Ambient solid-state triplet–triplet annihilation upconversion in ureasil organic–inorganic hybrid hosts

Triplet–triplet-annihilation upconversion (TTA-UC) has attracted significant attention as an approach to harvest low energy solar photons that cannot be captured by conventional photovoltaic devices. However, device integration requires the design of solid-state TTA-UC materials that combine high upconversion efficiency with long term stability. Herein, we report an efficient solid-state TTA-UC system based on organic–inorganic hybrid polymers known as ureasils as hosts for the archetypal sensitiser/emitter pair of palladium(ii) octaethylporphyrin and diphenylanthracene. The role of the ureasil structure on the TTA-UC performance was probed by varying the branching and molecular weight of the organic precursor to tune the structural, mechanical, and thermal properties. Solid-state green-to-blue UC quantum yields of up to 1.86% were observed under ambient conditions. Notably, depending on the ureasil structure, UC emission could be retained for >70 days without any special treatment, including deoxygenation. Detailed analysis of the structure-function trends revealed that while a low glass transition temperature is required to promote TTA-UC molecular collisions, a higher inorganic content is the primary factor that determines the UC efficiency and stability, due to the inherent oxygen barrier provided by the silica nanodomains.


Fabrication of organic-inorganic hybrid ureasils
Ureasils were prepared via a two-step sol−gel process.ICPTES is mixed with JEFFAMINE ® in a molar ratio of 2:1 for bis-functionalised amines and 3:1 for tris-functionalised amines in THF.The reaction mixture was refluxed at 70 °C for 24 h to obtain the organic−inorganic hybrid precursor, diureapropyltriethoxysilane (d-UPTES) or triureapropyltriethoxysilane (T-UPTES) in solution.In the second step, gelling reagents were sequentially added (ethanol, HCl (0.5 M) and water) to either d-UPTES or T-UPTES and thoroughly mixed.The molar ratio of ICPTES:ethanol:HCl:water used was 176:350:1:265.For chromophore doped ureasils, stock solutions of DPA and PdOEP were dissolved in THF and added to the mixture before the gelling reagents were added.The resulting mixture was poured into a polypropylene mould and gelled into free-standing monoliths via the sol-gel process.The mould was sealed with Parafilm M ® to ensure slow evaporation of excess THF in the samples over 1-2 days, followed by further oven drying at 40°C for 1-2 days, until the excess THF had evaporated.
The shrinkage of each ureasil was calculated by the original volume of solution before gelation and the final volume of the solid-state gel host after THF evaporation is complete.The shrinkage is calculated and then used to reverse engineer the desired final concentration of DPA and PdOEP in the dry gel.For example, if the ureasil shrinks by 50%, the original solution should be 50% diluted in chromophore concentration before gelation.
The weight percentage of silica was calculated according to Table S1, based on the molecular weight of O-Si-O linkages over the molecular weight of the entire polymer chain and urea linkage.The calculation assumes that the monolith is completely dry and gives an indication of the ratio of organic to inorganic domains and their subsequent contribution to the material properties.

Ultraviolet/visible transmittance spectroscopy
UV/Vis transmittance spectra were measured with a PerkinElmer Lambda 750 spectrophotometer using wavelength scan with a resolution of 1 nm at a scan speed of 267 nm/min and a slit width of 2 nm.Solid-state samples (thickness = 1 mm) were directly mounted to the sample holder.

Steady-state photoluminescence (PL) spectroscopy
Steady-state PL spectroscopy was performed on a Fluorolog-3 spectrophotometer (Horiba Jobin Yvon).Solid-state emission spectra were recorded using the front-face configuration.
The excitation and emission slits were adjusted so that the maximum PL intensity was within the range of linear response of the detector and were kept the same between samples if direct comparison between the emission intensity was required.Emission and excitation spectra were corrected for the wavelength response of the system and the intensity of the lamp profile over the excitation range, respectively, using correction factors supplied by the manufacturer.Photoluminescence quantum yields (PLQYs) were measured using a Quanta-phi integrating sphere (Horiba Jobin Yvon) mounted on the Fluorolog-3 spectrophotometer.

