Open Access
Add to BibSonomy
Abstract
The ability to 3D print optical elements will greatly expand the accessibility of optical fabrication. Here, we report on two fabrication techniques for plano-convex lens files using a consumer-grade lithographic printer. Lenses were post-processed using a simple spin coating technique with the resin used in the printing process or by curing directly on glass concave lenses. Average RMS roughness values were between 13 and 28 nm and RMS wavefront deviations were between 0.297 and 0.374 wave for spin-coated lenses. The average roughness RMS for the glass-cured lenses was 6 nm and the average form RMS was 0.048 wave.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Grinding and polishing are traditional manufacturing techniques for the fabrication of optical elements. These elements have conventionally been produced from glass, crystal, or metal materials, with material selection based on the desired portion of the electromagnetic spectrum [1–3]. Injection molded optics have become popular as a cost-effective way to mass produce optical components using optical quality polymers, especially for the production of aspheric elements [4–6]. Though all of these manufacturing techniques can be used to create high quality optics, they all struggle to achieve widespread accessibility. Grinding and polishing can be expensive in terms of both production cost and time. Furthermore, these methods often will require multiple expensive machines for the manufacturing of components of differing specifications. While molded optics are cost effective at scale, initial costs for mold development are high, limiting the effectiveness of this manufacturing method for prototyping and specialized manufacturing [3]. Diamond turning, while not a traditional choice for direct fabrication of refractive lens components, can be utilized to produce customized prototype parts. However, this method of manufacturing can be both time consuming and expensive, limits material selection to those that are diamond-turnable, and creates parts with increased scattering; limiting the effectiveness of parts manufactured with this method [7].
Additive manufacturing techniques have grown in popularity since their emergence in the 1990’s [8] and are now used in a wide variety of fields [9–13]. Several reasons exist for this growing prevalence. Parts can be produced in an inexpensive manner after the initial fixed cost of the 3D printer, designs can be modified and built on site on an as needed basis, and parts can be made in arbitrary geometries that may not be possible using traditional manufacturing techniques. Additionally, new materials and printing processes are continuously being developed to expand the capabilities of additive manufacturing [14–17]. However, additive manufacturing techniques were not developed for use in optical fabrication. Therefore, a major drawback to using traditional extrusion and lithographic based printers for optical production has been the relatively rough final part surface resulting from the layer-by-layer production process [18,19]. For these reasons, current research into adapting these printing methods for use in optical production has focused on post-processing of parts [20] and introducing modifications to the printing process [21]. Newer printers that utilize two-photon based polymerization have much higher resolutions and can produce parts with micron-level detail, sufficient to achieve optically smooth surfaces [22]. Unfortunately, the extremely high resolution of these printers corresponds to inefficient printing speeds and limits the effectiveness of these machines to small optics [23,24]. Additionally, the cost of these machines is around two-orders of magnitude higher than the costs of conventional lithographic and extrusion based units. The increased cost and limitations on part size relegates the use of two-photon methods to specialized laboratory environments. Finally, recent efforts have been made to produce 3D printed optics using ink-jet printing equipment developed for optical fabrication [25]. However, even commercially manufactured optics fabricated using these machines struggle to achieve good surface quality [26].
Stereolithography is a subset of additive manufacturing that uses curable UV resin and optics to build up parts in a layer-by-layer process [27]. This approach to 3D printing is particularly appealing for the production of optics due to both the availability of clear printing materials and the relatively higher resolution these printers can achieve when compared to extrusion-based printing. Commercially available lithographic printers can achieve resolutions up to 25 microns. Additionally, these printers are much cheaper and print with much faster speeds than two-photon based methods. The ability to 3D print optical elements with a consumer printer will greatly expand accessibility of optical element fabrication by keeping required costs and infrastructure down.
Here, we present an example and optimization of the process to create plano-convex lenses with a consumer-grade Form 2 stereolithography printer (Formlabs, MA). Parts are post-processed using the same resin from which they are composed, eliminating the need for external consumable materials. Post-processing of convex surfaces of lenses was performed by spin coating. Alternatively, a secondary finishing method was also investigated in which convex sides of lenses were finished utilizing a concave lens with matching radius of curvature. Flat portions of the lenses for both methods were cured on a glass substrate that was removed at the conclusion of the curing process. Resin properties including index of refraction, transmission spectrum, and autofluorescence were measured during the production and evaluation process. Two plano-convex lenses of differing specifications were produced and processed and these lenses were then compared to glass lenses with matching diameters and focal lengths. Lenses were evaluated using white light interferometry to compare surface roughness to that of glass lenses. Surface shape was compared to control lenses using Fizeau interferometry. Results indicated lenses fabricated by utilization of a conjugate concave lens achieved low form RMS values, with an average of 0.048 wave. As a comparison, glass lens with the same design parameters had a form RMS of 0.023 wave. The spin coating of lenses resulted in worse quality, but allowed for the production of freeform elements. Finally, images of a negative 1951 United States Air Force resolution target were captured using glass and 3D printed lenses and compared.
