Apparatuses, systems and methods for three-dimensional printing

Abstract

The present specification contemplates apparatuses, methods and systems including a volumetric additive manufacturing (VAM) printer with a high angular dose delivery rate including projections optimized for low-viscosity printing conditions. High-fidelity VAM printing in poly(ethylene glycol) diacrylate PEGDA with viscosities as low as 12 cP is disclosed.

Description

APPARATUSES, SYSTEMS AND METHODS FOR THREE-DIMENSIONAL PRINTINGPriority ClaimThe present specification claims priority to Canadian Patent application 3247914, filed June 28, 2024. The entire contents of the foregoing are incorporated herein by reference.Field

[0001] The present disclosure relates generally to digital fabrication and more particularly relates to apparatuses, systems and methods for 3D printing.BackgroundTomographic volumetric additive manufacturing (VAM) is a high-speed 3D printing technique that overcomes many of the challenges faced by conventional layer-by-layer based approaches. However, unlike other vat photopolymerization techniques, in prior art earth-based printing VAM typically uses much higher viscosity resins limiting compatibility with more commonly available lower- viscosity materials. See for example, US 2020 / 0361152 which utilizes resins with at least 1000 cP.Summary

[0002] An aspect of the specification provides a method for three-dimensional printing, the method including: receiving, at a controller, a plurality of multi-angle projection patterns based on a digital model; rotating, under the control of the controller, a build chamber containing a curable material, emitting, under the control of the controller, each projection pattern into the build chamber during the rotation, such that the curable material receives projection energy according to the multi-angle projection patterns; coordinating the rotating with the emitting such that each projection pattern is directed into the build chamber at a corresponding angular position to achieve uniform dose accumulation in the curable material at spatial locations defined by the digital model;and, terminating the rotating and the emitting upon determining that a print-completion condition of the digital model is met.

[0003] An aspect of the specification provides a method further including: receiving a raw digital model in a mesh-based format and generating the plurality of multi-projection angle patterns based on the raw digital model.

[0004] An aspect of the specification provides a method wherein the mesh-based format is based on one of the formats known as STL (.stl), OBJ (.obj), or 3MF (.3mf).

[0005] An aspect of the specification provides a method, wherein the plurality of projection patterns is generated using a feedback loop to reduce a dose error across voxels of the digital model.

[0006] An aspect of the specification provides a method, wherein the feedback loop is combined with histogram equalization to increase overall light dose intensity and reduce in-part dose spread.

[0007] An aspect of the specification provides a method, wherein the digital model includes a voxelized representation of a three-dimensional object, and each voxel is associated with a target light dose.

[0008] An aspect of the specification provides a method, wherein the digital model is generated from the raw digital model by applying a projection pattern optimization operation to generate the plurality of multi-angle projection patterns, the operation including: applying proportional-integral feedback to iteratively update voxel intensities in the digital model based on a comparison between calculated and target dose values; and adjusting the multi-angle projection patterns by performing histogram equalization of pixel intensity values across the projection set.

[0009] An aspect of the specification provides a method, wherein the curable material includes poly(ethylene glycol) diacrylate (PEGDA) having a viscosity as low as 12 cP.

[0010] An aspect of the specification provides a method, wherein the PEGDA has a molecular weight of less than about 700 Da.

[0011] An aspect of the specification provides a method wherein the curable material includes a photoinitiator and monomer(s) and, possibly, a solvent. The monomer(s) may include triethylene glycol dimethacrylate, hexyl acrylate, trimethylolpropane triacrylate, trimethylolpropane trimethacrylate, pentaerythritol tetraacrylate, pentaerythritol triacrylate, 2-hydroxyethyl methacrylate, 2-hydroxyethyl acrylate, ethyl methacrylate, hydroxypropyl methacrylate, hydroxybutyl methacrylate, ethylene glycol methyl ether methacrylate, 1,6-hexanediol diacrylate, 2-hydroxy acrylate, isobornyl acrylate, glycidyl acrylate, glycidyl methacrylate, methacrylate, acrylate, 2-phenoxy ethylacrylate, tert-butyl acrylate, n-butyl acrylate, ethyl acrylate, benzyl acrylate, methyl acrylate, lauryl acrylate, vinyl acrylate, isobutyl acrylate, (2-methoxyethyl) acrylate, 2-ethylhexyl acrylate, ethylene glycol phenyl ether acrylate, acrylic acid, methacrylic acid, hexyl acrylate, hexyl methacrylate, pentaerythritol tetraacrylate, 1 ,4-butanediol diacrylate, 1 ,20-decanediol dimethacrylate, 1,3 -butanediol dimethacrylate, 1 ,4-butanediol diacrylate, 1,4- butanediol diacrylate, 1 ,4-butanediol dimethacrylate, 1,5 -pentanediol dimethacrylate, 1,6- hexanediol dimethacrylate, ethylene glycol diacrylate, ethylene glycol dimethacrylate, or any mixture thereof.

[0012] An aspect of the specification provides a method, wherein the curable material has a of viscosity less than about lOOOcP.

[0013] An aspect of the specification provides a method, wherein the build chamber is rotated through an angular extent of approximately 1 / n of a full rotation to deliver a complete set of the multi-angle projection patterns, where n is the number of projection patterns spaced about the build chamber.

[0014] An aspect of the specification provides a method, wherein the projecting step includes emitting projection patterns from a plurality of projectors spaced at equal angular intervals about the build chamber, each projector emitting a subset of the projection patterns in phase with the rotation of the build chamber.

[0015] An aspect of the specification provides a method, wherein the coordinating of the rotating and the emitting includes: receiving angular position data from a rotation sensor; and transmitting projection patterns to the plurality of projectors based on the angular position data such that eachprojector emits a projection pattern phase-aligned to its respective position around the build chamber.

[0016] An aspect of the specification provides a 3D printer system including a controller, a turntable and at least one projector according to any of the foregoing methods. An aspect of the specification provides a controller according to 3D printer system.

[0017] An aspect of the specification provides a non-transitory computer-readable medium configured to store programing instructions according to the foregoing and executable the controller.

[0018] An aspect of the specification provides a printed digital model produced according to any of the foregoing. An aspect of the specification provides a mold for a product produced from the printed digital model.Brief Description of the Figures

[0019] The present specification includes the attached Figures, in which:

[0020] Figure 1 shows a schematic representation of a three-dimensional printing system in accordance with an embodiment.

[0021] Figure 2 shows schematic representation of the controller of Figure 1.

[0022] Figure 3 shows a flowchart depicting a method for three-dimensional printing in accordance with another embodiment.

[0023] Figure 4 shows the system of Figure 1 when operated according to an example performance of the method of Figure 3.

[0024] Figure 5 shows the rapid volumetric printer. Figure 5 (a) shows the printer system. Figure 5 (b) shows an verhead view of the printer. To achieve high-speed printing, three projectors are spaced at 120 degree increments about the vial. Figure 5 (c) shows a lock diagram of the printer hardware.

[0025] Figure 6 Shows an increasing print rate through histogram equalization of tomographic projections. Histograms of 8-bit intensity values for tomographic projection sets computed with Figure 6 (a) OSMO and Figure 6 (b) PIHE. Sample tomographic projections for a 3DBenchy computed using OSMO and PIHE are shown in Figure 6 (c) and Figure 6 (d) respectively. (3DBenchy can be found at https: / / www.3dbenchy.com / )

[0026] Figure 7 shows a comparison of low-viscosity optimization and object-space model optimization. Figure 7 (a) shows a target design based on 3DBenchy. Figure 7 (b) shows a lice of target design indicated by the plane in Figure 7 (a). Scale bar is 5 mm. Figure 7 (c) shows convergence of VER and IPDS over 30 iterations for both optimization methods. Figure 7 (d) shows ose histogram for PIHE. Figure 7(e) shows thresholded dose at different values to show the print dynamics. Figure 7 (f) shows the same as in Figure 7 (d) but for OSMO. Figure 7 (g) is the same as in Figure 7 (e) but for OSMO.

