Investigation of Hole-Transfer Dynamics through Simple EL De-Convolution in Non-Fullerene Organic Solar Cells

In conventional fullerene-based organic photovoltaics (OPVs), in which the excited electrons from the donor are transferred to the acceptor, the electron charge transfer state (eECT) that electrons pass through has a great influence on the device’s performance. In a bulk-heterojunction (BHJ) system based on a low bandgap non-fullerene acceptor (NFA), however, a hole charge transfer state (hECT) from the acceptor to the donor has a greater influence on the device’s performance. The accurate determination of hECT is essential for achieving further enhancement in the performance of non-fullerene organic solar cells. However, the discovery of a method to determine the exact hECT remains an open challenge. Here, we suggest a simple method to determine the exact hECT level via deconvolution of the EL spectrum of the BHJ blend (ELB). To generalize, we have applied our ELB deconvolution method to nine different BHJ systems consisting of the combination of three donor polymers (PM6, PBDTTPD-HT, PTB7-Th) and three NFAs (Y6, IDIC, IEICO-4F). Under the conditions that (i) absorption of the donor and acceptor are separated sufficiently, and (ii) the onset part of the external quantum efficiency (EQE) is formed solely by the contribution of the acceptor only, ELB can be deconvoluted into the contribution of the singlet recombination of the acceptor and the radiative recombination via hECT. Through the deconvolution of ELB, we have clearly decided which part of the broad ELB spectrum should be used to apply the Marcus theory. Accurate determination of hECT is expected to be of great help in fine-tuning the energy level of donor polymers and NFAs by understanding the charge transfer mechanism clearly.


Introduction
Organic photovoltaic cells (OPVs) based on the bulk-heterojunction (BHJ) blend of semiconducting polymers and small molecule acceptors are considered the most promising candidate as a future mobile energy source due to their advantages, including light weight, flexibility, and low-cost fabrication process [1][2][3][4][5].The power conversion efficiency (PCE) of single-junction polymer solar cells has steadily increased and recently exceeded 18%.This was due to the introduction of non-fullerene acceptors (NFAs) [6][7][8].The major advantage of NFAs is their relatively low energy loss, which can be as low as 0.4-0.5 eV [9,10].It is generally believed that this energy loss is associated with the small charge transfer offset required to drive charge separation.In other words, charge separation efficiency is directly influenced by charge transfer offsets.Therefore, finding accurate information about the charge transfer (CT) state is essential for improving photovoltaic performance.
In the BHJ system based on a fullerene acceptor (f-BHJ), the electron transfer state ( e E CT ) was determined by Marcus theory, which is based on the mirror image relationship Polymers 2023, 15, 4042 2 of 14 between the photovoltaic external quantum efficiency (EQE PV ) and electroluminescence (EL) emission [11,12].All the charges that contribute to the onset part of the EQE PV are initially excited at the donor site, and then electrons are transferred to the acceptor through the e E CT .Thus, if the EQE PV is measured with a very sensitive tool such as Fourier transform photocurrent spectroscopy EQE (FTPS-EQE), it is expected that the CT-bands are visible in the low-energy part of the EQE PV (see Figure 1a) [13,14].In this framework, e E CT was extracted by fitting the CT-band from the EQE PV and projecting it on the EL spectrum (Figure S1 and Table S1).In the BHJ system based on a low bandgap non-fullerene acceptor (nf-BHJ), however, the electron density excited in the acceptor itself is already high, and the low-energy onset part of EQE PV is also formed only by the contribution of the acceptor.Therefore, e E CT created near the lowest unoccupied molecular orbital (LUMO) levels of donor and acceptor has little influence on the device performance, and there is no CTband feature in the EQE PV onset (see Figure 1b) [13].Rather, in this non-fullerene acceptor system, the energy offset associated with the hole transfer occurring on the highest occupied molecular orbital (HOMO) side of the donor and acceptor is more important.Accordingly, many previous studies to date have discussed the hole transfer on the HOMO side rather than electron transfer on the LUMO side [15][16][17][18][19].Of course, there may be a change in photocurrent due to the influence of e E CT in operation under the saturation mode, but the influence of e E CT is negligible because there are no electrons transferring from the donor in the V OC formation near the EQE onset we consider.
efficiency is directly influenced by charge transfer offsets.Therefore, finding accurate information about the charge transfer (CT) state is essential for improving photovoltaic performance.
In the BHJ system based on a fullerene acceptor (f-BHJ), the electron transfer state ( e ECT) was determined by Marcus theory, which is based on the mirror image relationship between the photovoltaic external quantum efficiency (EQEPV) and electroluminescence (EL) emission [11,12].All the charges that contribute to the onset part of the EQEPV are initially excited at the donor site, and then electrons are transferred to the acceptor through the e ECT.Thus, if the EQEPV is measured with a very sensitive tool such as Fourier transform photocurrent spectroscopy EQE (FTPS-EQE), it is expected that the CT-bands are visible in the low-energy part of the EQEPV (see Figure 1a) [13,14].In this framework, e ECT was extracted by fitting the CT-band from the EQEPV and projecting it on the EL spectrum (Figure S1 and Table S1).In the BHJ system based on a low bandgap non-fullerene acceptor (nf-BHJ), however, the electron density excited in the acceptor itself is already high, and the low-energy onset part of EQEPV is also formed only by the contribution of the acceptor.Therefore, e ECT created near the lowest unoccupied molecular orbital (LUMO) levels of donor and acceptor has little influence on the device performance, and there is no CT-band feature in the EQEPV onset (see Figure 1b) [13].Rather, in this nonfullerene acceptor system, the energy offset associated with the hole transfer occurring on the highest occupied molecular orbital (HOMO) side of the donor and acceptor is more important.Accordingly, many previous studies to date have discussed the hole transfer on the HOMO side rather than electron transfer on the LUMO side [15][16][17][18][19].Of course, there may be a change in photocurrent due to the influence of e ECT in operation under the saturation mode, but the influence of e ECT is negligible because there are no electrons transferring from the donor in the VOC formation near the EQE onset we consider.A study by J. Zhang et al. [20] showed that the minimum HOMO offset needed to achieve the most efficient exciton dissociation and photovoltaic performance was ~40 meV, and the method for relative comparison of HOMO offset energy was demonstrated by Y. Xie et al. [21].However, the former was just an estimate of the minimum value obtained through a simple comparison of the HOMO level difference and photovoltaic performance [20], and the latter was a relative comparison of estimated values obtained by determining EQEEL difference [21].Thus, the discovery of a method to determine hole A study by J. Zhang et al. [20] showed that the minimum HOMO offset needed to achieve the most efficient exciton dissociation and photovoltaic performance was ~40 meV, and the method for relative comparison of HOMO offset energy was demonstrated by Y. Xie et al. [21].However, the former was just an estimate of the minimum value obtained through a simple comparison of the HOMO level difference and photovoltaic performance [20], and the latter was a relative comparison of estimated values obtained by determining EQE EL difference [21].Thus, the discovery of a method to determine hole transport energy offsets by accurately measuring the E CT on the HOMO side ( h E CT ) remains an open challenge.In recent research by J. Wu et al., it was pointed out that the key factor of the high efficiency of the PM6:Y6 system was because of the low energetic disorder due to the low energy offset [22].Therefore, the accurate determination of h E CT is essential for Polymers 2023, 15, 4042 3 of 14 achieving further enhancements in the non-fullerene organic solar cell since it can be a guideline for the design of a higher-performing novel material.
In this work, we suggest a simple method to determine the exact h E CT level via deconvolution of the EL spectrum of the BHJ blend (EL B ).In general, the onset part of the EQE of the BHJ blend with NFA is formed solely by the contribution of the acceptor itself.In an EL measurement, electrons are injected directly into the acceptor LUMO, and holes are initially injected into the HOMO of the donor and then transferred to the acceptor through h E CT .Thus, EL B has constraints on both the singlet recombination of the acceptor and the radiative recombination via h E CT .Therefore, by deconvolution of EL B , it is possible to extract out the part of CT-band contribution.Through this, we can clearly decide which part of the broad EL B spectrum should be used to apply the Marcus theory.To generalize, we have applied our EL B deconvolution method to nine different BHJ systems consisting of the combination of three donor polymers (PM6, PBDTTPD-HT, PTB7-Th) and three NFAs (Y6, IDIC, IEICO-4F).