2. Methods
Two plano-convex lenses were fabricated to match commercially available lenses from Thorlabs, and the design specifications for these lens files are listed below in Table 1. Glass lenses were purchased from Thorlabs to provide standards with which to compare created plastic lenses. For each lens file, 3D printed lenses were made in batches of eight lenses at a time and then compared with glass lenses. Additionally, two different test pieces, an 8 mm by 15 mm by 3 mm sample block and a 35 mm by 8 mm by 5 mm sample block were produced for material characterization testing. Figure 1 below shows components that were examined in the study. Lenses were produced using a Form 2 commercial lithographic printer and corresponding Clear Resin (Formlabs, MA).
Fig. 1. Images of all tested components. Glass lenses are shown to the left of the line while 3D printed lenses are located to the right of the line. A sample block used in the study for material characterization is also shown.
Table 1. Specifications for Lenses Tested in Study
The Clear Resin material was characterized for transmission, refractive index, and autofluorescence. Transmission testing was performed on an 8 mm by 15 mm by 3 mm sample block that was finished after printing by coating the top 15 mm by 8 mm surface with 350 µL of resin and spinning at 1600 RPM for 10 seconds and then curing for 6.5 hours. The bottom surface was finished by placing it on a glass slide coated with approximately 90 µL of resin. The slide was placed in a vacuum chamber for 1.5 hours to remove bubbles between the glass and printed surface and then cured for 6.5 hours. Index of refraction characterization was conducted on a 35 mm by 8 mm by 5 mm block, post-processed by placing one of the 35 mm by 8 mm surfaces on approximately 120 µL of resin on a glass slide and cured for 6.5 hours. Slides were then placed in the freezer for 15 minutes and the sample blocks could be removed from the slides due to material differences in thermal expansion. Refractive index of the clear resin was measured on an Atago Multi-Wavelength Abbe DR-M4 refractometer, allowing for measurement at four different wavelengths (486, 546, 589, and 656 nm). Prior to taking each measurement, calibration was performed using a test piece of known refractive index (n = 1.6199). A thin layer of monobromonapthalene was used as the contact liquid and all measurements were taken at room temperature. Measurements were taken on three different sample blocks, each of which was fabricated using a different resin cartridge to ensure consistency between batches. Transmission of the cured resin was measured across a range of wavelengths (300-1050 nm) using a Cary 50 UV-Vis spectrophotometer. Measurements were taken using two different sample blocks and each block was measured three times. All transmission data was then averaged and compared to PMMA transmission and transmission of a commercial UV curable optical resin (Luxexcel). Autofluorescence emission was measured on two different sample blocks using a Horiba Fluorolog-3 spectrofluorometer. Excitation wavelengths from 300-750 nm spaced at four nm intervals were provided and emission data was collected from 300-1000 nm at two nm intervals. Autofluorescence results were then compared to those taken from measurement of the commercial UV curable resin.
The optimal fabrication process for lenses was dependent on the curvature of the lens. In general, most variables remained the same between lens files, with only slight differences in the spin profile between lenses. Best results for both lenses were seen when printing at an approximately 60° angle relative to the print bed (as seen in Fig. 2) with convex surfaces facing the back of the printer. Following the printing process, lenses were washed in 99% isopropyl alcohol (IPA) for 30 minutes to remove excess resin and allowed to dry for an additional 30 minutes in air at room temperature. Lenses were post-cured on printing supports using a heated UV curing chamber (Formlabs, MA) at 60 C for 15 minutes.
Fig. 2. Printing angles examined on 15 mm focal length lenses. From left to right: 50°, 60°, and 70° printing angles.
