Two-dimensional materials in perovskite solar cells

Modern electronic and optoelectronic devices are formed by a sequence of layers in a multimaterial/multijunction combination. Even Complementary Metal-Oxide Semiconductor (CMOS) technology, dominated for many decades by silicon and its native oxide SiO2 is moving toward a multimaterial structure where germanium (for strain engineering), alternative oxides (for high dielectric constant insulators) and metal gates are efficiently combined with silicon. Owing to the increased number of (hetero)junctions in these multimaterial structures, a fundamental role is played by the interface between the different layers. Thus, it should not be a big surprise that also the halide perovskite (HP) photovoltaic features strongly rely on interfaces properties. In fact, a typical perovskite solar cell (PSC) is composed by an HP photon absorbing layer, sandwiched between selective charge transport layers (CTLs) and electrodes for negative and positive charge extraction. These materials are selected carefully by taking into account their compatibility with the underneath layers [1], their charge mobility [2], and the energy level alignment with the perovskite absorber [3]. Similarly to other multimaterial/multijunction devices, interfaces in PSCs play a crucial role in setting device performances and stability: charge transfer at the interfaces [4], interface band alignment [5], interfacial vacancies [6], defects due to poor adhesion between layers [7] and energy barriers are all controlled by interfaces. However, HP are peculiar materials bridging the conventional semiconductor with ionic conductor. In fact, HP can be considered as crystalline liquid [8, 9] where the sublattice of organic cation liquid is confined in nanoscopic pore defined by the sublattice of corner-sharing PbX6 [3]− (X= I, Br, or Cl) octahedral [10]. The crystal-liquid duality has been discussed in relation with HP to explain the large polaron formation and screening of charge carriers, leading to defect tolerance, moderate charge carrier mobility, and radiative recombination properties [[10]. Beside these physical properties the existence of a ‘liquid’ behavior is also manifested by the mobility of ions into the crystal structure caused by the migration of halide-ion vacancies V–X (X= I, Br, Cl) [11, 12]. For this reason HP are considered solid-state mixed conducting materials where both electrons and ions contribute to the charge transport [12]. Halide ion migration, as well as organic/inorganic cation and Pb migrations, [13, 14] are shown to be related to hysteresis, giant capacitance and stability of devices [15]. Besides these intrinsic species, also extrinsic ions coming from adjacent layers (Li, Na, OH−, H etc) can migrate across the perovskite, impacting the device operation [16]. The relation between solar cell degradation and ion migration has been addressed by several authors, however many of the reactions are reversible [17] and ion migration will not induce degradation by per-se unless the ion is leaving the perovskite layer or the solar cells itself. In this respect the use of insulating interface layers such as poly(methyl methacrylate) (PMMA), polyethylene oxide (PEO), Polystyrene (PS) can prevent ion migration, passivate the defects of perovskite and, consequently, increase the stability of the cell [18]. Another peculiar characteristic of HP used in solar cells is related to the architecture of the cell. In fact, together with classical thin film photovoltaic where a planar configuration of the different layer forming the cell is used [19], PSC can be in the form of mesoscopic structure where the HP absorber is nucleated into and above a mesoporous transporting layer, typically TiO2 [20]. Such n-i-p mesoporous configuration is very effective and was providing the most efficient PSCs. Here, an interface with larger equivalent surface between the mesoporous transporting layer and the perovskite is formed with respect to planar configuration and any interface instability could be easily enhanced. Interface instability is typical in heterojunction devices and was, for example, one of main problem in the development of high-k silicon mosfets [21]. Heterointerfaces are critical region of matching between two different material where electrical (charge transfer, interlayer resistance, defects, charge accumulation and recombination), chemical (miscibility, wettability, surface tension etc.), structural (strain, dislocation,

defects) and mechanical (adhesion/delamination [7]), problems could arise. In HP devices this problem is further emphasized because of the possible presence of an organic/inorganic heterojunction typically fabricated at low temperature without a specific epitaxy.