Device Characterization
The power conversion efficiencies of the devices were measured by current densityvoltage (J-V) curves using a Keithley 2401 (Keithley instruments, Cleveland, OH, USA) source measurement unit under AM 1.5 G (100 mW/cm 2 ) illumination from a solar simulator (McScience, Suwon, Republic of Korea).The simulated light intensity was calibrated using a standard Si solar cell.External quantum efficiency (EQE) was measured using a solar cell spectral response/QE/IPCE (IQE-200B, Newport Co., Irvine, CA, USA).The light intensity at each wavelength was calibrated using a standard Si solar cell.
The absorption spectra of each BHJ layer were measured by using a UV-Vis spectrophotometer (Cary 5000, Agilent Technologies Inc., Santa Clara, CA, USA).Electroluminescence (EL) signals passed through the integral sphere and were recorded by using a high-sensitivity spectrophotometer (MAYA 2000 Pro, Ocean Insight Inc., Orlando, FL, USA).The wavelength of the PL excitation source was 632 nm.The integral sphere and spectrophotometer were calibrated to collect the radiated photon flux from the device by using a standard halogen calibration light source (HL-3-plus-INT-CAL, Ocean Insight Inc., FL, USA).
Fourier transform photocurrent spectroscopy (FTPS) was used to perform measurements in an FTPS setup built in-house, which consisted of a Fourier-transform infrared spectrometer (INVENIO-R, Bruker Co., Billerica, MA, USA) equipped with quartz beam splitter.The photocurrent produced by the solar cell under illumination was amplified using a low-noise preamplifier (SR570, Stanford Research Systems, Sunnyvale, CA, USA) and fed back into the external detector port of the FTIR.
The steady-state photoluminescence (PL) spectra were obtained using a multichannel spectrophotometer (QEPro, Ocean Insight Inc., FL, USA).In PL measurements, acceptors and BHJs were excited from active layer side using a continuous wave diode laser (Nd-YAG tuned at 532 nm).To calculate the hole-transfer yields, PL spectra were divided by the absorption intensity at the 532 nm.Femtosecond transient absorption (TA) spectra were obtained using a homemade TA measurement system comprising a femtosecond Ti:sapphire regenerative amplifier system (Hurricane, Spectra-Physics Inc., Milpitas, CA, USA) with a 1 kHz repetition rate, an optical parametric amplifier (OPA-800CF, Spectra-Physics Inc., CA, USA), and multichannel spectrometers (QEPro and NIRQUEST, Ocean Insight Inc., FL, USA).The pump pulse wavelength was 630 nm with a power density of ~1 µJ/cm 2 .