[0027] Figure 8. Demonstration of low-viscosity printing for three different target designs. Figure 8 (a) A 3DBenchy part. Figure 8(b) through Figure 8(d) shows photographs of 3D benchmark printed using Polyethylene Glycol Diacrylate700 or PEGDA700, PEGDA575, and PEGDA250 respectively. Figure 8(e) through Figure 8(g) shows signed distance function (SDF) of the micro- CT isosurface and the target design in Figure 8 (a). Figure 8 (h) shows a microfluidic splitter. Figure 8 (i) through Figure 8(k) is the same as in Figure 8 (b) through Figure 8(d) but for the part in (h). (1-n) Same as in Figure 8(e) through Figure 8(g) but for the part in Figure 8(h). Figure 8(o) shows a gyroidal lattice. Figure 8(p) through Figure 8(r) is the same as in Figure 8(b) through Figure 8(d) but for the gyroidal lattice in Figure 8 (o). Figure 8(s) through Figure 8(u) is the same as in Figure 8(e) through Figure 8(g) but for the part in Figure 8 (o). All scalebars are 5 mm.

[0028] Figure 9 shows Optical scatter imaging of print volume at different times during the printing process of a 3DBenchy. The exothermic reaction of polymerization causes the parts to rise in the print volume following gelation. Figure 9 (a) shows the larger in-part dose spread of OSMO causes different part regions to form at different times. This results in the hull (1) of the 3DBenchy forming first, and rising before the additional dose required to form the cabin (2) to form. Figure 9 (b) shows the photograph of 3DBenchy printed using the method in Figure 9 (a). Figure 9 (c) shows he reduction of in-part dose spread in PI (no histogram equalization) results ina sharp onset of gelation for all part regions but at a comparable time to OSMO, (d) Photograph of 3DBenchy printed using the method represented in Figure 9 (c). Figure 9 (e) is the ame as in Figure 9 (c) but with PIHE. Figure 9 (f) shows a photograph of 3DBenchy printed using the method in Figure 9 (e). Scalebars are 5 mm.

[0029] Figure 10 shows example printing of hydrogel lattice structures with PIHE. Figure 10 (a) shows a photograph of as-printed part in cylindrical water bath. Inset: Digital model of the target design. Figure 10 (b) shows an as- printed part on microscope slide in air. Figure 10 (c) is the ame part as in Figure 10 (b) but after 120 minutes of drying at STP. Figure 10 (c) is the ame part as in Figure 10 (b) but after rehydration with red-colored water. All scale bars are 5 mm.Detailed Description

[0030] Figure 1 shows a three-dimensional (3D) printing system 100 according to a non-limiting example embodiment. System 100 includes three projectors 104, namely, projector 104-1, projector 104-2 and projector 104-3. (Collectively, projectors 104; and, generically, projector 104). In a present embodiment each projector 104 is disposed on the same plane, about one-hundred- and-twenty degrees apart from each other, and therefore spaced substantially equally, around the periphery of a build chamber 108. Build chamber 108 can be implemented as clear cylindrical vessel that holds a curable material 112. Build chamber 108 is affixed to a turntable 116 that can rotate build chamber 108 at a given angular velocity co.

[0031] A light source 118, such as a red light emitting (LED) diode is disposed above build chamber 108 to illuminate build chamber 108 during rotation. Furthermore, each projector 104 projects a respective cone 120 towards the middle of build chamber 108, where an object can be formed or “3D printed” according to the present specification. The cones 120 thus converge on a target zone 124 substantially in the middle of build chamber 108. As will be elaborated further below, the formation of the object from curable material 112 is a result of the targeted energy fromeach projector 104, which causes a portion of curable material 112 to cure in the shape of a predefined 3D object within target zone 124.

[0032] A controller 132 connects to projectors 104 and turntable 116 via a bus 136 for transmitting control signals and receiving feedback signals. A laptop computer 140 connects to controller 132 to provide a human-machine interface (HMI) for system 100, and to provide specific instructions to controller 132 and receive logs or other data regarding the operation of projectors 104 and / or turntable 116. Laptop computer 140 can be implemented as any type of computing device that functions as an HMI such as desktop computer, tablet computer, mobile phone, work station or the like. Optionally, a network 144, such as the Internet or an intranet or combination thereof, can connect to laptop computer 140 to expand the data management functions of system 100.

[0033] Collectively, projectors 104, build chamber 108, curable material 112, turntable 116, light source 118, controller 132 and bus 136 are referred to herein as a volumetric additive manufacturing (VAM) printer 138.

[0034] System 100 is shown again in Figure 2 for further illustration. Notably Figure 2 includes a top view of build chamber 108 and turntable 116 showing projectors 104 around the periphery of build chamber 108. Light source 118 is not shown in Figure 2. A schematic diagram of a nonlimiting example of the internal components of controller 132 is also shown.

[0035] Controller 132 may include at least one input device 204 for providing input signals to a processor 208, which in turn may send output signals to at least one output device 212. Input device 204 may include traditional input mechanisms like keyboards or mice. Output device 212 may be a display or could be a speaker for providing real-time feedback. Processor 208 may be implemented as a plurality of processors, one or more multi-core processors, field programmable gate arrays (FPGAs) or specialized hardware accelerators, such as Graphics Processing Units (GPUs), Tensor Processing Units (TPUs), or similar parallel processing units optimized for handling large-scale data and / or complex machine learning models. In certain embodiments, processor 208 may include a memory-centric or near-memory compute architecture, where processing elements are co-located with memory cells to minimize data transfer latency and power consumption, which is beneficial for applications involving high-speed inference. Processor 208 may be configured to execute different programming instructions, including those optimized forAl and machine learning tasks, responsive to input received via the one or more input devices 204 and to control the one or more output devices 212 to generate output on those devices.

[0036] To fulfill its programming functions, processor 208 is configured to communicate with one or more memory units, including non-volatile memory 216 and volatile memory 220. Non-volatile memory 216 can be based on any persistent memory technology, such as an Erasable Electronic Programmable Read-Only Memory (EEPROM), flash memory, solid-state hard disk (SSD), other types of hard disks, or combinations thereof. In embodiments using a memory-centric architecture, memory 216 may include processing units embedded within or adjacent to memory cells to reduce latency and improve energy efficiency in data handling, supporting high-performance inference tasks. Non-volatile memory 216 may also be described as a non-transitory computer-readable medium, with multiple types of non-volatile memory 216 provided as needed.

[0037] Volatile memory 220 can be based on any random access memory (RAM) technology, such as Double Data Rate (DDR) Synchronous Dynamic Random-Access Memory (SDRAM). In certain embodiments, volatile memory 220 may incorporate high-speed, low-latency configurations, supporting rapid data access and processing required for real-time inference tasks. Other types of volatile memory 220, such as Low-Power DDR (LPDDR) or High-Bandwidth Memory (HBM), are also contemplated to optimize performance in high-computation environments.

[0038] Processor 208 connects to an interface 232. Interface 232 connects to projector 104 and controller 132 via bus 136, and to laptop computer 140. Thus, various peripherals may connect to controller 132 via interface 232 thereby potentially obviating the need for input device(s) 204 and / or output device(s) 212 altogether. The interface 232, when connected to network 144 via laptop computer 140, may facilitates any external interactions such as firmware updates for controller 132, uploading models to be printed in build chamber 108, or sending logging data to a central server (not shown).

[0039] Programming instructions in the form of software modules 224 are typically maintained, persistently, in non-volatile memory 216 and used by the processor 208 which reads from and writes to volatile memory 220 during the execution of software modules 224. Various methods discussed herein can be coded as one or more software modules 224. One or more tables ordatabases 228 are maintained in non-volatile memory 216 for use by software modules 224. Software modules 224 may include execution logic for distributed transactions, real-time coordination of dependent tasks, synchronization management, and execution prioritization. Databases 228 may store execution logs, state dependencies, transactional data, and historical processing records. Software modules 224 also include executable code for the various methods, and their variants, described herein. Databases 228 may store execution logs, transaction records, processing history, and other structured data relevant to distributed execution workflows.

