Patent Application
20220347929
The present invention relates to a system and method to improve the quality such as the resolution and fidelity of a 3D object in volumetric tomographic printers for the volumetric fabrication of three-dimensional objects or articles from photoresponsive materials, by use of a measurement of the 3D object in-situ followed by a corrective action, which is a significant improvement over prior art. In particular, the present invention is refated to volumetric manufacturing systems wherein the articles or objects being fabricated are imaged and monitored in real-time.
The working principle of volumetric tomographic printing (WO 2019/043529 A1) is entirely different than the traditional layer-by-layer approach (i.e. 3D printing with the formation of one layer over the other) in conventional additive manufacturing (AM).
In conventional additive manufacturing, a three-dimensional object is fabricated either by pointwise scanning of the object volume or in a layer-by-layer fashion. An example is stereolithography (SLA) (see for example U.S. Pat. No. 5,344,298), where the object is formed one layer at a time by the solidification of a photocurable resist under light irradiation before application of a subsequent layer. The successive layers of the object can be defined for example by scanning a laser beam point-by-point, as suggested in U.S. Pat. No. 5,344,29, or by digital light processing (DLP) technology, as described in U.S. Pat. No. 6,500,378.
In these methods, the photopolymerization process of the successive object's layers is controlled by using highly absorbing and strongly reactive resins. The layer thickness typically ranges from 10 μm to 200 μm. To cure well-defined layers of resins, a large concentration of one or more photoinitiators and dyes are typically included in resins so that they are highly absorbing and strongly reactive (T. Baldacchini, Three-Dimensional Microfabrication Using Two-Photon Polymerization, William Andrew, 2015). Thus, using highly absorbing inks is beneficial in SLA and DLP as it prevents the exposure of an already processed layer by the next layer being formed, which could result in manufacturing artefacts, a phenomenon referred as overcuring in additive manufacturing.
To overcome the geometric constraints and throughput limitations of layer-by-layer light-based AM techniques, namely digital-light processing (DLP) and stereolithography (SLA), multi-beam AM techniques have been proposed (Shusteff, M. et al., One-step volumetric additive manufacturing of complex polymer structures, Sci Adv 3, eaao5496- (2017); Kelly, B. E. et al., Volumetric additive manufacturing via tomographic reconstruction, Science 363, 1075-1079 (2019); Loterie, D., Delrot, P. & Moser, C., VOLUMETRIC 3D PRINTING OF ELASTOMERS BY TOMOGRAPHIC BACK-PROJECTION (2018), preprint DOI: 10.13140/RG.2.2.20027.46889). These techniques are subsequently referred to as volumetric tomographic printing or tomography-based additive manufacturing (tomographic additive manufacturing).
In tomography-based additive manufacturing methods, the object is not formed by sequentially curing layers of a photopolymer, but rather a volume of transparent photoresponsive material is irradiated from multiple angles with computed patterns of light, which results in the local accumulation of light dose and the consequent simultaneous solidification of specific object voxels, in order to fabricate a three-dimensional object in a single step. The main advantages of this method compared to existing methods are its very rapid manufacturing time (down to a few tens of seconds), and its ability to print complex hollow structures without the need for support structures as required in layer-by-layer manufacturing systems.
To achieve a correct three-dimensional light dose deposition in the build volume, the two-dimensional light patterns projected from multiple angles must illuminate the entire build volume. Thus, resins with a low-absorptivity at the illumination wavelength are used. Typically, a correct light dose deposition is achieved in tomographic additive manufacturing methods with an attenuation length of the illuminating light at 1/e equal to the diameter of the build volume, which sets an upper limit on the photo-initiator concentration.
Though such volumetric part generation allows processing more viscous resins than in existing DLP and SLA techniques (>4 Pa·s) that yields higher throughput (>105 mm3/hour), the smallest feature size demonstrated by multi-beam AM is currently limited to approximately 300 μm.
As opposed to DLP and SLA, where the polymerization extent of a layer is controlled by using highly absorbing resins, volumetric AM requires transparent resins which results in a reduced spatial and temporal control of the photopolymerization process and consequently limits the achievable printing resolution.
Therefore, the tomographic photopolymerization process has to be monitored to achieve high-resolution manufacturing of objects or articles.
As a consequence, there is a need for a robust, industrially applicable method and system to image and monitor the articles or objects being fabricated in cylindrical build volumes of various diameters via tomography-based additive manufacturing.
The present invention circumvents all of the previous shortcomings of 3D objects printed with a volumetric approach such as with tomographic back-projection.