Background of h E CT Determination
We first analyzed the BHJ systems based on the representative non-fullerene acceptor Y6 (Figure 2a) with three donor polymers: PM6, PBDTTPD-HT, and PTB7-Th (Figure S2). Figure 2b shows the current density (J)-voltage (V) characteristics of OPVs.The parameters related to the performance of these OPVs are listed in Table S2.The acceptor Y6 has a lower bandgap and higher absorption coefficient than those of the three donors we used, as shown in Figure 2c.Thus, in the low-energy onset part of EQE spectra, there is no photocurrent contribution from the donor.All the contributions originated from Y6 only (Figure S4), which means that all photoexcited electrons are created in the acceptor and then extracted to the cathode directly.In the case of holes, they would be transferred to the donor through the h E CT level and then extracted to the anode electrode (Figure 2d).Therefore, as similar to the FTPS-EQE onset of the f-BHJ system in which electrons generated from the donor are transferred to the acceptor has the information e E CT , the FTPS-EQE onset of nf-BHJ system will have information on h E CT .In the case of the f-BHJ system, since the e E CT bands are quite visible in the low-energy part of the FTPS-EQE spectrum, after determining the reorganization energy (λ) by fitting the e E CT signal of the EQE, the fitting can be applied to the EL spectrum (Figure S1).However, in the case of the nf-BHJ systems, since there is no CT band feature in FTPS-EQE onset, we designed a way to secure the information of the h E CT band from the EL spectrum and reflect it to FTPS-EQE.
Under the EL measurement mode, the electrons are directly injected into the LUMO of the acceptor, while the holes are injected into the HOMO of the donor and transferred to the acceptor through h E CT .Thus, the hole injection can be hindered because the HOMO offset acts as a potential barrier.In EL B , of course, the EL emission from the singlet of the acceptor will be dominant.However, since there is the possibility of radiative recombination between electrons at LUMO of acceptor and holes at the h E CT , both radiative charge recombination from singlet excitons and h E CT states are superimposed in the EL B spectrum.Our first goal is to extract the contribution of the h E CT in EL B to apply the Marcus theory from EL to FTPS-EQE.To achieve this, the exact position of h E CT emission should be addressed.To estimate the underlying contribution from h E CT , we have simply subtracted the acceptor EL (EL A ) spectrum from EL B . Figure 2e-g  Under the EL measurement mode, the electrons are directly injected into the LUMO of the acceptor, while the holes are injected into the HOMO of the donor and transferred to the acceptor through h ECT.Thus, the hole injection can be hindered because the HOMO offset acts as a potential barrier.In ELB, of course, the EL emission from the singlet of th acceptor will be dominant.However, since there is the possibility of radiative recombina tion between electrons at LUMO of acceptor and holes at the h ECT, both radiative charg recombination from singlet excitons and h ECT states are superimposed in the ELB spectrum Our first goal is to extract the contribution of the h ECT in ELB to apply the Marcus theor from EL to FTPS-EQE.To achieve this, the exact position of h ECT emission should be ad dressed.To estimate the underlying contribution from h ECT, we have simply subtracted the acceptor EL (ELA) spectrum from ELB. Figure 2e-g shows the difference in the EL spec trum obtained from PM6:Y6, PBDTTPD-HT:Y6, and PTB7-Th:Y6, respectively.The differ ence in the EL spectrum (shaded blue area) represents the EL contribution from h ECT.

EL Deconvolution
To determine the exact peak position, Gaussian fittings for ELB were performed based on the information about the components obtained from ELB-ELA.In the case of PM6:Y ELB (Figure 2h), the spectrum was deconvoluted into three components.The peak in th center (1.372 eV) is obviously the singlet radiative recombination between the LUMO and HOMO of the acceptor.The peaks on both the low-and high-energy side are caused b the effective h ECT band.It is likely that h ECT does not exist at a finite level.It has a broad