[0040] Overall, the controller 132 may be implemented using cloud computing platforms such as Microsoft Azure™ or Amazon Web Services (AWS)™. Furthermore engine 212 may also be implemented as virtual machines.

[0041] By the same token, a plurality of controllers 132 may be provided. For example, controller 132 can be implemented as a plurality of separate local computing devices that can distribute the functions of controller 132. For example a Raspberry-Pi™ controller can be connected to each projector 104. Further, a device such as an Arduino Due™ maybe used with a camera (not shown) or other input sensor to ascertain rotation speed of turntable 116. Thus, it is to be understood that the overall hardware configuration of controller 132 is not particularly limited.

[0042] A person of skill in the art will recognize that the core elements of processor 208, input device 204, output device 212, non-volatile memory 216, volatile memory 220 and network interface 232, as described in relation to controller 132, have analogues in the different form factors of different computing machines such as those that can be used to implement laptop computer 140.

[0043] Furthermore, it will be understood that in other variants, laptop computer 140 and controller 132 may be combined into a single computing device.

[0044] Figure 3 shows a flowchart depicting a method for three-dimensional printing indicated generally at 300. Method 300 can be implemented on system 100, though persons skilled in the art may implement variations thereof, including omitting certain blocks, performing steps in parallel, or reordering the sequence based on system configurations. For explanatory purposes, method 300 is described in the context of system 100, where it operates as, for example, software applicationsexecutable by, and maintained within persistent storage, of laptop computer 140 and persistent storage controller 132, interacting with other nodes as illustrated in Figure 3.

[0045] Initially, a high level overview of method 300 will be discussed. Thereafter, a detailed elaboration of various blocks in method 300 will be provided, including variations.

[0046] Block 304 comprises receiving a raw digital model of an object to be printed. To elaborate, this model can be received in any suitable file format. As will be discussed in greater detail below, it is contemplated that block 304 can include, but is not limited to, any pre-existing type of digital model to be printed according to any existing tomographic two-dimensional (2D) additive layering techniques. Referring now to Figure 4, in the context of the example system 100, a digital model 404-A of a ship 408-A is received at laptop computer 140. Model 404-A can come from any source and be inputted into laptop computer 140. According to the example in Figure 4, model 404-A is a pre-existing model as stored somewhere on network 144 and is loaded onto laptop computer 140.

[0047] Block 308 comprises generating projections based on the received model. These projections may be generated using one or more optimization algorithms which will be discussed in greater detail below. Referring to Figure 4, digital model 404-B is created by laptop computer 140 to create digital model 404-B, which modifies the electronic model of ship 408-A into the electronic model of ship 408-B. Model 404-B is structured to be instructive to controller 132, which in turn controls turntable 116 and projectors 104 according to the instructions represented by digital model 404-B. To briefly elaborate, block 308 comprises generating a plurality of multiangle projection patterns respective to different voxels within target zone 124, whether the complete pattern of voxels combine to result in physically generating ship 408-C.

[0048] Block 312 comprises controlling the turntable and, optionally, one or more illumination sources. In the context of the example of Figure 4, block 312 is performed under the control of controller 132. Block 312 thus contemplates rotating, under the control of the controller, a build chamber containing a curable material, As will be discussed in greater detail below, block 312 comprises activating turntable 116 at a defined rotational speed (or angular velocity) co. A camera (not shown) or other sensor may be included to measure the rotational speed and deliver it back to controller 132. Light source 118 is activated at this point, if not already activated. Block 312 includes setting the rotational speed co of the build chamber 108 to correspond with projectiondelivery timing (discussed at block 316), and enabling consistent illumination conditions across the build volume according to the chemical characteristics of curable material 112.

[0049] Block 316 comprises controlling the projectors to emit the generated projections into the target zone 124 of build chamber 108. Block 316 is performed under the control of controller 132. Block 316 thus contemplates emitting, under the control of the controller, at least one projection pattern into the build chamber during the rotation, such that the curable material receives projection energy according to the multi-angle projection patterns. As will be discussed in greater detail below, each projector 104 may be phase-offset according to its angular position relative to the build chamber 108, with the timing and sequence of projections governed by synchronized control signal from controller 132 according to digital model 404-B.

[0050] Block 320 comprises determining whether the printing process is complete. As will be discussed in greater detail below, this determination may be based on cumulative rotation, total projection delivery, elapsed time, or other programmable criteria associated with successful print completion. If the determination is “no” at block 320, then block 312 and block 316 continue, with ongoing coordination of rotating (block 312) and the emitting (block 316) from the projectors such that each projection pattern is directed into the build chamber at a corresponding angular position to achieve uniform dose accumulation in the curable material 112 at spatial locations defined by the digital model 404-B.

[0051] According to the example in Figure 4, the result of successful completion of block 320 is the formation of a physical copy of a ship 408-C within target zone 124, that realizes the model from ship 408- A as modified into ship 408-B.

[0052] Block 324 comprises removing the formed object. To elaborate, this may include pausing the turntable 116, ceasing projection emission from projectors 104, and initiating post-processing steps such as rinsing, curing, or visual inspection. In the example of Figure 4, block 324 comprises removing the now-solid ship 408-C from build chamber 108, leaving any residual curable material 112, still in liquid form, behind.

[0053] It is also to be understood that various blocks within method 300 itself may be implemented as multiple embodiments that each stand alone. For example, block 304 and block 308 may standalone. Block 304 and block 308 comprise the generation of digital model 404-B from model 404- A. Digital model 404-B is specifically configured for controller 132 working in conjunction with build chamber 108, curable material 112, projectors 104, light source 118 and turntable 116. Thus, other ways of generating digital model 404-B are contemplated, that are independent from converting a preexisting 2D model such as model 404-A. Likewise, block 312 and block 316 may stand alone. Block 312 and block 316 comprise realization of ship 408-C as turntable 116 and projector 104 are controlled by controller 132 according to digital model 404-B.

[0054] Having provided a high-level overview of method 300, a more detailed discussion follows elaborating on various blocks.

[0055] The present specification notes that low-viscosity poly(ethylene glycol) diacrylate (PEGDA) has seen wide usage in bioprinting techniques but has, until this specification, eluded printing in volumetric additive manufacturing (VAM) under standard Earth gravity conditions. (Prior work using low- viscosity resins in VAM has largely been limited to microgravity environments, where buoyancy and flow-related constraints are absent.) Using a VAM printer (such as printer 138) with a high angular dose delivery rate, as well as tomographic projections optimized for low-viscosity printing conditions, we demonstrate high-fidelity VAM printing in Polyethylene Glycol Diacrylate (PEGDA) with viscosities as low as 12 centipoise (cP). Microcomputed tomography imaging of printed parts reveal close-to voxel resolution limited performance. Furthermore, we have demonstrated the first direct printing of a low- viscosity hydrogel in VAM. The proposed method expands the viscosity range, and in turn the catalogue of materials accessible to VAM, giving this printing modality the broadest viscosity range of any vat photopolymerization technique.

[0056] For context, the present specification builds on known tomographic volumetric additive manufacturing (VAM), a type of 3D printing in which parts are fabricated in a photoresin suspension using the concepts of computed tomography (CT). VAM falls under the category of vat photopolymerization techniques. However, in contrast to these layer-by-layer approaches, VAM parts are printed volumetrically enabling the rapid fabrication of parts with enhanced design freedom [1,2]. For instance, VAM has been used to fabricate glass parts [3], optical elements[4], and medical applications [2,5-7]. In VAM, since parts are fabricated in a resin suspension, theviscosity must be much higher than other vat photopolymerization techniques (1000 - 100,000 cP) to prevent any relative motion between the writing beams and the voxels containing the fabricated object [8]. During the photopolymerization process, voxels that receive sufficient light dose to gel will undergo a density change that is dependent on the rate of polymerization. Computational fluid dynamics simulations by Salajeghe et al. [9,10] demonstrated that in the absence of heating effects, the increase in density during gelation will result in sedimentation of the printed part. However, when heating effects are included, they observed that the density change is reduced leading to reduced sedimentation. Furthermore, if the rate of heating is large enough then the part may initially float due to thermal expansion of the printed part. This effect was experimentally observed in Waddell et al.

