Orange/Red Benzo[1,2-b:4,5-b′]dithiophene 1,1,5,5-Tetraoxide-Based Emitters for Luminescent Solar Concentrators: Effect of Structures on Fluorescence Properties and Device Performances

Luminescent solar concentrators (LSCs) are a class of optical devices able to harvest, downshift, and concentrate sunlight, thanks to the presence of emitting materials embedded in a polymer matrix. Use of LSCs in combination with silicon-based photovoltaic (PV) devices has been proposed as a viable strategy to enhance their ability to harvest diffuse light and facilitate their integration in the built environment. LSC performances can be improved by employing organic fluorophores with strong light absorption in the center of the solar spectrum and intense, red-shifted emission. In this work, we present the design, synthesis, characterization, and application in LSCs of a series of orange/red organic emitters featuring a benzo[1,2-b:4,5-b′]dithiophene 1,1,5,5-tetraoxide central core as an acceptor (A) unit. The latter was connected to different donor (D) and acceptor (A′) moieties by means of Pd-catalyzed direct arylation reactions, yielding compounds with either symmetric (D–A–D) or non-symmetric (D–A–A′) structures. We found that upon light absorption, the compounds attained excited states with a strong intramolecular charge-transfer character, whose evolution was greatly influenced by the nature of the substituents. In general, symmetric structures showed better photophysical properties for the application in LSCs than their non-symmetric counterparts, and using a donor group of moderate strength such as triphenylamine was found preferable. The best LSC built with these compounds presented photonic (external quantum efficiency of 8.4 ± 0.1%) and PV (device efficiency of 0.94 ± 0.06%) performances close to the state-of-the-art, coupled with a sufficient stability in accelerated aging tests.


DFT and TD-DFT computational investigation
. Bond lengths (Å) and dihedral angles (degrees) of S0 and S1 (in brackets) optimized geometries of all compounds of the BDT series.

Synthesis of the oxidized BDT central core 1
The synthesis of the new compounds started with the preparation of the oxidized central BDT core. According to some previously reported procedures, [1][2][3] starting material benzo[1,2-b:4,5-b′]dithiophene-4,8-dione (A) was first reduced with metallic zinc and then alkylated to yield intermediate B. The latter was then oxidized with m-CPBA to give the corresponding sulfur tetraoxide derivative 1 in good yield (Scheme S1).
Scheme S1. Synthesis of the oxidized BDT central core 1.

4,8-Bis((2-ethylhexyl)oxy)benzo[1,2-b:4,5-b']dithiophene (B)
Benzo[1,2-b:4,5-b′]dithiophene-4,8-dione (A, 50 mg, 0.23 mmol, 1.0 eq) was dissolved in dry and degassed N,N-DMF (5 mL). Then, Zn powder (120 mg, 1.84 mmol, 8.0 eq, previously activated) and KOH (130 mg, 2.30 mmol, 10.0 eq) were added to the reaction mixture and the latter was heated to 80°C for 16 h under an inert atmosphere of N2. The mixture slowly turned from dark blue to rotten green. 2-Ethylhexyl bromide (193 mg, 1.0 mmol, 4.4 eq) was then added, and the suspension was stirred for an additional 16 h. After cooling to room temperature, the mixture was filtered over a short pad of Celite © to remove the excess Zn, washed with HCl 1M (10 mL) and water (10 mL), and extracted with CH2Cl2 (3 × 10 mL). The combined organic layers were dried over Na2SO4 and concentrated under vacuum. The residue was purified by flash column chromatography (SiO2, Petroleum Ether/ CH2Cl2 6:1 to 4:1) to yield product 2 (83 mg, 0.19 mmol, 81%) as a yellow oil. 1  Compound B (60 mg, 0.13 mmol, 1.0 eq) was dissolved in CH2Cl2 (5 mL) and the solution was cooled to 0°C. Then, a second solution of m-CPBA (140 mg, 0.81 mmol, 6.0 eq) in CH2Cl2 (10 mL) was added dropwise to the first one. The resulting mixture was shielded from the light. The yellow solution was allowed to warm up to room temperature and stirred for 72 h. The reaction was then quenched with H2O, extracted with CH2Cl2 (3 x 10 mL) and carefully washed with a solution of NaHCO3 (3 x 10 mL). The combined organic layers were dried over Na2SO4 and concentrated under vacuum. The residue was purified by flash column chromatography (SiO2, Petroleum Ether / CH2Cl2 1:1) to yield product 3 (51 mg, 0.1 mmol, 74%) as a pale-yellow solid. 1    a In good agreement with literature data. [4] b Absolute QY determined using an integrating sphere. c Stokes shift; d In these solvents, due to the low fluorescence yield and longer emission wavelengths of BDT-H1, the reported λ emi max and Stokes shift values are likely overestimated, as a result of the application of the correction function necessary to compensate the sensitivity loss of the photomultiplier in the red wavelength region. e No spectrum could be recorded due to the insufficient solubility of BDT-AA in this solvent.

