Open Access
Add to BibSonomy
Abstract
We characterize three commercial resins suitable for three-dimensional two-photon printing of mm3 volume micro-optical components for visible light –IP-S, IP-n162, and IP-Visio– under different print modes and post-processing conditions. Due to the combination of cured resin absorption and bulk scattering, we find a maximum total printed thickness of 4 mm (or greater) for at least 50% transmittance of red light, up to 2 mm for green light, and large maximum thickness variation for blue light (0.1 to 1 mm).
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Three-dimensional (3D) printing (and more general additive manufacturing methods) has created a revolution in rapid-prototyping and on-demand part creation, with printers ranging from hobbyist desktop systems to industrial prototyping and manufacturing. This technology further enables new devices and structures to be designed and fabricated that cannot be built using any other method. Most instruments are limited to minimum feature sizes on the order of 100 $\mu$m due to limitations in materials and curing methods, nozzles, or tolerances and accuracy in positioning hardware. A recent review of additive manufacturing for optical components is provided in [1]. Printing bulk (transmission) optics can result in high light scattering with common 3D printing methods such as fused deposition modeling (filament melting), due to bulk inhomogeneity and surface roughness, or UV stereolithography, due to surface roughness. Systematic studies of light transmittance, including scattering, have been performed using relatively opaque materials with a thermal filament printer [2]. High quality cm-scale optics, such as lenses, can be produced by ink-jet techniques with UV curing [3,4]. Three-dimensional printing methods have also been used for tissue phantoms, where scattering and/or fluorescing elements are deliberately incorporated into stock materials to mimic the optical properties of tissue [5–7].
The two-photon printing (2PP) method [8] specifically has been reviewed in [9,10]. In 2PP, the light source is typically a near-infrared (NIR) femtosecond laser, where two-photon exposure facilities spatially localized polymerization within a 3D resist or resin volume. With $\mu$m to sub-$\mu$m lateral and vertical feature sizes [11], 2PP opens up the capability to create high-resolution complex multi-surface micro-optics [12–16] of $\sim$mm$^{3}$ volumes, and has additional advantages of surface smoothness [9,17]. In part because of higher absorption losses due to incorporation of photoinitiators in the near-UV (around 400 nm) for 2PP resins [18,19], most micro-optics/long optical path structures and devices have been printed for NIR applications, with comparatively less work performed for visible wavelengths. In this paper, we focus on optical absorption and scattering of three commercial 2PP printing resins that are suitable for visible micro-optics. Using the Nanoscribe Photonics Professional GT2 system with DeScribe software, we studied how these optical properties are affected by print mode and post-processing, and we establish some upper bounds on the total thickness of printed micro-optical components.
2. Two-photon 3D printing
2.1 Print resins
There are a number of photoresins that can be used for 2PP. For visible micro-optics, three appropriate resins provided by Nanoscribe are IP-S, IP-n162 and IP-Visio. The optical properties (refractive index and absorption coefficient) of these resins depend upon the degree of polymerization, which depends upon exposure method (UV or two-photon), dose, and any post-exposure heat treatment [20–23]. Under UV development, the IP-S resin is slightly absorbing for blue light (about 0.3 mm$^{-1}$ near 450 nm wavelength [23]), but is otherwise extensively used for 3D printed optical devices. IP-n162, previously denoted PO4, has the advantage of higher refractive index (1.652 versus 1.522 at 450 nm) [23], while IP-Visio in intended for use in applications that require low autofluorescence, in part due to reduced absorption [23].
2.2 Process parameters
The Photonics Professional system has two fill modes. In solid mode, every voxel within the structure volume is two-photon printed (or polymerized). In this mode, we chose a slicing distance (distance between layers) of 1 $\mu$m with a simplification tolerance (minimal distance between two printing points) of 0.05 $\mu$m. The hatching distance refers to the distance between lines scanned by the laser in a single plane, and was set to 0.5 $\mu$m. Two options are available for the scan mode. Galvo mode was chosen as it is $\sim$100$\times$ faster than piezo mode. However, we used piezo mode along the $z$-axis for higher resolution, and the printed block size (before stitching) was about 285 $\mu$m $\times$ 285 $\mu$m $\times$ 200 $\mu$m. With IP-S and IP-n162 resins and a 25$\times$ objective, we used a scan speed of 100 mm/sec at 100% and 60% laser power, respectively, with a print time per mm$^{3}$ of about 9.1 hours in solid mode. With IP-Visio, scan speed of 55 mm/sec and 100% laser power were used, with a print time of 17 hours/mm$^{3}$.