[0011] for a part printed in poly(ethylene glycol) diacrylate (PEGDA) (molecular weight 700, viscosity p = 120 cP). To prevent density-related effects from shifting the part during printing, Wadell et al. conducted printing in a micro-gravity environment and observed improved print fidelity. For on-Earth VAM, the lowest reported viscosity resin used that lead to a successful print was p = 1100 cP as reported by earlier work from our group using a blend of Diurethane Dimethacrylate (DUDMA) and PEGDA700 in an 8:2 wt ratio

[0012] . During the approximately 1 minute print time, no sedimentation or floating of the printed part was observed.

[0057] The present specification introduces tomographic volumetric additive manufacturing (VAM) of low viscosity PEGDA in an example embodiment of curable material 112. PEGDA has found extensive use in biomedical applications such as cell culture studies

[0013] drug delivery applications

[0014] , and tissue engineering [15-17]. This success can be attributed to the attractive properties of PEGDA, such as its cyto-compatibility and hydrophilicity

[0018] . With the advent of additive manufacturing (AM), PEGDA has been a popular material in 3D bioprinting applications using techniques such as digital light processing (DLP) [19-22] and Stereolithography Apparatus (SLA)

[0023] . In particular, the freeform design nature of AM has enabled the bioprinting of complex geometries, such as sinus implants

[0021] , scaffolds for tissue engineering

[0016] , bone engineering

[0024] . Both DLP and SLA printing modalities print the objects in a layer-by-layer fashion, necessitating the requirement of photoresins with low viscosity. For bioprinting in PEGDA, this typically requires that the molecular weight be less than 1000 Dalton (Da)

[0025] . Lower molecular weight PEGDA are useful for applications requiring high strength, since they have a higher Young’s moduli and tensile strength because they retain less water than their higher molecular weight counterparts

[0026] . VAM printing of low-viscosity PEGDA-containing hydrogel wasrecently demonstrated by Riffe et al.

[0027] In this work, the hydrogel was mixed with gelatin such that at room temperature the material is sufficiently viscous for VAM printing. Removal of the printed part from the semi-solid photoresin was achieved by heating the material and liquefying the uncured gelatin network. As a result, this approach requires characterization of gelatin concentration as well as additional post-processing steps (e.g., heating) that add complexity to the printing process.

[0058] Tomographic printing of low-viscosity photoresins for curable material 112 has additional advantages beyond printing of PEGDA. First, the material catalogue for VAM is now larger enabling possible new application areas, as well as leveraging existing materials in the more widely adopted DLP and SLA modalities. Second, the use of low-viscosity materials simplifies the printing process as it is easier to remove partially-cured resin from as-printed parts. This is especially important in VAM since unlike DLP and SLA, all regions of the print volume receive some level of light dose. This becomes problematic for negative features, such as interior voids (e.g., hollow tubes) or exterior recesses (e.g. teeth on a gear), that may gel due to unwanted polymerization before removal. Third, lower viscosity materials produce smoother surfaces since uncured photoresin is more easily removed and also due to larger dose-diffusion lengths stemming from increased oxygen-induced polymerization inhibition. In a recent report

[0028] , oxygendiffusion was shown to smoothen features (on the projector pixel scale) in VAM-printed parts. While not advantageous for high-resolution printing, increased smoothness due to oxygendiffusion could be useful in lens fabrication applications for which VAM has a demonstrated capability [3,4,29]. Finally, in vat-photopolymerization methods like VAM, the reuse of exposed photoresins can lower material costs. After printing, regions exposed to light will have depleted levels of photoinitiator and oxygen that require replenishment prior to being used again for printing. Low-viscosity photoresins are advantageous in this aspect, as they are easier to mix enabling faster and more uniform redistribution of un-consumed photoinitiator and oxygen to the photoresin.

[0059] In this work, we present a two-pronged approach to enable printing of low-viscosity PEGDA in VAM. Using a triple projection VAM printer (e.g. Printer 138) capable of a high angular dose delivery rate (< 10 seconds print time), coupled with a new tomographic optimization method, we demonstrate on-Earth printing of photoresins with viscosities as low as 12 cP. Wevalidate our print quality using micro computed tomography and show close-to voxel resolution limited performance across varying geometries. We conclude by discussing the importance of optimization on print quality, followed by showing the potential the proposed method has on future bioprinting applications by demonstrating direct printing of a low-viscosity hydrogel material.

[0060] 2. Materials and methods

[0061] 2.1. Tomographic printer for rapid printing

[0062] The design of a high-speed tomographic printer (such as printer 138) is ultimately limited by the frame rate of the projection source (or sources, such as projectors 104) and the angular sampling requirement of the tomographic printing process. For a typical tomographic printer using telecentric projection and non-diverging beams, the angular sampling requirement (in radians) to utilize the full resolution of the projector is given by 80 < x / R, where R is the radius of the geometry and x is the size of the writing beam at that radius. For the parts printed according to the present specification, including system 100 and method 300, , R = 8.5 mm, and x = 0.0966 mm, resulting in a 80 = 0.66°. To account for the etendue of the projection source which will further enlarge the pixel size, we increase this value to 80 = 1°. The frame rate of a typical DLP projector is 60 Hz. Therefore, the maximum vial rotation rate to achieve the above angular sampling requirement is 60 deg / second, or 6 seconds per rotation. In practice, multiple vial rotations are used during printing to achieve a uniform dose delivery to the print volume and to achieve high print fidelity. As a result, printing time is on the order of 30 - 60 seconds. In printing with low-viscosity resins this is problematic as the light dose must be quickly delivered before buoyancy-related effects shift the partially-cured parts away from the writing beams.

[0063] As shown in Figure 5(a), we have addressed the sampling issue by using three synchronized WinTech 4710 DLP projectors (i.e. to implement projectors 104) equally spaced about the print volume or target zone 124. The frame rate of each projector is still 60 Hz, however now the vial (e.g. build chamber 108) only needs to undergo 1 / 3 of a rotation to achieve an equivalent 1- complete rotation of dose. With an angular sampling of 1 deg, this means that a complete dose can be delivered as quickly as 2 seconds. In practice, the angular sampling can be relaxed further to 4 deg (resulting in a build chamber 108 rotation rate of 240 deg / s) to achieve good printing in the lo west-viscosity resin studied. An Overview Titan servo motor for turntable 116 rotates the buildchamber 108. A block diagram of how synchronous control of the rotation stage and the projection patterns was achieved is shown in Figure 5(c), which generally shows the same structure as Figure 2, except that controller 132 is divided into separate components. Specifically, an Arduino Due microcontroller 132a-4 was used to read the rotation stage angle of build chamber 108 and relay this information to three separate Raspberry Pi microcomputers (ie. Raspberry Pi 132a- 1; Raspberry Pi 132a-2; and Raspberry Pi 132a-3) that are responsible for controlling each of the three projectors (104), respectively. For every 1 -degree increment of rotation, each Raspberry Pi 132a transmits a corresponding tomographic projection to its respective projector 104. To account for the different position of each projector 104, the tomographic projections are each shifted by a corresponding angle such that each projector is in phase with the build chamber 108. That is, at time = 0 seconds, projector 104-1, projector 104-2, and projector 104-3 transmit tomographic projections at theta = 0°, 120°, and 240° respectively. To minimize any angular velocity gradient that may form among different axial layers in the print volume, the build chamber 108 is slowly accelerated over three rotations prior to light exposure.