Printing supports were clipped and lenses were spin coated with the same resin used in the printing process. Best results were seen when depositing approximately 600 µL of resin on the lenses before spinning. Highest quality 25 mm focal length lenses were produced by spinning the lenses at 1600 RPM for 9 seconds. Similarly, the highest quality 15 mm focal length lenses were produced by spinning at 1000 RPM for 9 seconds. A secondary method for coating of convex sides of lenses was also investigated using commercially available concave glass lenses with matching radius of curvature and diameter to the 25 mm focal length plano-convex lens file. A small amount of resin was placed on the concave glass surface and the convex 3D printed surface was pressed onto the glass surface to ensure that resin coated the entire convex surface. The 3D printed lens was cured on top of the glass surface, and the glass lens could be removed following the curing process by freezing the lens structure. Lenses for both the spin coating and glass curing technique were cured in UV light for 6.5 hours with the heater element of the curing chamber turned off. After curing of the curved surface, flat surfaces were finished by placing the flat portion of the lens on approximately 75 µL of resin on a glass slide. The slide was then placed in a vacuum chamber for 1.5 hours to remove bubbles that formed between the 3D printed surface and the glass slide. The lens was cured on the slide for an additional 6.5 hours. Following the second cure, the slide was placed in a freezer for 15 minutes, which allowed for separation of the lens from the slide due to differences in thermal expansion between the two materials.
After fabrication, a Zygo NewView 5000 (Zygo, CT) white light interferometer was used to characterize surface roughness and radius of curvature. Acquired measurements used a 10X Mirau objective with 1.3x zoom, corresponding to a measurement area of .22 mm2 (.55 mm x .4 mm). The Zygo micro.app software was used to assess cured surfaces with a 20 µm vertical scan. Built-in interpolation was also used to measure radius of curvature of each 3D printed lens over the same .22 mm2 area and this measurement was recorded. Measured roughness profiles had the radius of curvature subtracted out and a 2.51 µm low-pass filter was utilized to remove noise from the data. The following characteristics of each lens was recorded following the removal of the spherical component: the peak to valley height (PV), root mean squared roughness (RMS), and roughness average (RA). Surface quality was characterized using a Zygo PTI 250 Fizeau interferometer. All lens surfaces were measured using a F/# 4.8 reference sphere at 655 nm. Acquisition and post-processing settings remained the same for each lens measurement. Measurement position was adjusted based on observed interference pattern and the axial position of the measurement head was adjusted to remove z-axis offset. For each lens, peak to valley (PV), root mean squared (RMS), astigmatism magnitude, and coma magnitude were recorded. The optical resolution of created lenses was then evaluated using a negative 1951 USAF resolution target. The target was illuminated by a white light source, and then imaged through the 3D printed lens. The imaging setup provided 1x magnification and was constructed such that it consisted solely of the illumination setup, the created lens, and the camera. Target images were recorded on a 12-bit, monochromatic Flea3 CCD camera (FLIR, OR). Due to the naturally occurring chromatic aberrations in a singlet lens, a 500 nm narrow-band filter (10 nm FWHM) was placed directly behind the illumination system.
3. Results
3.1 Material characterization
Material study of the clear resin resulted in measurements for refractive index, transmission spectrum, and autofluorescent emission spectra. Measurements for the refractive index of the resin were performed on samples made from three different resin cartridges to ensure consistency. The results of the refractometer measurements are plotted in Fig. 3. All refractive index measurements were limited to the 486-656 nm spectrum range due to the set of narrow-band filters compatible with the refractometer.
Fig. 3. Refractive index measurements of cured clear resin material. Error bars indicate standard deviations.
The transmission of fabricated test blocks was measured from 300-1050 nm. The average transmission measurements are displayed below in Fig. 4, along with results for blocks of PMMA and a commercially available optical resin (Luxexcel). To be consistent with previously reported results, transmission plots show combined material transmission and Fresnel reflection related losses. We estimate, however, that average material related absorption and scattering loses are approximately 3% for a 3 mm thick sample. Results show that there is not a significant difference in transmission between Clear Resin and other optical polymers that are in use, as the Clear Resin material averages 89.5% transmission from 500-1050 nm. However, while all three polymers display a sharp decrease in transmission towards the UV end of the spectrum, this transition occurs at higher wavelengths for the Clear Resin material.
The same samples used for transmission testing were also used to measure autofluorescence of the Clear Resin material. Excitation wavelengths between 300-750 nm were supplied and corresponding emission data between 300-1000 nm was collected. The results are displayed below in Figs. 5(a) and 5(b), comparing the autofluorescent data between the Luxexcel material and the used SLA material. Due to the higher fluorescence intensity of the SLA material, it appears that optical components made from this material will cause relatively high fluorescent background for use in applications with wavelengths shorter than 600 nm.