EL Deconvolution
To determine the exact peak position, Gaussian fittings for EL B were performed based on the information about the components obtained from EL B -EL A .In the case of PM6:Y6 EL B (Figure 2h), the spectrum was deconvoluted into three components.The peak in the center (1.372 eV) is obviously the singlet radiative recombination between the LUMO and HOMO of the acceptor.The peaks on both the low-and high-energy side are caused by the effective h E CT band.It is likely that h E CT does not exist at a finite level.It has a broad distribution near the acceptor's HOMO.The h E CT levels near the acceptor's HOMO are like a degenerate state, so it will be indistinguishable.Therefore, the h E CT band can be divided into three regions (Figure 2d).What we are interested in is the energy range that is lower than the acceptor's HOMO (higher effective h E CT ), which affects energy loss.The h E CT levels in the energy region that is higher than the acceptor's HOMO (lower effective h E CT ) will be excluded from the analysis because this part has no effect on the energy loss of the solar cells.Moreover, there is some overlap with the tail of the donor's EL (EL D ).The results of the same Gaussian fittings for PBDTTPD-HT and PTB7-Th are shown in Figure 2i,j, respectively.In the case of the PBDTTPD-HT:Y6 blend, EL spectrum features are similar to the PM6:Y6 blend due to the similar energy band configurations.However, in the case of the PTB7-Th:Y6 system, it was determined that the contribution of singlet excitons was lower than that of h E CT .Simply put, this means that the energy loss caused by h E CT is greater in the PTB7-Th:Y6 blend.
, where E g PV is the photovoltaic gap determined from the derivatives of EQE PV [25].Based on the h E CT position obtained through EL B deconvolution, it required as much reorganization energy (λ) to match the onset part of FTPS-EQE (Figure 3).The ∆ h E CT values obtained in this analysis were 0.030 eV for PM6:Y6, 0.058 eV for PBDTTPD-HT:Y6, and 0.044 eV for PTB7-Th:Y6.In many previous studies, E CT determination was performed using the maximum peak of EL B [26][27][28].However, as confirmed in the previous discussion, the maximum peak region of the EL B contains a lot of the singlet contribution of the acceptor.Although a Gaussian fit can be matched to the FTPS-EQE spectrum, even if the CT state is analyzed based on the maximum of EL B , the extracted CT state value was nearly the same as or even larger than the bandgap of the Y6 acceptor (see Figure S5 and Table S3).Furthermore, many researchers simply considered E CT to be the same as qV OC in the Shockley-Queisser (SQ) limit (qV OC SQ ).However, these two values have different theoretical backgrounds [29][30][31][32].In our study, we compared the qV OC SQ values obtained through energy loss analysis with h E CT (Figure S6).The results clearly showed that h E CT and qV OC SQ have different values (Table S4).The results of the total energy loss analysis indicate that PM6:Y6 forms a large V OC compared to PBDTTPD-HT:Y6, which has similar energy band configuration due to the low ∆ h E CT , radiative energy loss (∆E 2 = qV OC SQ − qV OC rad , where qV OC rad is qV OC in the radiative limit), and non-radiative energy loss (∆E 3 = qV OC rad − qV OC nonrad , where qV OC nonrad is qV OC in the non-radiative limit; details are presented in Appendix B).In the case of PTB7-Th:Y6, although hole transfer is slightly better than the PBDTTPD-HT:Y6 due to the lower ∆ h E CT , PTB7-Th:Y6 has a large V OC loss due to relatively large values of ∆E 2 and ∆E 3 .
the PTB7-Th:Y6 system, it was determined that the contribution of singlet excitons was lower than that of h ECT.Simply put, this means that the energy loss caused by h ECT is greater in the PTB7-Th:Y6 blend.

Marcus Theory and Energy Loss Model
Now, we can apply the Marcus theory (Appendix A) to determine the exact Δ h ECT (=Eg PV − h ECT), where Eg PV is the photovoltaic gap determined from the derivatives of EQEPV [25].Based on the h ECT position obtained through ELB deconvolution, it required as much reorganization energy (λ) to match the onset part of FTPS-EQE (Figure 3).The Δ h ECT values obtained in this analysis were 0.030 eV for PM6:Y6, 0.058 eV for PBDTTPD-HT:Y6, and 0.044 eV for PTB7-Th:Y6.In many previous studies, ECT determination was performed using the maximum peak of ELB [26][27][28].However, as confirmed in the previous discussion, the maximum peak region of the ELB contains a lot of the singlet contribution of the acceptor.Although a Gaussian fit can be matched to the FTPS-EQE spectrum, even if the CT state is analyzed based on the maximum of ELB, the extracted CT state value was nearly the same as or even larger than the bandgap of the Y6 acceptor (see Figure S5 and Table S3).Furthermore, many researchers simply considered ECT to be the same as qVOC in the Shockley-Queisser (SQ) limit (qVOC SQ ).However, these two values have different theoretical backgrounds [29][30][31][32].In our study, we compared the qVOC SQ values obtained through energy loss analysis with h ECT (Figure S6).The results clearly showed that h ECT and qVOC SQ have different values (Table S4).The results of the total energy loss analysis indicate that PM6:Y6 forms a large VOC compared to PBDTTPD-HT:Y6, which has similar energy band configuration due to the low Δ h ECT, radiative energy loss (ΔE2 = qVOC SQ −qVOC rad , where qVOC rad is qVOC in the radiative limit), and non-radiative energy loss (ΔE3 = qVOC rad −qVOC non- rad , where qVOC nonrad is qVOC in the non-radiative limit; details are presented in Appendix B).In the case of PTB7-Th:Y6, although hole transfer is slightly better than the PBDTTPD-HT:Y6 due to the lower Δ h ECT, PTB7-Th:Y6 has a large VOC loss due to relatively large values of ΔE2 and ΔE3.In general, two methods are mainly utilized to determine ΔE3.When qVOC SQ and qVOC rad are derived theoretically based on measured JSC and FTPS-EQE, ΔE3 is determined by simply subtracting the qVOC rad from the VOC of the solar cells.Alternatively, ΔE3 can also be calculated using the theoretical approach given by ΔE3 = −kBTln(EQEEL), where kB is the Boltzmann constant, T is temperature, and EQEEL is the external radiative efficiency.Considering that the ELB of the nf-BHJ system consists of EL from the acceptor's single In general, two methods are mainly utilized to determine ∆E 3 .When qV OC SQ and qV OC rad are derived theoretically based on measured J SC and FTPS-EQE, ∆E 3 is determined by simply subtracting the qV OC rad from the V OC of the solar cells.Alternatively, ∆E 3 can also be calculated using the theoretical approach given by ∆E 3 = −k B Tln(EQE EL ), where k B is the Boltzmann constant, T is temperature, and EQE EL is the external radiative efficiency.Considering that the EL B of the nf-BHJ system consists of EL from the acceptor's single recombination and EL from h E CT , once EQE EL obtained based on EL B is used, both the contribution of the acceptor and the contribution of h E CT are already included in the ∆E 3 calculation.The fact that the non-radiative voltage loss can be further split into these two contributions was already recognized and reported in previous studies [20].However, the problem is to accurately extract the contribution of the acceptor from only EL B .It is expected to be very difficult to extract only the contribution of acceptors in a general situation where EL A and h E CT bands overlap each other.In particular, since there is also a degeneracy part between the h E CT band and acceptor HOMO, it is currently impossible to completely separate the two components.Therefore, to obtain the exact ∆ h E CT , the application of Marcus theory through our EL deconvolution is the most accurate method so far.The effect of h E CT on energy loss is already included in recombination loss (∆E 3 ) if EL B is used to extract EQE EL .On the other hand, the effect of e E CT on energy loss is mainly reflected in ∆E 1 .∆E 1 should be selected with the larger of e E CT and qV OC SQ (see Figure 1a).Of course, since there is a possibility of recombination between an electron at the e E CT level and a hole at the HOMO of the donor, the effect of e E CT is also included in ∆E 3 if EL B is used to extract EQE EL .