[0064] 2.2. Optimization of tomographic projections for low-viscosity materials

[0065] In this specification, we introduce an optimization procedure that minimizes the in-part dose spread (IPDS) such that all in-part voxels (e.g. that contribute to formation of an object such as ship 408-C) receive the same dose, and in turn gel within curable material 112 at substantially the same time. This is similar to the in-part dose range introduced by Rackson et al.

[0030] but instead of quantifying the range of in-part dose, instead the spread (specifically the standard deviation) is computed. In this way, the impacts of buoyancy-related effects on print quality are reduced or even minimized since all voxels will gel simultaneously (or substantially so) enabling complete delivery of light dose before those voxels translate from the writing beams respective to each cone 120. Consequently, the proposed method is designed for a binary target dose distribution.

[0066] The method used to generate tomographic projections considered the object-space model optimization (OSMO) algorithm of Rackson et al.

[0030] as well as proportional-integral-derivative controller (PID) used in industrial control systems

[0031] . However Rackson et al

[0030] and other industrial control systems

[0031] were simply inadequate and did not work. In this specification, we enhanced the OSMO approach of iteratively modifying a voxelized object Q (in object space)such that the corresponding tomographic projections will produce a dose M that matches the target design. We retain the two user-defined parameters to threshold the dose: 1. A low value, Dlow, representing a maximum value of dose that the print volume can receive before the onset of gelation. 2. A high value, Dhigh, representing a minimum dose value that will result in sufficient gelation to produce the solid part. However, instead of sequentially optimizing the object based on first the low dose threshold followed by the high dose threshold, resulting in two dose calculations per iteration, the specification optimizes both at the same time, resulting in a single dose computation per iteration. The error for iteration u between the dose, M, and these parameters are as follows:Errors

[0067]

[0068] The values iinand ioutare the linear indices for voxels that should gel (i.e. inside the part) and those that should not gel (i.e. outside the part), respectively. Since dose thresholding is retained, neither in-part dose above the high-threshold or out-of-part dose below the low-threshold are penalized during the optimization. In this work, we introduce a new set of equations to update Q.

[0069]

[0070] Here, we present in these teachings a proportional-integral (PI) feedback. The terms kpand ki are the proportional and integral coefficients respectively, which were set to kp= 1 and ki = 0.5 for this work. The proportional term is identical to OSMO except that it is computed for low andhigh thresholds at the same time. However, during initial testing we observed that the dose for certain in-part voxels would not reach Dhigh, but instead plateau at a lower dose value. This is problematic in printing in VAM with low-viscosity materials because the effects of buoyancy will cause gelled voxels to translate from their initial location. As such, it is necessary that all in-part voxels converge to Dhigh so that the onset of gelation occurs uniformly for all in-part voxels. This steady-state error observed in P-type feedback loops (e.g. OSMO) is the exact problem that a proportional-integral (PI) feedback loop compensates for, and is the reason why we have implemented it here.

[0071] To reduce or minimize the impact of buoyancy effects on print quality, we increase or maximize the gelation rate by increasing the overall intensity of the tomographic projections using histogram equalization during optimization. Eet T and c represent the mean intensity and standard deviation for the tomographic projection set T. The first step of histogram equalization is the removal of intensity values below a lower threshold, T - 3o, as shown in Eq. (5). Following this thresholding operation, the tomographic projections are then scaled by an upper threshold, T + 3o, and clipped to the maximum intensity value (e.g. 255 for the 8-bit compression used in this work), as shown in Eq. (6).

[0073] Since the optimization is built upon a PI feedback loop with histogram equalization, we refer to this method as proportional-integral histogram equalization (PIHE). In Figure 6(a) and Figure 6(b) are shown histograms of the 8-bit pixel intensity for tomographic projection sets of a 3DBenchy

[0032] computed using OSMO and PIHE respectively. In OSMO, most pixels have low intensity values due to large positive outliers in the tomographic projections. In contrast, our histogram equalization approach populates more of the higher-pixel values, increasing the overall intensity of the tomographic projections and light dose delivery rate. The effect can be visualized more easily in Figure 6(c) and Figure 6(d) where we show sample tomographic projections forboth methods. As a result of this process, histogram equalization on the tomographic projection results in a general increase in pixel value, and clipping of some pixels at the low- and high- extremes. An example for a typical 3D printing benchmark, called a 3DBenchy, is shown in Figure 6. To account for the non-telecentric optical configuration, we utilize 3D ray-tracing (3DRT) that we introduced in an earlier report

[0012] but now supports graphical processing unit (GPU) acceleration enabling a significant speed-up in computation. For example, for the projections used in this work (image pixel size = 96.6 pm, cubic voxel side length = 96.6 pm), optimization was completed in under 10 minutes on a NVIDIA RTX 2060 m GPU, with a per-iteration computation time under 20 seconds.

[0074] The impact of such an optimization approach on print dynamics is shown in Figure 7. Here, we compute the tomographic dose delivered to the print volume for a 3DBenchy using PIHE and OSMO methods. In Figure 7(a) is the target design (e.g. Ship 408-C) , and Figure 7(b) is a slice of the target at the indicated plane in Figure 7(a). The voxel error rate (VER) and the IPDS are plotted in Figure 7(c). Here, we use the definition from Rackson et al.

[0030] where the VER is defined as the ratio of the number of out-of-part voxels with dose exceeding the lowest in-part voxel dose. For both optimization approaches, VER and IPDS converge towards zero, with PIHE achieving both a lower VER and IPDS. In Figure 7(d) and Figure 7(f) and Figure 7e) and Figure (g) are the dose histograms and thresholded dose for OSMO and PIHE respectively. In the dose histograms, we observe a narrower IPDS (shown by the dashed line, referenced from the mean in-part dose D) in agreement with the IPDS curve in Figure 7(c). We can observe the impact of this in the thresholded dose. To make a fair comparison between both methods, we threshold the dose at three different values such that the number of gelled voxels, Ngel, is equal. As a consequence of the differing optimization approaches, this results in slightly different values for the dose threshold. In OSMO, we see that the central region of the part forms first followed by exterior features (such as the feature indicated by the white circle) forming later. In contrast, in PIHE there is a redistribution of dose resulting in both central and exterior features receiving dose. We investigated this redistribution of dose further by plotting the computed dose at three different points in the Target Slice (numbers 1, 2, and 3 in Figure 7(b)) and show the results in Figure 7(h). Small features like 1 and 3 rapidly converge to Dhigh using PIHE optimization, whereas dose computed using OSMO plateau at a lower value. These results are typical for PI- and P- feedbackloops respectively. Notably, bulk features like feature 2 converged to the same dose value for both methods suggesting that these features have minimal steady-state error.

[0075] 2.3. Materials

[0076] The three standard photoresins and one hydrogel photoresin studied in this specification as different for curable material 112 are all based on poly(ethylene glycol) diacrylate (PEGDA). For the standard resins, PEGDA with molecular weights of 250 Da, 575 Da, and 700 Da were mixed with 1.75 mM Ethyl (2,4,6-trime-thylbenzoyl) phenylphosphinate (TPO-L) as a photoinitiator. The hydrogel photoresin was synthesized by mixing 65 % (w / v) PEGDA (molecular weight = 700 Da) with deionized water and 1.51 mM Lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) (95 %) as the photoinitiator. PEGDA variants and LAP were purchased from Sigma Aldrich, and TPO- L was purchased from Oakwood Chemical. Refractive indices were measured using a Schmidt- Haensch ATR-BR refractometer. The viscosity (at standard temperature and pressure, STP) was measured using a TA Instruments Discovery HR20 Rheometer. The measured refractive index (at 405 nm) and STP viscosities for the PEGDA materials studied in this work are listed in Table 1.Table 1

[0077] 2.4. X-Ray Micro-CT measurements

[0078] Printed parts were scanned using a Skyscan 1275 (Bruker) x-ray micro-computed tomography (micro-CT) machine with a pixel resolu-tion of 25 pm and reconstructed using Skyscan NRecon (Bruker). Signed-distance fields were extracted from the scans using MeshLab.