Fig. 5. Autofluorescence of printed materials. Luxexcel material is shown in panel A, Clear Resin material is shown in panel B.
3.2 Optimal printing methodology
Several different variables were examined in the manufacturing process. In general, lenses should be printed at an angle to balance the impact of layer lines and deformations. Printing lenses parallel to the printing bed results in lenses with prominent layer lines that prevent the coating of the convex surface while preserving desired lens curvature. Printing lenses perpendicular to the printing bed resulted in lenses with larger deformations and correspondingly worse form measurements. Additionally, consideration must be given to support structures to avoid placing supports on the lens surfaces. Angles of 50° to 70° relative to the print bed were found to balance deformations and support structure requirements, whereas lenses printed at angles less than 50° had support marks on the lens surface and lenses printed above 70° had large form deformations. Results, as shown in Fig. 6, indicated a trend towards lower printing angles leading to smoother surfaces. However, lenses printed at 60° had the best form measurements, which can be seen in Fig. 7. This is likely due to the requirement to use a lower density of supports for the 50° lenses such that no support marks are left on the planar surface of the lenses. It is hypothesized that the lower amount of supports used lead to lower form measurements, as there were not as many supports providing stability to the lens during the printing process, leading to a higher prevalence of deformations.
Fig. 6. Impact of printing angle on final surface roughness. Error bars indicate standard deviation.
Fig. 7. Impact of printing angle on final lens form measurements. Error bars indicate standard deviation
Lenses printed at the highest resolution setting on the printer (25 µm) had better form measurements than those printed at lower resolutions (50 and 100 µm). The resolution at which the design is converted to .stl file must be high enough to avoid the presence of triangular patterning on the surface; however, further improvement in .stl conversion resolution beyond the resolution limit of the printer did not have a significant impact on final element quality. Lens .stl files were uploaded to the manufacturers slicing program and the automated slicing was left unedited.
3.3 Optimal coating method determination
Static coating (placing resin on lens prior to spin) was preferable to dynamic coating (applying resin after initiating spinning). Placing lenses in a vacuum chamber prior to curing of resin on convex surfaces lead to poor surfaces and the reemergence of layer lines. We hypothesize that this is due to the resin running off the sides of the lens during the degassing period. While most variables remained unchanged between lens files, the optimal spin profile did change between different lens files. Optimal spin coating settings were determined by comparing interferometry data between profiles. Large outliers due to the presence of macro-level surface defects (bubbles or uncoated patches on lenses) were removed for data analysis; however, we do note the number of outliers recorded with each coating profile. As can be seen in Figs. 8 and 9, the spin speed that produced the best average lens was 1600 RPM for the 15 mm focal length lens files. However, when analyzing each individual lens fabricated across the different spin speeds, the overall best quality lens (lowest observed values for roughness RMS and form RMS) was produced with a 1000 RPM spin speed even though coating at this speed produced a much wider range of lenses. Since the 1000 RPM spin speed produced the best lens, this coating speed was used for the fabrication of lenses for imaging experiments. Additionally, no outliers out of the eight lenses made with this spin speed were seen as opposed to the other speeds that provided between one to three outliers. For the 25 mm focal length lens file, spinning at 1600 RPM for approximately 9 seconds yield smooth lenses with accurate shape. Additionally, this coating profile produced the fewest outliers with one of the eight displaying large surface defects, while all other speeds contained at least two outliers. Results from measured lenses can be seen in Figs. 8 and 9 below.
Fig. 8. Impact of coating speed on the average lens surface RMS roughness for 15 and 25 mm focal length lenses. Error bars indicate standard deviations.
Fig. 9. Impact of spin speed on final average form RMS measurements for 15 and 25 mm focal length lenses. Error bars indicate standard deviations.
There were small differences in tested spin speeds (800, 1000, 1300, and 1600 for 15 mm focal length lenses and 1000, 1300, 1600, and 2000 for 25 mm focal length lenses) due to the differing curvature of the files. This decision was made because initial testing of shorter focal length lenses at high spin speeds indicated that it became easier for resin to become dislodged due to the steeper curvature. This creates patches that lack any coating and cause both roughness and form deviations. Figure 10 below shows 15 mm focal length lenses at different time points in the fabrication process.
Fig. 10. Steps in fabrication of 15 mm focal length lens. From left to right: lens on support after printing, lens removed from support, lens with convex side coated, and fully coated lens.