One of the major motivations for 3D printing is rapid prototyping of novel structures. With solid mode, the print time can be long, especially for large structure volumes that accompany micro-optics designs. To decrease the total printing time, the surfaces of the structure can be two-photon printed instead of printing the entire volume. DeScribe calculates the shell that defines the structure’s surfaces. Moreover, it defines an interior scaffold to increase the mechanical stability of the structure during printing and development. The stability of the structure depends on the thickness of the shell, type of scaffold (triangular, rectangular or hexagonal), wall to floor spacing, and scaffold thickness. These parameters can be optimized to decrease print time and improve the mechanical stability of the printed part, but also have an effect on light transmittance. Print time per mm$^{3}$ was around 2 hours in shell mode in our studies. The final structure has unpolymerized resin enclosed by the outer shell and interior scaffold. The shell prevents developer from reaching the unpolymerized interior of the structure.
Values of the various writing parameters used are listed in Appendix A of Supplemental Materials.
2.3 Post-processing
Shell mode-printed devices typically undergo UV exposure after developing in order to polymerize the interior liquid volumes. However, the refractive index (RI) of resin polymerized under TPA may not be the same as that polymerized under UV, depending upon the UV dose [23]. Generally, polymerization conversion under 2PP is in the 30% to 50% range [24]. Post-processing is up to the user, where combinations of UV dose and heat treatment can lead to comparable conversion and RI to that of two-photon polymerization [23]. Any RI mismatch between the two-photon printed scaffolding and UV exposed interior volume (potentially with lower polymerization conversion, and lower RI) leads to light scattering within the volume.
In our post-processing steps, a UV light source (Litex 680A from Dentamerica) with a power density of 600 mW/cm$^{2}$ was used, with a sample to source distance of 1 cm. Heat treatment was performed in a lab oven at $\sim$60 $^{\circ }$C.
2.4 Samples tested
We first characterized homogeneous samples, starting with resins in the liquid state, and then resins that have undergone UV exposure and heat treatment. For these samples, IP-S, IP-n162 or IP-Visio liquid resin was placed in-between two glass slides. The thickness of the resin was maintained at 700 $\mu$m by silicon pieces also placed in-between the glass slides.
We next characterized solid mode-printed slabs. These slabs were printed with a surface area of 2$\times$2 mm$^{2}$ and 500 $\mu$m thickness, and were left on the ITO coated glass substrates they were printed on. The substrate allows for easy sample handling and mounting in the optical test bench. With solid-mode printed structures, there is no liquid resin inside, but post-processing UV exposure was performed for slabs printed in IP-Visio, for which maximum polymerization conversion could not be obtained at intensity and scan settings below the bubble formation threshold. On all solid-mode samples, 20 and 40 minute heat treatment steps were performed to assess any homogenization effects on RI within the volume.
Shell mode slabs of the same dimensions were printed, and underwent post processing steps of 10 and 20 minute UV exposure and 20 and 40 minute heat treatment to fully polymerize unexposed liquid resin and reduce inhomogeneities in RI, unless otherwise noted.
3. Optical characterization
3.1 Measurement methods
Absorption and extinction measurements were made with a fiber-coupled white light source, integrating sphere, and fiber-coupled spectrometer. Light from the source fiber was relayed to the sample plane with a 75 mm focal length lens, with a spot size on the sample of about 1 mm. Light transmission through the sample was then collected through the 6 mm entrance aperture of the integrating sphere, which was connected by fiber to the spectrometer.
Direct transmittance (extinction) measurements - $I_{\rm ext}(\lambda )$ - were made by adjusting the separation between the sample and integrating sphere such that only the directly transmitted beam was collected, with a collection numerical aperture less than 0.04 (see illustration of the light collection path in Appendix B). For liquid and UV-cured resin characterization, the sample bulk was homogeneous, so there was minimal light scattering. For printed samples, light scattering needed to be considered. For these samples, a second measurement was made - $I_{\rm sca}(\lambda )$ - in which the integrating sphere was butted against the sample to collect directly transmitted plus forward scattered light. These measurements were normalized by reference measurements - $I_{\rm ref}(\lambda )$ - made by light transmittance through two air-spaced glass slides (for liquid and UV-cured resin samples) or through an ITO-glass substrate (for printed samples). All light intensity measurements were dark corrected.
