Deuteration of Perylene Enhances Photochemical Upconversion Efficiency Andrew Danos,y Rowan W. MacQueen,y Yuen Yap Cheng,y Miroslav Dvoˇr´ak,z,k Tamim A. Darwish,{ Dane R. McCamey,x and Timothy W. Schmidt∗,y School of Chemistry, UNSW Sydney, NSW 2052, Australia, Department of Physical Electronics, Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University in Prague, V Holesovickach 2, 180 00 Prague, Czech Republic, National Deuteration Facility, Bragg Institute, Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, NSW 2232, Australia, and School of Physics, UNSW Sydney, NSW 2052, Australia E-mail: [email protected] ∗To whom correspondence should be addressed ySchool of Chemistry, UNSW Sydney, NSW 2052, Australia zDepartment of Physical Electronics, Faculty of Nuclear Sciences and Physical Engineering, Czech Tech- nical University in Prague, V Holesovickach 2, 180 00 Prague, Czech Republic {National Deuteration Facility, Bragg Institute, Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, NSW 2232, Australia xSchool of Physics, UNSW Sydney, NSW 2052, Australia kSchool of Chemistry, The University of Sydney, NSW 2006, Australia 1 Abstract Photochemical upconversion via triplet-triplet annihilation is a promising technol- ogy for improving the efficiency of photovoltaic devices. Previous studies have shown that the efficiency of upconversion depends largely on two rate constants intrinsic to the emitting species. Here we report that one of these rate constants can be altered by deuteration, leading to enhanced upconversion efficiency. For perylene, deuteration decreases the first order decay rate constant by 16±9 % at 298 K, which increases the linear upconversion response by 45±21% in the low excitation regime. The effect is explained by a balance between the competing effects of changing the density of states and Frank-Condon factors relevant to intersystem crossing. TOC Graphic (in final size). H H H H H H D D D D H H D D ) –1 H H H H D D D D D D decay rate (s The global transition towards renewable energy is the most pressing environmental and moral imperative of the 21st century. Only direct solar collection has the capacity to satisfy the large and growing global energy demand in its own right, although it is more commonly used to complement other renewable energy sources.1,2 At present, the two preeminent technologies for solar collection are photovoltaics (PV) and solar thermal. The widespread deployment of both technologies is currently limited by their higher cost of electricity production compared with traditional combustion of fossil fuels, although price parity is rapidly approaching in certain contexts.3,4 These costs are largely attributed to fixed manufacture and installation, and so the overall cost of electricity generated can be reduced by improving operational 2 (a.) TPTBPt (b.) S1 N N perylene Pt N N 2. ISC k S1 2 T1 T1 T1 T1 5. 3. TET 4. TTA k1 1. S0 S0 S0 S0 Figure 1: The mechanism of TTA-UC. (a.) 1. Photons are absorbed to take sensitizers from the ground to the excited singlet state. 2. Intersystem crossing (ISC) takes sensitizers to their lowest triplet state. 3. The triplet energy is transferred to the emitter molecule, which may undesirably decay with a rate constant k1. (b.) 4. Emitter triplets annihilate to produce a higher energy singlet with rate constant k2. 5. Higher energy photons are emitted. efficiency. The electrical conversion efficiency of PV devices, and consequently the cost of the elec- tricity they produce, is limited by the optical properties of the absorbing material, usually a semiconductor.5,6 Photons with energy below the bandgap of the absorbing material are un- able to generate photocurrent, which for silicon represents 20% of incoming solar photons.7 One way to access and harness the energy of these below-bandgap photons is photochemical upconversion (UC) via triplet triplet annihilation (TTA).8{14 Significant progress has been made in recent years to improve PV applications of UC,8,15{19 to develop integrated UC-PV devices,20 and to better understand the underlying TTA process.21{25 3 The process by which sub-bandgap photons can be reclaimed by UC is shown in Figure 1. Low energy photons are absorbed by a sensitizing species, which is chosen such that it undergoes efficient intersystem crossing (ISC) to the first excited triplet state (T1). Excita- tion is then transferred to an emitting species via triplet energy transfer (TET), a Dexter process which preserves multiplicity.26{28 In this way, sustained excitation of the sensitizer population can generate a population of excited T1 emitters. When two excited emitter molecules undergo TTA, their total energy is redistributed to promote one molecule to the first excited singlet state (S1) and demote the other to the ground state (S0). The TTA process is spin allowed as it preserves