Perovskite solar cells (PSCs) have great potential for future application. However, the commercialization of PSCs is limited by the prohibitively expensive and doped hole-transport materials (HTMs). In this regard, small molecular dopant-free HTMs are promising alternatives because of their low cost and high efficiency. However, these HTMs still have a lot of space for making further progress in both efficiency and stability.
1. Introduction
Solar photovoltaic (PV) technology, which converts sunlight directly to electricity in a green, safe and efficient way, plays a central role in fulfilling the net-zero targets (for greenhouse gas emission, especially CO
2) set for around 2050
[1]. In the current PV market, crystalline silicon (c-Si) solar cells are in a dominant position. Despite an 88% decline in their levelized cost of energy (LCOE) between 2010 and 2021
[2], fossil fuels still accounted for a global share of 62% in electricity generation in 2021
[3]. Hence, continuous efforts to reduce the LCOE of solar PV are needed to enhance their cost competitiveness as compared to fossil fuels.
Perovskite solar cells (PSCs) are a new type of PV technology developed in recent decades in which the absorber material is composed of metal halide perovskites with a general chemical formula of ABX
3. Generally, the A site is a monovalent cation, such as [HC(NH
2)
2]
+ (FA
+), CH
3NH
3+ (MA
+) or Cs
+. The B site is mainly Pb
2+, and the X site is a halide ion (I
−, Br
− or Cl
−). To fabricate a PSC, the perovskite layer is usually sandwiched between an electron-transport layer (ETL) and a hole-transport layer (HTL), constituting a n–i–p device or p–i–n device, which depends on the relative position of the charge carrier transport layers (CTLs). Upon illumination, the photogenerated electrons and holes in the perovskite layer will be selectively transported to the corresponding electrode through CTLs, and the charge carrier recombination at the two perovskite interfaces can be mitigated. Therefore, CTLs are indispensable for efficient PSCs despite the fact that metal halide perovskites possess the property of ambipolar charge transport. Practically, the CTL, especially the HTL, plays a key role in the development of PSCs. In 2009, Miyasaka et al. first introduced MAPbI
3 and MAPbBr
3 nanocrystals to sensitize a meso-TiO
2 electrode in dye-sensitized solar cells (DSCs), obtaining a power conversion efficiency (PCE) of 3.81% and 3.13%, respectively
[4]. However, perovskite nanocrystals degraded very fast in the liquid electrolyte. In 2012, Park et al. used solid spiro-OMeTAD (2,2′,7,7′-tetrakis[N,N-di(4-methoxyphenyl)amino]-9,9′-spirobifluorene) to replace the liquid electrolyte and achieved a solid DSC with PCE of 9.7%
[5]. Thereafter, PSCs have embarked on a fast track of efficiency development
[6][7][8]. Very recently, a certified PCE of 25.7% and 32.5% has been reported for single-junction PSCs and perovskite-silicon tandem solar cells
[9][10], respectively, demonstrating perovskite PV as an efficient solar PV technology.
In addition to PCE, manufacturing cost and long-term stability are two other key factors determining the LCOE of a certain photovoltaic technology
[11]. While the manufacture of PSCs is compatible with low-temperature (<300 °C) and high-volume processing technologies
[12], its manufacturing cost strongly depends on the materials
[13]. For modules containing expensive HTL and metal electrodes, the manufacturing cost could be significantly higher than that of the c-Si PV
[11]. In recent PSCs with recorded PCEs, they all use spiro-OMeTAD or PTAA (poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine]) as the HTL
[14][15][16]. However, the synthesis and purification procedures of spiro-OMeTAD and PTAA are tedious, making them prohibitively expensive for future commercialization
[17].
Moreover, both spiro-OMeTAD and PTAA generally need p-doping due to their low intrinsic hole mobility (~10
−5 cm
2 V
−1 s
−1)
[18]. Upon doping, the oxidized HTM molecules will increase the charge carrier density and thereby the conductivity, whereas it will induce a series of stability issues
[18][19]. For example, it was found that the oxidized spiro-OMeTAD molecules will degrade or be reduced under thermal stress
[20][21]. In addition, the Lewis acid p-dopant in the HTL may attack the perovskite surface, resulting in rapid performance degradation
[22]. Therefore, the usage of p-doped HTL limits the lifetime of the PSCs. Studies on cost analysis have shown that the LCOE of perovskite PV strongly depends on the lifetime
[23][24]. Meng et al. estimated the dependence of LCOE on lifetime by assuming a module efficiency of 19%
[25]. A LCOE of ~0.15 USD kW
−1 h
−1 was calculated for the module with a lifetime of 5 years, which is much higher than the c-Si PV.