Owing to the general problem of heterointerfaces and the specific characteristics of HP, a pivotal role of interface engineering (IE) has recently gained attention as a strategy to master the efficiency and stability performance of PSC [22]. More in details, IE aims to improve the charge injection/collection at perovskite/CTLs interfaces by modifying the chemical/physical properties of involved layers at the interface. This is possible by inserting inter or buffer layers [23,24] or simply by changing the composition and eventually the morphology of CTLs [25].
Interface layers offer an additional degree of freedom in the device design permitting to (i) decouple the adjacent layers, (ii) tune physical, chemical and structural interface properties without impacting on the material choice of the adjacent layers, (iii) reduce the degradation avoiding cross contamination of the layers, (iv) guide the epitaxial growth of two adjacent layers improving their structural properties. Several strategies have been developed so far to master interface properties by using small molecules, polymers, self-assembling monolayers, inorganic thin film, nanomaterials with both chemical and physical deposition techniques [26].
Among these approaches, the use of two-dimensional (2D) materials such as graphene and related materials (GRM) and transition metal dichalcogenide (TMD), has been proposed and demonstrated to be very effective.
2D materials have gained importance due to their unique physical and chemical properties such as tailored electrical conductivity, mechanical robustness, high optoelectronic tunability, easy production and solution processability [27]. A vast library of 2D materials exists with several thousand materials from metals to semiconductors and insulators that can be dispersed in several solvents and optimized for printing processes. Moreover, 2D materials can vary their properties if stacked with other 2D materials by generating novel material structures with new properties [28,29]. Large amounts of 2D materials with unusual optoelectronic properties can be produced on-demand by liquid phase exfoliation or by ball-milling technique [30], making their use suitable for large-scale industrial production. 2D materials could suffers from restacking, resulting from strong interlayer interaction and high surface energy. When this occurs, 2D materials can be functionalized with small molecules, metal nanoparticles, self-assembled monolayers, ionic liquids and polymers [31][32][33], retaining the potential advantages offered by their 2D nature. This confers them wide processability in organic and polar solvent and/or the ability of patterning using photolithographic techniques. Nevertheless, Chemical Vapour Deposition techniques are also available for well controlled 2D materials making the possible fabrication choice quite large.
IE with 2D materials is one of the most promising strategy to boost both PCE and stability of PSC [34] and has been extended also to other devices based on HP [35]. One of the main potentialities of 2D materials is the possibility to easily tailor their electronic structure, such as Work Function (WF) [36] or band gap [37], by operating a specific choice of the 2D material, by proper functionalization and eventually by inducing a quantum confinement playing with the lateral dimension of the 2D material [38]. Fine-tuning of WF and band gap allows to obtain appropriate energy level alignment leading to an ideal energy offset between perovskite active layer and CTLs, eventually inducing built-in potential for efficient charge collection at the electrodes. From the photovoltaic performance point of view, a good energy level alignment is paramount to achieving ideal open-circuit voltage (V OC ), by minimizing the carrier recombination losses and reducing the hysteretic behaviour of PSC. It also contributes to optimal charge extraction, leading to high short circuit current density (J SC ) and larger fill factor (FF). As an example, the possibility to chemically or thermally modify GRMs allows their oxygen vacancy/defect concentration to be changed, varying their optoelectronic properties [27]. This permits GRMs to be used as electron or hole transporting layers as well as interlayers at the perovskite/CTL interface [22].
The synergy between HP and 2D materials has recently been revealed. A beneficial interaction between CH 3 NH 3 PbI 3 (MAPbI 3 ) HP and graphene has been outlined by the theoretical work of Volonakis and Giustino [39,40]. DFT (Density Functional Theory) calculations shows that (i) the interface between MAPbI 3 and graphene is stable, (ii) the octahedral tilt of the MAPbI 3 is reduced pushing the I atoms away from the graphene while attracting the Pb atoms (figure 1) and (iii) the resulting interfacial potential step, of the order of 0.3 eV, is expected to assist electron transfer from MAPbI 3 to graphene, and to act as a barrier for photogenerated holes [39,40]. The impact of graphene on the octahedral phase has been verified experimentally showing that transition between tetragonal to octahedral is inhibited by the presence of graphene [25]. An 'healing effect' of graphene preventing carrier trapping near the surface of metal halide perovskites has been also theoretically demonstrated: when the surface of MAPbI 3 is covered with a graphene sheet, the presence of I, Pb and/or methylammonium (MA) vacancies is reduced and consequently the charge trap density reduces [41]. The presence of graphene within the mesoporous layer can increase the device lifetime under different stress test conditions [42]. In fact, graphene can locally reduce the light-induced perovskite degradation and the consequent release of iodine species, which diffuse across the interfaces and cause the modifications at the gold electrode (Au-I bonding) and the TiO 2 (Ti-I bonding) interfaces [43]. In addition, an improved stability of the carrier temperature was demonstrated when graphene is embedded in mesoscopic electron transporting layer, reflecting the superior stability of perovskite nanocrystals in the mesoporous region [44].
