Solvent Engineering for Layer Formation Control with Cost-Effective Hole Transport Layer in High-Efficiency Perovskite Solar Cell

Abstract

Among hole transport materials (HTMs), 2,2′,7,7′-Tetrakis(N,N-di-p-methoxyphenylamino)-9,9′-spirobifluorene (spiro-OMeTAD) is the most frequently adopted, due to its suitable energy band level in conventional-type perovskite solar cells (PSCs). However, the high price of spiro-OMeTAD is an obstacle faced in its research and commercialization. In our previous work, we introduced a low-cost HTM, (E,E,E,E)-4,4′,4″,4‴-[Benzene-1,2,4,5-tetrayltetrakis(ethene-2,1-diyl)]tetrakis[N,N-bis(4-methoxyphenyl)aniline] (α2); however, it was immiscible in the conventional solvent chlorobenzene, leading to the adoption of dichloromethane (DCM) as an alternative. Nevertheless, its high vapor pressure led to poor reproducibility, limiting its practical applicability. To address this issue, we investigated alternative solvents to DCM to facilitate the application of α2 to dichloride alkane materials, from 1,2-dichloroethane (DCE) to 1,4-dichlorobutane. In these materials, DCE exhibits the most superior properties in terms of layer formation control, due to its vapor pressure in spin-coating. Accordingly, a PSC containing α2-DCE HTL showed high performance, with 1.15V of open-circuit voltage and a 22.7% power conversion efficiency.

1. Introduction

Perovskite solar cells (PSCs) are emerging as next-generation photovoltaics, due to these cells having a tunable band gap and a simplified fabrication process compared to silicon-based solar cells, while also achieving a high power conversion efficiency (PCE) of 27% [1,2,3,4,5,6]. However, the substantial cost associated with hole transport materials (HTMs) remains a critical barrier to both the commercialization and continued development of PSCs. Among various HTMs, 2,2′,7,7′-tetrakis[N,N-di(4-methoxyphenyl)amino]-9,9′-spirobifluorene (spiro-OMeTAD) is the most extensively utilized, due to its well-aligned energy levels with perovskite absorbers. Nevertheless, its widespread implementation is hindered by its high cost, primarily attributed to its complex synthetic pathway. Consequently, significant research efforts have been directed toward the development of cost-effective alternatives [7,8,9,10,11]. One promising candidate is (E,E,E,E)-4,4′,4″,4‴-[benzene-1,2,4,5-tetrayltetrakis(ethene-2,1-diyl)]tetrakis[N,N-bis(4-methoxyphenyl)aniline] (α2), which exhibits excellent potential for achieving high-performance PSCs. In our previous investigations of α2, it was discovered that dichloromethane (DCM) serves as a viable alternative to chlorobenzene (CB), the conventional solvent for HTMs, as CB fails to dissolve α2 at room temperature [12]. The α2-DCM system demonstrated superior hole mobility, enabling high-efficiency PSCs with a recorded PCE of 20.18%. However, achieving consistent deposition of the α2-DCM-based hole transport layer (HTL) remains a challenge due to the high vapor pressure of DCM. Its excessively volatile nature disrupts the reproducibility of spin-coated films by inducing variations in solution concentration through rapid evaporation or solvent dripping [13,14,15,16,17]. Therefore, identifying alternative solvents is imperative to address this issue. From a chemical standpoint, the carbon chain length of organic solvents significantly influences their physical properties, including their boiling point and vapor pressure [18,19,20,21,22,23]. To retain chemical compatibility with α2 while addressing the challenges posed by DCM, we systematically explored dichlorinated alkane solvents (DASs) of varying molecular sizes as alternatives to DCM, including 1,2-dichloroethane (DCE), 1,3-dichloropropane (DCP), and 1,4-dichlorobutane (DCB). Given that the α2 layers formed with each DAS exhibited distinct morphological and electronic properties, we conducted optical characterization, atomic force microscopy (AFM), and scanning electron microscopy (SEM) analyses to assess their impact on film quality. Furthermore, photoluminescence (PL) measurements revealed that carrier mobility varies depending on the solvent employed, thereby influencing overall device performance. As a result, this study demonstrates that DCE exhibits superior characteristics as a solvent for α2-based HTLs, offering advantages in solubility, film uniformity, cost-effectiveness, and device performance over the other investigated DASs. The FAPb (I_0.99Br_0.01)₃-based PSC incorporating an α2-DCE layer achieved the highest PCE of 22.7% and an open-circuit voltage (V_OC) of 1.15 V.