Light transmittance through the sample is given by
where $\chi$ is a correction factor that takes into account the different reflectance coefficients between the reference and sample measurements. The calculation of extinction (absorption plus scattering) is via the attenuation coefficient where $L$ is the sample thickness. Finally, (forward) scattered light is calculated from The liquid and UV-cured resin samples were homogeneous, so that $I_{\rm ext} \approx I_{\rm sca}$ and $S \approx 0;$ the attenuation coefficient was the absorption coefficient for these samples. All reported values were based upon the average of measurements from three different locations on the same sample.Two different reflectance correction factors were used. For measurements of resin between glass slides, the reference measurement had four air-glass reflections ($R_{\rm a}\hbox{-}{\rm g}$), while the sample measurement had two air-glass reflections and two glass-resin reflections ($R_{\rm g}\hbox{-}{\rm r}$). Hence, this correction factor is given by
The RI for IP-S, IP-n162 and IP-Visio were calculated from the Cauchy fits from [23], where the fittings were based upon RI measurements of UV plus heat cured resins. It should be noted that these RI values may not accurately represent the resin RI values at the interfaces of our samples. We performed a sensitivity analysis, detailed in Appendix C, to quantify the uncertainty in our transmittance and extinction values introduced by uncertainty in RI. For the UV-cured liquid resin samples, we estimated an RI uncertainty bounded by 0.07 based upon comparison of RI values between partially and fully UV-cured IP-S [23], leading to an uncertainty in the absorption coefficient of up to 0.003 mm$^{-1}$, which is much less than the values reported in Fig. 1.
Fig. 1. Absorption coefficients for a) IP-S, b) IP-n162 and c) IP-Visio resins in liquid state and in polymerized state under UV curing for 10 and 20 minutes, a) and b), or UV curing for 30 and 60 minutes, c), and after subsequent heat treatment at about 60$^{\circ }$ for 20 and 40 minutes. The uncertainty in absorption coefficients in the cured states is <0.003 mm$^{-1}$.
Similarly, for all printed samples, the resin at the interfaces was polymerized by TPP, with an RI uncertainty bounded by 0.05 based upon comparison of RI values between TPP and Cauchy fit for IP-S [23]. Here, the uncertainty in extinction is 0.02 mm$^{-1}$ and in transmittance is <1%, which are less than the values reported in Figs. 2, 3, and 4.
Fig. 2. Attenuation coefficient, transmittance, and scattering spectra for a,d,g) IP-S, b,e,h) IP-n162, and c,f,i) IP-Visio resins after solid mode printing, and after subsequent heat treatment at about 60$^{\circ }$ for 20 and 40 minutes. IP-Visio is not fully cured after two-photon printing, leading to post-process UV curing of 30 and 60 minutes. Uncertainty in extinction is <0.02 mm$^{-1}$ and uncertainty in transmittance/scattering is <1%.
Fig. 3. Attenuation coefficient, transmittance and scattering spectra for a,c,e) IP-S and b,d,f) IP-n162 resins after shell mode printing, and after subsequent UV exposure for 10 and 20 minutes (30 and 60 minutes for IP-Visio) and heat treatment at about 60$^{\circ }$ for 20 and 40 minutes. Uncertainty in extinction is <0.02 mm$^{-1}$ and uncertainty in transmittance/scattering is <1%.
Fig. 4. a) Transmittance and b) scattering spectra for different shell-mode scaffolding parameters with IP-S resin. Homogeneous until cell volumes are 38% for T1, 45% for P1, and 78% for P3. Post-processing is with 10 minutes of UV exposure (10 UV) and 20 minutes of UV plus 40 minutes of heat treatment (40 Heat). Uncertainty in transmittance/scattering is <1%.
3.2 Results
3.2.1 Resins
The absorption of these resins has been measured previously in the liquid state and after UV curing and heat treatment [23]. Figure 1 shows our measured absorption coefficients for the IP-S, IP-n162, and IP-Visio resins. The initial measurement was with liquid resin, and subsequent measurements were made after 10 and 20 minutes of cumulative UV exposure (30 and 60 minutes for IP-Visio due to its lower UV absorption), and then after 20 and 40 minutes of cumulative heat treatment.
Absorption increases in the polymerized state in the visible range for IP-S and IP-n162, but decreases slightly for IP-Visio. All three resins have higher absorption for blue light ($\sim$450 nm) compared to green ($\sim$540 nm) and red ($\sim$620 nm) light, resulting from the photoinitiator, while absorption increases again in the 750 to 850 nm range due to the monomer. Comparing IP-S and IP-n162, the latter shows higher absorption for blue light. For green light, absorption in the polymerized state is also higher than for IP-S, whereas for red light, absorption is lower for IP-n162. With IP-Visio, importantly, absorption is lower in the polymerized state than for the other resins, which is in part responsible for its lower autofluorescence.
