Triplet Upconversion under Ambient Conditions Enables Digital Light Processing 3D Printing

The rapid photochemical conversion of materials from liquid to solid (i.e., curing) has enabled the fabrication of modern plastics used in microelectronics, dentistry, and medicine. However, industrialized photocurables remain restricted to unimolecular bond homolysis reactions (Type I photoinitiations) that are driven by high-energy UV light. This narrow mechanistic scope both challenges the production of high-resolution objects and restricts the materials that can be produced using emergent manufacturing technologies (e.g., 3D printing). Herein we develop a photosystem based on triplet–triplet annihilation upconversion (TTA-UC) that efficiently drives a Type I photocuring process using green light at low power density (<10 mW/cm2) and in the presence of ambient oxygen. This system also exhibits a superlinear dependence of its cure depth on the light exposure intensity, which enhances spatial resolution. This enables for the first-time integration of TTA-UC in an inexpensive, rapid, and high-resolution manufacturing process, digital light processing (DLP) 3D printing. Moreover, relative to traditional Type I and Type II (photoredox) strategies, the present TTA-UC photoinitiation method results in improved cure depth confinement and resin shelf stability. This report provides a user-friendly avenue to utilize TTA-UC in ambient photochemical processes and paves the way toward fabrication of next-generation plastics with improved geometric precision and functionality.


Instrumentation
UV-Visible Absorption Spectrometer.Steady-state optical characterization was accomplished by collecting UV-visible absorption spectra on an Ocean Optics (QE PRO-ABS) Fiber Optic Spectrometer utilizing deuterium-tungsten halogen light sources (DH-2000-BAL).600 μm fiber-optic cables (QP600-025-SR) were coupled to the detector with a slit width of 10 μm.Dilute absorption data was collected using quartz cuvettes (1 cm path length) and an Ocean Optics sample holder (qpod2e).
Emission Spectrometer.Fluorescence and phosphorescence spectra were recorded on a Horiba Fluorolog-3 spectrofluorometer.The optical upconversion measurements were performed with a custom-built setup that uses a Coherent Verdi V-10 532 nm laser as an excitation source and Ocean Optics Flame spectrometer as a detector (Figure S1).The emission was collimated and focused with two plano-convex lenses to a 600 μm fiber-optic cable (QP600-025-SR) connected to the spectrometer.The intensity of the laser excitation was modulated with a Newport variable neutral density filter and by changing the laser power.The excitation power was measured with a Thorlabs SV-120VC power sensor and PM100D console and the beam diameter was 2.25 mm according to the manufacturer's specification.Fluorescence lifetimes were recorded using a time-correlated single photon counting module of the Fluorolog-3 spectrofluorometer with a Horiba NanoLED pulsed diode laser (375 nm, 1 MHz pulse frequency).