Upconversion (UC) quantum yield
The UC emission spectra, phosphorescence spectra, threshold intensity ( !" ), UC quantum yield (UCQY) and UC fluorescence lifetime of all samples were measured with an FLS 1000 TCSPC spectrometer (Edinburgh Instruments Ltd.).Samples were excited with a 532 nm laser (MGL-III-532, 200mW), with a short-pass filter (cut-off 500 nm, Thorlabs) applied in front of the detector.The laser power was adjusted using Thorlabs PM100A Power Meter Console combined with a S120VC Si photodiode power sensor (range: 200-1100 nm).
The UCQY was measured with an integrating sphere (SNS125 5-inch sphere, three windows, International Light Technologies, Figure S1).The sample was cut into a 1 cm 2 square and loaded at the center of the sphere by a sample holder.A baffle is placed in front of the observation window, which blocks any scattering and reflection of the laser from the sample surface.The angle of the sample holder is adjustable.The normal direction of the sample holder is 22.5˚ to the excitation beam line, which leads the reflection of the laser to the inner surface of the sphere.
The laser power was measured with a photodiode before each UCQY measurement.Both the emission of the sample (380-500 nm) and scattering of the laser beam (530-534 nm) were measured.A neutral density filter (O.D.=3.0) was placed before the excitation beam for the scattering intensity measurements.Six data sets were collected to calculate the UCQY of each sample: 1. sample in the path of the beam -"in fluorescence"; 2. sample in scattering; 3. sample facing away from beam -"out of fluorescence", 4. sample out of scattering; 5. empty sphere fluorescence; 6. empty sphere scattering.
Three sets of data were obtained for three 1 cm 2 sections of ureasil sample.The parallel data sets were calculated separately, which gives three UCQY results for each sample, and the reported UCQY is the average of these data, along with the standard deviation of the measurements.During the 'sample in beam' measurement, the sample was facing toward both the excitation window and the observation window, while in the sample-out mode, the holder was turned 180˚ to have the back of the holder facing the windows.For the fluorescence measurement (Data sets 1 and 3), the bandwidth was 1 nm for the detector, and the scan step was 1 nm per data point with duration of 1 second, scanned from 380 nm to 500 nm.For the scattering measurement (Data sets 2 and 4), the bandwidth was 1 nm for the detector, and the scan step was 0.1 nm per data point with duration of 0.1 second, scanned from 530 nm to 534 nm.The transmittance of the filter at the excitation wavelength was measured with a UV-Vis absorption spectrometer (DS5, Edinburgh Instruments Ltd.), taking the average over 10 parallel measurements.The empty-sphere data sets (Data sets 5 and 6) were collected at the beginning of the measurement, under the same conditions of the sample-in measurement, which were shared in all calculations of samples measured in the same day.During the calculation, all data were corrected by the transmittance of each filter used, and normalized based on the slit-width, scan step and the scan duration used.
The UCQY was calculated using the experimental approach described by Porrès et al. 1 and the following formulae: 2,3 where A is the percentage of the photons absorbed directly by the sample, which is corrected by removing the secondary absorption from the sphere-reflected photons: where E is the integrated photon counts from emission spectra, and L is the integrated photon counts from the scattering spectra.x delineates sample, while b is blank.In means the sample was in the path of the excitation beam, and out delineates the sample is out of the beam line.
A quantum yield is defined as the ratio of absorbed to emitted photons, meaning the UCQY is limited to 50% since this is a bimolecular process.While some papers report this as a normalised value, UCQY is reported to its un-normalised value here, to a maximum of 50%.

Time-resolved emission measurements
Fluorescence decay measurements were performed using the multi-channel scaling (MCS) method on a FLS1000 PL spectrometer.The emission decay was recorded using a highspeed photomultiplier tube (PMT-980) equipped with TCC2 counting electronics.The UC fluorescence decay profile was measured using 532 nm laser excitation (MGL-III-532 laser) and a visible PMT-980 detector, with a short-pass filter (cut-off 500 nm, Thorlabs) applied in front of the detector.The plus repetition rate was 12.5 kHz (80 ms), the laser power was set to its maximum, and the pulse width was gradually increased until more than 1000 counts/s was observed at 440 nm.For phosphorescence lifetimes, a long-pass 550 nm filter was used in place of the 500 nm cut off filter and the repetition rate increased to1 kHz (1 ms).The instrument response function (IRF) was measured using with a SiO2 particle suspension solution (Ludox® colloidal silica) using a neutral density filter (OD=3) in front of the excitation source, and with no emission filter.
Individual tail-fits were applied to each emission decay curve using the FAST software package (Edinburgh Instruments) using a multiexponential decay function: where ai and tI are the amplitude and lifetime of the ith component, respectively, and ai is normalized to unity for n components.The goodness of fit was assessed using the reduced chi-square statistics, χ 2 , and the randomness of the residuals. 5To facilitate comparison of the emission decay behaviour between different samples, the average lifetime 〈〉 and fractional contributions (fi) are also reported, which for a bi-exponential decay is given by: where (Eq.S5)