3.4 Roughness and radius of curvature measurements
Radius of curvature, peak to valley, root mean squared, and Ra were measured for reference glass lenses and manufactured lenses using white light interferometry. Summarized results are listed below. Minimum, maximum, and average values are listed to indicate the range of quality that may arise in the fabrication process. For the 15 mm focal length lenses, a batch of eight lenses made by spinning at 1000 RPM spin speed for 9 seconds is shown in Table 2. For the 15 mm focal length file, radius of curvature measurements of the eight lenses varied between 0.8%-12% of the nominal value, however the average variation (1%) was similar to the variation of the glass lens (0.8%). Average roughness measurements differed from the glass lens by about a factor of seven for the root mean squared and Ra measurements and a factor of about four for the PV measurements. It is also noted that the process leads to the possibility of significant outliers and poor-quality lenses, as indicated by the maximum measurement of the eight lenses below.
Table 2. Measured roughness values for 15 mm 3D lenses in comparison to glass references.
Table 3. Roughness data for 25 mm focal length lenses in comparison to glass reference lenses.
For the 25 mm focal length lenses, a batch of eight lenses made using a 1600 RPM spin speed for 9 seconds is shown in Table 3 (one lens was damaged in the analysis phase so data is shown for only seven lenses). Similar to the 15 mm focal length lenses, there was a wide variation in radius of curvature measurements, but the average value is similar to that of the glass lens. Additionally, we note that longer focal length lenses fabricated using the spin coating process tend to be smoother than shorter focal length lenses. As mentioned above, this may result from the difficulties inherent in evenly coating a more severely curved surface. Additionally, a batch of lenses made using the glass curing method is shown in Table 3 (Eight lens were created; however, one outlier was excluded from data analysis). Noticeably, lenses fabricated in this fashion outperform those made using the spin coating method, albeit at the loss of geometrical freedom. Figure 11 below compares the average roughness RMS values between commercial glass lenses, those made using the glass curing method, and those using the spin coating method.
Fig. 11. Surface roughness between the two different finishing methods and glass reference lenses. Error bars indicate standard deviations.
Fig. 12. Surface roughness profile of spin coated (A), glass cured (B) and glass reference lenses (C). All surfaces are shown at the same scale for ease of comparison.
Figures 12(a)–12(c) below shows the 2D surface profiles for a 25 mm focal length spin coated lens, glass cured lens, and glass reference lens. For all lenses, the measured radius of curvature has been removed. The spin coated lens has deviations of 97 nm while the glass lens deviates by 49 nm. The 2D map of the spin coated lens appears to show the effect of layer lines inherent in the printing process, as visible rings encircle the apex of the lens. Noticeably, these rings are not as prominent in the surface profile of the glass cured lens or the glass reference lens.
3.5 Form measurements
Wavefront deformations were measured using Fizeau interferometry. Summarized statistics for peak to valley, root mean squared, astigmatism magnitude, and coma magnitude are given in Table 4 and Table 5. Results in wave are rounded to the nearest hundredth. Again, the minimum, maximum, and average values from the batch are given to show the variance of the fabrication process. 3D printed spin coated lenses in general display significantly larger aberrations than glass equivalents. In contrast to the trends seen in roughness measurements, shorter focal length lenses tend to have better form measurements than longer focal length lenses. We theorize that this may be due to the printer being less accurate in faithfully producing gradual curves. All 3D printed lenses display large astigmatism magnitudes, which is likely inherent in the printing process due to similar findings in other 3D printed component study [20,26]. In general, 3D spin coated lenses differ in PV and RMS measurements by about one order of magnitude and astigmatism magnitude by about two orders of magnitude. Coma magnitude was found to be higher in the longer focal length lenses, with values about an order of magnitude higher when compared to the 15 mm focal length lenses. Figure 13 compares the average form RMS for glass cured and spin coated lenses in comparison to glass lenses.
Table 4. Form measurements for 15 mm 3D printed lenses in comparison to glass references.
Table 5. Form measurements for 25 mm 3D printed lenses in comparison to glass references.
Fig. 13. Form RMS between two different finishing methods and glass reference lens. Error bars indicate standard deviations.
Representative 2D wavefront plots are shown in Figs. 14(a) and 14(c) for a 25 mm focal length spin coated lens, glass cured lens, and a glass reference equivalent. Plots are shown at the same scale for ease of comparison between the methods.
Fig. 14. Form profiles for spin coated lens (A) in comparison to glass cured lens (B) and glass reference lens (C). All plots have the same scaling.