Such encouraging results pushed the scientific community to address new classes of 2D materials with optoelectronic properties customized for specific PSC structures and for specific interfaces. In this context, several efforts have been devoted to varying the WF of 2D materials to better match with the PSC interface requests [45]. Even though this strategy has been demonstrated to be compatible with PSC, the tunability of WF was quite restricted to values above 4 eV, consequently limiting its applicability [46].
The use of GRMs and other 2D materials for a direct tuning of WF is not limited to CTLs but can be also extended directly to the perovskite. Mastering the perovskite WF without changing other properties such as band gap and optical absorption, is of paramount importance for band alignment with CTLs and consequently for the enhancement of charge transfer and reduce interface losses. Such approach has been recently demonstrated by employing a new class of 2D materials, namely the transition metal carbides, nitrides and carbonitrides (MXenes). MXenes, with a general formula M n+1 X n T x (n = 1, 2, 3), where M represents an early transition metal, X is carbon and/or nitrogen, and T x stands for surface terminations (such as OH, O, and F), came out as a promising class of 2D materials in particular for what concerns WF. During the synthesis of MXenes, their surfaces are naturally functionalized, which changes the electrostatic potential near the surfaces and affects the electronic structures, significantly shifting the WF [47]. As density functional theory predicts, surface termination strongly influences the density of states [48] and the WF of MXenes [49] which can range from 1.6 eV (for OH-termination) to 6.25 eV (for O-termination) [49]. This opens new exciting opportunities for MXene applications in optoelectronics and photovoltaics. Ti 3 C 2 T x MXenes were incorporated into the perovskite absorber layer showing an improved morphology of the perovskite and an enhanced PCE with respect to the reference cell without MXenes (+12%) [50]. In 2019, Yang and coworkers [51] proposed SnO 2 -Ti 3 C 2 MXene nanocomposite as electron transport layers (ETLs) in low-temperature processed planar PSCs. The authors demonstrated the role of Ti 3 C 2 MXene nanosheets in WF shifting of ETL providing superior charge transfer paths that permits to enhance the PCE of MXene-based PSCs with respect to the reference cells (+6.5%). Ti 3 C 2 T x MXene with various termination groups (T x ) has been used to engineer the perovskite/ETL interface and, when dispersed into the perovskite to tune the WF of the perovskite absorber [52]. The combined action of WF tuning and IE can lead to substantial performance improvements in MXene-modified perovskite solar cells, as shown by the +26% increase of power conversion efficiency (that was above 20%) and hysteresis reduction with respect to reference cells without MXene. This finding has been supported by DFT calculation showing the large and non-linear variation of perovskite WF induced by MXenes with different termination groups [52]. The MXene-PSC strategy has just begun, however, its impact on WF tuning and IE can inspire innovative efficient designs of PSCs and other perovskite-based devices such as light emitting diodes (LEDs) and detectors. Moreover, the proposed approach can be easily scaled-up to large area modules and panels, since the MXene doping involves only the precursor solutions and it does not impact the deposition method. Figure 2 and table 1 summarize the use of 2D materials in PSC showing the maximum efficiency reached so far by using a specific 2D material in a specific part of the cell. According to the scientific investigation, we consider the use of 2D material as interface layer in several parts of the cells, as a dopant for the layers, as a CTL replacing the bulk material and as electrode.