Hence, the development of low-cost and dopant-free hole-transport materials (HTMs) is of significant importance for the commercialization of perovskite PV, and it has become a research hotspot. Generally, dopant-free HTMs can be classified into three categories, including polymers, small molecules, and inorganic compounds. Among them, polymeric HTMs possess the highest PCE for dopant-free HTMs
[16]. However, their large-scale application is limited by their extremely high cost
[26]. Inorganic HTMs seem to be ideal candidates for dopant-free HTMs because of advantages such as high hole mobility, good stability, and low cost, while the efficiency of inorganic HTMs is still lagging behind that of their organic counterparts
[27]. Furthermore, recent studies indicate that the interaction between inorganic HTMs and perovskites may accelerate the aging of PSCs
[22][28]. The above limitations of polymeric and inorganic HTMs have paved the way for the development of small molecular HTMs
[29]. Recently, dopant-free small molecular HTMs have achieved impressive progress. In p-i-n PSCs, a PCE of 24.34% was achieved by a small molecule coded as BTP1
[30], approaching the state-of-art PTAA-based device (25.0%)
[16]. More importantly, the synthetic cost of BTP1 is significantly lower than that of PTAA (8.98 vs.~2700 USD/g). In the n-i-p device, a small molecule, BDT-DPA-F, achieved a PCE of 23.12%
[31]. The efficiency was slightly lower than the highest record (24.6%) achieved by P3HT (poly(3-hexylthiophene))
[32], but the BDT-DPA-F-based device exhibited ultralong stabilities under both operation and thermal aging conditions. In addition, a phthalocyanine derivate coded as SMe-TPA-CuPc obtained a similar PCE of 23.0% and retained 96% of the initial PCE after being thermally stressed at 85 °C for 3624 h
[33]. In this respect, small molecular HTMs have demonstrated their great potential as efficient, low-cost, and stable HTMs.
Despite the encouraging progress of small molecular HTMs, their efficiency is still lagging behind the state-of-art doped HTMs. After numerous efforts devoted to the optimization of molecular design, there are dozens of new HTMs with higher hole mobility than the doped spiro-OMeTAD in efficient PSCs (10
−4~10
−3 cm
2 V
−1 s
−1)
[34]. However, dopant-free HTMs, even for those with suitable energy levels as obtained from cyclic voltammetry (CV) in solution, often suffer from inferior open-circuit voltage (
VOC)
[35][36][37]. Generally, the
VOC of PSCs is determined by the energy level alignment and the charge carrier recombination
[38]. For small molecular HTMs, their lowest unoccupied molecular orbital (LUMO) energy levels are generally lower than that of spiro-OMeTAD, resulting in less sufficient electron blocking at the interface and thereby more charge carrier recombination
[37][39][40]. However, the charge carrier recombination cannot explain the difference in
VOC alone as demonstrated by Gelmetti et al.
[41]. In their work, three TAE coded HTMs were compared with spiro-OMeTAD. The hole mobilities of three TAE HTMs are not quite distinct from that of spiro-OMeTAD. Furthermore, according to the result from CV measurements, both TAE-1 and TAE-4 exhibited more suitable highest-occupied molecular orbital (HOMO) energy levels than spiro-OMeTAD, i.e., larger
VOC would be expected for the former two. However, all three TAE HTMs showed much lower
VOC in comparison to spiro-OMeTAD, and the difference in average
VOC can be up to 170 mV for TAE-4. The comprehensive analysis on charge carrier recombination demonstrated that the largely different
VOC could not be ascribed to the recombination. The contact potential difference (CPD) measured by Kelvin probe force microscopy (KPFM) revealed that distinct HTMs deposited on perovskite resulted in different shifts in the vacuum level (V
L). Especially, the V
L was upward shifted to as large as ~200 meV upon the deposition of TAE-3, implying a same shift in HOMO of TAE-3 compared to the valence band (VB) of perovskite. This result demonstrates that the energy level alignment can be quite different from that predicted by CV measurements. For organic semiconductors, molecular orientation, which is a feature of molecular stacking in the film, plays a determinant role in the interface energy level alignment
[42]. Furthermore, molecular orientation of organic semiconductors has significant influence on hole mobility, demonstrated in the field of organic field-effect transistors (OFET)
[43][44].