Figure S1.
Upconversion setup used to measure the upconversion quantum yield and intensity threshold in 2-phenoxyethyl acetate.The laser beam is reflected with mirrors M1 and M2 through the variable neutral density (ND) filter to the sample cuvette.The emission is collected at a 45-degree angle with lenses L1 and L2 through an optical fiber to the spectrometer.
Transient Absorption Spectrometer.Triplet lifetimes were measured using a Magnitude Instruments enVISion transient absorption spectrometer using a 532 nm pulsed excitation laser (800-1000 Hz, 0.23-0.4µJ/cm 2 pulse fluences, measured with Thorlabs SV-120VC power sensor and PM100D console).The spectrometer can be used to measure both transient absorption and time-resolved photoluminescence.
Real-Time Fourier Transform Infrared (RT-FTIR) Spectrometer.RT-FTIR was recorded utilizing an INVENIO-R FT-IR Spectrometer from Bruker (Figure S2) and controlled via OPUS Spectroscopy Software.A liquid nitrogen cooled (LN-MCT Mid) detector was used for measurements.Surface FTIR characterization of solids and liquids under an argon atmosphere during illumination is accomplished using a GladiATR Illuminate accessory (SKU 026-1800) from PIKE Technologies (Figure S2) that couples to the Bruker INVENIO through Quick-Lock recognition.Samples are irradiated through the bottom of a diamond crystal using an LED transmitted through a 3 mm liquid light guide (LLG3-4Z) from ThorLabs.Illumination occurs from under a diamond attenuated total reflectance (ATR) element, which sits in a heated plate with a maximum temperature of 210 °C.The angle of incidence of the IR probe beam is 45°, the diamond crystal surface is 3 mm, and the spectral range is 4000 to 30 cm -1 .The sampling depth was calculated to be 2.6 µm using an on-line calculator from PIKE Technologies (https://www.piketech.com/pikecalc/).FTIR characterization of thin films (100 µm) was performed using a transmission accessory (A043-N/Q) from Bruker (Figure S3).Samples sit horizontally and are irradiated using an LED transmitted through a 5 mm liquid light guide (LLG5-4Z) from ThorLabs.S4).Calibration was done using Canon S60 (low-viscosity) and S600 (high-viscosity) oils as standards and data collected and analyzed using the TRIOS software.The minimum torque is 1 nN.m for oscillation and 3nN.m for steady sheer, both have a maximum of 200 mN.m with 0.1 nN.m resolution.The frequency range is 1.0×10 -7 -100 Hz.The angular velocity range is 0 -300 rad/s.The maximum normal force is 50 N with a sensitivity of 0.005 N and resolution of 0.5 mN.

Figure S4.
Rheometer with light guide accessory used for gel point and cure depth experiments.DLP 3D Printing.3D printing was performed using a custom-made, DLP-based 3D printer (Monoprinter, MA, USA) (Figure S5).Detailed information about the 3D printer can be found in our previous report.S1 The Printer was equipped with an exchangeable visible LED centered at 525 nm (green, Luminus PT-120-G) with a full width at half maximum (FWHM) of 34 nm (Figure S51).The projector resolution was 1920 × 1080 pixels, with each pixel being 20 μm × 20 μm at the image plane.The minimum achievable layer thickness is 25 μm.Here, prints were primarily performed using a layer thickness of 100 μm.The DLP employed here follows a layer-by-layer additive manufacturing process, where each 100 μm layer is cured between an upper cured layer or build platform and a transparent fluorinated polymer film (Teflon FEP film, DuPont, 127 μm thick) as the base of the resin tank that provides a non-stick and somewhat flexible surface for detachment.

Light Emitting Diodes (LEDs)
All LEDs used were purchased from Mightex Systems.The product numbers for the three LEDs are LCS-0405-12-22 (405 nm), LCS-0530-15-22 (530 nm), and LCS-0525-60-22 (525 nm) and powered by an LED Controller (SLC-MA02-U).Green LEDs were equipped with a 525×25 nm bandpass filter from Edmund Optics (#87-789).Lightguide Adapters from ThorLabs were used to irradiate resins via either a 3 mm liquid light guide (LLG-3-4Z) when using the ATR accessory or a 5 mm liquid light guide (LLG5-4Z) when using the transmission accessory.The emission profile for each LED was measured using a calibrated UV-Vis Ocean Insight system.Irradiation intensities were measured with a Thorlabs PM100D photometer equipped with silicon-based photodiode power sensor (S130C, Thorlabs) prior to each experiment.

Absorption & Emission
The absorption spectra of PtOEP (Figure S6), DPA (Figure S7), BAPO, and Ivocerin ® (Figure S8) were measured using a 1 cm cuvette.Samples used to record emission spectra of PtOEP (Figure S6) and DPA (Figure S7) were prepared in a 1 cm cuvette at a concentration that gave a maximum absorbance of 0.1 in the excitation and emission range to exclude inner filter effects.The PtOEP sample was bubbled with nitrogen for 20 minutes prior to measuring.Emission spectra of the 405 nm and 525 nm LEDs measured with the Ocean Optics (QE PRO-ABS) Fiber Optic Spectrometer are shown in Figure S9.