So far, the most efficient PSC exploiting the properties of 2D materials has been obtained by using TiS 2 as interlayer between the SnO 2 ETL and the multication HP, achieving a PCE = 21.73%. The improvement of the device performance was attributed to two aspects, namely (i) a better alignment of energy levels due to TiS 2 interlayer and (ii) the passivation of the trap states. This result is very promising showing that the emerging 2D materials could improve further the efficiency of PSC. We should point out that results of figure 2 and table 1 refer to single junction PSCs. However, the tailoring of electronic/optical properties with 2D materials can be extended to tandem junction such as HP/silicon solar cells. In this context, graphene has been use as contact layer exploiting its good conduction and transparency properties [85] as well as dopant of TiO 2 ETL in mechanically stacked two-terminal tandem cell able to push the efficiency up to 26.3% [86].
Besides improving of efficiency, 2D materials have been shown to be extremely effective for stability of HP. The main observation came by the seminal work of Z. Fan and co-workers [87] showing that the thermal degradation (85 • C) of MAPbI 3 crystalline structure shows a gradual evolution from tetragonal MAPbI 3 to trigonal lead iodide layered crystals with a fixed crystallographic direction. The degradation starts at the surface of MAPbI 3 and progresses in a layer-by-layer fashion (see figure 3). To prevent this degradation, the same authors proposed to protect the HP surface with a 2D material, namely h-BN. They showed that h-BN/HP/h-BN stack does not show structural change after 30 min of continuous heating at 85 • C, in stark contrast to the rapid emergence of the trigonal phase within 1 min and the complete transformation to trigonal PbI 2 within 7 min of the same thermal treatment for the non-encapsulated HP. Following this discovery, BN has been used to prevent degradation of inorganic HP (CsPbI 3 ) laser devices [88].
Beside the surface morphology protection, 2D materials are also very effective to prevent ion migration [14]. To clarify this role, Rand and co-worker interfaced 2D single crystal perovskite with a graphene layer in a field-effect transistor (FET) configuration [89]. They revealed that iodide loss is an important degradation Table 1. Numerical value of maximum efficiency of PSC obtained by using 2D materials in specific part of the cells (as for figure 2). References to the paper where the work is performed are reported.   pathway of 2D perovskite single crystals which can be suppressed by covering perovskites with graphene, thus significantly improving 2D perovskite stability. The ability of 2D material to prevent the ion diffusion and the reaction between adjacent layer has been exploited by Graetzel group in one of the most successful attempt to improve the stability of mesoscopic PSC [90]. Here, the operational stability (light soaking at maximum power point and temperature equal to 60 • C) of a mesoscopic PSC with Copper thiocyanate (CuSCN) as Hole transporting Layer (HTL) and Au as top electrode was increased from few hours till 1000 h by placing Reduced Graphene Oxide (RGO) interlayer between CuSCN and Au (figure 4).
MoS 2 was also used to improve the operational stability of inverted PSC beyond the state of art [91]. The photovoltaic performances of the HTL/MoS 2 /PSK/ETL retained 80% of their initial performance after 568 h of continuous stress test under 1 SUN light and at MPP in ambient conditions.
Regarding inverted PSCs, a two-fold approach has been recently implemented by using ultra-thin Bi 2 Te 3 flakes to dope the ETL and to form, at the same time, a protective interlayer above the ETL. Bi 2 Te 3 -based PSCs reached a PCE up to 19.46% and retained more than 80% of their initial PCE, after a burn-in phase, over 1100 h under continuous 1 Sun illumination [92].