2. Materials and Methods

2.1. Materials

α2 (>98.0%, HPLC), CB (98.0%), DCM (99.5%), GC (stabilized with 2-methyl-2-butene), DCP (>98%), DCB (>98%, GC), N,N-dimethylformamide (DMF, >99.5%, GC), N-methylpyrrolidone (NMP, >99.0%, GC, low water content), lead(II) iodide (PbI₂, 99.99%, metals basis, for perovskite precursor), and lithium bis(trifluoromethanesulfonyl)imide (Li-TFSI, >99.0%) were procured from TCI Chemicals (Tokyo, Japan). Formamidine hydroiodide (FAI, >99.5%) and methylamine hydrochloride (MACl, >99.5%) were supplied by Lumtec (New Taipei City, Taiwan). Acetonitrile (ACN, anhydrous, 99.8%), DCE (99.8%, anhydrous), isopropyl alcohol (IPA, 99.5%, anhydrous), lead (II) bromide (PbBr₂, >98%), and 4-tert-butylpyridine (tBP, 98.0%) were obtained from Sigma-Aldrich (St. Louis, MI, USA). N-octylammonium iodide (OAI, >99%) was purchased from Greatcell Solar (Queanbeyan, Australia). Tin (IV) oxide (SnO₂, 15% in H₂O colloidal dispersion) was sourced from Alfa (Ward Hill, MA, USA). All chemicals were utilized as received, without additional purification.

2.2. Preparation of Precursors

The SnO₂ precursor solution was prepared by diluting a commercial SnO₂ colloidal dispersion in deionized (DI) water at a 1:1 v/v% ratio. To prepare 1 mL of 1.35 M FAPb (I_0.99Br_0.01)₃ perovskite precursor solution, 232.2 mg of FAI, 22.3 mg of PbBr₂, 594.35 mg of PbI₂ (5 mol% excess), and 22.8 mg of MACl (25 mol%) were dissolved in a DMF/NMP co-solvent mixture (9:1 v/v%). For the preparation of 1 mL of α2 solution, 40 mg of α2, 19 μL of Li-TFSI/ACN stock solution (340 mg/mL), and 19 μL of tBP were dissolved in 1 mL of DCM, DCE, DCP, or DCB, respectively. The α2 solutions, except for the one with DCM, were heated on a hot plate at 50, 65, and 105 °C until completely dissolved. OAI was dissolved in IPA at a concentration of 10 mM.

2.3. Device Fabrication

ITO glass substrates (2 × 1.5 cm²) were sequentially cleaned using an ultrasonic bath with DI water, acetone, and IPA, with each step conducted for 15 min. The cleaned substrates were subjected to UV–ozone treatment for 15 min. For the electron transport layer (ETL), 100 μL of SnO₂ solution was dispensed onto the cleaned substrates, spin-coated at 5000 rpm for 30 s, and subsequently annealed at 150 °C for 15 min. After annealing, the SnO₂-coated substrates underwent an additional 10 min UV–ozone treatment. The SnO₂ layer deposition was performed under ambient conditions. The perovskite layer was deposited by spin-coating 50 μL of the perovskite precursor solution onto the SnO₂-coated substrates at 3000 rpm for 30 s. During the final 15 s of the spin-coating process, 30 μL of CB was dispensed as an anti-solvent. The coated substrates were then annealed at 150 °C for 15 min to induce perovskite crystallization. For surface passivation, 40 μL of OAI solution was spin-coated onto the perovskite layer while spinning at 3000 rpm for 30 s. The HTL was dynamically spin-coated at 2500 rpm for 30 s, using 60 μL of α2 solution. Each solution was applied onto the perovskite layer while in a hot and transparent state. A gold electrode (50 nm thickness) was thermally evaporated onto the HTL using a KVE-T2000 (Korea Vacuum Tech, Gimpo, Republic of Korea). The perovskite layer and HTL were fabricated under a controlled dry air atmosphere with 5% relative humidity. The active area of the device was 0.0256cm².