Fourier transform infrared (FTIR) spectroscopy
FTIR spectra were obtained using a Nicolet iS5 (ThermoFisher instruments) spectrometer fitted with an attenuated total reflectance (ATR) diamond press at room temperature over a range of 4000-400 cm -1 with a resolution of 0.5 cm -1 , averaging 64 scans.Deconvolution of the Amide I region (1800-1600 cm -1 ) was performed using a non-linear least square (NLLS) fit to a Gaussian function in Origin 2019 (OriginLab) software.The quality of all fits was assessed qualitatively by comparison of the cumulative fit peak (given by the sum of the intensity of each of the fitted peak at a given x value) to the raw data.

Mechanical testing
4-Point flexural testing and uniaxial tensile testing were performed on a Tinius Olsen 1ST using a 25 N load cell.Samples were cut into tensile specimens using a HPC Laser Ltd.Laserscript LS3060 to produce a 15 mm × 3 mm testing area with a 4.5 mm gripping area at the ends.In the 4-point geometry, samples were placed on supports 12 mm apart and loaded from above at a speed of 1 mm/min by two loading points separated by 4 mm.The force and corresponding displacement on these loading points was recorded until the point at which the loading points were compressing rather than bending the sample.In the uniaxial tensile geometry, samples were gripped at both ends and pulled apart at a speed of 2 mm/min until failure was observed.The force and corresponding displacement of the grips was recorded until failure.The datasets were analysed in Origin 2019 (OriginLab) to generate the elastic and flexural moduli of the samples from the gradient of the initial straight region of the stressstrain (first 100 data points) or force-extension curves (first 500 data points), respectively.In the case of the 4-point geometry, the experimental data was first transformed by multiplying the force required for deflection by a sample-dependent constant, k, and plotted against the position of the grips.This constant was determined from  = 0 ./ ( 01 1, where a is the distance between loading points (4 mm) and I is the second moment of area of the sample about the bending neutral axis.I was calculated from  = +" ( 23 , where b is the width of the sample and h is the sample thickness.Repeats were taken of each sample.

1.9.
Dynamic mechanical analysis (DMA) Dynamic mechanical analysis was performed on a TA Instruments Q800 DMA in 4-point bending mode over the temperature range -100-300°C at a heating rate of 5 °C/min.The samples were cut to dimensions 50 mm × 8 mm × 1-2 mm to fit the double cantilever.

Powder X-ray diffraction (PXRD)
PXRD was performed on a B3 Bruker D8 Advance.The samples were exposed to the Cu Kα radiation (λ = 1.54 Å) at room temperature and measured using silicon substrate to minimise background scattering.The investigated 2θ range was 0°−50°.

