Volumetric Additive Manufacturing: How To Use The VAMtoolbox For Computed Axial Lithography

Volumetric additive manufacturing utilizes tomographic light projections to cure an entire three-dimensional object simultaneously within a rotating volume of photopolymer resin. By leveraging the VAMtoolbox, researchers and engineers can import target computer-aided design models, simulate ray-tracing behavior through cylindrical boundary interfaces, and execute optimization algorithms to compute the precise dynamic intensity maps required for high-fidelity printing. Achieving optimal print quality relies on balancing spatial discretization resolutions, mathematical optimization iterations, and resin-specific threshold parameters.

Pre-Operation Setup and Computational Infrastructure

Before generating projection patterns for computed axial lithography, you must establish a robust computing environment capable of executing high-dimensional matrix operations. The VAMtoolbox is a MATLAB-based open-source software suite designed to solve inverse tomographic optimization problems. Because these calculations require substantial memory and processing power to compute forward and back-projections through three-dimensional voxel spaces, your workstation must be specifically configured for parallel computing and matrix accelerations.

Essential Software and Hardware Prerequisites

• Host Workstation: Multi-core processor (minimum 8 physical cores, recommended 16 or more) paired with at least 32 gigabytes of system memory (RAM). A dedicated graphics processing unit supporting parallel computing architectures is highly recommended for acceleration of volumetric ray-tracing operations.
• Software Environment: MATLAB environment (version R2021a or newer) equipped with the Image Processing Toolbox, Optimization Toolbox, Parallel Computing Toolbox, and the Signal Processing Toolbox.
• The VAMtoolbox Repository: The latest stable release of the VAMtoolbox source files cloned or downloaded into a dedicated workspace directory on your local machine, with all subdirectories added to your active MATLAB environment search path.
• Target Geometric Files: Watertight 3D models in standard stereolithography (STL) or wavefront object (OBJ) file formats, verified for non-manifold edges and self-intersections prior to processing.
• Polymer and Optical Calibration Benchmarks: Measured material metrics including the critical gelation dose threshold of your photoinitiator system, the target dynamic range of the digital micromirror device (DMD) projector, and the precise refractive index of both the glass print vial and the uncured photopolymer resin at the operating laser wavelength.

Technical Step-by-Step Optimization Workflow

Generating volumetric projection sequences involves importing geometry, configuring the computational space, simulating light propagation, running an optimization solver, and exporting the resulting intensity maps. Follow this precise protocol to convert a physical target design into highly accurate optical projection frames.

Step 1: Geometry Import and Voxelization

The initial step requires converting your continuous boundary representation model into a discrete three-dimensional grid of voxels. This voxel grid serves as the mathematical baseline for the light dose optimization algorithm.

Begin by setting your MATLAB working directory to the root of your VAMtoolbox installation. Load your STL file using the native mesh loading utility functions built into the toolkit. Once the mesh coordinates are loaded, define the target voxel grid size. Typical grid setups range from 150 to 512 voxels per dimension, depending on the scale of your target print and available GPU memory. Run the voxelization function, which employs a ray-casting method to classify voxels as internal (value of 1) or external (value of 0).

To ensure the physical print survives processing, you must verify that fine features are mapped to a sufficient number of voxels. If your target geometry has fine channels, ensure the voxel scale allows at least three to five voxels across the narrowest feature. Add spatial padding of at least 10 to 20 empty voxels around all boundaries of your target object inside the matrix to prevent edge-boundary artifacts during computational rotation.

Warning: Insufficient spatial padding around the boundary of your target voxel grid will cause the tomographic backprojection equations to clip at the edges. This clipping introduces severe high-frequency noise and artificial high-dose concentrations along the outer perimeter of your physical print vial.

Step 2: Optical and Refractive Index Matching Alignment

Computed axial lithography projects light patterns through a cylindrical glass vial containing photopolymer resin. Because light bends when moving between media of different refractive indices, you must configure the optical path matrix to prevent severe geometric distortions.

In the initialization parameters of your workspace script, declare the refractive index values for the three primary components: the immersion bath medium, the glass vial wall, and the liquid photopolymer resin. For optimal results, use index-matching fluids like glycerol or specific oils to match these values as closely as possible.

In the VAMtoolbox configuration structure, set the outer vial radius, the wall thickness, and the inner resin radius in millimeters. The toolkit uses these values to build a cylindrical ray-tracing model. If your experimental printer does not use an index-matched immersion bath, you must enable the refraction correction routines within the toolbox setup. This setting calculates non-parallel ray paths through the curved interfaces rather than assuming straight parallel light pathways.