2. Charge Carrier Transport and Transfer
In organic semiconductors, charge carrier transport is realized via the intramolecular and intermolecular delocalization of π electrons
[45]. In polymeric organic semiconductors, long-range intramolecular delocalization is possible due to their elongated π-backbone. Differently, charge carrier transport in small molecular organic semiconductors is more dependent on the intermolecular delocalization. In this respect, charge carrier transport in organic solids depends on the packing details of molecules, which determine the intermolecular π-π overlap. For example, lamellar packing exhibits larger π-π overlap than herringbone packing
[45]. In the lamellar packing motif, the π-π overlap degree varies upon the slip displacement and slip angle, which depends on the molecular structure
[46][47]. Although charge carrier transport depends on the packing details of molecules, the most efficient charge carrier transport channel is in the direction parallel to the π-π stacking orientation
[48][49].
Due to the anisotropy of charge carrier transport, molecular orientation is an important factor that should be taken into consideration when organic semiconductors are applied in electronics or optoelectronics
[50]. Upon being deposited onto the substrate, molecules may adopt face-on or edge-on stacking with respect to the substrate. In the face-on configuration, the π-backbone stacks in a manner parallel to the substrate surface. In contrast, π-backbones in edge-on configurations stand on the substrate. Hence, the face-on orientation is preferred in solar cells and OLEDs, while edge-on is the optimal molecular orientation for OFETs due to the in-plane distribution of drain and source electrodes. The mismatch between molecular orientation and current flow direction of the device may result in significantly retarded charge carrier transport. Noh et al. demonstrated that an edge-on oriented porphyrin film exhibits 100 times higher mobility than the face-on oriented one in OFET
[51]. In PSCs, several times higher hole mobility was observed in the HTM film with dominant face-on orientation compared to the ones with edge-on orientation
[33][52].
In addition to the charge carrier mobility, charge carrier transfer dynamics at the interface also strongly depend on molecular orientation. Similar to the intramolecular charge transport between the electron donor part and the electron acceptor part
[53], the interfacial charge transfer depends on the effective electronic coupling,
Veff, at the interface
[54]. For the charge transfer between diabatic states
i and
f, the relationship between transfer rate
ki→f and
Veff is described as
ki→f=2πћV2eff(i→f)ρ(Ef), where
ρ(Ef) is the density of states
[54]. Upon depositing a molecule onto a substrate, the interfacial
Veff depends on molecular orientation. Taking the sexithiophene (6T) molecule on Au as an example, density functional theory (DFT) calculations demonstrate that both HOMO and the lowest unoccupied molecular orbital (LUMO) of face-on oriented 6T to delocalize into the Au substrate, while it was not observed in the edge-on system
[55]. The faster charge carrier transfer at the interface then can be obtained in the face-on system due to the strong electronic coupling
[55][56][57][58]. In 2020, Igci et al. reported three small molecular HTMs coded as CI-B1, CI-B2 and CI-B3, whose conductivities were 7.98 × 10
−8 S cm
−1, 5.53 × 10
−7 S cm
−1, and 9.60 × 10
−7 S cm
−1, respectively
[59]. The lowest conductivity of CI-B1 is expected to be not unfavorable for hole extraction
[60]. However, both steady-state photoluminescence (ssPL) and time-resolved photoluminescence (TRPL) measurements demonstrated that the CI-B1 HTM possessed the fastest hole extraction due to its face-on orientation.
3. Interfacial Energy Level Alignment
In PSCs, the energy level alignment at the perovskite/CTL interfaces plays an essential role in determining the final device performance. Firstly, the energy offset should be large enough for efficient charge carrier extraction. Taking the perovskite/HTL interface as an example, the valence band maxima (VBM) of the perovskite layer should be lower than the HOMO energy level of HTL. Otherwise, a hole extraction energy barrier will form at the interface, resulting in charge carrier recombination and thereby inferior device performance
[61]. Westbrook et al. suggested that the energy offset of ~0.07 eV is enough for efficient hole transfer at the perovskite/HLT interface
[62]. After the minimum offset being guaranteed, the lower the HOMO of HTL, the larger the
VOC that can be achieved
[63][64].