2.4. Measurements and Characterizations

Photocurrent density–voltage (J-V) characteristics were measured under AM 1.5G 1 sun (100 mW/cm²) illumination using an LED solar simulator (LSH-7320, class ABA, Newport, RI, USA) and a solar cell I-V test system (T2003, Ossila, Sheffield, UK). Incident photon-to-electron conversion efficiency (IPCE) spectra were acquired using a K3100 system (McScience, Suwon, Republic of Korea) to evaluate the quantum efficiency of each device across different wavelengths. To investigate the surface morphology of α2 layers fabricated with different DASs, optical microscopy was performed using a BX53M microscope (Olympus, Tokyo, Japan), and atomic force microscopy (AFM) analysis was conducted using an MFP-3D Origin+ system (Oxford Instruments, Abingdon, UK). Scanning electron microscopy (SEM) (SU-8600, Hitachi, Tokyo, Japan) was employed to examine cross-sectional structures. UV-Vis absorption spectra of α2 layers were measured using UV-Vis spectroscopy (UV-3600i Plus, Shimadzu, Kyoto, Japan) to assess the optical properties of the films. Photoluminescence (PL) and time-resolved photoluminescence (TRPL) measurements were conducted using an FLS1000 spectrometer (Edinburgh Instruments, Livingston, UK) to analyze carrier recombination dynamics.

3. Results

First, we investigated the solubility of α2 in various DASs to determine their suitability for solution processing, as shown in Figure 1 and

Table 1. Information on the DASs and the dissolution onset of α2.

Solvent Name	Boiling Point (°C)	Vapor Pressure (mmHg, at 25 °C)	Dissolution Onset Temperature of α2 (°C)	Price per 250 mL ($)
Dichloromethane (DCM)	40	430	25	15
1,2-dichloroethane (DCE)	83	87	50	40.25
1,3-dichloropropane (DCP)	122	18.2	65	273
1,4-dichlorobutane (DCB)	155	4.2	105	31

. Each solvent exhibits distinct physicochemical properties, including vapor pressure, dissolution temperature, and cost, which collectively influence its applicability in spin-coating processes. DCM, possessing the lowest dissolution temperature (25 °C) and the highest solubility for α2, is the most cost-effective option. However, its exceptionally high vapor pressure (430 mmHg) leads to rapid solvent evaporation, making it challenging to maintain a stable solution concentration during thin-film deposition. In contrast, DCB has the lowest vapor pressure (4.2 mmHg), which enhances process control by mitigating solvent loss. It is also relatively inexpensive; however, its significantly elevated dissolution temperature (105 °C) severely restricts its practicality, as α2 may precipitate upon cooling, thereby compromising the uniformity of the deposited film. Meanwhile, DCE and DCP exhibit moderate vapor pressures (87 mmHg and 18.2 mmHg, respectively), which are more conducive to controlled film formation. Their dissolution temperatures (50 °C for DCE and 65 °C for DCP) are higher than that of DCM, but substantially lower than that of DCB, providing a balanced compromise between solubility and process stability. However, DCP is the most expensive among the tested DASs, which may limit its practicality. Considering its optimal balance of vapor pressure, solubility, and economic feasibility, DCE is identified as the most suitable solvent for α2 processing.