Surface profiles of spin coated lenses show a high degree of systemic and random error as shown in Fig. 15. Errors for the 15 mm focal length lenses all appear to have similar profiles indicating a source of systemic error in the process. Unfortunately, once supports are removed from the lens surface, it becomes difficult to track lens orientation on the printer. Further study is needed to conclusively state how the systematic error occurs in relation to the printing direction. In contrast, the 25 mm focal length lenses seem to be dominated by random errors, as the profiles do not share obvious similarities.
Fig. 15. (A): Systematic error in fabrication of 15 mm lens files. (B): Random error in the fabrication of 25 mm focal length lens files.
3.6 Imaging performance
While characterization measurements of surfaces using white light and Fizeau interferometry can provide a reasonable ability to determine the optical capabilities of an element, experimental imaging measurements were also performed to ensure that surface quality is the overall deterministic factor for 3D printed lenses. Experimental imaging was performed using a 1951 USAF resolution target. Images for the 15 mm focal length lens are shown in Fig. 16 along with the corresponding 3D printed lens image. Images for the 25 mm focal length lenses are shown in Fig. 17, showing results with both glass cured and spin coated lenses. Imaging experiments were set up such that they provided for 1x magnification. Lenses used in the imaging setup were those that had the smallest wavefront RMS values. The glass 15 mm focal length lens was able to achieve a resolution of 143.7 lp/mm whereas the spin coated version achieved 90.5 lp/mm and was limited by the astigmatism in the lens. The glass 25 mm focal length lens was able to achieve 90.5 lp/mm, the glass-cured lens resolved 90.5 lp/mm and the spin coated version was able to achieve 80.6 lp/mm. It can be seen that the 25 mm focal length spin coated lens shows significantly more background than the corresponding 15 mm focal length spin coated lens. We believe that this likely stems from the superior form measurements of the 15 mm focal length lenses. Interestingly, both spin coated lenses showed a halo effect that limited resolutions achieved. However, resolutions achieved with the spin coated lenses are sufficient for many applications that do not require high-resolution imaging such as illumination applications. 3D printed lenses that were finished using the glass curing method were able to achieve better resolution limits than their spin coated counterparts and are a more logical choice for applications requiring higher resolution when a glass counterpart to use in the curing process is available.
Fig. 16. 15 mm Focal Length Lens. Panel A shows imaging using glass lens purchased from Thorlabs. Panel B shows imaging using 3D printed spin coated lens.
Fig. 17. 25 mm Focal Length Lens. Panel A shows imaging using glass lens purchased from Thorlabs. Panel B shows imaging using 3D printed glass cured lens. Panel C shows 3D printed spin coated lens.
3.7 Fabrication costs
There is an initial fixed cost for the printing setup totaling roughly $4000. Additionally, the spin coater available for this study had an approximate price of $5000. Although having this quality of spin coater is not necessary for the application and more affordable devices should be suitable. The material cost of each lens for either the spin coating method or the glass curing method was estimated at $0.24 (glasses lenses matching these designs cost $22.94 and $20.89 respectively). Commercial lenses that were used as a mold (conjugate surface) for the glass curing method had a low cost of $18, and could be cleaned and reused to make additional lenses without any observed reduction in quality. The low cost of the method makes conjugate curing an attractive option for affordable low volume production of 3D printed polymer lenses. However, we would anticipate most adaptations of these 3D fabrication techniques to focus on complex lens geometries or larger lenses with associated higher commercial costs.
4. Discussion
We have described a process for fabricating 3D printed lenses using an inexpensive, consumer grade stereolithographic printer. Lenses were put through simple post-print processing and their roughness, wavefront aberrations, and imaging capabilities were compared to glass lens components of the same geometries. Roughness measurements indicated that surface Ra values for most spin coated lenses were between 10 and 30 nm, which was within one order of magnitude compared to the glass polished lenses but comparable with plastic molded lenses. Measurements taken with a Fizeau interferometer indicated that spin coated lenses had form measurements that were worse than reference lenses. Root mean squared data indicated deviations on the order of tenths of a wave. Lenses displayed a high level of astigmatism that indicates a limitation of the fabrication process. Roughness measurements for the lenses that were cured on matching concave lenses were between 3 and 14 nm, which is comparable to the glass reference lenses. Likewise, form RMS measurements for this group of lenses were between 0.021 and 0.097 wave, and on average was double the form RMS of the glass lenses but about an order of magnitude better than the spin coated lenses.