the overall spin of the pair. The promoted emitter molecule rapidly fluoresces to produce a photon of higher energy than those initially ab- sorbed by the sensitizers. These upconverted photons can be utilised by PV devices, adding to their photocurrent and improving their electrical conversion efficiency.8,15,18,20 Previous studies have shown that the efficiency of UC depends on two rate constants intrinsic to the emitter molecule.22,25 The kinetics of UC is governed by the rate equation: p d I d[T] UC / = k [S] − k [T] − k [T]2 (1) dt dt φ 1 2 where IUC is the intensity of UC emission, [T] is the concentration of emitter molecules in the T1 state, [S] is the concentration of sensitizer molecules in their S0 state, kφ is the rate constant of excitation of sensitizer molecules, and k1 and k2 are the first and second-order decay rate constants for excited emitters, respectively. Under low excitation and steady state conditions (relevant to solar applications and where first-order decay dominates), the rate equation for IUC yields 2 2 k2kφ[S] ηΦPL IUC / 2 (2) 2k1 where η is the efficiency of the TTA process for two excited emitters to produce an excited 29 singlet, and ΦPL is the fluorescence quantum yield of the emitter. Equation 2 provides the roadmap for improving the PV enhancing capabilities of UC 4 devices. Significant research effort has been made to identify and synthesize high perfor- mance emitters (those with high k2,ΦPL, and η; and low k1), although a rational approach to improving each of these parameters simultaneously has not yet been developed. Previous studies of polycyclic aromatic hydrocarbons (PAHs) such as perylene in low temperature crystals or solids have shown that deuteration reduces k1, as measured by the phosphores- cence lifetime.30{32 In research and in applications, small PAHs such as perylene make up the majority of UC emitters due to their chemical robustness and desirable UC parameters. 9,13,14 For this reason, confirming that deuteration causes an increase in UC performance by reducing k1 is of general interest and wide applicability. Experimental The hydrogen isotopologue of perylene was purchased from L. Light & Co. Ltd., while perylene-d12 with 98% isotopic purity was purchased from Sigma Aldrich. The sensitizer used for both emitters was Pt(II) meso-tetraphenyl tetrabenzoporphine (TPTBPt), purchased from Frontier Scientific. Upconverting solutions with 1 mM of either perylene isotopologue and 0.1 mM TPTBPt were prepared in toluene for the collection of action spectra. For UC kinetics, the concentra- tions were 2.5 mM of emitter and 0.25 mM of sensitizer. These solutions are referred to as P0 and P98 according to their degree of emitter deuteration. In order to ensure identical concen- trations of sensitizer in comparative samples, the solutions were prepared by dispensing equal volumes from a common sensitizer stock solution. Each sample was rigorously deoxygenated by three freeze-pump-thaw cycles at 3×10−3 mbar and 77 K before measurement. Action spectra were collected using a lock-in pump-probe method described in Ref.29 Perylene fluorescence was detected at 470 nm and repeated at several pump powers. The pump beam itself was passed through a 570 nm long-pass filter so as not to cause any direct excitation of the emitter. The experimental conditions were identical between P0 and P98 5 UC samples. Time-resolved photoluminescence spectra were collected in the same way as previous ki- netic studies.24 An Acton/Princeton iCCD spectrograph/camera was used to collect UC flu- oresence following excitation of the sample by the 615 nm output of a TOPAS OPA pumped with a Clark MXR CPA 2210 femtosecond laser running at 1 kHz repetition rate. The elec- tronic gating of the CCD was set to acquire in 1 µs intervals from 100 ns to 900.1 µs after the laser pulse, and to collect for 1000 shots at each time-point. Measurements were repeated at several pulse energies in the range of 150 nJ to 12 µJ using neutral density filters to attenuate the excitation. Results and Discussion Action spectra The raw action spectra were processed by scaling the P0 traces up so that the direct fluo- rescence response peak at 445 nm was the same height for P0 and P98 for scans with the same pump power. Since the electronic properties of perylene are unaffected by deuteration (experimentally confirmed by their identical absorbance and fluorescence spectra, included in the Supporting Information), we expect this peak to be the same for both emitters, and attribute these small scaling factors to variability in detector sensitivity. The scaling factors are tabulated in the Supporting Information, with none larger than 4%. As the P0 traces were scaled up to the P98 traces, the effect of this correction is to make the reported relative efficiency enhancement a conservative underestimate.