The invention herein disclosed provides higher resolution. Experimental demonstration showed a resolution of 80 μm with volumetric production of centimeter-scale acrylic and silicone parts by feedback-enhanced tomographic reconstruction.
In detail, the present invention is related to a method for monitoring the generation of a three-dimensional object being formed in a tomographic additive manufacturing system from simulated tomographic two-dimensional back-projections of a desired 3D article, the method comprising the steps of:
Preferably, the variants i) to iv) are carried out automatically.
According to a preferred embodiment, the present invention is related to a method for obtaining a set of corrected projection patterns in a volumetric back-projection printer, the method comprising the steps of:
Preferably, two-dimensional projections of said desired 3D article are computed from an original 3D digital object, more preferably a 3D CAD object.
Preferably, the images of said object are captured by said imaging system concurrently with illuminating said container with the two-dimensional light patterns.
Preferably, the method further comprises the step of capturing reference background images at multiple angles of the container comprising the photoresponsive material, preferably by rotation of the container, before said container is illuminated with the two-dimensional light patterns, wherein in the step of determining from said images the shape or extent to which the object has been formed the reference background images are subtracted from the captured images of said object being formed, so as to detect parts of the resin container that became polymerized using a threshold. In particular, from such detected polymerized parts a 3D map is created by a tomographic back-projection method, and new set of projection patterns is generated by comparing the created 3D map with the original 3D digital object, preferably 3D CAD object, wherein the new set of projection patterns is used either for printing the same object (5) by restarting the method, using the first printed object (5) as a sacrificial print, or for implementing a real-time correction feedback system.
Preferably, from detected polymerized parts a map of the regions of the object is generated that have already been formed as seen from the angles at which each captured image and optionally background image was recorded, and
Preferably, the imaging system comprises a structured illumination for generating the captured images. In particular, a difference is extracted between the captured images of the object being formed and the desired 3D article, said difference being used to generate the new set of projection patterns by reverse tomographic back-projection. Especially preferred, said extracted difference is used for determining whether the object has been completely manufactured or whether the shape of the object is correct, wherein the method is stopped if complete manufacturing of the object is determined, or illuminating the container with two-dimensional light patterns continues if the shape of the object is correct or no complete manufacturing of the article is determined. According to another preferred variant of this embodiment comprising structured illumination, each of said measurements of the two-dimensional projections of the object is obtained by performing a differential overlay of a first two-dimensional projection measured with said imaging system at an angle by a second two-dimensional projection previously measured with said imaging system from said same angle.
Preferably, the beam is modulated by projection patterns in a DLP modulator.
Also preferably, the beam and either the structured illumination or a measuring beam enter the container at an angle of 90°.
The present invention is furthermore related to a method of printing an object in a volumetric back-projection printer, the method comprising the steps of
The present invention is in addition related to a method of imaging an object in a volumetric back-projection printer for monitoring the production process of a three-dimensional article.
The present invention is furthermore related to a system for performing the method of the present invention defined above, comprising
Preferably, said display unit comprises a component selected from the group consisting of a liquid-crystal display illuminated by light with one or more colors, an organic light-emitting diode display with one or more colors, and a light-emitting diode display.
Preferably, said resin container (is arranged in an index matching liquid bath.
Preferably, said imaging system and said unit for providing a light beam to be guided into the resin container are arranged such that an angle, preferably of 90°, is formed between the beam or the structured illumination and the beam when entering the resin container.
These and other features, aspects and advantages of the present invention will become better understood with regard to the following non-limiting description, appended claims and accompanying drawings where:
The present invention is related to volumetric tomographic manufacturing systems wherein the articles or objects being fabricated are imaged and monitored, preferably in real-time. In tomographic additive manufacturing, monitoring the photopolymerization process leading to the formation of a three-dimensional article is critical for obtaining a high geometric fidelity with respect to its digital three-dimensional model and to adapting to different resins' reactivity. The present invention discloses an apparatus and methods to monitor tomographic additive manufacturing of three-dimensional articles, preferably with a structured illumination.
Volumetric back-projection printing, including the method of obtaining simulated tomographic two-dimensional back-projections of a desired 3D article and generating two-dimensional light patterns from said two-dimensional back-projections, and respective printer suitable for said printing technique have been described in e.g. WO 2019/043529 A1 or US 2018/0326666 A1. In tomographic volumetric additive manufacturing, a volume of photoresponsive material is illuminated from many directions with patterns of light. These patterns of light are computed with an algorithm similar to that used in X-ray computed tomography, also known as medical CT scanners. These algorithms are known to the skilled person.