RT-FTIR Monitoring of Photopolymerization Kinetics
Photopolymerization kinetics was monitored via the loss of the monomer C=C stretching band (802-818 cm -1 ) for ATR and (6105-6214 cm -1 ) for transmission mode FTIR.Maximum monomer conversion (ρ max ) is the total conversion reached after 110 seconds of irradiation found as an average of the final 2 seconds; rate of polymerization (r p ) is a linear fit of points between 4 seconds and the time required to reach a monomer conversion of

5
; inhibition time (t inh ) is the time required to pass 2% monomer conversion after the LED is turned on.

Upconversion Quantum Yield
Optical upconversion measurements (Figure S13) were performed by collecting the emission from the front face of the 1 mm path length cuvette at approximately a 45-degree angle due to the high optical densities of the resins.The upconversion samples were thoroughly purged of oxygen by bubbling with nitrogen for at least 30 minutes and adding 30 mM of oleic acid as an oxygen scavenger S2 to improve the stability of the sample over the long measurement times.The upconversion quantum yield ( UC ) in 2-phenoxyethyl acetate was determined using Rhodamine 6G in ethanol (fluorescence quantum yield,  ref , of 0.95) S3 as a reference and using Equation S1: S4

Transient Absorption Spectroscopy
The triplet lifetimes of DPA and PtOEP were measured via transient absorption spectroscopy upon irradiation with a 532 nm pulsed laser.PtOEP phosphorescence decay was recorded at the peak wavelength of 645 nm.DPA triplet decay was recorded by monitoring the transient absorption of PtOEP sensitized DPA at 475 nm.S5 The resulting transient photoluminescence decays were fitted with a single exponential function in Mathworks MATLAB R2022B software to obtain the triplet lifetimes of PtOEP.The triplet decay traces of DPA were measured under two low fluences of 0.23 µJ/cm 2 and 0.31 µJ/cm 2 .The resulting decays were analyzed to yield both the spontaneous triplet decay rate (k An T ) and the fraction () of triplets that undergo second-order decay, in this case via TTA.The analysis is based on solving the following differential rate equation that governs the triplet decay: S6,S7 This equation has the following analytic solution: where and equates to the fraction of annihilator triplets undergoing TTA at time zero.
Thus, fitting equation S3 to the annihilator triplet decays yields both the spontaneous triplet decay rate and the fraction of annihilating triplets at time zero, which can be used to qualitatively compare the rate of TTA between the acetate and acrylate systems.The fitting was performed on OriginLab Origin 2024 by globally sharing k An T between the decays measured in each solvent at different fluences while allowing  to float.These normalized decays and the respective fits are shown in Figures S18 (acrylate) and S19 (acetate).As expected,  increases in both solvents as fluence increases.More notably,  is significantly higher in acrylate, indicating that the rate of TTA is considerably faster.Figure S20.Quantum yield determination of the upconversion system in proxy solvent (2-phenoxyethyl acetate) using the upconverted emission measurement setup (Figure S1).

Fluorescence Lifetime
The fluorescence decays of DPA directly excited by a 375 nm laser were recorded using time-correlated single photon counting (Figure S21).The fluorescence decays were fitted with a single exponential function deconvoluted with the instrument response function in the Horiba DAS6 software and the fit residuals were inspected to evaluate the goodness of the fit.We find that the emission lifetime of DPA in 2-phenoxyethyl acrylate and 2-phenoxyethyl acetate is nearly identical (6.82 ns) indicating that DPA's S1 state is not quenched by 2-phenoxyethyl acrylate.For this reason, we hypothesize that charge transfer from DPA's triplet pair state prepared by TTA is responsible for photopolymerization in the absence of BAPO (Figure S11).