Absorption efficiency calculation
To provide a quantitative evaluation of the spectral match between the absorption of the organic fluorophores prepared in this study and the emission of the solar simulator lamp used to characterize the LSCs (see below, Figure S11), absorption efficiency (ηabs) was calculated according to the method described by Debije et al. in their reference work. [5] The expression used for calculation was: (Eq. S1) Where Sso(λ) is the emission spectrum of the light source and A(λ) is the absorbance of the fluorophore dispersed in the polymeric matrix, both wavelength-dependent. The result is a numerical value that is dependent on luminophore concentration. The integration limits were selected to account for the spectral range of the luminophore and the emission range of the light source. The curves were plotted against the maximum absorbance at each concentration (Amax), and fitted to the following exponential functions, as previously described: [6] = (1 − − max )  Figure 9a for completeness, but no fitting was carried out, since the spectral shape was altered by the effect of the microcrystalline aggregates of the luminophore, causing an excessive tailing of the spectra at the longest wavelengths ( Figure  S8).

Measurement of external (ηext) and internal (ηint) photon efficiency
All measurements were performed by using a commercially available system (Arkeo, Cicci research s.r.l., Grosseto, Italy) containing a CMOS-based spectrometer with a symmetrical Czerny-Turner optical bench connected to an integrating sphere ( Figure S10).

Figure S10. Photos of the experimental setup utilized for the ηint, ηext and ηdev determination
The illumination source was an ORIEL® LCS-100 solar simulator 94011A S/N: 322 (1 Sun, AM 1.5G). An integrating sphere of 5 cm of diameter and 1 cm of aperture was placed along the edge of the glass plate. To avoid the collection of the stray light, the sphere was covered by an opaque plastic holder with a rectangular aperture of 1 cm (the width of the sphere aperture) × 3 mm (i.e., the thickness of the LSC slab). The integrating sphere was moved along the side of the LSC until all the slab edge had been scanned. The spectrally-resolved edge output photon count was collected from the CMOS-based spectrometer and calibrated into optical power (W) and then in irradiance. Aimed at limiting reflections of unabsorbed light, an absorbing matte black background was placed in contact with the LSC rear side. The illumination source was kept close and perpendicular to the center of the LSC front surface to minimize the divergence of the excitation beam and to avoid the direct illumination of the integrating sphere. A series of 3-5 measurements were repeated to allow the integration sphere to collect the maximum single-edge output power.
The optical performances of LSCs were evaluated in terms of the external (ηext) and internal (ηint) photon efficiency, calculated from equations S4 and S5, respectively: Where: a) n = 4, λ1 = 300 nm and λ2 = 1100 nm; b) The number of edge-emitted photons was obtained from the sum of the output power spectra measured for each edge of the LSC; c) The total number of photons incident on the front surface of the LSC was obtained from the input power spectrum of the light source incident on the illuminated surface area of the LSC ( Figure S11); d) The number of total absorbed photons was obtained by convoluting the absorption spectrum of the LSC and the input power spectrum of the light source incident on the illuminated surface area of the LSC. Such value was also obtained by the difference between the incident input power and the power transmitted by the LSC.