For a certain organic molecule, the transformation in molecular orientation can induce sizable variations in HOMO levels
[65], and the energy level alignment at its interface will change accordingly. In PSCs, the impact of molecular orientation on energy level was impressively revealed by Zhou et al.
[52]. The ultraviolet photoelectron spectroscopy (UPS) measurements revealed that the HOMO level of HTL changed from −5.03 to −5.42 eV for the film with 0% and 59.6% face-on orientation, respectively. The similar downward shift of HOMO level was also observed for the a phthalocyanine HTM when molecular orientation was transformed from edge-on to face-on
[33]. The dependence of HOMO level on molecular orientation can be rationalized by the change in surface dipole
[66]. For a face-on oriented molecule, its π-electron cloud will be exposed to the vacuum with the positively charged backbone lying below, i.e., an intramolecular dipole pointing toward the substrate is formed. Consequently, the position of HOMO level will be lower than that of the molecule in gas phase. In contrast, the direction of the surface dipole will be reversed for edge-on-oriented molecules if it is the H atom exposed to the surface
[42].
4. Device Stability
Due to the removal of dopants, PSCs based on dopant-free HTMs generally exhibit enhanced humid and thermal stability
[67][68][69][70]. Comparing the face-on and edge-on orientations, the former one is more conducive to the hole extraction and transport as discussed above. Therefore, minimized charge accumulation at perovskite/HTL can be obtained
[71]. On the one hand, reduced charge accumulation is beneficial for suppressing charge carrier recombination and thereby higher
VOC [72]. On the other hand, while the underlying mechanism is complicated, the charge accumulation has been related to the photo-induced degradation of PSCs
[73]. Byeon et al. analyzed the degradation process of PSCs working under light illumination, and it was identified that the degradation started from the interface with accumulated charges
[74]. DuBose et al. suggested that the localized holes at the iodide sites would induce the formation of I
−, resulting in halide segregation
[75]. In addition, enhanced ion migration was observed for HTMs with lower hole mobility
[76]. Recently, based on the result of theoretical calculations, Tong et al. proposed that both the diffusion barrier and the migration length for I
− migration could be reduced by the injected holes
[77]. In this respect, the face-on orientation is more desirable for the operational stability of PSCs. For example, Cheng et al. reported three HTMs with the ratio of face-on varied in the trend of BDT-DPA-F > BDT-TPA-F > BDT-TPA. After a continuous maximum power point (MPP) tracking for 1200 h, the BDT-DPA-F-based PSC maintained 82.6% of its initial PCE, while the BDT-TPA-F and BDT-TPA-based PSCs showed PCE retentions of 74.3% and 69.4%, respectively
[31], which is consistent with the trend of the ratio of face-on. Enhanced operational stability was also observed for SMe-TPA-CuPc-based PSCs when molecular orientation was transformed from edge-on to face-on
[33].
Mechanical stability is another issue needed to be addressed for long-life PSCs, especially for the flexible ones
[78][79]. It has been demonstrated that the perovskite/HTL interface is the most mechanically vulnerable part in the PSC, and strengthened interface adhesion is desirable
[80][81]. In comparison to the edge-on orientation, the face-on orientation is beneficial for a larger contact area between HTM molecules and perovskite, and thereby, a stronger interface interaction can be realized
[82]. In 2021, Javaid et al. calculated the adsorption energies of metal phthalocyanines on a MAPbI
3 surface. It was revealed that the adsorption energy could be up to about 2.6 eV in a face-on adsorption case, while it could be below 0.4 eV in the edge-on case
[83]. In 2017, Kim et al. obtained a tetra-tert-butyl substituted copper (II) phthalocyanine HTL with face-on orientation via engineering the interface structure. In the tape test, both spiro-OMeTAD and PTAA were removed by the adhesive tape. However, only the Au electrode was detached for the phthalocyanine based PSC, which demonstrated the stronger adhesion of phthalocyanine HTL to perovskite surface
[84].