IDIC-Acceptor-Based nf-BHJ Systems (Larger HOMO Offset)
Next, we applied the EL deconvolution method to another NFA system, a BHJ blend with IDIC.The chemical structure and J-V characteristics of the IDIC system are shown in Figure 4a,b.The HOMO level of IDIC is similar to that of Y6, but it has a larger bandgap.Therefore, the absorption of IDIC largely overlaps with the donor polymers (Figure 4c).In particular, the absorption of IDIC completely overlaps with PTB7-Th absorption.These characteristics are directly reflected in the composition of EQE (Figure S7).For PM6:IDIC and PBDTTPD-HT:IDIC, though, the onset of EQE PV is based on the photo-carriers created by the absorption of the acceptors.In the case of the PTB7-Th:IDIC, both donor and acceptor appear to be involved.Figure 4d-f shows the difference between EL B and EL A .The peculiar point is that the contribution of the effective h E CT was symmetric with respect to the HOMO of the acceptor in the case of Y6, whereas the IDIC system has an asymmetric shape where the lower-energy side is dominant.Thus, the effective h E CT region is formed only on the upper side of the CT-band.In the case of PM6:IDIC and PBDTTPD-HT:IDIC, the maximum peak of EL B was the same as that of EL A , but in the case of PTB7-Th:IDIC, the maximum part of EL B was completely different from that of EL A .It can be seen that the CT-band contribution is much larger in PTB7-Th:IDIC only by a simple comparison of these EL deconvolutions.With this simple EL deconvolution analysis, it can be simply confirmed that the h E CT band contribution in PTB7-Th:IDIC is much larger compared to PM6:IDIC and PBDTTPD-HT:IDIC.Exact ∆ h E CT obtained through accurate Gaussian fitting (Figure 4g-i) and Marcus theory (Figure 4j-l) for PM6:IDIC, PBDTTPD-HT:IDIC are shown in Table 1.The overall energy loss analysis results for the IDIC system are summarized in Figure S8 and Table S5.Note that it is impossible to distinguish whether it is ∆ h E CT or ∆ e E CT only with the current method in the case of PTB7-Th:IDIC blend.Of course, this discernment is also not possible with the conventional method (Gaussian fitting of FTPS-EQE and then matching to the maximum of EL B ).The conventional method also gave unacceptable values (Figure S9 and Table S6).