To investigate the formation of α2-HTLs on photoelectronic devices, transparent α2-solutions with DASs were prepared at their respective dissolution onset temperatures and spin-coated onto perovskite layers. The resulting surfaces were examined using optical microscopy at ×20 magnification. As shown in Figure 2, α2-DCM and α2-DCE exhibited smooth and uniform morphologies, whereas α2-DCP and α2-DCB presented irregularly distributed spots. Notably, the density of these spots increased significantly in α2-DCB. These features are presumed to have arisen from the localized aggregation of α2, which likely occurred due to a decrease in solvent temperature prior to complete evaporation during spin-coating, impeding uniform film formation and promoting solute precipitation.

For a more detailed surface characterization, AFM analysis was performed on the films, and the results are presented in Figure 3 and

Table 2. Specific surface data of each α2-HTL from AFM.

Solvent	Average Roughness (Ra)	RMS Roughness (Rrms)	Maximum Height (Ht)
DCM	40.40 nm	11.53 nm	82.75 nm
DCE	44.61 nm	12.65 nm	93.08 nm
DCP	58.9 nm	15.57 nm	136.1 nm
DCB	62.9 nm	18.32 nm	266.2 nm

. The average roughness (R_a), root mean square roughness (R_rms), and maximum surface height (H_t) exhibited an increasing trend with the chain length of DASs. DCM produced the smoothest surface, with R_a = 40.40 nm, R_rms = 11.53 nm, and H_t = 82.75 nm. In the case of α2-DCE, a slight increase was observed, with values of R_a = 44.61 nm, R_rms = 12.65 nm, and H_t = 93.08 nm. However, the roughness parameters increased markedly for DCP and DCB, with R_a values of 58.9 nm and 62.9 nm, R_rms values of 15.57 nm and 18.32 nm, and H_t values of 136.1 nm and 266.2 nm, respectively. These pronounced increases are attributed to the formation of localized surface irregularities in α2-DCP and α2-DCB.

To further examine the layer morphology, cross-sectional SEM imaging was conducted on ITO/SnO₂/perovskite/α2-HTL/Au devices. All fabricated devices maintained consistent thicknesses of ITO (~140–150 nm), SnO₂ (~40–50 nm), perovskite (~500–550 nm), and Au (~40–50 nm). However, the α2-HTL thickness varied considerably depending on the DASs employed. As depicted in Figure 4, the DCM-based HTL exhibited the greatest thickness (~524 nm), while the values decreased progressively with DCE (218 nm), DCP (117 nm), and DCB (69 nm). This analysis suggests that surface roughness is significantly influenced by HTL thickness, which, in turn, is dictated by the solvent properties of DASs. However, the sharp increase in roughness from DCE to DCP is primarily attributed to the aggregation of α2 particles, resulting in a substantial increase in maximum surface height, rather than uniform film thickening.

To investigate the optical characteristics of α2-based HTLs formed with different DASs, UV-Vis absorption spectra were measured with ITO/SnO₂/perovskite/α2-HTLs. As shown in Figure 5a, all samples exhibit nearly identical absorption profiles, indicating that the choice of solvent did not significantly influence the overall optical absorption of the perovskite/HTL structure. Additionally, photoluminescence (PL) and time-resolved PL (TRPL) measurements were conducted on ITO/SnO₂/perovskite/α2-HTL structures to further examine carrier dynamics. According to Figure 5b, the PL intensity gradually decreased in the order of α2-DCM, α2-DCB, α2-DCP, and α2-DCE. Conversely, as displayed in Figure 5c, TRPL decay followed the order α2-DCB, α2-DCP, α2-DCM, and α2-DCE. Specific lifetime data were calculated with Equation (1), as given below.