ATR-FTIR Characterization (Inert)
Photopolymerization optimization-1.The effect of photoinitiator (BAPO) concentration and light intensity on the polymerization kinetics were examined (Figures S22-S23).Photopolymerization controls.The effect of removing one or more of the photosystem components on the polymerization kinetics was examined (Figures S24-S25).Alternate photoinitiator.The impact on polymerization kinetics due to replacing BAPO with Ivocerin ® as an alternate photoinitiator was examined (Figure S26).See Figure S8 for the UV-vis absorption spectrum of Ivocerin ® .

UV/Vis Photodegradation (Ambient)
Optimized resins for both TTA-UC and Type II systems were prepared in a 50/50 wt% mixture of 2phenoxyethyl acrylate and TMPTA.The resin was loaded between glass slides with 100 μm spacers blanking with the same resin without the PS (PtOEP).Samples were then irradiated with a 525 nm LED (10 mW/cm 2 ) collecting spectra at 1.3 second intervals.A time series of normalized absorbance spectra are shown below.

Transmission FTIR Characterization (Ambient)
Photopolymerization optimization-3 for 3D printing.The effect of BAPO concentration and light intensity on the polymerization kinetics in the presence of the TMPTA crosslinker under ambient conditions were examined via transmission RT-FTIR (Figures S44-S45).Threshold intensity was determined under ambient conditions at several intensities between 2 mW/cm 2 and 250 mW/cm 2 .The complete data is provided in Figures 6A and 6B in the main manuscript and Table S8, with selected intensities plotted in Figure S46.

Rheology Studies
First, the gel point was found using an oscillation test (20 mm stainless steel upper plate, 100 µm gap, 1.0 strain%, 10 rad/s angular frequency, torque 10.0 µN•m) irradiating with a 525 nm LED (10 mW/cm 2 ) starting at 10 s and finding the modulus crossover (Figure S47).In a standard cure depth experiment the geometry gap was set to 500 µm.In each run the resin was loaded onto the lower acrylic plate of the photorheology accessory and the upper plate was brought down to a gap height of 500 µm.The LED was triggered to shine through the acrylic plate and into the sample for incrementally shorter times.After each irradiation period, the upper plate was raised and excess monomer gently wicked off of the top of the sample with a Kim-wipe.Then the upper plate was lowered at 1 µm/s while measuring the axial force (N).The cure depth was taken as the gap height at a force of 2 N, which was found to be the force giving the most consistent measure of depth when curing through the 500 µm gap.S8 The light intensities used were 0.1 mW/cm 2 , 0.5 mW/cm 2 , and 10 mW/cm 2 for the Type I, Type II, and TTA-UC resins (Figure S49), respectively, which were selected because they gave comparable rates of polymerization as determined through previous ATR-FTIR experiments.   .Data collected via FTIR-ATR was used to determine light intensities for the three respective photosystems that provided similar rates of polymerization ~220 mM/s (TTA-UC, 10 mW/cm 2 ; Type II, 0.5 mW/cm 2 ; Type I, 0.1 mW/cm 2 ).These light intensities were used for subsequent photorheology cure depth experiments.S51).The resulting ratio between the absorbed photons (in a 100 µm thick sample slice) that can initiate polymerization and the total photon flux emitted by the LED was used to calculate the total transmittance of light in the sample.We found a good match between the transmittance when using either BAPO (78%) or PtOEP (79%) and thus we expect the light penetration to be comparable in all three systems.The same overlap was used to quantify ε avg (the average molar absorptivity), following that described previously (Figure S52).Based on these results, the predicted D p values based on equation 4 in the main manuscript for Type I was 920 µm and for the Type II and TTA-UC resins was 892 µm.

3D Printing
Time-based 3D printing array for exposure time optimization.The below resolution print was done at a layer thickness of 100 μm using a custom 1-channel DLP 3D printer with a 525 nm LED (Figure S53) to qualitatively identify the optimal exposure time.Each resolution print contains a set of squares (5 -115 sec in 10 sec increments) varying exposure time/layer (Figures S54a and S54b).

Figure S2 .
Figure S2.ATR-FTIR accessory allowing for LED irradiation at sample interface.Glass cover used for inert sample environment.Type-B LED (530 nm) used for intensities <50 mW/cm 2 , Type-H LED (525 nm) used for intensities >50 mW/cm 2 .A 525×25 nm bandpass filter used in both configurations.