IEICO-4F-Acceptor-Based BHJ Systems (Smaller HOMO Offset)
What if the acceptor's HOMO is lower than the donor's HOMO while having a low bandgap?We selected IEICO-4F as a material that satisfies these conditions (Supplementary Figure S3).The chemical structures and J-V characteristics of the IEICO-4F system are shown in Figure 5a,b.Since IEICO-4F has a lower bandgap than Y6, there is no overlap with that of absorption of the donor like Y6 (Figure 5c), and the EL B and EL D are also Polymers 2023, 15, 4042 8 of 14 separated.In the IEICO-4F system, the onset of EQE is formed by only the contribution from the acceptor.Figure 5d-f shows the difference between EL B and EL A .In the case of IEICO-4F, unlike IDIC, the difference between EL B and EL A does not appear at low energy but is formed in a higher-energy region (Figure 5d-f).The low-energy part of EL B exactly overlaps with the EL A .Thus, the effective h E CT region is formed only on the lower side of the CT band.The EL B is perfectly simulated with the EL A spectrum and one additional Gaussian (Figure 5g-i).The absence of the low-energy CT-band leads to a complete overlap of EQE fitting and EL fitting in Marcus theory analysis for PM6:IEICO-4F and PBDTTPD-HT:IEICO-4F, as shown in Figure 5j,k.The lower effective h E CT than the HOMO of the acceptor indicates an insufficient HOMO offset for charge separation.The low J SC in PM6:IEICO-4F and PBDTTPD-HT:IEICO-4F are attributed to this charge-separation problem.Only PTB7-Th:IEICO-4F has a proper condition for h E CT formation.As a result, PTB7-Th:IEICO-4F shows the best photovoltaic performance.In fact, the values of J SC are quite predictable by simple comparison of the lower effective h E CT size.For PM6:IEICO-4F, which has a relatively large size of lower effective h E CT , the smallest J SC was shown, and in PTB7-Th:IEICO-4F, which has a relatively small size of lower effective h E CT , the largest J SC was formed.Sufficient hole transfer seems to be made through the degeneracy region.It is possible that in nf-BHJ, a certain level of hole offset may not be a prerequisite.Lastly, the overall energy loss composition of PTB7-Th:IEICO-4F was similar to that of PM6:Y6.Nevertheless, the lower photovoltaic performance of PTB7-Th:IEICO-4F relative to that of PM6:Y6 is attributed to the fact that the absorption coefficient of IEICO-4F itself is lower than that of Y6.

IEICO-4F-Acceptor-Based BHJ Systems (Smaller HOMO Offset)
What if the acceptor's HOMO is lower than the donor's HOMO while having a low bandgap?We selected IEICO-4F as a material that satisfies these conditions (Supplemen tary Figure S3).The chemical structures and J-V characteristics of the IEICO-4F system ar shown in Figure 5a,b.Since IEICO-4F has a lower bandgap than Y6, there is no overla with that of absorption of the donor like Y6 (Figure 5c), and the ELB and ELD are als  is possible that in nf-BHJ, a certain level of hole offset may not be a prerequisite.Lastly the overall energy loss composition of PTB7-Th:IEICO-4F was similar to that of PM6:Y6 Nevertheless, the lower photovoltaic performance of PTB7-Th:IEICO-4F relative to that o PM6:Y6 is attributed to the fact that the absorption coefficient of IEICO-4F itself is lowe than that of Y6.

Hole Decay and Transfer Time at h ECT
The effect of the configuration change of the hole-transfer state on the hole-transfe time was investigated through transient absorption (TA) measurement.Figure 6 show the decay profiling of the hole-transfer time obtained from the TA measurements for ac ceptors and BHJ films used in this study.Detailed values of the hole-transfer time and PL quenching efficiency are listed in Table 2.In the case of the BHJ system with a Y6 acceptor the measured hole-transfer time was much faster than the TA decay time of the Y6 film (~211 ps).The hole-transfer time was the fastest in PM6:Y6 (~2.8 ps), became slower in PBDTTPD-HT:Y6 (~5.1 ps), and showed the lowest value in PTB7-Th:Y6 (~7.0 ps).Thi order of hole-transfer time is well correlated with the order of the h ECT area size obtained in EL deconvolution (see Figure 2h-j).The fast hole-transfer property delivers a lowe recombination rate.Thus, the PL-quenching efficiency was the highest in PM6:Y6 (93%) and it was the lowest in PTB7-Th:Y6 (78%), which has the slowest hole-transfer rate (Fig ure S18).In the BHJ system with the IDIC acceptor (Figure 6e-h), the PBDTTPD-HT:IDIC exhibited the fastest hole-transfer time of ~1.8 ps, even faster than PM6:Y6.This is proba bly because the upper h ECT region acting as an obstruction region of hole transfer i smaller.However, the overall PCE was lower than that of Y6-based devices.This is be cause of the decreased JSC due to the large bandgap characteristics of IDIC.The longes hole-transfer time and the lowest PL-quenching efficiency were observed in PTB7 Th:IDIC, with the largest upper h ECT region.For the BHJ systems with IEICO-4F, a pro longed hole-transfer time of ~13.0 ps was observed for PM6:IEICO-4F, and even the hole transfer time of pristine IEICO-4F was remarkably faster (~71 ps) than other acceptor