3.2.2 Solid mode printed samples
Absorption measurements on thin (2 to 4.5$\mu$m) two-photon printed samples of IP-S have been reported previously [21], but to our knowledge, absorption and scattering have not been studied for these resins at visible wavelengths after two-photon printing (but we note recent work that deliberately designed scattering micro-structures for two-photon printing [27]). Figure 2 shows the extinction, transmission and scattering spectra for IP-S, IP-n162, and IP-Visio solid-mode printed slabs. Note that the IP-Visio sample underwent additional UV curing, which demonstrates that IP-Visio did not reach saturated conversion under the two-photon writing parameters (see Appendix A for writing parameter settings).
From Fig. 2, it is interesting to notice that changes in extinction (and transmittance) due to post-processing heat treatment are not obvious for IP-S and IP-n162, but the effects of heat treatment are more evident in the scattering spectra for IP-S and IP-n162. IP-Visio experienced increased transmittance and the greatest drop in scattering under post-processing, mainly due to the fact that additional UV curing was required, after which scattering remains slightly higher than that of IP-S. Overall, heat treatment appears to provide more homogeneous refractive index within the volume, as evidenced by the reduced light scattering (all resins). Because the measured intensity of scattered light depends upon a combination of bulk inhomogeneity (including spatial structure, which can lead to coherence effects, and RI contrast), bulk absorption, and wavelength, it is difficult however to ascribe any one effect of heat treatment to the reduction in scattered light. For example, scattering typically scales inversely with wavelength [28] (rigorous scattering analysis is beyond the intent of this paper), but measured scattered light in the blue/near-UV was low because of high bulk absorption; at around 450 nm and below for IP-S and IP-n162 specifically, the high absorption caused the large drop in scattering, while this absorption edge for IP-Visio is closer to 410 nm.
Even though scattered light % of IP-n162 is highest, direct light transmittance is equal to or higher than IP-S throughout the visible for the same solid mode parameters, and comparable to IP-Visio, except in the blue, showing that solid-mode IP-n162 a viable option for micro-optics in the green-red if light scattering were not a concern. IP-Visio is clearly the best option for visible light applications if scattering were not a concern. If low scattering, or high transmittance to scattering ratio, is required, then heat-treated IP-S solid-mode would be the best option among the three.
3.2.3 Shell mode printed samples
Shell-mode printing used a triangular scaffold, with wall spacing $W_s=20\,\mu$m, floor spacing $F_s=25\,\mu$m, and wall/floor thickness $T=5\,\mu$m. Figure 3 shows the extinction, transmission and scattering spectra for shell-printed slabs. Now, the effects of post-processing are evident across all three measures, due to the changes in absorption (and RI) of the resins in going from liquid to polymerized states. It is not surprising that light scattering is strong in shell-mode printed devices prior to curing, and that scattering undergoes strong reduction with curing due to equalizing of RI between the scaffold and unprinted volume. However, it is surprising that light scattering is lower for cured shell-mode IP-n162 slabs than solid mode slabs (and lower than cured shell-mode IP-S and IP-Visio), results that cannot be explained by the differences in absorption. Part of the explanation lies in the fact that post-process curing of the liquid resin results in more homogeneous polymer volume than polymerization via printing. Another factor could be closer matching of 2PP versus UV-cured RI for IP-n162. Another interesting comparison is the very similar effects of post-processing on the scattering of IP-n162 and IP-Visio slabs, where only small changes in scattering are observed across the different post-processing steps, whereas scattering in the IP-S slab drops significantly after heat treatment, also leading to a jump in transmittance.
Another surprising result is that direct light transmittance of post-processed shell mode slabs is comparable to, or higher than, solid mode slabs. The higher transmittance over solid-mode printing is especially significant as shell-mode printing is roughly 4-5$\times$ faster. Comparing the resins, transmittance for IP-n162 remains higher than IP-S, and IP-Visio in the green-red, except for wavelengths below $\sim$470 nm, where it falls below IP-S, and for wavelength below 550 nm, where it falls below IP-Visio. This is expected based upon the results of Section 3.2.1. In contrast to the conclusions for solid-mode printing, with shell mode and subsequent post-process curing, IP-n162 is better both in terms of light transmittance and low scatter for wavelengths above $\sim$470 nm, while IP-S has slightly higher light transmittance for shorter wavelengths, and IP-Visio has obvious advantages in terms of transmittance and scattering in the blue.
