“Any sufficiently advanced technology is indistinguishable from magic,” wrote Arthur C. Clarke. And with computed axial lithography, or CAL, we’ve finally created magic.
CAL is a layerless 3D printing process invented at the University of California, Berkeley in 2019. There’s no growth, no buildup — parts are made entirely at once. The common question, after a few moments of trying to comprehend what was just said, is “How?”
How It Works
First is the printing material, typically a UV light-cured resin whose chemistry is similar to those used in stereolithography (SLA) printing. The simplest material is comprised of a monomer and a photoinitiator. A monomer is a molecule that can react to grow long molecule chains and intertwine with itself, creating a polymer. The reaction is started by the photoinitiator, a special molecule that will react to a certain wavelength of light. The materials can get much more complex than this, having multiple monomers, or more molecules that react to light differently, or even tiny silica/glass particles. CAL has already been demonstrated with over 60 different materials.
Next is the mechanical setup of the printer. The illustration above shows almost exactly what the patent looks like. The printing material is put into a clear cylindrical glass container (D). There’s a rotation element (C) that rotates the container, and a projector (A) that shines the images (B) that will create the part. Everything else is extra and is just used to improve the printing process. This could include lenses that will better control the quality of the light rays, or an index matching box that will prevent the light from refracting when hits the edge of the rounded glass vial.
The magic of CAL is almost all within the images the projector shows, which have their roots in computed tomography, aka a CT scan. In a medical CT scan, a 2D fan of X-rays is directed through a slice of a patient’s body. Then the X-ray source is rotated to a small degree, and the process is repeated. After several rotations and complex math, a clean slice of this person can be reconstructed — computed by determining how the X-rays absorbed energy through the patient and then using some filtering techniques. Now the spinning X-ray source is moved up and down the person, to get multiple slices. Once the length of the person is fully scanned, the slices can be put together to make a full 3D model of the patient, insides and all.
CAL is essentially a CT scan in reverse. It starts with a 3D model and computes the slices, from multiple angles, that it would take to form that 3D model. This process begins with breaking down the model into voxels, little cubes that are essentially 3D pixels. Then, like a CT scan, a slice is taken from the model. Starting from 0 degrees, a one-dimensional row of varying intensity pixels is calculated, based mainly on how dense the object is viewed from that angle. This is repeated after rotating in small increments, for 180 degrees. (Not 360° — the row of pixels would be identically calculated when viewed from 180°, so there is no need to calculate twice.) The combination of these 1D rows of pixels from various angles is called a sinogram. Then, like in a CT scan, this is done for multiple slices of the model until sinograms are created for all slices.
Now the first version of the projected image can be made. By taking all the 1D rows of pixels from a certain angle (such as 30°) and stacking them on each other, a 2D image is created that accounts for 3D space. However, the first version of this image is going to be blurry and distorted in the wrong ways. Through filtering, optimization methods, and other math tricks, a high-quality 2D image is created. This image can then be projected into our vial, to start forming our part.
But to form the desired geometry, many angles of these 2D projections are needed, which are combined into a 3D projection video (above). A typical video created for projection will have 180 unique frames/angles. This is just the start of how these videos are created. CAL also computes a lot of physics that is involved with the printing process — how the light refracts when hitting the surface, how molecules move around in the printing material when parts form, etc. — and the video can be modified and distorted to account for these effects. There are few physical effects that can’t be corrected for.
Prints in Seconds
The three core parts of CAL are now complete: the material, the mechanical setup, and the special projection videos. Now begins the printing. First the vial will start rotating about its central axis, then the projection video is illuminated through the vial, with the focus traditionally being at the center of the vial. However, because light is passing through the entire volume of the vial at each angle, everywhere within the part is starting to be formed. The material needs a certain dosage of light before it starts to become solid. As the vial is rotated, only the spots where enough combined light has passed through will become solid. This truly causes the part to form all at once, sometimes in as little as 10 seconds.
CAL parts do have to undergo post-processing. The part is removed from the vial with tweezers, washed in a solvent to remove its excess resin, and then exposed to more light under a strong UV LED lamp to reach its full mechanical properties.
Because of the unique way the parts are formed, CAL can do several unique things, such as overprinting. Overmolding is when an existing part, such as a screwdriver bit, has plastic molded around it, creating a screwdriver handle. CAL can do the same thing with printing, forming 3D geometry over existing parts. The examples above were printed in just 20 seconds (rather than 1 minute), sacrificing quality for speed.
CAL also does not need support structures to create parts. As the part forms, it is upheld by the printing material itself; this is why CAL typically uses much more viscous materials than SLA. A common CAL material has viscosity similar to molasses, or even higher.
CAL in Zero Gravity
High-viscosity materials tend to have stronger and more desirable printing properties. However, if there’s a need for lower-viscosity material, several different techniques can be used, the most exciting of which is simply removing gravity, by using CAL in space! CAL has already been demonstrated in microgravity (the more scientific term for “zero gravity”), on parabolic airplane flights (above) where it printed over 400 parts. While the microgravity environments only last for 20 seconds on the parabolic plane, CAL is able to completely form parts. The next step for the space version of CAL is to be tested in suborbital space, and then eventually on a space station!
What’s Next
Since the technology’s invention in 2019 by Brett Kelley and others under the supervision of Dr. Hayden Taylor, much has happened. Not only has the technology been tested in microgravity, but pure glass parts with feature sizes 5x smaller than the width of a human hair have been made! Much more is in the works as well, from optical lens printing, to printing in a half-meter diameter vial, to recyclable materials.
CAL has grown beyond UC Berkeley as well, with more than a dozen institutions researching and expanding the technology. And it’s an open source project, so you can try it yourself. The future of CAL seems limitless.
This article appeared in Make: Volume 88.
ADVERTISEMENT