$$\mathrm{I}(\mathrm{t})={\mathrm{A}}_1·{\mathrm{e}}^{-{\displaystyle \frac{\mathrm{t}}{\mathsf{\tau }1}}}+{\mathrm{A}}_2·{\mathrm{e}}^{-{\displaystyle \frac{\mathrm{t}}{\mathsf{\tau }2}}}$$

(1)

Here, A₁ and A₂ represent the amplitudes corresponding to the decay components τ₁ and τ₂, respectively. The parameter τ₁ is attributed to carrier trapping at defect states, while τ₂ is associated with radiative recombination. The average lifetime τ_avg was subsequently calculated using Equation (2).

$${\mathsf{\tau }}_{\mathrm{a}\mathrm{v}\mathrm{g}}={\displaystyle \frac{{\mathrm{A}}_1{\mathsf{\tau }}_1^2+{\mathrm{A}}_2{\mathsf{\tau }}_2^2}{{\mathrm{A}}_1{\mathsf{\tau }}_1+{\mathrm{A}}_2{\mathsf{\tau }}_2}}$$

(2)

As shown in

Table 3. Specific lifetime data of HTLs formed with each α2-DAS.

HTL	A1	A2
	τ1 (ns)	τ2 (ns)	τavg (ns)
DCM	2466.81	29.89
	6.74	74.88	14.82
DCE	1903.47	207.01
	5.52	16.20	8.10
DCP	1879.18	20.07
	8.40	111.88	21.29
DCB	1464.76	612.73
	24.09	141.49	107.53

, the calculated average lifetimes were 14.82, 8.10, 21.29, and 107.53 ns for DCM, DCE, DCP, and DCB, respectively. Among the α2-HTLs, α2-DCE exhibited the fastest decay behavior. Although the TRPL decay profile of α2-DCM is faster than that of α2-DCP, the significantly higher PL intensity observed in α2-DCM implies that non-radiative recombination, particularly at the perovskite/HTL interface, was more pronounced in this condition. While α2-DCB showed a PL intensity comparable to that of α2-DCP, its TRPL decay was the slowest among all HTLs, indicating inferior carrier extraction.

We fabricated PSCs with the structure ITO/SnO₂/FAPb (I_0.99Br_0.01)₃/OAI/α2-HTLs/Au, and evaluated their performance. As shown in Figure 6, all α2-DAS-based PSCs, except for α2-DCB, exhibited a short-circuit current density (J_SC) of approximately 24–25 mA/cm² and a fill factor (FF) ranging from 0.75 to 0.8. However, only α2-DCE and α2-DCP achieved high open-circuit voltages (V_OC), recording 1.1–1.5 V. These variations were reflected in the PCE, with α2-DCE and α2-DCP demonstrating efficiencies of approximately 20–22%. The shortest lifetime achieved by α2-DCE suggests its more efficient carrier extraction, which is consistent with its superior device performance. Notably, despite the presence of unidentified surface spots and slower TRPL decay than that of α2-DCM, which are expected to negatively impact device performance, α2-DCP exhibited high efficiency. In contrast, α2-DCB-based PSCs showed the lowest performance among α2-DASs, with significant deviations in J_SC (13–18.5 mA/cm²), V_OC (0.7–0.9 V), FF (0.5–0.55), and PCE (5–9%). The highest-performing device, as presented in Figure 7a and

Table 4. The best data from each PSC with an α2-HTL. The hysteresis index (HI) was calculated using Equation (3).

Sample (HTL)	Scan Direction	JSC from J−V (mA/cm2)	JSC from IPCE (mA/cm2)	VOC (V)	FF	PCE (%)	HI
α2-DCM	Forward	24.71	23.43	1.02	0.76	19.19
	Reverse	24.72		1.01	0.68	16.87	−0.14
α2-DCE	Forward	24.88	24.43	1.15	0.80	22.71
	Reverse	24.71		1.13	0.76	21.42	−0.06
α2-DCP	Forward	24.61	24.22	1.15	0.79	22.29
	Reverse	24.49		1.14	0.77	21.50	−0.04
α2-DCB	Forward	18.75	18.49	0.90	0.55	9.16
	Reverse	17.70		0.93	0.53	8.73	−0.05