The tomographic approach delivers a 3D light dose in such a way that when an individual voxel obtains enough dose, it turns from liquid to solid. This individual voxel immediately scatters light because of the refractive index difference between the solid and the surrounding liquid. To avoid light scattering, all 3D voxels must transition from liquid to solid at the same time. The requirement that all 3D voxel forming the object must obtain the same light dose instantaneously is quintessential to the accurate formation of the 3D object.
Optical beam non-uniformity, resin non-uniformity, vial non-uniformity, as well as the increase of chemical reaction kinetics that are dependent on local temperature because of the exothermic polymerization reaction, makes the system intractable in an open loop configuration to obtain the highest resolution and fidelity of the 3D printed object. Because of these reasons, a feedback system is not a “nice to have” feature, but a “must have” for tomographic printing.
Monitoring and thus detecting the fabrication of the tomographically-produced objects allows, for instance, to automatically stop the exposure of the build volume, if the resin being used has a different reactivity than resins known to the operator. In addition, monitoring and detecting the fabrication of the tomographically-produced objects allows to automatically stop the exposure of the build volume, if an object is formed faster or more slowly than theoretically expected. Furthermore, monitoring the fabrication of the tomographically-produced objects allows to measure the geometric fidelity of the formed object to the digital model and therefore to adjust the light patterns illuminating the build volume, either in real-time or from one print to another.
Imaging transparent objects being polymerized in a transparent photoresponsive liquid or gel may be achieved by measuring a small refractive index change between the solid phase of the object and the liquid or gel phase of the unreacted photoresponsive material over a centimeter-scale volume. This refractive index change between the monomer and polymer form of the resin is typically 0.01 to 0.05 for acrylic resins (Refractive index of methacrylate monomers & polymers. Technical bulletin [online]. Esstech Inc., 2010 [retrieved on 2020 Mar. 19]. Retrieved from https://www.esstechinc.com/refractive-index-of-methacrylate-monomers-polymers/).
According to the present invention, it was found that measuring a refractive index change between the solid phase of the object and the liquid or gel phase of the unreacted photoresponsive material can be used for establishing a robust, industrially applicable method and system to image and monitor the articles or objects being fabricated in cylindrical build volumes of various diameters via tomography-based additive manufacturing.
Measuring a small refractive index can be performed with advanced optical techniques such as interferometry or Schlieren imaging.
Though interferometric imaging is the most sensitive of the aforementioned techniques, it is not easy to adapt it to the measurement of tomographic photopolymerization process, since it requires a complex optical setup, a clean coherent light source as well as a compensation path to ensure accurate measurements (Srivastava, A., Muralidhar, K. & Panigrahi, P. K. Comparison of interferometry, schlieren and shadowgraph for visualizing convection around a KDP crystal. Journal of Crystal Growth 267, 348-361 (2004)). In tomography-based additive manufacturing, the objects are produced in a rotating cylindrical glass vial, the build volume. To perform interferometric measurements, it is necessary to equalize the measurement path and the blank compensation path. However, compensating for the cylindrical aberration induced by the build volume demands additional optics or placing the build volume in an immersion bath filled with a refractive-index matching liquid. A liquid compensation path is not practical within an industrial-grade apparatus, as the liquid could evaporate or spill on the inner sensitive parts of the machine. Furthermore, interferometric measurements are affected by their high sensitivity to thermal gradient, like the ones generated during the exothermic photopolymerization onset of objects in tomography-based additive manufacturing.
Schlieren imaging, in which refractive-index gradients are measured, is not practically adapted to monitoring tomography-based additive manufacturing. Schlieren imaging requires optically filtering the light field unrefracted by the density gradient. In the same way as interferometric imaging, the cylindrical aberration induced by the build volume would require an optical compensation to correctly filter out the unrefracted rays to perform Schlieren imaging. Such a compensation cannot accommodate different build volume diameters or requires to use an immersion bath filled with a refractive-index matching. Thus, Schlieren imaging is not straightforwardly applicable within an industrial-grade tomographic printer.
In detail, the present invention is related to an apparatus used for monitoring and measuring tomographic additive manufacturing of a three-dimensional article.
According to a first embodiment of the present invention, monitoring is performed using a shadowgraphy method.
In detail, the method according to the first embodiment for obtaining a set of corrected projection patterns 9 in a volumetric back-projection printer 1, the method comprising the steps of:
Said object to be formed is a three-dimensional object/article.