Recently, 2D organic-inorganic halide perovskites have been attracting considerable attention because of their unique performance and enhanced stability for photovoltaic solar cells [93], photoluminescent devices [94,95] and photodetector [96]. The layered 2D perovskites show a wide tunability of optoelectronic properties [97] both from chemical and physical point of view [98]. In fact, the band gap can be systematically tuned from the UV to the near-infrared region even by controlling the nanoscale quantum confinement. Moreover, the 2D halide perovskites are direct band gap semiconductors regardless of the thickness and the number of metal halide octahedral layers between the two layers of the larger cations (mostly large size or long chain organic molecules). Lastly, the 2D perovskite showed increased stability that has been attributed to stronger van der Waals interaction between the capping organic molecules and the [PbI 6 ] unit, since the organic molecule are more strongly bound to the 2D perovskite than their MAI based 3D-structured counterpart. In addition, the longer alkyl chain of bulkier organic component ensures higher perovskite hydrophobicity when compared to the MA + ions by improving the moisture stability [99]. Although the first generation 2D PSCs have shown relatively lower photovoltaic performance [100], recent reports suggest that they are also capable of achieving high power conversion efficiency well beyond 20% [101]. In this context, a synthetic protocol using Butylammonium (BA) as spacer has been developed in order to achieving the layered perovskites in the pure form, with a reduced number of layer, named 2D/quasi-2D perovskite. A record PCE of 18.2% was reported for 2D perovskites employing a BA-derivative as spacer cations [102], while the unencapsulated devices sustain over 82% of their initial efficiency after 2400 h under relative humidity of ≈40%. The robust performance of perovskite solar cells results from the quasi-2D perovskite films was demonstrated to be related to the hydrophobic nature, the high degree of electronic order and high crystallinity.
Mixed composite of 3D and 2D perovskite phases, known as 2D/3D perovskites emerged as a class of excellent photovoltaic material with long-term stability. The 2D/3D multi-dimensional perovskite interface revealed a bright example of IE by increasing the efficiency and stability of perovskite photovoltaics. In fact, the 2D layer acts as a protective interlayer preserving the efficient 3D perovskite form degradation. 2D perovskite has been included in an effective way in the 3D perovskite composition (2D/3D perovskite) [70] as well at the interface with HTL [103] and ETL [73] with efficiencies well exceeding 20% and extended operational stability under continuous light soaking (1-SUN) with respect to 3D perovskite. Grancini et al demonstrated the unique combination of 2D/3D perovskite with a carbon-based architecture resulting in a large area modules able to sustain a continuous light-soaking for 10.000 h without degradation of PCE [104].
The scaling-up of the perovskite PV technology, together with high efficiency, low production costs and long-term operational stability, is one the main challenges to push perovskite solar cell toward industrial levels. A successful scaling-up from laboratory scale cells to large area modules and panels accounts for a uniformly coat compact perovskite absorber (no pinholes), good grain morphology and optimized interfaces between adjacent layers. In fact, it is imperative to develop scalable fabrication processes capable of retaining the ability of nucleation and subsequently crystal growth processes.
The impact of 2D materials on efficiency and stability of PSCs and the ability to adopt conventional HP printing processes for the deposition of such 2D materials has been shown to be valid also for large area devices such HP solar modules [62]. Recently, graphene and functionalized MoS 2 has been used for engineer both interfaces between HP and electrodes in large aera modules, achieving PCEs of 13.4% and 15.3% on active areas of 108 cm 2 and 82 cm 2 , respectively (figure 5). These modules showed a remarkable stability under prolonged (>1000h) thermal stress test at 65 • C (ISOS-D2), representing a crucial advancement in the exploitation of perovskite photovoltaic technology [105]. Notably, the proposed graphene IE strategy coupled with the use of stable polymeric hole transporting layer, allowed to realize perovskite large area module (active area >80cm 2 ) complying with prolonged 80 • C thermal stress as requested by standard ISOS D2 protocol [106].

Conclusions
Two-dimensional materials ranging from graphene to transition metal dichalcogenides and MXenes, including also 2D halide perovskites, have been demonstrated to be very effective in mastering efficiency and stability of perovskite solar cells and modules. 2D materials used as interface layers or dopant improve charge transfer, increase mobility, improve band alignment, tune work-function and limit ion diffusion. The beneficial role of 2D materials on efficiency and stability coupled to the availability of a wide library of 2D material and to the possibility to adopt conventional large area printing processes represent the hint for the development of a stable perovskite photovoltaic technology and to reduce the time to the market of perovskite solar cell technology. Moreover, some 2D materials have proven to be successful in replacing transporting layer and electrodes. This, combined with the recent development of high absorbing 2D halide perovskite and the realization of semi-transparent 2D electrodes based on GRM could pave the way for a full 2D photovoltaics [107].