3.2.4 Effects of shell-mode scaffold
Because of the many parameters involved in shell-mode printing, further optimizations (in terms of reducing bulk scattering) are possible. Here, based on transmittance and scattering, we compare the triangular scaffold of the previous section (‘T1’) to two planar scaffolds for IP-S. In principle, printed structures with the greatest homogeneous volume fraction (i.e. volume of resin that is cured in post-processing) within the scaffold unit cell should have the lowest scattering. Volume fraction calculations are detailed in Appendix D. We printed two additional slabs in shell mode using a planar scaffold. Sample ‘P1’ refers to planar scaffold with wall spacing $W_s=20\,\mu$m and floor spacing $F_s=25\,\mu$m, resulting in 84% UV volume per unit cell. Sample ‘P3’ had a lower scaffold density, with wall spacing 60 $\mu$m and floor spacing of 75 $\mu$m, and 94% UV volume. Figure 4 shows the transmittance and scattering data for this experiment. Transmittance was higher for P3 as it has a greater UV-cured volume (lower scaffold density) compared to P1. Likewise, scattering was lower (by more than a factor of two) for P3 because of its lower scaffold density, and is nearly comparable to scattering from IP-S solid mode printing. It is interesting to note that the scattering difference between P1 and P3 dropped upon full curing. This phenomenon can be explained by the refractive index mismatch between the 2PP and UV volumes reducing due to curing. It is also interesting that the transmittance of the fully-cured T1 and P1 samples are nearly the same (the main difference occurring around 450 nm), which is due to similar scaffold densities (83% for T1 and 84% for P1).
4. Implications for printed optics
We can quantify the maximum optical thickness (or path length) under the criterion of minimum transmittance of 50%. This thickness is plotted in Fig. 5 for resin and processing combinations that achieve maximum transmittance, and compared to the maximum thicknesses as limited by absorption in homogeneous UV-cured resin. For IP-S and IP-n162, printed thickness for blue light is limited to about 0.5 mm, whereas thicknesses greater than 1 mm can be realized with IP-Visio. For proper combinations of resin, print mode, and post-processing, thicknesses of greater than 2 mm can be achieved for green light and greater than 4 mm for red light. The cross-over wavelengths at which absorption and scattering equally contribute to the maximum thickness (i.e. equal absorption length and scattering mean free path) are about 420 nm for IP-S, 460 nm for IP-n162, and 400 nm for IP-Visio shell, whereas this cross-over doesn’t occur for IP-Visio solid; the shorter cross-over wavelengths for IP-Visio are due to its lower absorption at short wavelengths. For wavelengths longer than these cross-over points, scattering largely determines the maximum thickness. It is generally understood that the boundary between the single and multiple-scattering regimes is $\alpha L = 1,$ or $T < 0.37$ [2], so the use of Eqn. 2 for establishing the thickness bounds is reasonable.
Fig. 5. Maximum thickness for solid and shell mode printed optical parts for the three resins, compared to polymerized resin, which represents the absorption-limited thickness. Here, IP-S solid mode is with no post-processing, IP-S shell mode is cured with 20 minutes of UV and 40 minutes of heat, IP-n162 solid mode is with 40 minutes of heat, IP-n162 shell mode is with 20 minutes of UV and 20 minutes of heat, IP-Visio shell and solid modes are with 60 minutes of UV and 20 minutes of heat.
We test these predictions by printing $\sim$2 mm thickness compound optical devices for evaluation. These devices are printed versions of glass devices designed for multi-site delivery of visible light to tissue for all-optical neural manipulation and recording [29,30]. Advantages to 3D printing of these devices are the direct monolithic integration of microlenses for light coupling into the tissue-penetrating probes and better geometrical control over probe diameter and length and complex tip shape (a flat tip is used here to simplify interpretation). 6$\times$6 optical probe arrays of 85 $\mu$m diameter and 1.5 $\mu$m length were printed on a 400 $\mu$m pitch. Light coupling occurs via a coaxially-aligned 6$\times$6 spherical microlens array of 100% fill fraction, spaced by a 0.4 mm thick solid backplane. Lens radii of curvature are 282 $\mu$m and 292 $\mu$m for IP-S and IP-n162, respectively, with sag depths of 282 $\mu$m and 216 $\mu$m from lens center to corner (along the 282 $\mu$m diagonal). Even with the large spherical aberration of these designs, the circle of least confusion for parallel incoming rays is less than the probe diameter, so complete light coupling is expected. Images of the IP-S device are shown in Fig. 6.
Fig. 6. Images of the 6$\times$6 array of penetrating optical probes printed in IP-S solid mode. Here, tips are flat to facilitate optical testing, actual tips are shaped for light distribution control and tissue insertion. a) shows a top view of the device, while b) shows an edge of the underside, which is covered by a 6$\times$6 microlens array of 100% fill-fraction for free-space light coupling.