An additional imaging system 6, 7, 8 for monitoring comprises the measuring beam 6, the camera 7 and the lens system 8. The forming object 5 is illuminated with said beam 6 and imaged onto the camera 7 via the lens system 8 of appropriate magnification and aberration correction. The plane of best focus can be chosen to be in the middle of the container 3, but is not limited to. The measuring beam 6 may be expanded by lenses (not shown in
For each rotation angle of the container 3, an image is recorded on the camera 7. The wavelength of the measuring beam 6 is different than the wavelength of the beam 2a used to initiate polymerization of the object 5, such that it does not cause polymerization of the photoresponsive material 103 in the container 3. In
According to a variant of the system of the first embodiment shown in
In
First, in step 2.1 a three-dimensional computer aided design of the object 5 is processed to generate a set of projection patterns 9 in the tomographic back-projection volumetric printer 1. This step can be adapted to any type of volumetric printer such as, but not limited to multi-beam volumetric printers.
In step 2.2, a set of reference images (background images) Ib,g,i (x,y,ai) are recorded upon rotating the build volume before starting the actual exposure, in step 2.3, with the light wavelength 2a sensitive to the resin.
In step 2.4, the camera 7 of the imaging system 6, 7, 8 records intensity images Ii (x,y,ai,t) of the build volume in synchronization with the rotation of the resin container 3. The intensity imaging system 6, 7, 8 is used with any resin material that scatters light upon solidification in such a way that the solidified area shows up as dark in the intensity image. If the object 5 being formed (i.e. solidified) scatters light weakly so as to remain transparent, a phase imaging system can be substituted to the intensity imaging system 6, 7, 8 illustrated in
The build volume is illuminated from the back (transmission imaging) by an expanded and collimated laser beam 6 with a wavelength different than the wavelength of the beam 2a used for polymerizing the resin.
Then, a new set of images are recorded for each angle, filtered and down-sampled to reduce noise and speed up processing respectively. As seen in
Then, in step 2.5 the difference ΔIi (x,y,t)=Ii (x,y,ai,t)−Ib,g,i (x,y,ai)) between the new set of images and the set of background images is computed on each subsequent turn.
A threshold ΔIi (x,y,t) is then used in step 2.6 to detect which parts of the volume became solid (i.e. scattered light) at which time. Spatially, this information is still two-dimensional at this point because it is derived from 2D transmission images of the entire build volume.
In order to build a three-dimensional map of the time needed for solidification, in step 2.7 the said difference images are back-projected using the tomographic algorithm to obtain the time values into a 3D grid. When a transmission image at a particular angle and time showed ‘solid pixels’ (as measured by the thresholding procedure), the time of solidification was recorded in the 3D volume in a line corresponding to these pixels and oriented along the direction of projection. If a later transmission image from any other angle shows no solidification or a transition to solid state at a later time for the same group of pixels in 3D, then the recorded solidification time was increased for those pixels. The resulting 3D volume of ‘solidification times’ directly gives the required intensity correction for the next print. Indeed, the dose D is related to intensity I and time t as D=I·t. If a given part of the object takes a longer time t to solidify, the intensity I in this part simply needs to be increased proportionally to make it solid at the same time as the rest of the model. After this 3D intensity adjustment, in step 2.8 corrected Radon projections (corrected projection patterns) are calculated from the corrected model of the object using the same procedure as described before.
Thus, according to the first embodiment the new set of projection patterns 9 is preferably generated by the following steps:
Said new set of projection patterns 9 is used for adjusting the illumination dose of the beam 2a.
Thus, according to this variant of the first embodiment, the first print made by illuminating the resin container 3 with the beam 2a using the SLM projection patterns computed from the 3D digital object, preferably 3D CAD object is a sacrificial print that is used to obtain the new set of projection patterns 9.
In the case when a sacrificial print cannot be done or to further increase the fidelity of the print, a real-time correction feedback system can be implemented (
Thus, according to the first embodiment, the present invention is related to a method of printing an object 5 in a volumetric back-projection printer 1, the method comprising the steps of
According to the first embodiment of the present invention, a system for obtaining a set of corrected projection patterns 9 and for printing an object 5 in a volumetric back-projection printer 1, comprises
Said resin container 3 may be arranged in an index matching liquid bath 4; however, this is not mandatory.
Preferably, said imaging system 6, 7, 8 and said unit for providing a light beam 2a to be guided into the resin container 3 are arranged such that an angle is formed between the beam 6 and the beam 2a when entering the resin container 3. More preferably, said angle is 90°.
According to a second embodiment of the present invention, the apparatus comprises a device for performing a structured illumination for monitoring, for example, but not limited to, an LCD display or a Digital Light Processor (DLP), a camera objective with optionally an adjustable numerical aperture diaphragm, and a camera.
For clarification, the term structured illumination used in this invention has the same meaning as in the state of the art. It means any two dimensional spatially varying light intensity.