Optimization of sensitizer/emitter ratio
The effect of sensitizer/emitter (S/E) concentration was investigated in the tU(3000) host to optimize the TTA-UC performance based on the upconversion quantum yield (FUC) and upconversion lifetime (τUC).The results are summarized in Table S2.It was found that doubling the concentration of both chromophores produced the same TTA-UC efficiency within error.The effect of doubling or halving the concentrations of S/E was also investigated and, in both cases, showed that doubling/halving the ratio between them lowered the TTA-UC efficiency from 16% to 12%.When DPA concentration was doubled to 20 mM, the TTA-UC efficiency rose to 20±5% from 16%, but the total number of UC emission counts decreased, which suggests that self-annihilation may be occurring.
Analysis of the upconversion lifetime data (Figure S3, Table S3) confirmed these trends.All decay curves required a bi-exponential fit to the data, with the quality of fit evaluated by the evenness of the residuals.The intensity-averaged upconversion lifetime was found to be longest at an S:E ratio of 1:100, as seen in the 10 mM DPA/0.1 mM PdOEP and 20 mM DPA/0.2 mM PdOEP systems.These also have a larger contribution from the first lifetime component (t1) than other samples.Doubling or halving the emitter concentration leads to a reduction in the average UC lifetime.The 20 mM DPA/0.1 mM PdOEP system had a high UCQY but a reduced lifetime compared to 10 mM DPA/0.1 mM PdOEP system.This is likely due to increased self-annihilation at elevated emitter concentrations.Shorter lifetimes were also observed when doubling or halving the sensitizer concentration.Based on these observations, we selected the DPA (10 mM) and PdOEP (0.1 mM) as the emitter/sensitizer concentration to be used in all samples.Table S3.Summary of fitting parameters obtained for double exponential tail fits to the upconversion lifetime decay curves (Figure S3) at different sensitizer/ emitter concentrations.λex = 532 nm, λem = 440 nm.Table S4.Photoluminescence quantum yields (FPL) of emitter-doped and sensitizer/emitter-doped ureasils.Figure S7.Upconversion emission spectra of all DPA/PdOEP doped ureasils excited with a 532 nm laser at 1000 mW cm -2 .These spectra are from one measurement and serve as a demonstration of the UC emission spectrum.Repeat measurements were taken to yield the average UCQY and associated error.

Figure S8.
Phosphorescence emission spectra of all ureasils (λex = 532 nm at 1,000 mW cm -2 , λem = 600-800 nm) doped with 0.1 mM PdOEP only (black lines) and doped with 10 mM DPA and 0.1 mM PdOEP (red lines).These spectra are from one measurement and serve as a demonstration of the emission spectrum.Repeat measurements were taken to yield the average TTET and associated error.

Calculation of TTET efficiency
Equation S5 was used to calculate the TTET efficiency from the phosphorescence spectra of ureasil samples doped with just the sensitizer or the sensitizer/emitter pair.An average of three emission scans were taken for each doped ureasil to obtain an error from the standard deviation.4000), (c) tU(3000) and (d) tU(5000).Labels A-E refer to datasets for repeat samples which yield an average constant and associated error.DU(4000) showed the lowest flexural modulus, as found by a linear fit to the first 100 points of data.tU(3000) had the highest modulus due to higher crosslinking density and inorganic silica content.

PXRD: determination of structural unit distance and coherence length
Powder X-ray diffraction was used to analyze the inorganic domains in ureasil hosts.A typical ureasil diffraction pattern consists of a primary band (Band 1, Figure S15) centred at ~20°, with a shoulder at ~15-18°(Band 2, Figure S15). 10,11Ordering within the siliceous domains gives rise to the main peak, while in-plane ordering within the siliceous domains is ascribed to the lower intensity shoulder peak. 12The coherence length along which the structural unit survives, L, can be calculated from the primary peak using the Scherrer equation: 13  = 6.89: where λ is the wavelength of the X-ray radiation (Cu, λ = 1.5406 nm) and A is the full-widthhalf-maximum of the Bragg peak in radians.
The structural unit distance, d, is calculated using Bragg's law: where n is a positive integer (in this case, n=1), λ is the X-ray wavelength and θ is the angle of incident radiation.

Figure S1 .
Figure S1.Integrating sphere set-up used to determine UCQYs.

Figure S3 .
Figure S3.Upconversion decay curves (open symbols), fits (solid lines) and residuals of PdOEP/DPAdoped ureasil tU(3000) at varied concentrations, obtained from excitation at 532 nm and detection at 440 nm (with 550 nm short-pass filter).All fits were obtained via tail-fitting to a bi-exponential function.

Figure S4 .
Figure S4.Photographs of all four doped ureasils exhibiting green-to-blue upconversion in ambient conditions under excitation with a 532 nm laser pointer.

Figure S5 .
Figure S5.Optical properties of (0.1 mM)/DPA (10 mM) in the solid-state DU(2000) host.The absorption (solid black line, normalised to DPA maximum) and emission spectra (normalised to peak maximum for each component) for DPA (dashed blue lines) and PdOEP (dashed pink lines) obtained upon excitation at 532 nm are shown.The normalised PdOEP absorption band for this sample (solid pink lines) is also included to facilitate comparison.

