The PSC’s working mechanism includes the absorption of photons ^[39] from the absorber/perovskite and the transfer and extraction of charge carriers to the respective electrodes. Specifically, photons’ absorption is followed by the presence of excited electrons in the conduction band and holes in the valence band of the perovskite. The electrons are injected into ETL and migrate to the anode. Subsequently, the electron passes through the external circuit to the cathode. At the same time, the holes are transferred through the hole transport layer (HTL) and migrate to the cathode, where electrons and holes recombine (

Figure 2b). Charge carrier extraction occurs at the perovskite/ETL and perovskite/HTL interfaces and, then, at the ETL/anode and HTL/cathode interfaces. Thus, the overall photovoltaic performance of PSCs is greatly affected by the properties of these interfaces. Two types of PSC device architectures exist, depending on the ETL and HTL ordering. The case of the ETL, which is an n-type semiconductor, being placed before the perovskite and the HTL, which is a p-type semiconductor, after perovskite attributes to the n–i–p structure, while the opposite attributes to the p–i–n structure ^[40][41][42]. In the case of the n-i-p type PSCs, planar ^[43][44] or mesoporous ^[14][38] configurations can be distinguished, depending on the morphology of the ETL material (

Figure 2c,d). Nevertheless, in addition to the perovskite absorber deposited on the ETL, a complete PSC device also comprises the electron collection substrate, usually a conducting oxide glass such as fluorine-doped tin oxide (FTO) or indium tin oxide (ITO) and the hole transporting layer (HTL) in contact with a metallic (Au, Ag) or carbon upper electrode ^[45][46].

Nowadays, PSCs have already reached a high efficiency in a very short period of research activity; however, there are other factors that also need to be addressed. First of all, during the charge carriers’ transfer and extraction through the successive layers, excess of carriers are not collected and recombine (radiatively or nonradiatively) at the interfaces ^[47][48]. Other challenges in the field of PSCs are the thermal and chemical stability ^[49] of the constituent materials and the overall device, hysteresis phenomenon ^[50][51], intrinsic and surface defects ^[52][53], enhanced charge carriers’ mobility and lead’s toxicity. Towards the mitigation of these problems a number of material and interface engineering approaches have been proposed including the integration of reduced graphene oxide as additive in the ETL, the perovskite and the HTL, the passivation of the ETL’s surface with metallated porphyrins and organic dyes, the modification of the titania ETL with transition metals, such copper and niobium, the passivation of perovskite layer’s surface with formamidinium iodide solution in isopropyl alcohol, and the addition of 4-tert-butylpyridine (tBP) in perovskite precursor as surface modification agent ^{[17][18][19][20][21][22][23][24][25][26]}. The incorporation of graphitic carbon nitride has been proved to be an efficient method for controlling crystal growth, passivating defects and reducing charge carriers’ recombination rate, and, thus, confronting many of the above-mentioned challenges.

3. Properties of g-C₃N₄

Intense theoretical and experimental interest has been focused on materials with tunable properties. Among these, carbon nitrides have attracted great attention due to their relative stability, structure and physicochemical properties ^[54]. In particular graphitic carbon nitride (g-C

₃

₄

₃

₃

₆

₇ (melem) planar units create a layered structure similar to graphite. The typical XRD peaks are placed at 2θ ~27° and 2θ ~13° ^[55], from which the first is related to interlayer stacking of the (002) melem planes and the second is attributed to the (100) planes corresponding to inplane ordering of the nitrogen-linked heptazine units. As in the case of graphite, the C

₃

₃

₆

₇

² C-N hybridization presenting intrinsic structural vacancies surrounded by nitrogen atoms ^[56]. Depending on the precursor and the conditions of the pyrolysis used, the 

g

₃

₄ powders have the possibility to result in different structures from highly crystalline samples to bulk (high temperature) and tubular (low temperature) samples ^[57][58]. In fact, DFT calculations were performed on carbon nitride single-layer to evaluate the structural and optoelectronic properties. The obtained results are consistent with a layered structure of slightly buckled heptazine cores, where the observed distortion can be attributed to electrostatic repulsions on the triazine-rings involving the two-coordinated N atoms with lone pair electrons ^[59]. Density of states (DOS) calculations (

Figure 3

₃

₄

Figure 3b) ^[60][61].

Figure 3.