For this application, both high light transmittance and low light scatter are important. We measured devices printed in IP-S solid mode and IP-n162 shell mode (T1), as shown in Fig. 7, but we note that no further optimizations were made other than the post-processing. Two measurements were made using a weakly-focused beam from 50 $\mu$m core fiber-coupled LED sources, with the beam spot size of 150 $\mu$m nearly filling the maximum microlens inscribed circular cross-section. The first measurement was total light transmission, while the second measurement was transmittance just through the tip (where scattered light was blocked by placing an aperture at the tip); the difference is total scattered light.
Fig. 7. Transmittance and total forward scattered light measurements for compound micro-optical devices consisting of an array of 1.5 mm-long optical probes (i.e. cylindrical waveguides of 80 $\mu$m diameter). Light coupling is facilitated by an aligned array of microlenses, spaced by a 0.4 mm solid backplane. a-c) Device printed with IP-S solid mode, and subjected to post-processing of 20 minutes UV exposure and 40 minutes of heat treatment. d-f) Device printed with IP-n162 shell mode (T1), and subjected to 20 minutes of UV, and then 40 minutes of heat treatment. For the measurements after final post processing, transmittance predictions are shown as solid black lines, the lower bound representing the IP-S solid mode and IP-n162 shell mode predictions from slab measurements, and the upper bound set by the two-partition model.
As expected, neither device is suitable for blue light, and post processing increased light transmittance and decreased scattered light at all wavelengths, qualitatively following the predictions of Fig. 5. However, the predictions systematically underestimate the tip transmittance, which is especially inaccurate for blue light. There is a key difference between this device geometry and the slabs from which the limits were determined - the 1.5 mm long waveguides. Within the waveguides, light that is predominantly forward-scattered (either in the bulk or at the surface) is guided and contributes to the measured transmittance. Assessing the contribution of waveguided scattered light requires more sophisticated modeling of light transport based on the internal RI structure to determine the scattering coefficient and anisotropy factor at each wavelength. The RI structure of shell-mode printed waveguides is more complex than solid mode, as the shell-mode waveguide consists of a two-photon printed outer shell (of 12.5 $\mu$m thickness) and internal scaffolding, and UV-cured internal scaffold volumes. Here, we employ a highly-simplified two-partition model to determine if waveguided scatter can explain the measured results. This model treats the device as the combination of a 0.5 mm-thick slab (i.e. the 0.4 mm backplane and $\sim$0.1 mm average lens thickness) and a 1.5 mm waveguide. In the slab, predicted extinction is based upon the slab measurements. In the waveguide, we make the assumption that all scattered light is captured, which clearly serves as an upper bound, such that attenuation is due only to cured resin absorption. Nevertheless, the measurements of tip transmittance more reasonably correspond to the upper bounds set by the simplified two-partition model than to the slab-only predictions (as indicated by the black lines in Fig. 7). This model further suggests that forward scattered light is the dominant source of tip transmittance in the blue, and that the fraction of light scattered in the forward direction generally decreases with increasing wavelength.
Figure 7 shows that the IP-n162 shell mode device is promising for green, and especially red, light applications, with 50% or higher tip transmittance and total external scatter less than half the tip transmittance. Straightforward optimizations in shell mode scaffold parameters and/or use of other resins should improve performance in terms of increased tip transmittance and reduced scattering. IP-Visio is especially promising for use with blue light whereas the high index of IP-n162 provides an advantage over other resins in terms of decreased radius of curvature for the same lens power, which will benefit optimized free-form lens designs. More sophisticated analysis would need to be performed to model light transport and guide further design and process optimizations, such as ray tracing with suitable scattering models or Monte Carlo analysis using the radiation transport equation [31]. In summary, it is clear that the ultimate performance of a complex printed device needs to be evaluated consistent with the actual application requirements.
5. Conclusions
Somewhat unexpectedly, there is no one resin/print mode/post-processing recipe that is optimal for micro-optics applications across the entire visible range. The lowest overall bulk scattering occurs with IP-S solid mode with heat treatment (but optimized shell-mode printing can also achieve low scattering), while the overall highest transmittance occurs with IP-S from low scaffold density shell mode printing (lower scaffold density should also improve transmittance for the other resins). For applications with blue light, IP-Visio shell mode shows the highest transmittance, also making IP-Visio the best choice for broadband applications. It is clear that shell-mode printing enables greater thickness (in terms of the 50% transmittance constraint), which is significant since shell-mode facilitates more time-efficient printing of large volumes. It is also clear that additional optimizations of the scaffold geometry may further improve transmittance (and reduce scattering), within the limits of maintaining structural integrity prior to post-process UV exposure.