Step 3: Setting Up Optimization Solvers and Convergence Bounds

The core challenge of volumetric additive manufacturing is solving an inverse problem: finding a sequence of two-dimensional projection patterns that, when integrated over a full 360-degree rotation, deliver an optical dose exceeding the gelation threshold inside the target geometry while remaining safely below the threshold everywhere else.

You must select the mathematical solver that matches your engineering requirements. The VAMtoolbox offers several optimization approaches:

1. Filtered Backprojection (FBP): A fast analytical method modeled after medical CT reconstruction. It provides quick calculations but struggles with high-contrast sharp edges and complex internal cavities.
2. Iterative Gradient Descent: A optimization routine that minimizes a cost function by iteratively adjusting projection intensities. This method is excellent for highly detailed objects with complex internal structures.
3. Alternating Direction Method of Multipliers (ADMM): A highly advanced optimization solver that excels at handling strict non-negativity constraints (since projectors cannot emit negative light intensity) and binary threshold constraints.

For most high-precision projects, configure the iterative ADMM solver. Set your target projection angle count; a standard baseline is 360 projections spaced evenly at 1-degree increments around a full rotation. Define the step size and the maximum iteration limit. For standard prints, 30 to 50 iterations provide a solid balance between accuracy and computational time.

Step 4: Configuring Photopolymer Dose Response Thresholds

Liquid resins do not react linearly to light accumulation. They require a specific cumulative energy dose before polymer chains cross-link into a solid gel. You must program this non-linear behavior into the solver to ensure your printed parts match the desired geometry.

Configure the mathematical thresholding function within the VAMtoolbox parameters. The toolkit models the resin activation curve using a smoothed Heaviside step function or a sigmoid mathematical curve. Enter your resin's measured critical gelation dose, typically specified in millijoules per square centimeter, along with the corresponding oxygen inhibition constant.

By defining the contrast parameters of this sigmoid function, you instruct the solver to penalize any light projections that deposit a transition dose (near the gel point) at the boundaries. This forces the algorithm to produce high-contrast, steep dose profiles at the object's edges, resulting in sharper physical parts with clean surfaces.

Pro-Tip: If you are using a highly oxygen-inhibited acrylate resin, set your optimization sigmoid slope parameter to a higher value. This models the sharp, threshold-like curing behavior of these materials, enabling the solver to design sharper features at the outer boundaries of your printed parts.

Step 5: Executing the Solver and Verifying Reconstructed Dose

Run the optimization sequence in MATLAB. As the algorithm executes, monitor the convergence metrics displayed in your command window. The solver will track both the objective function value and the constraint violations across each step.

Once computation is complete, the VAMtoolbox generates a reconstructed three-dimensional dose map. Use the built-in slice viewer to inspect this virtual print. Examine cross-sections along the XY, YZ, and XZ planes to verify that the high-dose regions match your target geometry.

Analyze the histogram of the computed dose distribution. A successful optimization run shows a clear, bimodal distribution of energy: one peak representing voxels safely below the gelation threshold (the liquid resin), and a second peak representing voxels safely above the threshold (the solid object), with minimal voxels landing in the intermediate transition zone.

Step 6: Formulating and Exporting the Projection Videos

The final step converts the optimized projection intensity matrices into physical files that your projector can play in synchronization with your print vial's rotation.

Specify your projector's native resolution, such as 1920 by 1080 pixels, and define how your computational voxels map to physical DMD pixels. The VAMtoolbox scales, centers, and pads your optimized 2D intensity matrices to fit this output area. Scale the resulting intensity values to match your projector's color depth, mapping the values to an 8-bit scale from 0 (completely black) to 255 (maximum projector brightness).

Export the compiled sequence of images as an uncompressed multi-frame TIFF stack or a lossless MP4 video container. Ensure your export configuration does not apply lossy compression, as video compression artifacts can alter the carefully optimized light intensities and degrade the resolution of your final print.

VAMtoolbox Parameter Reference and Algorithmic Specifications

The following table outlines the primary control variables within the VAMtoolbox workspace. Adjusting these parameters allows you to balance physical printing accuracy with computational processing time.

Troubleshooting Common VAM Projection Anomalies

When working with tomographic printing systems, discrepancies between your digital CAD files and the physical results often stem from setup errors or miscalibrated optimization settings. Below are common failure points along with instructions on how to resolve them.