Figure S3 .
Figure S3.Bruker Transmission FTIR accessory allowing for top-down irradiation of a sample.Rheometer.The Discovery Hybrid Rheometer (DHR) 20 from TA Instruments (serial # 5343-0845) equipped with a HR x0 Upper Peltier Plate System (part # 534050.901)and 25 mm stainless steel geometry

Figure S7 .
Figure S7.Absorption (solid yellow) and emission (dashed blue) spectra of DPA in 2-phenoxyethyl acrylate.The absorbance maximum occurs at 376 nm and phosphorescence maxima at 413 and 433 nm.
S1)   Here,  ref and  UC are the absorbances of the reference and upconversion sample at 532 nm,  ref and  UC are the integrated emission spectra of the reference (500 to 720 nm) and upconversion fluorescence (380 to 530 nm), and  ref and  UC are the refractive indices of the solvents (1.51 for 2-phenoxyethyl acetate and 1.36 for ethanol), respectively.The Rhodamine 6G sample was prepared in a 1 cm pathlength cuvette at a concentration that gave a maximum absorbance of 0.1 to minimize the inner filter effect.

Figure S11 .
Figure S11.Image of the upconversion mixture without BAPO, showing polymer formation occurred after high-intensity (1000 mW/cm 2 ) 532 nm laser excitation over the course of 2 minutes.

Figure S13 .
Figure S13.Emission spectra of the upconversion system without BAPO in the acrylate monomer at time points between 1 and 120 s with an excitation wavelength of 532 nm at an intensity of 1000 mW/cm 2 .

Figure S16 .
Figure S16.PtOEP phosphorescence decay in 2-phenoxyethyl acrylate with DPA monitored at 645 nm fitted with a single exponential function that yields a triplet lifetime of 1.88 µs.Pulse fluence was 0.40 µJ/cm 2 .

Figure S17 .
Figure S17.PtOEP phosphorescence decay in 2-phenoxyethyl acetate with DPA monitored at 645 nm fitted with a single exponential function that yields a triplet lifetime of 1.85 µs.Pulse fluence was 0.40 µJ/cm 2 .

Figure S21 .
Figure S21.Fluorescence lifetime measurements of DPA in both acrylate and acetate after 375 nm excitation and corresponding fits to the data.IRF = instrument response function.
Photopolymerization optimization-2.The influence of overall photosystem concentration and light intensity on polymerization kinetics was examined (FiguresS27-S30).

Figure S33 .
Figure S33.Upconversion resin comprised of PtOEP (0.01 mol%), DPA (0.10 mol%), and BAPO (0.50 mol%) in 2-phenoxyethyl acrylate.Sample irradiated with a 525 nm LED and monitored via FTIR-ATR under an argon atmosphere.The initial rate of polymerization was plotted against intensity within the sublinear regime between 50 and 200 mW/cm 2 .Table S6.Summary of TTA-UC to Type I photopolymerization kinetics as a function of light intensity measured using RT-FTIR and provided in Fig. 5A and 5B and Fig. S31-S33.The [PtOEP] was 600 µM and the ratio of PtOEP:DPA:BAPO was 1:10:50.Values are averages from triplicate measurements (or more) with ±1 standard deviation from the mean.

Figure S34 .
Figure S34.Representative traces of Type I resin comprising BAPO (0.50 mol%) in 2-phenoxyethyl acrylate and irradiated with a 525 nm LED between 0.001 and 1 mW/cm² while monitoring via FTIR-ATR under an argon atmosphere.

Figure S35 .
Figure S35.Plot of initial rate vs. excitation intensity for the Type I resin from 0.001-1 mW/cm².The trendline represents a power law fit to the data and shows a sublinear relationship between r p and I ex .