Experimentally determined resolutions of created lenses were compared to glass lenses. As expected from the interferometry measurements, coated lenses displayed a lower resolution than that of glass lenses. From data collected, it appears that this 3D production process favors the creation of lenses with shorter focal lengths, corresponding to a more curved surface. Additionally, general trends indicate a balance between longer focal length lenses having smoother surfaces, and shorter focal length lenses having better form measurements. 3D printed lenses that were cured on glass concave lenses performed similarly to reference glass convex lenses. Analysis of the glass cured lenses indicated very low form RMS, with averages below 0.05 wave. Material study of the resin used here indicates a refractive index similar to that of PMMA and slightly different than the supplier’s listed index of refraction (1.5304). Transmission measurements through a 3 mm thick sample indicated a similar level of light transmission to that of PMMA or other optical polymers.
3D printing is an intriguing option for the future fabrication of optics as it allows for the quick creation of optics in a variety of geometries that may be difficult to produce in conventional methods. Additionally, while this study focused only on one type of material, there exists many different materials for stereolithographic printing which may have suitable material properties for the fabrication of optics. Initial efforts to adapt this methodology to different surface geometries, including biconvex, aspheric, and dispersive elements have indicated promise. However, the fabrication of plano-convex components allows for easier and more reliable characterization, as aspheric elements present much greater characterization requirements. Therefore, only plano-convex lenses were examined in this study. In general, however, production of aspheric lenses can be carried out with the exact same fabrication process, as 3D printing is not bound to any particular lens shape. Selection of post-processing method should be made with consideration to desired final element shape. Plano-convex, plano-concave, and biconvex element production is possible with either post-processing method presented here and selection should be based on available commercial conjugates and needed surface quality. Aspheric element fabrication favors post-processing using spin coating unless a conjugate aspheric surface were to be manufactured using a more traditional optical fabrication method. If necessary these reference surfaces can be made with fabrication methods including diamond turning or free form grinding and polishing in order to create suitable conjugates. Lenses with radius of curvatures between 7.7 and 51.5 mm have been successfully fabricated using the spin coating method; however, wavefront error correlates with increasing radius of curvature. Further investigation is needed to provide conclusive recommendations for adapting this technique for lenses outside of this curvature range. The coating process is similar for all lens files, although in general, lower radius of curvatures will not perform as well at high spin speeds. Together the spin coating and glass curing post-processing approaches to 3D printed elements should provide a platform for testing and quick redesign of components at a low cost, and expand the accessibility of optical element creation.
Funding
National Science Foundation (1648451); Precise Advanced Technologies and Health Systems for Underserved Populations (PATHS-UP) Engineering Research Center.
Disclosures
Dr. Tomasz Tkaczyk has financial interests in Attoris LLC focusing on applications and commercialization of hyperspectral imaging technologies.
References
1. D. Malacara and B. J. Thompson, Handbook of Optical Engineering (Marcel Dekker, Inc, 2001), pp. 915–957.
2. H. H. Karow, Fabrication Methods for Precision Optics (Wiley & Sons, 2004).
3. M. P. Schaub, The Design of Plastic Optical Systems (SPIE, Bellingham, WA, 2009).
4. S. Bäumer, Handbook of Plastic Optics, (Wiley-VCH, 2005).
5. L. Dick, “High Precision Freeform Polymer Optics,” Opt. Photonik 7(2), 33–37 (2012). [CrossRef]
6. G. Cogburn, A. Symmons, and L. Mertus, “Molding aspheric lenses for low-cost production versus diamond turned lenses,” Proc. SPIE 7660, 766020 (2010). [CrossRef]
7. A. Kotha and J. E. Harvey, “Scattering effects of machined optical surfaces,” Proc. SPIE 2541, 54–65 (1995). [CrossRef]
8. E. Sachs, M. Cima, P. Williams, D. Brancazio, and J. Cornie, “Three Dimensional Printing: Rapid Tooling and Prototypes Directly from a CAD Model,” J. Eng. Ind. 114(4), 481–488 (1992). [CrossRef]
9. C. M. B. Ho, S. H. Ng, K. H. H. Li, and Y. Yoon, “3D printed microfluidics for biological applications,” Lab Chip 15(18), 3627–3637 (2015). [CrossRef]
10. S. J. Trenfield, A. Awad, A. Goyanes, S. Gaisford, and A. W. Basit, “3D Printing Pharmaceuticals: Drug Development to Frontline Care,” Trends Pharmacol. Sci. 39(5), 440–451 (2018). [CrossRef]
11. M. Vukicevic, B. Mosadegh, J. K. Min, and S. H. Little, “Cardiac 3D Printing and Its Future Directions,” JACC Cardiovasc Imaging 10(2), 171–184 (2017). [CrossRef]
12. S. V. Murphy and A. Atala, “3D bioprinting of tissues and organs,” Nat. Biotechnol. 32(8), 773–785 (2014). [CrossRef]
13. C. L. Ventola, “Medical Applications for 3D Printing: Current and Projected Uses,” Pharm. Therapeutics 39(10), 704–711 (2014).