Undercured Features and Internal Structural Weakness

• Root Cause: The target dose threshold in your optimization settings was set too high, or your algorithm's convergence limit stopped the solver before it could deliver enough energy to internal structures. This issue can also occur if the optimization model fails to account for light attenuation as it travels deeper through the resin vial.
• Actionable Fix: Lower the target threshold parameter in your configuration file by 10% to 15%. Additionally, enable the Beer-Lambert law attenuation compensation setting in your VAMtoolbox configuration, providing the measured optical absorption coefficient of your resin to ensure the solver boosts projection intensities for deeper features.

Surface Bloating and Loss of Fine Cavities

• Root Cause: This is often caused by setting the optimization sigmoid slope parameter too low, which models a soft polymerization transition. This allows intermediate light doses to accumulate in areas that should remain liquid, curing those features shut.
• Actionable Fix: Increase the sigmoid slope parameter to a higher value to enforce a sharper threshold boundary in the solver. You should also verify that your physical resin's oxygen inhibition layer is performing as expected; if necessary, add a small amount of photoinhibitor (such as Tempo) to your formulation to physically sharpen its curing threshold.

Geometric Distortion and Focal Smearing

• Root Cause: A mismatch between the refractive index values entered into the VAMtoolbox and the physical components of your printing setup. This error causes light rays to bend along paths different from those calculated by the solver, blurring features along the outer radius of the print zone.
• Actionable Fix: Use a refractometer to measure the refractive indices of your photopolymer resin, glass vial, and immersion bath at your printer's operating wavelength. Enter these exact values into the software's optical initialization script and rebuild your projection matrices with the refraction correction routine fully enabled.

Severe Striping Artifacts and Surface Ridges

• Root Cause: An insufficient number of angular projection steps, or a mismatch between the rotational speed of your vial and the playback frame rate of your projector.
• Actionable Fix: Increase the projection angle parameter to at least 360 frames (1-degree increments) or 720 frames (0.5-degree increments). Verify that your projector is playing back the exported video at an exact, uncompromised frame rate that matches your physical rotation platform's speed.

Frequently Asked Questions

What is the difference between Filtered Backprojection and iterative algorithms in the VAMtoolbox?

Filtered Backprojection is a fast, direct mathematical method that calculates projection patterns in a single pass. However, it cannot easily prevent negative intensity values, which can lead to overcuring and background noise. Iterative algorithms, such as those using the Alternating Direction Method of Multipliers, run multiple optimization loops to refine the projection patterns. This approach enforces strict constraints—such as ensuring no projection has negative light values—resulting in much sharper, more accurate prints for complex shapes.

How do I handle refractive index mismatch when setting up my projection geometry?

To prevent refractive index mismatch, place your cylindrical print vial inside a flat-walled square container filled with an index-matched fluid, such as a water-glycerol mixture. You must then measure the refractive indices of your resin, glass vial, and surrounding fluid, and enter these exact values into the VAMtoolbox's optical configuration script. The software will run ray-tracing calculations to adjust the projection patterns and compensate for any remaining light bending at the circular boundaries.

Can the VAMtoolbox handle multi-material or graded volumetric prints?

The standard VAMtoolbox is built to optimize single-material prints with a single, uniform curing threshold. To print multi-material objects, you must write custom scripts that segment your target model into separate material zones, each with its own target dose and threshold. You can then run independent optimization routines for different light wavelengths or project patterns into a multi-component resin system.

What are the memory requirements for high-resolution 3D reconstructions?

Memory usage scales rapidly with your voxel grid size. A grid of 128 cubed voxels is computationally lightweight and runs easily on standard laptops. However, a high-resolution grid of 512 cubed voxels requires handling very large matrices during optimization, which can easily demand over 32 gigabytes of system RAM and several gigabytes of dedicated video memory. For these high-resolution prints, you should run the optimization on a dedicated workstation with a high-performance, CUDA-enabled GPU.

Why does my printed part look bloated compared to the digital CAD target?

Bloated prints usually occur when the light dose in non-target areas rises above the resin's gelation threshold. This is often caused by light scattering within the resin, or by using a solver configuration that does not penalize intermediate, low-level light exposure. To fix this, increase the threshold contrast setting in the VAMtoolbox to sharpen the calculated boundaries, and verify that your physical resin contains enough light absorber to control scattering.

Accelerate Your Volumetric Printing Research

Streamline your additive manufacturing workflows by integrating advanced tomographic optimization models into your design pipelines. By mastering the computational parameters of the VAMtoolbox, you can unlock ultra-fast, layerless 3D printing capabilities for complex engineering components and advanced optical devices.