 (

a

) g-C

₃

N

₄

 band structure from density of states (DOS) calculations and (

b) analysis of partial density of states (PDOS) comprising C and N orbitals. Reproduced with permission from ^[61], Copyright 2018, Elsevier.

) analysis of partial density of states (PDOS) comprising C and N orbitals. Reproduced with permission from [42], Copyright 2018, Elsevier.

So far, numerous methods have been used in order to modify the surface of bulk graphitic carbon nitride for different purposes. Chemical and thermal exfoliation ^[62] has been used in order to increase the active surface area as well as the optical properties. It becomes apparent that the exfoliation of bulk graphitic carbon nitride into 2D structure yields in an elevated bandgap, in agreement with the observed color modification of the samples to light yellow (from dark ones). ^[63][64]. Doping with heteroatoms ^[65] leads to an enhancement of the electrical conductivity and photoactivity due to the substitution of C or N by B, S, and Fe atoms of different valence in the honeycomb lattice. Such substitution introduces supplementary states in the density of states and usually involves nitrogen and boron as doping elements as their valence number is closer, in comparison to carbon atoms ^[66]. Finally, the introduction of functional groups leads to improved electrochemical properties as a result of changes in conductivity resulting from delocalized π-electrons, adsorbed species and defect sites ^[67]. Thus, the research community has focused on graphitic carbon nitride (g-C

So far, numerous methods have been used in order to modify the surface of bulk graphitic carbon nitride for different purposes. Chemical and thermal exfoliation [88] has been used in order to increase the active surface area as well as the optical properties. It becomes apparent that the exfoliation of bulk graphitic carbon nitride into 2D structure yields in an elevated bandgap, in agreement with the observed color modification of the samples to light yellow (from dark ones). [89,90]. Doping with heteroatoms [91] leads to an enhancement of the electrical conductivity and photoactivity due to the substitution of C or N by B, S, and Fe atoms of different valence in the honeycomb lattice. Such substitution introduces supplementary states in the density of states and usually involves nitrogen and boron as doping elements as their valence number is closer, in comparison to carbon atoms [92]. Finally, the introduction of functional groups leads to improved electrochemical properties as a result of changes in conductivity resulting from delocalized π-electrons, adsorbed species and defect sites [93]. Thus, the research community has focused on graphitic carbon nitride (g-C

₃

N

₄) as a metal free, nontoxic and VLA (visible light activated) photocatalyst with a unique 2D structure, outstanding chemical stability and tunable physicochemical properties ^[68][69]. Compared to other materials reported in literature for PSCs engineering, the advantages of g-C

) as a metal free, nontoxic and VLA (visible light activated) photocatalyst with a unique 2D structure, outstanding chemical stability and tunable physicochemical properties [55,56]. Compared to other materials reported in literature for PSCs engineering, the advantages of g-C

₃

N

₄

 can be attributed to its tunable structural and optoelectronic properties which can be optimized to the needs of the PSC technology. The carbon nitride material, characterized by easy and low-cost fabrication, also disposes high chemical and thermal stability that is required for the long-term operation of robust PSC devices. In general, the incorporation of bulk or modified g-C

₃

N

₄

 into solar cells is desirable in order to achieve highly stable and efficient performances.

4. Incorporation of g-C₃N₄ in PSCs

Recent studies have shown that nitrogen-rich precursors demonstrated enhanced PCE for PSCs. It seems that the incorporation of nitrogen functionalities has resulted in effective passivation of organic–inorganic halide perovskite materials ^[68][69]. In 2017, Lee et al. ^[70] reported that the defect sites of methylammonium lead tribromide (MAPbBr

Recent studies have shown that nitrogen-rich precursors demonstrated enhanced PCE for PSCs. It seems that the incorporation of nitrogen functionalities has resulted in effective passivation of organic–inorganic halide perovskite materials [55,56]. In 2017, Lee et al. [94] reported that the defect sites of methylammonium lead tribromide (MAPbBr

₃) were passivated through coordinate bonding between the nitrogen atoms and undercoordinated lead ions, by using amine-based treatment on top of the perovskite film. In another study by Hsieh et al. ^[68], they used urea and thiourea, previously dissolved in DMF at a concentration of 25 wt% and stirred overnight, as additives in MAPbI