[0079] 3. Results and discussion

[0080] 3.1. Low-viscosity printing of varying geometries

[0081] To demonstrate the versatility of PIHE, three different test geometries of different digital models 404 were printed in PEGDA with molecular weights of 250 Da, 575 Da, and 700 Da. For purposes of clarity we refer to these as PEGDA250, PEGDA575, and PEGDA700 respectively. These geometries are shown in Figure 8(a) Figure 8( h) and Figure 8(1) and are as follows: 1. A 3DBenchy (e.g. Ship 408-C) (height = 12 mm), a standard test part used in fused deposition molding 3D printing

[0032] . 2. A microfluidic 3-1 splitter, with a channel inner diameter of 2 mm. 3. A 3x3 lattice of gyroid unit cells (side length 4 mm, wall thickness 0.9 mm). Dose threshold values of Dlow = 0.7 and Dhigh = 0.9 were used for the micro-fluidic splitter and gyroidal lattice, whereas Dlow = 0.8 and Dhigh = 0.9 was used for the 3DBenchy.

[0082] In this work, we found that lowest printable viscosity resin was dependent on the target design. For both PEGDA575 and PEGDA700, we observed good conformity of the printed part to the target design for all three geometries. For the part printed in PEGDA250 we only see faithful reproduction of the microfluidic splitter (Figure 9(k)) and gyroidal lattice (Figure 8(r)). The 3DBenchy (Figure 8(d)) is missing the bow (indicated by the white arrow in Figure 8(d)). The print fidelity for each part was assessed using micro computed tomography (pCT) and the computed signed-distance functions (SDF) are shown in Figure 8. With exception of the 3DBenchy printed in PEGDA250, all parts had SDF values within + / - 0.5 mm. The average root mean square (RMS) of the SDF for all three geometries was 141 pm, 154 pm, and 227 pm for PEGDA700, PEGDA575, and PEGDA250 respectively. In comparison to the most well- resolvable central voxel used in this work (side length = 96.6 pm), these RMS values are slightly larger indicating we have achieved close to voxel resolution-limited performance for this printer configuration. Individual RMS values for each geometry and resin are tabulated in Table 2 for each of the prints in Figure 8.Table 2

[0083] For PEGDA575 and PEGDA700, we observed that the print time was geometry dependent and not dependent on the different viscosities. However, for PEGDA250 we observed print times that were -30-50 % longer than the higher molecular weight variants. Since all three variants have the same type and concentration of photoinitiator, we believe this difference is due to increased polymerization-inhibition rate due to oxygen diffusion and reduced polymerization in PEGDA250

[0033] . In a previous study it was shown that diffusion of oxygen during the printing process causes smaller feature sizes to gel at a slower rate than larger features

[0028] . Due to the binary target dose optimization used in this work, the pre-correction introduced in Orth et al. to account for dose diffusion cannot be used here since it would require a grayscale target dose optimization.

[0084] A criticism of the proposed printer configuration for the embodiment in printer 138 is the need for multiple projection sources implemented as projectors 104. The primary reason was to meet the angular sampling requirements of the tomographic reconstruction process and the frame rate limitation of available low-cost projection sources. It is therefore to be understood that the number of projectors 104 could be reduced in favor of increased vial rotation rate (e.g. 720 degrees / second for a single projector system) recognizing a potential for a corresponding decreased print quality due to an increased angular velocity gradient in the resin as well as increased part sedimentation and / or floating

[0010] . However, in certain applications such a print quality can still be useful and accordingly systems like system 100 are contemplated that utilize one or two projectors 104. Additional projectors 104 could also be used.

[0085] 3.2. Comparison of OSMO and PIHE

[0086] In Figure 9, we compare the print dynamics for a 3DBenchy (e.g. Ship 408-C) printed in PEGDA250 made with OSMO and PIHE. In Figure 9(a) and Figure 9(c) are shown photographs of the optical scatter signal

[0035] at the onset of gelation, as well as half and a full vial rotation afterwards. Surprisingly, in both optimization approaches the printed part (e.g. Ship 408-C) floats (moves towards top of Figure 9) as opposed to sinks. Although gelation results in an increase in density of the printed part relative to the surrounding liquid, the strong exothermal reaction of gelation causes the part to rise upwards. This effect has been observed in an earlier VAM study

[0011] and was attributed to either thermal expansion of the printed part (resulting in a net density decrease) or drag caused by convection currents. Results from recent computational fluid dynamicsindicate that for a large heating rate combined with a sufficiently large coefficient of thermal expansion that the part would initially float due to a temporary net decrease in density [9].

[0087] For the part printed with OSMO, we observe the hull and cabin roof form first (labels (1) and (2) in Figure 9(a)), but the supports (label (3)) do not form before the part rises away from the writing beams. In OSMO, any in-part dose that exceeds the high threshold Dhigh is not penalized. As a result, the in-part dose can span [Dhigh, 1] and we observe that only the hull forms (Figure 9(b)). For the part printed with PI (no histogram equalization), we observe that the hull (1), cabin roof (2), and supports (3) all form prior to the part rising from the writing beams. This is a result of PI minimizing the IPDS, resulting in more uniform onset of gelation throughout the printed part. Finally, in Figure 9e), Figure 9(f) are shown the results for the same part printed with PIHE. Print quality is highest with PIHE, with the chimney forming (4), which is attributed to the higher printed rate afforded by histogram equalization (9.3 s vs 18 s).

[0088] The print dynamics observed for OSMO can also be observed when using the traditional filtered back projection method with negativity constraint [2]. In an earlier work from our group

[0035] we observed that the hull of the boat formed first prior to the cabin. This was acceptable for the much more viscous resins used in that work (9000 cP) since voxels did not move during the printing process.

[0089] 3.3. Potential for bioprinting

[0090] The advantages of VAM in printing of biocompatible materials have been demonstrated in earlier studies [5,27]. In both of these works, the authors were able to demonstrate fabrication of constructs which are difficult to fabricate using conventional layer-by-layer means, such as a human auricle model and fluidic ball-cage valve [5], as well as complex periodic structures

[0027] . However, this impressive fabrication was enabled by using gelatin materials that are in a semisolid phase at room temperature. This sacrificial gelatin suspension is advantageous as it prevents cell sedimentation during printing but it requires that post processing be done at elevated temperatures to liquefy uncured resin and release the printed part. The present specification circumvents the need for gelatin materials, enabling simpler and faster post-processing since parts can be directly extracted from the resin bath. To demonstrate the potential of this method for bioprinting applications, gryoidal lattice structures were printed in a PEGDA-based hydrogel(viscosity = 27.60 cP) and show the results in Figure 10. Printing time was shorter than the part printed in non-hydrogel PEGDA700 (2.5 seconds vs. 5.7 seconds) which we attribute to increased absorption by the photo-initiator and lower viscosity of the material. After printing, the part was placed in a water bath for 10 minutes to remove any uncured photoresin (Figure 10(a)). The rehydration properties of the printed hydrogel were tested by first placing the part on a microscope slide (Figure 10(b)) and letting it air-dry at STP for approximately 2 hours (Figure 10(c)) followed by rehydration in a bath of water colored with water-soluble red dye (Figure 10(d)). After drying, we observed isotropic shrinkage of the part (22 % linear shrinkage in directions). Subsequent rehydration in red-colored water shows full expansion to original as-printed dimensions with no observable damage to the lattice structure.