As shown above in
However, the method according to the first embodiment of the present invention has drawbacks in determining the edges of manufactured three-dimensional articles when the refractive index difference between the unpolymerized photoresponsive material and the manufactured three-dimensional article is too small to induce significant refraction, e.g. below 0.01.
With the method according to the second embodiment of the present invention, an even more sensitive imaging apparatus to monitor tomographic additive manufacturing is provided.
According to the second embodiment, the present invention is related to a method of increasing the sensitivity of the structured illumination monitoring of tomographic additive manufacturing comprising the steps of:
For clarification, the term differential overlay used in this invention has the same meaning as in the state of the art. It means computing and displaying an overlay of a first image from which is subtracted a second image with a scaling factor applied on all the pixels of the second image.
In a preferred variant of the second embodiment, said photoresponsive material is not sensitive to the wavelength of said structured illumination.
In a preferred variant of the second embodiment, a camera lens 72 relays the image 73 of the structured illumination refracted by the photoresponsive material 103 and the object (three-dimensional article) 5 onto a camera sensor 74. The focus of the part of said structured illumination that propagated through said object (three-dimensional article) 5 is shifted because of the higher refractive index of said object (three-dimensional article) 5 with respect to said photoresponsive material 103. The focal shift of said part of the structured illumination creates an image 75 of said object (three-dimensional article) 5.
In a further preferred variant of the second embodiment, the numerical aperture of the camera lens 72 can be adjusted, which in turn adjust the depth of focus 76 of the camera 74.
In a further preferred variant of the second embodiment, both the object (three-dimensional article) 5 and the source 70 of the structured illumination 71 are at focus on the camera sensor 74.
In a further preferred variant of the second embodiment, the colors of the structured illumination 71 are displayed sequentially and captured by camera sensor 74 at a speed faster than the two-dimensional light patterns provided by the beam 2a. An enhanced contrast of the object created may be obtained by subtraction of said colored structured images because of the slight axial and lateral shift due to color.
In a further preferred variant of the second embodiment, an optical filter is placed between said camera lens 72 and said transparent vessel (container 3) to filter out the wavelength of said two-dimensional light patterns but still transmit the wavelength of said structured illumination.
In a further preferred variant of the second embodiment, a controller is used to synchronize the image acquisition by the camera 74 with the rotation of the photoresponsive material 103 and transparent vessel (container 3).
The flowchart of
The flowchart of
The flowchart of
The flowchart of
According to a third embodiment of the present invention, an accelerated correction procedure for the projected patterns is provided. According to this embodiment, each captured image of the object being formed is processed to detect which pixels of the image correspond to regions of the photoresponsive material that are already cured. This processing can be done for example by comparing each image to a reference image (background image) taken at the same angle before the start of exposure, and then applying a threshold to detect pixels that changed (pixels that “cured”), as explained in more detail for a previous embodiment. Other examples of processing include edge-detection filters or artificial intelligence-based image recognition. The background image and the captured image of the object being formed can be recorded with an imaging system like the imaging system 6,7,8 used in the first embodiment. However, the third embodiment is not limited with respect to the imaging system.
The processing yields a map of pixels that correspond to parts of the object being printed which are already cured, as seen from a specific rotation angle (i.e. the angle at which the camera image was recorded). Using this pixel map, the corresponding pixels at the corresponding angles in the set of projected light patterns are disabled (set to 0) or attenuated (set to a significantly lower value) in order to prevent over-curing at this location.
Thus, with this accelerated correction procedure, the camera images can be processed without knowledge of the shape of object, yet it still yields the desired corrections for the light patterns. More specifically, in this embodiment the camera images are neither compared to the two-dimensional projections nor to the three-dimensional model of the object, yet the method still produces corrected light patterns that avoid overexposure of parts of the object that are already cured during any point of the printing procedure. Because of the reduced amount of computations, this correction procedure can more easily be implemented with a high processing speed and in real-time.
It is understood that a camera image captured at a given angle can be used to correct one or multiple light projections at similar angles, depending on whether the rate of acquisition of the camera is equal to or different from the rate of projection of the light patterns. It is also understood that multiple camera images can be acquired at the same angle, or similar angles, in order to observe how the photoresponsive material changes over time as seen from this particular angle.
Thus, according to the third embodiment the method preferably comprises the following steps as illustrated in the flowchart of
The present invention is furthermore related to a system as already described above. In detail, the present invention is related to a system for obtaining a set of corrected projection patterns (9) and for printing an object (5) in a volumetric back-projection printer (1), comprising
| Number | Date | Country | Kind |
|---|---|---|---|
| 19181595.0 | Jun 2019 | EP | regional |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2020/064405 | 5/25/2020 | WO |