Hole Decay and Transfer Time at h E CT
The effect of the configuration change of the hole-transfer state on the hole-transfer time was investigated through transient absorption (TA) measurement.Figure 6 shows the decay profiling of the hole-transfer time obtained from the TA measurements for acceptors and BHJ films used in this study.Detailed values of the hole-transfer time and PL-quenching efficiency are listed in Table 2.In the case of the BHJ system with a Y6 acceptor, the measured hole-transfer time was much faster than the TA decay time of the Y6 film (~211 ps).The hole-transfer time was the fastest in PM6:Y6 (~2.8 ps), became slower in PBDTTPD-HT:Y6 (~5.1 ps), and showed the lowest value in PTB7-Th:Y6 (~7.0 ps).This order of hole-transfer time is well correlated with the order of the h E CT area size obtained in EL deconvolution (see Figure 2h-j).The fast hole-transfer property delivers a lower recombination rate.Thus, the PL-quenching efficiency was the highest in PM6:Y6 (93%), and it was the lowest in PTB7-Th:Y6 (78%), which has the slowest hole-transfer rate (Figure S18).In the BHJ system with the IDIC acceptor (Figure 6e-h), the PBDTTPD-HT:IDIC exhibited the fastest hole-transfer time of ~1.8 ps, even faster than PM6:Y6.This is probably because the upper h E CT region acting as an obstruction region of hole transfer is smaller.However, the overall PCE was lower than that of Y6-based devices.This is because of the decreased J SC due to the large bandgap characteristics of IDIC.The longest hole-transfer time and the lowest PL-quenching efficiency were observed in PTB7-Th:IDIC, with the largest upper h E CT region.For the BHJ systems with IEICO-4F, a prolonged holetransfer time of ~13.0 ps was observed for PM6:IEICO-4F, and even the hole-transfer time of pristine IEICO-4F was remarkably faster (~71 ps) than other acceptors (Figure 6i).This is because of the relatively large upper h E CT and the small degeneracy region, as predicted by the EL deconvolution result.Since most of the hole-transfer levels were created on the upper h E CT region that causes hole-transfer retardation, the hole-transfer characteristics were significantly degraded compared to other acceptor systems.However, PTB7-Th: IEICO-4F, with a relatively small upper h E CT region, showed a fast hole-transfer time and high hole-quenching efficiency.Thereby, PTB7-Th: IEICO-4F exhibited relatively higher performance within BHJ systems based on the IEICO-4F acceptor.
olymers 2023, 15, x FOR PEER REVIEW 10 of 1 (Figure 6i).This is because of the relatively large upper h ECT and the small degeneracy region, as predicted by the EL deconvolution result.Since most of the hole-transfer level were created on the upper h ECT region that causes hole-transfer retardation, the hole-trans fer characteristics were significantly degraded compared to other acceptor systems.How ever, PTB7-Th: IEICO-4F, with a relatively small upper h ECT region, showed a fast hole transfer time and high hole-quenching efficiency.Thereby, PTB7-Th: IEICO-4F exhibited relatively higher performance within BHJ systems based on the IEICO-4F acceptor.

Conclusions
In summary, we have presented a more accurate analysis method that can obtain th exact value of Δ h ECT in NFA-based OPV by extracting information on h ECT through EL deconvolution.We have applied our ELB deconvolution method on nine BHJ systems con sisting of a combination of three donor polymers (PM6, PBDTTPD-HT, PTB7-Th) and three NFAs (Y6, IDIC, IEICO-4F).To extract the exact Δ h ECT, the following conditions mus be satisfied to apply our ELB deconvolution method.(i) Absorption of the donor and ac ceptor must be separated sufficiently.Accordingly, (ii) the composition of EQE should b clearly divided into donor and acceptor contribution parts.In particular, the initial EQEP

Conclusions
In summary, we have presented a more accurate analysis method that can obtain the exact value of ∆ h E CT in NFA-based OPV by extracting information on h E CT through EL B deconvolution.We have applied our EL B deconvolution method on nine BHJ systems consisting of a combination of three donor polymers (PM6, PBDTTPD-HT, PTB7-Th) and three NFAs (Y6, IDIC, IEICO-4F).To extract the exact ∆ h E CT , the following conditions must be satisfied to apply our EL B deconvolution method.(i) Absorption of the donor and acceptor must be separated sufficiently.Accordingly, (ii) the composition of EQE should be clearly divided into donor and acceptor contribution parts.In particular, the initial EQE PV onset of an OPV with NFA should be formed only by the contribution of the acceptor.If these conditions are satisfied, through EL B deconvolution, it is possible to specify exactly

Figure 1 .
Figure 1.Difference between fullerene-based OSCs and non-fullerene-based OSCs near EQE onset (flat-band VOC condition).(a) Electrons are initially excited at donor polymer and then transferred to acceptor in f-BHJ OSCs.ECT level is visible in sensitive FTPS-EQE measurement.(b) Electrons are initially excited at acceptor near EQE onset.Holes are transferred to donor in nf-BHJ OSCs.ECT level is invisible even in sensitive FTPS-EQE measurement.

1 Figure 2 .
Figure 2. EL deconvolution analysis of the BHJ systems based on the Y6 acceptor.(a) Chemica structure of Y6 acceptor.(b) J-V characteristics of BHJ OPVs fabricated using Y6 acceptor.(c) Ab sorption coefficient spectra of polymer donors and Y6.(d) The schematic diagram of hole transfe in PV mode and charge recombination in EL mode.In EL mode, radiative recombination can occu in three ways: (i) recombination with acceptor singlet (mono-molecular recombination), (ii) recom bination with holes at upper effective h ECT, or (iii) recombination with holes at lower effective h EC (e-g) EL spectra of BHJ blend, donor only, and acceptor only; shaded blue area (ELB-ELA) represent the EL contribution by h ECT.(h-j) Simulation of EL spectrum using Gaussian function and ELA.

Figure 2 .
Figure 2. EL deconvolution analysis of the BHJ systems based on the Y6 acceptor.(a) Chemical structure of Y6 acceptor.(b) J-V characteristics of BHJ OPVs fabricated using Y6 acceptor.(c) Absorption coefficient spectra of polymer donors and Y6.(d) The schematic diagram of hole transfer in PV mode and charge recombination in EL mode.In EL mode, radiative recombination can occur in three ways: (i) recombination with acceptor singlet (mono-molecular recombination), (ii) recombination with holes at upper effective h E CT , or (iii) recombination with holes at lower effective h E CT .(e-g) EL spectra of BHJ blend, donor only, and acceptor only; shaded blue area (EL B -EL A ) represents the EL contribution by h E CT .(h-j) Simulation of EL spectrum using Gaussian function and EL A .