[0091] The present specification provides a novel hardware-software approach to achieve high- fidelity printing in low-viscosity PEGDA resins. In this specification, printing performance in photoresins with viscosities 100X smaller than typically used in tomographic VAM has been demonstrated. The proposed method utilizes a multiple projector system combined with a new tomographic optimization algorithm to achieve rapid and uniform dose delivery which is essential for printing in low-viscosity photoresins. Micro-computed tomography of printed constructs reveals good conformity to the target design, with an RMS SDF error close to the voxel resolution for this printer system. This indicates the versatility of the proposed method in using these previously unprintable materials in VAM. Furthermore, the demonstrated ability to directly fabricate a low-viscosity hydrogel showcases the potential of this technique for bioprinting applications. This specification broadens the range of materials available to VAM which will be invaluable for applications requiring the enhanced design freedom of volumetric printing but also the need for low-viscosity materials.

[0092] While certain embodiments have been discussed, a person of skill in the art will now recognize that variations, subsets and combinations thereof are contemplated. For example, while three projectors 104 are shown in Figure 1, in other embodiments, one or more projectors may be provided with the remainder of system 100 modified accordingly. While not presently known to the inventors, a single projection source may be possible where a projector can achieve a frame rate of 360 frames per, second or above, for 8 bit depth images. Thus, while such a projector may not be known as of the filing date of this specification, the inventors foresee the teachings hereinbeing applicable to such a future-developed projector. Thus, a presently preferred embodiment contemplates three projectors based on known technology.

[0093] In another variant, the present teachings can be applied to any curable material having a of viscosity less than about lOOOcP.

[0094] In another variant, the curable material can include a photoinitiator and monomer(s), wherein the monomer comprises triethylene glycol dimethacrylate, hexyl acrylate, trimethylolpropane triacrylate, trimethylolpropane trimethacrylate, pentaerythritol tetraacrylate, pentaerythritol triacrylate, 2-hydroxyethyl methacrylate, 2-hydroxyethyl acrylate, ethyl methacrylate, hydroxypropyl methacrylate, hydroxybutyl methacrylate, ethylene glycol methyl ether methacrylate, 1,6-hexanediol diacrylate, 2-hydroxy acrylate, isobornyl acrylate, glycidyl acrylate, glycidyl methacrylate, methacrylate, acrylate, 2-phenoxyethylacrylate, tert-butyl acrylate, n-butyl acrylate, ethyl acrylate, benzyl acrylate, methyl acrylate, lauryl acrylate, vinyl acrylate, isobutyl acrylate, (2-methoxyethyl) acrylate, 2-ethylhexyl acrylate, ethylene glycol phenyl ether acrylate, acrylic acid, methacrylic acid, hexyl acrylate, hexyl methacrylate, pentaerythritol tetraacrylate, 1 ,4-butanediol diacrylate, 1 ,20-decanediol dimethacrylate, 1,3- butanediol dimethacrylate, 1 ,4-butanediol diacrylate, 1 ,4-butanediol diacrylate, 1 ,4-butanediol dimethacrylate, 1,5 -pentanediol dimethacrylate, 1,6-hexanediol dimethacrylate, ethylene glycol diacrylate, ethylene glycol dimethacrylate, or any mixture thereof.

[0095] Furthermore, the present specification contemplates that a mold, such as a plastic injection mold, or other type of mold, can be created from the products derived from the digital models made according to this specification.

[0096] The foregoing example embodiments are representative and further variants, combinations and subsets will now be apparent to those of skill in the art, with the claims hereto defining the scope of the time-limited monopoly sought by this specification.

[0097] Citations.

[0098] (The contents of all of these are incorporated herein by reference.)

[0099] [1] B.E. Kelly, et al., Volumetric additive manufacturing via tomographic reconstruction, Science 363 (2019) 1075-1079.

[0100] [2] D. Loterie, P. Delrot, C. Moser, High-resolution tomographic volumetric additive manufacturing, Nat. Commun. 11 (2020) 852.

[0101] [3] J.T. Toombs, et al., Volumetric additive manufacturing of silica glass with microscale computed axial lithography, Science 376 (2022) 308-312.

[0102] [4] S. Peng, et al., Ultra-fast 3D printing of assembly — free complex optics with subnanometer surface quality at mesoscale. Int. J. Extrem. Manuf. 5 (2023) 035007.

[0103] [5] P.N. Bernal, et al., Volumetric Bioprinting of Complex Living-Tissue Constructs within Seconds. Adv. Mater. 31 (2019) 1904209.

[0104] [6] M. Xie, et al., Volumetric additive manufacturing of pristine silk-based (bio)inks, Nat. Commun. 14 (2023) 210.

[0105] [7] C. Darkes-Burkey, R.F. Shepherd, Volumetric 3D Printing of Endoskeletal Soft Robots, Adv. Mater. (2024) 2402217, https: / / doi.org / 10.1002 / adma.202402217.

[0106] [8] J. Madrid-Wolff, et al., A review of materials used in tomographic volumetric additive manufacturing, MRS Commun. 13 (2023) 764-785.

[0107] [9] R. Salajeghe, C.S. Kruse, D.H. Meile, D. Marla, J. Spangenberg, Invest. Influ. Therm. Meeh. Prop. Resin Sediment, rate Print. Geom. Vol. Addit. Manuf. Tech. (2022), https: / / doi.org / 10.26153 / TSW / 44203.

[0108]

[0010] R. Salajeghe, D.H. Meile, C.S. Kruse, D. Marla, J. Spangenberg, Numerical modeling of part sedimentation during volumetric additive manufacturing, Addit. Manuf. 66 (2023) 103459.

[0109]

[0011] T. Waddell, et al., Use of volumetric additive manufacturing as an in-space manufacturing technology, Acta Astronaut 211 (2023) 474-482.

[0110]

[0012] D. Webber, et al., Versatile volumetric additive manufacturing with 3D ray tracing, Opt. Express 31 (2023) 5531.

[0111]

[0013] S. Schneiderman, J.L. Farber, R. Baserga, A simple method for decreasing the toxicity of polyethylene glycol in mammalian cell hybridization, Somat. Cell Genet 5 (1979) 263-269.

[0112]

[0014] C. Yamini, K.L. Shantha, K.P. Rao, Synthesis and Characterization of Poly[N-Vinyl- 2-Pyrrolidone-Polyethylene Glycol Diacrylate] Copolymeric Hydrogels for Drug Delivery, J. Macromol. Sci. Part A 34 (1997) 2461-2470.

[0113]

[0015] E.M. Moore, J.L. West, Bioactive poly(ethylene glycol) acrylate hydrogels for regenerative engineering, Regen. Eng. Transl. Med. 5 (2019) 167-179.

[0114]

[0016] J. Kim, et al., Synthesis and evaluation of novel biodegradable hydrogels based on poly(ethylene glycol) and sebacic acid as tissue engineering scaffolds, Biomacromolecules 9 (2008) 149-157.

[0115]

[0017] X. Zheng Shu, Y. Liu, F.S. Palumbo, Y. Luo, G.D. Prestwich, In situ crosslinkable hyaluronan hydrogels for tissue engineering, Biomaterials 25 (2004) 1339-1348.

[0116]

[0018] A.N. Buxton, et al., Design and Characterization of Poly(Ethylene Glycol) Photopolymerizable Semi-Interpenetrating Networks for Chondrogenesis of Human Mesenchymal Stem Cells, Tissue Eng. 13 (2007) 2549-2560.

[0117]

[0019] M.T.H. Nguyen, et al., Preparation and Stability of PEGDA / GO Conductive Materials by DLP 3D Printing, Electron. Mater. Lett. 18 (2022) 275-281.

[0118]

[0020] L. Sheng, M. Li, S. Zheng, J. Qi, Adjusting the accuracy of PEGDA-GelMA vascular network by dark pigments via digital light processing printing, J. Biomater. Appl. 36 (2022) 1173— 1187.

[0119]

[0021] R. Mau, J. Nazir, S. John, H. Seitz, Preliminary Study on 3D printing of PEGDA

[0120] Hydrogels for Frontal Sinus Implants using Digital Light Processing (DLP), Curr. Dir. Biomed. Eng. 5 (2019) 249-252.