From [23], OrmoComp resin has significantly lower absorption throughout the visible than the resins measured here, but that reduced absorption would directly lead to increased writing times; that tradeoff needs to be weighed on an application basis. Finally, it is important to note that our measurements with printed slabs only take into account surface scattering at one interface (resin-air, assuming negligible scatter at the substrate-resin interface). Scattering due to surface roughness could pose additional limitations for micro-optics when significant interaction between light and printed sidewalls occurs (e.g. in waveguiding or light-pipe structures [32]), as in the optical probe array of Fig. 6, where additional characterization would be needed to assess surface scattering. Waveguiding of forward scattered light also needs to be considered in these designs.
Funding
National Science Foundation (1828480); National Eye Institute (1R21EY030717-01).
Acknowledgments
The authors thank Brian Baker at the Utah Nanofab for assistance with this work and Dr. Benjamin Richter of Nanoscribe GmbH for helpful discussions.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. A. Zolfaghari, T. Chen, and A. Y. Yi, “Additive manufacturing of precision optics at micro and nanoscale,” Int. J. Extreme Manuf. 1(1), 012005 (2019). [CrossRef]
2. S. M. Hamp, R. D. Logan, and J. A. Shaw, “Optical transmittance of 3D printing materials,” Appl. Opt. 60(22), 6573–6578 (2021). [CrossRef]
3. J. Gawedzinski, M. E. Pawlowski, and T. S. Tkaczyk, “Quantitative evaluation of performance of three-dimensional printed lenses,” Opt. Eng. 56(08), 1–13 (2017). [CrossRef]
4. 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]
5. P. Diep, S. Pannem, J. Sweer, J. Lo, M. Snyder, G. Stueber, Y. Zhao, S. Tabassum, R. Istfan, J. Wu, S. Erramilli, and D. Roblyer, “Three-dimensional printed optical phantoms with customized absorption and scattering properties,” Biomed. Opt. Express 6(11), 4212–4220 (2015). [CrossRef]
6. Y. Liu, P. Ghassemi, A. Depkon, M. I. Iacono, J. Lin, G. Mendoza, J. Wang, Q. Tang, Y. Chen, and T. J. Pfefer, “Biomimetic 3d-printed neurovascular phantoms for near-infrared fluorescence imaging,” Biomed. Opt. Express 9(6), 2810–2824 (2018). [CrossRef]
7. A. J. Ruiz, S. Garg, S. S. Streeter, M. K. Giallorenzi, E. P. M. LaRochelle, K. S. Samkoe, and B. W. Pogue, “3D printing fluorescent material with tunable optical properties,” Sci. Rep. 11(1), 17135 (2021). [CrossRef]
8. S. Maruo, O. Nakamura, and S. Kawata, “Three-dimensional microfabrication with two-photon-absorbed photopolymerization,” Opt. Lett. 22(2), 132–134 (1997). [CrossRef]
9. S. Maruo and J. T. Fourkas, “Recent progress in multiphoton microfabrication,” Laser & Photonics Rev. 2(1-2), 100–111 (2008). [CrossRef]
10. X. Wang, H. Yu, P. Li, Y. Zhang, Y. Wen, Y. Qiu, Z. Liu, Y. Li, and L. Liu, “Femtosecond laser-based processing methods and their applications in optical device manufacturing: A review,” Opt. Laser Technol. 135, 106687 (2021). [CrossRef]
11. H.-B. Sun, M. Maeda, K. Takada, J. W. M. Chon, M. Gu, and S. Kawata, “Experimental investigation of single voxels for laser nanofabrication via two-photon photopolymerization,” Appl. Phys. Lett. 83(5), 819–821 (2003). [CrossRef]
12. T. Gissibl, S. Thiele, A. Herkommer, and H. Giessen, “Two-photon direct laser writing of ultracompact multi-lens objectives,” Nat. Photonics 10(8), 554–560 (2016). [CrossRef]
13. 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]
14. M. A. Tadayon, S. Chaitanya, K. M. Martyniuk, J. C. McGowan, S. P. Roberts, C. A. Denny, and M. Lipson, “3D microphotonic probe for high resolution deep tissue imaging,” Opt. Express 27(16), 22352–22362 (2019). [CrossRef]