Measurement of device photovoltaic efficiency (ηdev)
The LSC photovoltaic efficiency is determined by attaching two Si-PV cells connected in series to an edge of the thin-film LSCs by using silicone grease. The performance of the assembled LSC-PV systems is assessed under standard illumination conditions by measuring the power conversion efficiency of the resulting LSC device (ηdev), defined as the electrical power effectively extracted from the PV cells (P out el) relative to the luminous power hitting the top surface of the LSC (P in opt): where ISC, VOC and ff are the short-circuit current, open-circuit voltage and fill factor of the edge-mounted PV cells, respectively, ALSC is the front-illuminated area of the LSC device, and P in opt is the incident solar power density expressed in mW cm −2 .
For the determination of ηdev, two Si-PV cells IXYS KXOB25-12X1F (22 x 7 mm, Voc = 0.69 V, Isc = 46.7 mA, ff > 70%, η = 25%) were connected in series ( Figure S12 shows the I/V curve of the two cells assembly under direct simulated solar illumination). The current/voltage characteristics of the LSC-PV system were determined with a precision source/measure unit (Keysight Technologies B2900 Series). Silicon was used to grease the LSC edge. A black matte layer was placed beneath the LSC with an air gap of about 2.5 mm during the measurements.

Accelerated photodegradation test
The experiment was conducted for 650 minutes total at a constant temperature of 70 °C on a BDT-H2containing PMMA film (2.2 wt.%). Wishing to comply with the ASTM G154 standard, [7] a home-made setup was used, which was composed of a LED tower (Cicci research s.r.l., Grosseto, Italy) as light source and an optical fiber connected to a spectroradiometer as the detector (CCARK.A.4.Spectroradiometer, Fiber Optic VIS/NIR spectrometer, 2048 pixels, grating VA (360-1100 nm), slit-50, OSC, DCL-UV/VIS), placed at a distance of < 1 cm from the polymeric film with a detection angle of ca. 35° ( Figure S13). The sample was placed on a controllable hot stage (originally used to carry out spatially resolved photoluminescence tests with a thermal module) to adjust its temperature during the experiment. Figure S13. Photograph of the home-built setup used for the accelerated ageing experiment.
The experimental temperature was set directly using the hot stage control and was checked by means of a FLIR thermal camera ( Figure S14). During the experiment, the entire setup was covered with a black blanket to exclude the external radiation. As incident light sources, two of the LEDs, named "Far UV" and "UV", were selected (95% irradiation in the 361-406 nm range). Their emission was calibrated with an integrating sphere, which was placed at the same distance from the source as that of the sample. After adjusting the LED intensity, we measured an irradiance of 38.43 W/m 2 in the selected area, to be compared with that of the AM 1.5G solar irradiation (33.45 W/m 2 ). The measurement thus took place at 1.15 Sun irradiation conditions. Figure S15. Emission spectrum of the light source used for the photostability experiment, compared to the AM 1.5G spectrum. The area integrated to determine the spectral flux is highlighted in yellow.
The emission spectra of the sample at the different experimental times and the progression of emission peak area with time are reported in Figure S16. To evaluate the real time aging period corresponding to the accelerated test, the method reported in the ASTM F1980 standard [8] was employed, which was based on the estimation of an accelerated ageing factor (AAF) by application of the following equation (Eq. S7). Where Q10 is an acceleration factor indicating how many times the photodegradation rate increases for each 10 °C temperature increase, TAA is the accelerated aging temperature (70°C) and TRT is room temperature (22°C).
Here, we assumed a conservative (and widely accepted) value of 2 for Q10, which equals to consider that the photodegradation rate doubles for each 10 °C temperature increase. By applying equation S7, a value of ca. 27.86 can be calculated for AAF. Then, the real time ageing period can be obtained from the product of the actual experiment duration and the AAF, according to Eq. S8.