[0121]

[0022] M.H. Khalili, et al., Nanoindentation Response of 3D Printed PEGDA Hydrogels in a Hydrated Environment, ACS Appl. Polym. Mater. 5 (2023) 1180-1190.

[0122]

[0023] Q. Chen, B. Zou, Q. Lai, K. Zhu, SLA-3d printing and compressive strength of PEGDA / nHAP biomaterials, Ceram. Int. 48 (2022) 30917-30926.

[0123]

[0024] LI. Preobrazhenskiy, A. A. Tikhonov, P.V. Evdokimov, A.V. Shibaev, V.I. Putlyaev, DLP printing of hydrogel / calcium phosphate composites for the treatment of bone defects, Open Ceram. 6 (2021) 100115.

[0124]

[0025] M. Hakim Khalili, et al., Additive Manufacturing and Physicomechanical Characteristics of PEGDA Hydrogels: Recent Advances and Perspective for Tissue Engineering, Polymers 15 (2023) 2341.

[0125]

[0026] R.H. Schmedlen, K.T. Nyugen, J.L. West, Tissue engineered vascular grafts: elastic polyethylene glycol hydrogel scaffolds for culture under pulsatile flow conditions, in: in Proceedings of the Second Joint 24th Annual Conference and the Annual Fall Meeting of the Biomedical Engineering Society] [Engineering in Medicine and Biology, vol. 1, IEEE, Houston, TX, USA, 2002, pp. 813-814.

[0126]

[0027] M.B. Riffe, et al., Multi-Material Volumetric Additive Manufacturing of Hydrogels using Gelatin as a Sacrificial Network and 3D Suspension Bath. Adv. Mater. 36 (2024) 2309026.

[0127]

[0028] A. Orth, et al., Deconvolution volumetric additive manufacturing, Nat. Commun. 14 (2023) 4412.

[0128]

[0029] D. Webber, et al., Micro-optics fabrication using blurred tomography, Optica 11 (2024) 665.

[0129]

[0030] C.M. Rackson, et al., Object-space optimization of tomographic reconstructions for additive manufacturing, Addit. Manuf. 48 (2021) 102367.

[0130]

[0031] G.C. Goodwin, S.F. Graebe, M.E. Salgado, G.C. Goodwin, Control System Design, Prentice Hall, Upper Saddle River, NJ, 2001.

[0131]

[0032] #3DBenchy. #3DBenchy (https: / / www.3dbenchy.com / ) .

[0132]

[0033] S. Kim, H. Lee, H. Choi, K.-Y. Yoo, H. Yoon, Investigation on photopolymerization of PEGDA to fabricate high-aspect-ratio microneedles, RSC Adv. 12 (2022) 9550-9555.

[0133]

[0034] T. Chen, et al., High-fidelity tomographic additive manufacturing for large-volume and high-attenuation situations using expectation maximization algorithm, Addit. Manuf. 80 (2024) 103968.

[0134]

[0035] A. Orth, et al., On-the-fly 3D metrology of volumetric additive manufacturing, Addit. Manuf. 56 (2022) 102869.

Claims

Claims1. A method for three-dimensional printing, the method comprising: receiving, at a controller, a plurality of multi-angle projection patterns based on a digital model; rotating, under the control of the controller, a build chamber containing a curable material, emitting, under the control of the controller, each projection pattern into the build chamber during the rotation, such that the curable material receives projection energy according to the multi-angle projection patterns; coordinating the rotating with the emitting such that each projection pattern is directed into the build chamber at a corresponding angular position to achieve uniform dose accumulation in the curable material at spatial locations defined by the digital model; and, terminating the rotating and the emitting upon determining that a print-completion condition of the digital model is met.

2. The method of claim 1 further comprising: receiving a raw digital model in a mesh-based format and generating the plurality of multi -projection angle patterns based on the raw digital model.

3. The method of claim 2 wherein the mesh-based format is based on one of the formats known as STL (.stl), OBJ (.obj), or 3MF (.3mf).

4. The method of any one of claims 1 to 3, wherein the plurality of projection patterns is generated using a feedback loop to reduce a dose error across voxels of the digital model.

5. The method of claim 4, wherein the feedback loop is combined with histogram equalization to increase overall light dose intensity and reduce in-part dose spread.

6. The method of any one of claims 1 to 5, wherein the digital model comprises a voxelized representation of a three-dimensional object, and each voxel is associated with a target light dose.

7. The method of any one of claims 1 to 6, wherein the digital model is generated from the raw digital model by applying a projection pattern optimization operation to generate the plurality of multi-angle projection patterns, the operation comprising: applying proportional-integral feedback to iteratively update voxel intensities in the digital model based on a comparison between calculated and target dose values; and adjusting the multi-angle projection patterns by performing histogram equalization of pixel intensity values across the projection set.

8. The method of any one of claims 1 to 7, wherein the curable material comprises poly(ethylene glycol) diacrylate (PEGDA) having a viscosity as low as 12 cP.

9. The method of claim 8, wherein the PEGDA has a molecular weight of less than about 700 Da.

10. The method of any one of claims 1 to 7 wherein the curable material includes a photoinitiator and monomer(s) and, possibly, a solvent. The monomer(s) may comprise triethylene glycol dimethacrylate, hexyl acrylate, trimethylolpropane triacrylate, trimethylolpropane trimethacrylate, pentaerythritol tetraacrylate, pentaerythritol triacrylate, 2-hydroxyethyl methacrylate, 2-hydroxyethyl acrylate, ethyl methacrylate, hydroxypropyl methacrylate, hydroxybutyl methacrylate, ethylene glycol methyl ether methacrylate, 1,6-hexanediol diacrylate, 2-hydroxy acrylate, isobornyl acrylate, glycidyl acrylate, glycidyl methacrylate, methacrylate, acrylate, 2-phenoxyethylacrylate, tert-butyl acrylate, n-butyl acrylate, ethyl acrylate, benzyl acrylate, methyl acrylate, lauryl acrylate, vinyl acrylate, isobutyl acrylate, (2-methoxyethyl) acrylate, 2-ethylhexyl acrylate, ethylene glycol phenyl ether acrylate, acrylic acid, methacrylic acid, hexyl acrylate, hexyl methacrylate, pentaerythritol tetraacrylate, 1 ,4-butanediol diacrylate, 1 ,20-decanediol dimethacrylate, 1,3 -butanediol dimethacrylate, 1,4-butanediol diacrylate, 1 ,4-butanediol diacrylate, 1 ,4-butanediol dimethacrylate, 1,5 -pentanediol dimethacrylate, 1,6-hexanediol dimethacrylate, ethylene glycol diacrylate, ethylene glycol dimethacrylate, or any mixture thereof.

11. The method of any preceding claim, wherein the curable material has a of viscosity less than about lOOOcP.

12. The method of any one of claims 1 to 11, wherein the build chamber is rotated through an angular extent of approximately 1 / n of a full rotation to deliver a complete set of the multi-angle projection patterns, where n is the number of projection patterns spaced about the build chamber.

13. The method of any one of claims 1 to 12, wherein the projecting step comprises emitting projection patterns from a plurality of projectors spaced at equal angular intervals about the build chamber, each projector emitting a subset of the projection patterns in phase with the rotation of the build chamber.

14. The method of any one of claims 1 to 13, wherein the coordinating of the rotating and the emitting comprises: receiving angular position data from a rotation sensor; and transmitting projection patterns to the plurality of projectors based on the angular position data such that each projector emits a projection pattern phase-aligned to its respective position around the build chamber.

15. A 3D printer system comprising a controller, a turntable and at least one projector according to any of the foregoing methods.

16. A controller according to claim 15.

17. A non-transitory computer-readable medium configured to store programing instructions according to any one of claims 1-14 executable the apparatus according to claim 14.

18. A printed digital model produced according to any one of the methods of claims 1-11.

19. A mold for a product produced from the printed digital model according to claim 16.

Images