15. S. Ristok, S. Thiele, A. Toulouse, A. M. Herkommer, and H. Giessen, “Stitching-free 3D printing of millimeter-sized highly transparent spherical and aspherical optical components,” Opt. Mater. Express 10(10), 2370–2378 (2020). [CrossRef]
16. K. Weber, Z. Wang, S. Thiele, A. Herkommer, and H. Giessen, “Distortion-free multi-element hypergon wide-angle micro-objective obtained by femtosecond 3d printing,” Opt. Lett. 45(10), 2784–2787 (2020). [CrossRef]
17. 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]
18. B. H. Cumpston, S. P. Ananthavel, S. Barlow, D. L. Dyer, J. E. Ehrlich, L. L. Erskine, A. A. Heikal, S. M. Kuebler, I. Y. S. Lee, D. McCord-Maughon, J. Qin, H. Röckel, M. Rumi, X.-L. Wu, S. R. Marder, and J. W. Perry, “Two-photon polymerization initiators for three-dimensional optical data storage and microfabrication,” Nature 398(6722), 51–54 (1999). [CrossRef]
19. S. C. Ligon, R. Liska, J. Stampfl, M. Gurr, and R. Mülhaupt, “Polymers for 3D printing and customized additive manufacturing,” Chem. Rev. 117(15), 10212–10290 (2017). [CrossRef]
20. T. Gissibl, S. Wagner, J. Sykora, M. Schmid, and H. Giessen, “Refractive index measurements of photo-resists for three-dimensional direct laser writing,” Opt. Mater. Express 7(7), 2293–2298 (2017). [CrossRef]
21. Y. Li, S. Park, M. McLamb, M. Lata, S. Schöche, D. Childers, I. D. Aggarwal, M. K. Poutous, G. Boreman, and T. Hofmann, “UV to NIR optical properties of IP-Dip, IP-L, and IP-S after two-photon polymerization determined by spectroscopic ellipsometry,” Opt. Mater. Express 9(11), 4318–4328 (2019). [CrossRef]
22. S. Dottermusch, D. Busko, M. Langenhorst, U. W. Paetzold, and B. S. Richards, “Exposure-dependent refractive index of Nanoscribe IP-Dip photoresist layers,” Opt. Lett. 44(1), 29–32 (2019). [CrossRef]
23. M. Schmid, D. Ludescher, and H. Giessen, “Optical properties of photoresists for femtosecond 3D printing: refractive index, extinction, luminescence-dose dependence, aging, heat treatment and comparison between 1-photon and 2-photon exposure,” Opt. Mater. Express 9(12), 4564–4577 (2019). [CrossRef]
24. L. J. Jiang, Y. S. Zhou, W. Xiong, Y. Gao, X. Huang, L. Jiang, T. Baldacchini, J.-F. Silvain, and Y. F. Lu, “Two-photon polymerization: investigation of chemical and mechanical properties of resins using Raman microspectroscopy,” Opt. Lett. 39(10), 3034–3037 (2014). [CrossRef]
25. https://refractiveindex.info/?shelf=glass&book=soda-lime&page=Rubinclear, “Optical constants of soda lime glass,”.
26. https://refractiveindex.info/?shelf=other&book=In2O3SnO2&page=Moerland, “Optical constants of In2O3-SnO2 (Indium tin oxide, ITO),”.
27. E. Marakis, R. Uppu, M. L. Meretska, K.-J. Gorter, W. L. Vos, and P. W. H. Pinkse, “Deterministic and controllable photonic scattering media via direct laser writing,” Adv. Opt. Mater. 8(24), 2001438 (2020). [CrossRef]
28. C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by Small Particles (Wiley-Interscience, 1998).
29. T. V. Abaya, M. Diwekar, S. Blair, P. Tathireddy, L. Rieth, and F. Solzbacher, “Deep-tissue light delivery via optrode arrays,” J. Biomed. Opt. 19(1), 015006 (2014). [CrossRef]
30. N. McAlinden, Y. Cheng, R. Scharf, E. Xie, E. Gu, C. F. Reiche, R. Sharma, P. Tathireddy, L. Rieth, S. Blair, and K. Mathieson, “Multisite microLED optrode array for neural interfacing,” Neurophotonics 6(03), 1 (2019). [CrossRef]
31. J. Jönsson and E. Berrocal, “Multi-scattering software: part i: online accelerated Monte Carlo simulation of light transport through scattering media,” Opt. Express 28(25), 37612–37638 (2020). [CrossRef]
32. K. Willis, E. Brockmeyer, S. Hudson, and I. Poupyrev, Printed Optics: 3D Printing of Embedded Optical Elements for Interactive Devices (Association for Computing Machinery, New York, NY, USA, 2012), pp. 589–598.