, was α2-DCE, achieving J_SC = 24.88 mA/cm² V_OC = 1.15 V, FF = 0.80, and PCE = 22.71% in the forward scan. The hysteresis index (HI) calculated using Equation (3) was −0.06 for α2-DCE.

$$\mathrm{H}\mathrm{I}=({\displaystyle \frac{{\mathrm{P}\mathrm{C}\mathrm{E}}_{\mathrm{R}\mathrm{e}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{e}}-{\mathrm{P}\mathrm{C}\mathrm{E}}_{\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{w}\mathrm{a}\mathrm{r}\mathrm{d}}}{{\mathrm{P}\mathrm{C}\mathrm{E}}_{\mathrm{R}\mathrm{e}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{e}}})}$$

(3)

α2-DCP exhibited the second-highest performance, closely approaching that of α2-DCE, while also maintaining a lower HI. The corresponding values for α2-DCP were J_SC = 24.61 mA/cm², V_OC = 1.15 V, FF = 0.79, and PCE = 22.29%, with slightly lower J_SC and FF compared to α2-DCE. The highest-performing α2-DCM device achieved a PCE of 19.19%, which aligns closely with previous studies. However, its performance was slightly lower, and it exhibited significant hysteresis due to the poor reproducibility caused by the high vapor pressure of DCM. α2-DCB, in contrast, recorded J_SC = 18.75 mA/cm², V_OC = 0.90 V, FF = 0.55, and PCE = 9.16%, the lowest among all the tested conditions.

To further investigate the photoelectronic properties of these devices, incident photon-to-electron conversion efficiency (IPCE) measurements were performed. As shown in Figure 7b, α2-DCE and α2-DCP exhibited comparable performance, with calculated J_SC values of 24.43 and 24.22 mA/cm², respectively. The IPCE spectrum of α2-DCM was slightly lower than that of α2-DCE and α2-DCP in the 500–800 nm range, with a corresponding J_SC of 23.43 mA/cm². In contrast, α2-DCB exhibited an IPCE of approximately 70%, except in the 420–500 nm region, resulting in the lowest J_SC of 18.49 mA/cm² among the samples.

4. Conclusions

This study investigated the feasibility of DCE as a solvent for α2-based hole transport layers, focusing on solubility, vapor pressure, film morphology, and photovoltaic performance. The limitations of DCM, including its high vapor pressure and poor film reproducibility, were addressed by comparing alternative DASs, such as DCE, DCP, and DCB. The results demonstrated that DCE provides the most favorable balance between solubility, vapor pressure, processability, and economic feasibility. While DCM exhibited the highest solubility for α2, its excessive vapor pressure led to inconsistent film formation. In contrast, DCE, with a moderate vapor pressure, enabled controlled deposition and resulted in uniform film morphology, as confirmed by AFM and SEM analyses. Photovoltaic characterization revealed that devices incorporating α2-DCE achieved the highest PCE of 22.71%, surpassing those using DCP (22.29%), DCM (19.19%), and DCB (9.16%). Despite the presence of surface irregularities, α2-DCP exhibited comparable efficiency to α2-DCE, whereas α2-DCB showed significant performance degradation due to poor carrier transport. However, despite its competitive efficiency, DCP remains limited as a practical solvent, due to its lower solubility for α2 and significantly higher cost compared to DCE. PL and TRPL measurements confirmed that α2-DCE facilitated the most efficient charge extraction, as indicated by its lowest PL intensity and fastest TRPL decay. Overall, DCE emerged as the most suitable solvent, offering superior film uniformity, enhanced carrier transport, and improved device performance compared to other DASs. These findings highlight the critical role of solvent selection in device fabrication, and suggest that DCE is a promising alternative to DCM for solution-processed α2-HTL.