Figure S37 .
Figure S37.Plot of initial rate vs. excitation intensity for the Type I resin from 0.01-2 mW/cm².The trendline represents a power law fit to the data and shows a sublinear relationship between r p and I ex .

Figure S39 .
Figure S39.Plot of initial rate vs. excitation intensity for the Type II resin from 2-20 mW/cm².The trendline represents a power law fit to the data and shows a strong sublinear relationship between r p and I ex when operating at high light intensities.This is hypothesized to arise from competitive degradation that occurs at high I ex values.
Shelf stability.The stability of the TTA-UC to Type I resin (FigureS40) and Type II resin (Figure41) were assessed by monitoring photopolymerization kinetics using RT-FTIR-ATR under an argon atmosphere over the course of 45 days.Periodic aliquots were removed from samples stored at room temperature in the absence of light.Data in this section was used to create Figure5Cin the main text.

Figure S40 .
Figure S40.Plot of monomer conversion vs. time for the TTA-UC resin.The samples were degassed before each trial and then irradiated with a 525 nm LED at 20 mW/cm².

Figure S41 .
Figure S41.Plot of monomer conversion vs. time for the Type II resin.The samples were degassed before each trial and then irradiated with a 525 nm LED at 20 mW/ cm².

Figure S42 .
Figure S42.PtOEP absorbance in the TTA-UC resin during 120 s of irradiation with the 525 nm LED.The peak corresponding only to PtOEP absorbance at 535 nm remained constant.

Figure S43 .
Figure S43.PtOEP absorbance in the Type II resin during 120 s of irradiation with the 525 nm LED.

Figure S47 .
Figure S47.Representative photorheology experiment of the TTA-UC resin to determine the gel point as defined by the crossover in storage and loss moduli.The 525 nm LED (10 mW/cm 2 ) was turned on at 10 s giving an effective gel point of 39.6 ± 2.4 seconds at a 100 µm gap.

Figure S48 .
Figure S48.Rheology experiments plotting cure depth (C d ) against the natural log of incident light exposure energy (E 0 ) used to determine depth of penetration (D p ) as the slope of the linear fit.

Figure S49
Figure S49.Data collected via FTIR-ATR was used to determine light intensities for the three respective photosystems that provided similar rates of polymerization ~220 mM/s (TTA-UC, 10 mW/cm 2 ; Type II, 0.5 mW/cm 2 ; Type I, 0.1 mW/cm 2 ).These light intensities were used for subsequent photorheology cure depth experiments.

E 0 =
incident exposure intensity; C d = cure depth Optical Density Matching.The optical densities of the resins over the LED emission range were estimated by determining the overlap between the photoinitiator (BAPO or PtOEP) absorption spectra and the LED emission spectra (Figures S50 and

Figure S50 .
Figure S50.Absorption spectrum of PtOEP (600 µM) in 2-phenoxyethyl acrylate at a thickness of 100 µM overlaid with the emission spectrum of the 525 nm LED with a 525×25 nm bandpass filter.ε avg value provided as an inset.

Figure S51 .
Figure S51.Absorption spectrum of BAPO (30 mM) in 2-phenoxyethyl acrylate at a thickness of 100 µM overlaid with the emission spectrum of the 405 nm LED.

Figure S52 .
Figure S52.Average extinction spectrum based on the overlap of either BAPO or PtOEP with the 405 nm and 525 nm LED, respectively.ε avg value calculated as the integral of the trace.

Figure S53 .
Figure S53.Emission profile of the Green LED (525 nm) in the DLP 3D printer

Table S3 . Summary of photopolymerization kinetics of optimized resin as measured using RT-FTIR and provided in Fig. 4B. Resin composition: [PtOEP] = 0.6 mM; [DPA] = 6 mM; [BAPO] = 30 mM. Values are averages from triplicate measurements (or more) with ±1 standard deviation from the mean.
I ex = excitation intensity; r p = rate of polymerization post-induction period; ρ max = maximum monomer conversion reached after 110 seconds of irradiation; t inh = inhibition time after turning the LED 'on' during which no monomer conversion occurs.