3 .
Marcus Theory and Energy Loss ModelNow, we can apply the Marcus theory (Appendix A) to determine the exact ∆

Figure 3 .
Figure 3. Δ h ECT was determined based on the Marcus theory for (a) PM6:Y6, (b) PBDTTPD-HT:Y6, and (c) PTB7-Th:Y6 BHJs.The position of h ECT emission was obtained from ELB deconvolution and then fit to FTPS-EQE onset part.The CT absorption (blue dotted line) and CT emission (red dotted line) were fitted using Gaussian function.

Figure 3 .
Figure 3. ∆ h E CT was determined based on the Marcus theory for (a) PM6:Y6, (b) PBDTTPD-HT:Y6, and (c) PTB7-Th:Y6 BHJs.The position of h E CT emission was obtained from EL B deconvolution and then fit to FTPS-EQE onset part.The CT absorption (blue dotted line) and CT emission (red dotted line) were fitted using Gaussian function.

Figure 4 .
Figure 4. EL deconvolution analysis of the BHJ systems based on the IDIC acceptor.(a) Chemica structure of IDIC acceptor.(b) J-V characteristics of OPVs fabricated using IDIC acceptor.(c) Ab sorption coefficient spectra of PM6, PBDTTPD-HT, PTB7-Th, and IDIC.(d-f) EL spectrum of blend donor, and acceptor; the difference in EL spectrum (shaded blue area) represents the EL contributio by h ECT.(g-i) Simulation of EL spectrum using Gaussian function and ELA.(j-l) Δ h ECT determinatio based on the Marcus theory.The CT absorption (blue dotted line) and CT emission (red dotted line were fitted using Gaussian function.

Figure 4 .
Figure 4. EL deconvolution analysis of the BHJ systems based on the IDIC acceptor.(a) Chemical structure of IDIC acceptor.(b) J-V characteristics of OPVs fabricated using IDIC acceptor.(c) Absorption coefficient spectra of PM6, PBDTTPD-HT, PTB7-Th, and IDIC.(d-f) EL spectrum of blend, donor, and acceptor; the difference in EL spectrum (shaded blue area) represents the EL contribution by h E CT .(g-i) Simulation of EL spectrum using Gaussian function and EL A .(j-l) ∆ h E CT determination based on the Marcus theory.The CT absorption (blue dotted line) and CT emission (red dotted line) were fitted using Gaussian function.

Figure 5 .
Figure 5. EL deconvolution analysis of the BHJ systems based on the IEICO-4F acceptor.(a) Chem ical structure of IEICO-4F acceptor.(b) J-V characteristics of BHJ OPVs fabricated with IEICO-4 acceptor.(c) Absorption coefficient spectra of PM6, PBDTTPD-HT, PTB7-Th, and IEICO-4F.(d-f) E spectrum of blend, donor, acceptor, and the difference in EL spectrum.(g-i) Simulation of EL spec trum using Gaussian function and ELA.One Gaussian fitting was obtained from ELB-ELA.(j-l) Δ h EC determination based on the Marcus theory.The CT absorption (blue dotted line) and CT emissio (red dotted line) were fitted using Gaussian function, but indistinguishable from singlet band.

Figure 5 .
Figure 5. EL deconvolution analysis of the BHJ systems based on the IEICO-4F acceptor.(a) Chemical structure of IEICO-4F acceptor.(b) J-V characteristics of BHJ OPVs fabricated with IEICO-4F acceptor.(c) Absorption coefficient spectra of PM6, PBDTTPD-HT, PTB7-Th, and IEICO-4F.(d-f) EL spectrum of blend, donor, acceptor, and the difference in EL spectrum.(g-i) Simulation of EL spectrum using Gaussian function and EL A .One Gaussian fitting was obtained from EL B -EL A .(j-l) ∆ h E CT determination based on the Marcus theory.The CT absorption (blue dotted line) and CT emission (red dotted line) were fitted using Gaussian function, but indistinguishable from singlet band.

E g PV qV OC Recombination Loss ΔqV OC rad + ΔqV OC nonrad Ground State Energy (eV) no E CT feature qV OC SQ E (eV) E (eV) Donor Acceptor h E CT h e E g PV qV OC Recombination Loss ΔqV OC rad + ΔqV OC nonrad Ground State Energy (eV) e E CT e -transfer h + transfer qV OC SQ
Figure 1.Difference between fullerene-based OSCs and non-fullerene-based OSCs near EQE onset (flat-band V OC condition).(a) Electrons are initially excited at donor polymer and then transferred to acceptor in f-BHJ OSCs.E CT level is visible in sensitive FTPS-EQE measurement.(b) Electrons are initially excited at acceptor near EQE onset.Holes are transferred to donor in nf-BHJ OSCs.E CT level is invisible even in sensitive FTPS-EQE measurement.

Table 1 .
Extracted h E CT values of each BHJ system obtained using EL B deconvolution method.

Table 2 .
Hole-transfer time and PL-quenching efficiency of BHJ used in this study.

Table 2 .
Hole-transfer time and PL-quenching efficiency of BHJ used in this study.