectral characterization of a 3D parameter set, conversion to perceptual brightness, and palette search or regression fitting for direct parameter assignment (Wang et al., 2020). For DMD-based TPL, the mapping function from input grayscale to pixel occupation is direct and spatial, with compensatory calibration to equalize intensity along the scan line (Zhong et al., 28 Dec 2025):

Input Gray Value ($G$)	Parameter Set (Nanopillar: $h$, $d$, $p$)	DMD: $N_{on}$ Mirrors
$g \in [0,1]$ or $G \in [0,G_{max}]$	$(h_i,d_i,p_i)$ from palette (minimizing $|L_i^* - 100g|$)	$$N_{on} = \mathrm{round}\left(\frac{G}{G_{max}} N_{total}\right)$$

Palette-based lookup is critical to guarantee monotonicity, dynamic range, and physical feasibility. Exposure-based methods rely on deterministic arithmetic tied to hardware properties, with empirical correction for spatial inhomogeneity.

5. Experimental Resolution, Performance, and Limitations

Parameter ranges and physical resolution are dictated by the writing system and material response. In nanopillar-based schemes (Wang et al., 2020):

• $h$ (height): $0.1–3.0\,\mu$m in 0.1 µm steps (≈100 nm axial resolution)
• $d$ (diameter): $\sim 200–450$ nm (controlled via exposure time)
• $p$ (periodicity): $0.8–1.6\,\mu$m
• Lateral pitch typically $3.2\,\mu$m
• At least 18 visually distinct gray steps, constrained by geometry and detection NA

In DMD-based Line-TF TPL (Zhong et al., 28 Dec 2025):

• Gray levels per line: up to $\sim 1,600$
• Smallest voxel sizes: lateral 75–84 nm, axial 99–106 nm
• Operational rates: up to $3.3 \times 10^7$ voxels/s
• No temporal grayscale modulation required, preserving maximum scan rate

In both modalities, the primary limitations are set by hardware precision (DMD refresh rates, piezostage calibration for nanopillars), data-transfer bottlenecks, photoresist response (threshold, shrinkage), and, at fine scales, stochastic noise in polymerization or spectrophotometric readout. Resonance-induced hue artifacts above critical $h$ (for nanopillars) or beam inhomogeneity (for DMD lines) can introduce non-monotonicity or reduce practical gray resolution.

6. Applications and Advanced Features

Grayscale voxel tuning is fundamental for applications demanding high-dynamic-range brightness modulation within a single layer or volume, including meta-optics, photonic structural coloration, security steganography, and 3D photonic elements. The ability to invert grayscale appearance under darkfield, and desaturate color at will through nanopillar geometry, enables information encoding at single-pillar scale. Continuous-scanning TPL with spatial grayscale control permits rapid, centimeter-scale patterning of detailed, 3D grayscale or contoured elements while avoiding stitching artifacts (Wang et al., 2020, Zhong et al., 28 Dec 2025).

Optimization for industrial deployment targets both throughput (bandwidth-matched data streaming, high DMD refresh rates) and gray fidelity (calibrated lookup, beam uniformity compensation). Trade-offs involve voxel size, achievable gray steps, and machine uptime, with grayscale tuning offering robust performance across applications where sub-diffraction, large-area, and high-dynamic-range 3D patterning are required.

7. Fabrication and Characterization Considerations

Key aspects include:

• Precise stage calibration (±50 nm for field stitching in nanopillar schemes).
• Photoresist formulation (e.g., IP-Dip, $n \approx 1.52$), chemical development protocol (PGMEA/IPA/nonafluorobutyl methyl ether), and drying to avoid structure collapse (Wang et al., 2020).
• Fast data streaming and DMD control (multi-zone memory, compression factors $n \gtrsim 2.66$), sustaining print rates near hardware limits (Zhong et al., 28 Dec 2025).
• Characterization via microspectrophotometry (400–700 nm), spatial color analysis (CIE1931, HSB/L*a*b*), and SEM cross-section to verify target voxel size and brightness.

Accurate grayscale mapping and artifact-free field stitching at high speed distinguish state-of-the-art grayscale voxel tuning as an enabling technology for next-generation 3D nanomanufacturing (Wang et al., 2020, Zhong et al., 28 Dec 2025).