14. J. H. Martin, B. D. Yahata, J. M. Hundley, J. A. Mayer, T. A. Schaedler, and T. M. Pollock, “3D printing of high-strength aluminium alloys,” Nature 549(7672), 365–369 (2017). [CrossRef]
15. J. R. Tumbleston, D. Shirvanyants, N. Ermoshkin, R. Janusziewicz, A. R. Johnson, D. Kelly, K. Chen, R. Pinschmidt, J. P. Rolland, A. Ermoshkin, E. T. Samulski, and J. M. DeSimone, “Continuous liquid interface production of 3D objects,” Science 347(6228), 1349–1352 (2015). [CrossRef]
16. D. Zhang, B. Chi, B. Li, Z. Gao, Y. Du, J. Guo, and J. Wei, “Fabrication of highly conductive graphene flexible circuits by 3D printing,” Synth. Met. 217, 79–86 (2016). [CrossRef]
17. C. Huber, C. Abert, F. Bruckner, M. Groenefeld, S. Schuschnigg, I. Teliban, C. Vogler, G. Wautischer, R. Windl, and D. Suess, “3D Printing of Polymer-Bonded Rare-Earth Magnets With a Variable Magnetic Compound Fraction for a Predefined Stray Field,” Sci. Rep. 7(1), 9419 (2017). [CrossRef]
18. Y. Wang, J. Gawedzinski, M. E. Pawlowski, and T. S. Tkaczyk, “3D printed fiber optic faceplates by custom controlled fused deposition modeling,” Opt. Express 26(12), 15362–15376 (2018). [CrossRef]
19. Y. L. Sung, J. Jeang, C.-H. Lee, and W.-C. Shih, “Fabricating optical lenses by inkjet printing and heat-assisted in situ curing of polydimethylsiloxane for smartphone microscopy,” J. Biomed. Opt. 20(4), 047005 (2015). [CrossRef]
20. N. Vaidya and O. Solgaard, “3D printed optics with nanometer scale surface roughness,” Microsyst. Nanoeng. 4(1), 18 (2018). [CrossRef]
21. X. Chen, W. Liu, B. Dong, J. Lee, H. O. T. Ware, H. F. Zhang, and C. Sun, “High-Speed 3D Printing of Millimeter-Size Customized Aspheric Imaging Lenses with Sub 7 nm Surface Roughness,” Adv. Mater. 30(18), 1705683 (2018). [CrossRef]
22. T. Gissibl, S. Thiele, A. Herkommer, and H. Giessen, “Sub-micrometre accurate free-form optics by three-dimensional printing on single-mode fibres,” Nat. Commun. 7(1), 11763 (2016). [CrossRef]
23. J. Li, P. Fejes, D. Lorenser, B. C. Quirk, P. B. Noble, R. W. Kirk, A. Orth, F. M. Wood, B. C. Gibson, D. D. Sampson, and R. A. McLaughlin, “Two-photon polymerisation 3D printed freeform micro-optics for optical coherence tomography fibre probes,” Sci. Rep. 8(1), 14789 (2018). [CrossRef]
24. S. Thiele, K. Arzenbacher, T. Gissibl, H. Giessen, and A. M. Herkommer, “3D-printed eagle eye: Compound microlens system for foveated imaging,” Sci. Adv. 3(2), e1602655 (2017). [CrossRef]
25. B. G. Assefa, M. Pekkarinen, H. Partanen, J. Biskop, J. Turunen, and J. Saarinen, “Imaging-quality 3D-printed centimeter-scale lens,” Opt. Express 27(9), 12630–12637 (2019). [CrossRef]
26. J. Gawedzinski, M. E. Pawlowski, and T. S. Tkaczyk, “Quantitative evaluation of performance of 3D printed lenses,” Opt. Eng. 56(08), 1 (2017). [CrossRef]
27. F. P. W. Melchels, J. Feijen, and D. W. Grijpma, “A review on stereolithography and its applications in biomedical engineering,” Biomaterials 31(24), 6121–6130 (2010). [CrossRef]