Figure S6 .
Figure S6.Normalized emission spectra of undoped ureasil samples as a function of excitation wavelength.The emission bands can be resolved into two components: a longer-wavelength blue component is ascribed to the donor-acceptor electron-hole recombination at the urea linkages via proton transfer, and the short-wavelength purple-band band is due to localized oxygen defects at the siliceous nanodomains.6,7

Figure S9 .
Figure S9.Decay curves of DPA/PdOEP doped ureasils.(a) Upconversion decay curves shown in Figure5awith scale adjusted to show the rise in emission prior to decay.(b) Phosphorescence decay curves (open symbols), fits (solid lines) and residuals for ureasil samples doped with sensitizer only (0.1 mM PdOEP).Measurement conditions: 532 nm laser excitation, detecting at 665 nm until1,000 counts, at 1 kHz with a 550 nm long-pass filter.

Figure S10 .
Figure S10.UV-Vis absorption spectra of DPA/PdOEP doped ureasils at increasing time intervals after gelation showing the photodegradation of DPA over time.

Figure S14 .
Figure S14.Tensile testing data from (a) DU(2000), (b) DU(4000), (c) tU(3000) and (d) tU(5000).Labels A-E refer to datasets for repeat samples which yield an average constant and associated error.DU(4000) showed the lowest flexural modulus, as found by a linear fit to the first 100 points of data.tU(3000) had the highest modulus due to higher crosslinking density and inorganic silica content.

Figure S15 .
Figure S15.PXRD spectra of undoped ureasils.The original processed spectrum (black line), band 1 contribution (blue line), band 2 contribution (green line) and cumulative fit of Gaussians (white circles) are all shown.

Table S1 .
Calculation of silica weight percentage (wt%) for organic-inorganic polymer hybrids.PEO = polyethylene oxide in JEFFAMINE precursor.b Calculated MW of the total organic polymer chain (including organic amide links formed in step 1 of synthesis).c MW of organic polymer divided by the number of linkages for each chain (2 for di-ureasils, 3 for tri-ureasils).
a d Molecular weight of inorganic -O-Si-O-linker.

Table S2 .
Investigation of optimum chromophore concentrations in ureasil tU(3000).Rough measurements were taken of key parameters to determine UC performance.
a Photoluminescence quantum yield of the emitter determined by the integrating sphere method (lex = 375 nm).c Upconversion quantum yield at excitation intensity of 1 W cm -2 (532 nm).c Upconversion emission intensity is the integrated area under UC emission (380-500 nm).d Average upconversion lifetime (lex = 532 nm, lem = 440 nm) determined from a bi-exponential fit.** Insufficient absorbance by PdOEP at this concentration to measure accurately.

Table S5 .
Summary of fitting parameters obtained for double exponential tail fits to the upconversion lifetime decay curves.λex = 532 nm, λem = 440 nm.Chi-squared, a measure of goodness of fit.b Normalized pre-exponential factor for ith component.c Fractional contribution of the ith component.d Lifetime of the ith component. a

Table S6a .
Summary of fitting parameters obtained for double exponential tail fits to the phosphorescence decay curves for DPA (10 mM)/PdOEP(0.1 mM)-doped ureasils.λex = 532 nm, λem = 665 nm.Chi-squared, a measure of goodness of fit.b Normalized pre-exponential factor for ith component.c Fractional contribution of the ith component.d Lifetime of the ith component. a
a Chi-squared, a measure of goodness of fit.b Normalized pre-exponential factor for ith component.c Fractional contribution of the ith component.d Lifetime of the ith component.

Table S10 .
Key photophysical parameters of tU(3000) doped with 10 mM DPA and 0.1 mM PdOEP, when prepared in either entirely ambient conditions, or when the doped t-UPTES solution was degassed with nitrogen bubbling for one hour prior to gelation in ambient conditions.excitation was used for all above measurements.a Average upconversion lifetime (lex = 532 nm, lem = 440 nm).b Average phosphorescence lifetime in presence of DPA (lex = 532 nm, lem = 665 nm).All measurements were performed under ambient conditions.Errors are the standard deviation of two measurements.