) were passivated through coordinate bonding between the nitrogen atoms and undercoordinated lead ions, by using amine-based treatment on top of the perovskite film. In another study by Hsieh et al. [55], they used urea and thiourea, previously dissolved in DMF at a concentration of 25 wt% and stirred overnight, as additives in MAPbI

_3−x

Cl

_x

 and MAPbI

₃

 perovskite solar cells. In both cases, the PCE was improved and the MAPbI

₃

 based PSC presented a PCE of 18.8% (from 14.6% in the reference cell). The authors characterized the additives containing perovskite absorbers using UV-vis spectroscopy, scanning electron microscopy and grazing incidence wide angle X-ray diffraction measurements and attributed the observed performance enhancement to reduced grain boundaries of MAPbI

₃

 hindering charge recombination and allowing the effective charge carriers (electrons and holes) transport. Similarly, by incorporating 4mol% of urea (a bifunctional nonvolatile Lewis base) in MAPbI

₃, Lee et al. ^[71] obtained significantly enhanced photoluminescence lifetime accompanied by suppression of the trap-related nonradiative recombination. XRD analysis and FTIR spectroscopy were used to monitor both the perovskite growth and intramolecular interactions. In parallel, time-resolved photoluminescence (TRPL) measurements permitted the determination of charge-carrier lifetime and trap densities. Thus, the interaction of urea with the perovskite precursors in solution was confirmed. The urea additive retards growth and enhances perovskite crystallinity, and then passivates defects in the absorber by precipitating at the grain boundaries.

, Lee et al. [95] obtained significantly enhanced photoluminescence lifetime accompanied by suppression of the trap-related nonradiative recombination. XRD analysis and FTIR spectroscopy were used to monitor both the perovskite growth and intramolecular interactions. In parallel, time-resolved photoluminescence (TRPL) measurements permitted the determination of charge-carrier lifetime and trap densities. Thus, the interaction of urea with the perovskite precursors in solution was confirmed. The urea additive retards growth and enhances perovskite crystallinity, and then passivates defects in the absorber by precipitating at the grain boundaries.

Graphitic carbon nitride is a pyrolysis product of nitrogen-rich precursors (

Figure 4

). Its incorporation in PSCs has attracted scientific attention only in the few last years. However, the up-to-now findings appear to be promising and of high interest for the emerging perovskite solar cells.

Figure 4.

 Schematic diagram of a typical pyrolysis procedure for g-C

₃

N

₄

 fabrication.

4.1. g-C

₃

N

₄

 Modified Perovskite Absorbers

In the few past years many studies have focused on the incorporation of g-C

₃

N

₄ into the perovskite layer as a passivation strategy for the improvement of the crystallinity, which is a complicated process, the conductivity and, thus, the overall performance of the devices. Towards this strategy, the work from Liao et al. ^[72] refers to the incorporation of a small amount of g-C

 into the perovskite layer as a passivation strategy for the improvement of the crystallinity, which is a complicated process, the conductivity and, thus, the overall performance of the devices. Towards this strategy, the work from Liao et al. [96] refers to the incorporation of a small amount of g-C

₃

N

₄

 derived by standard pyrolysis of dicyandiamide, into the perovskite ink, which revealed impressive PCEs up to 21.1%. The addition of g-C

₃

N

₄

 led to compact perovskite films with passivated defects, large grain sizes, improved conductivity, and high crystallinity (

Figure 5

a–e), as confirmed by SEM, FTIR, XPS, and c-AFM measurements, which led to devices with considerably low charge recombination rate. The g-C

₃

N

₄

 modified PSCs are characterized by extremely low hysteresis accompanied by enhanced stability. In fact, after 500 h continuous illumination under one sun, only about 10% decline in devices PCE was observed (

Figure 5

f).

Figure 5.

 Top-view SEM images of the perovskite films based on (

a

) control MAPbI

₃

 and (

b

) 0.1 wt% g-C

₃

N

₄

 modified MAPbI

₃

 film, conductive AFM images of (

c

) control MAPbI

₃

 and (

d

) 0.1 wt% g-C

₃

N

₄

 modified MAPbI

₃

 films with a scanning area of 10 × 10 μm

²

, (

e

) XRD patterns of control and 0.1 wt% g-C

₃

N

₄

 modified MAPbI

₃

 films, and (

f

) operational stability of the encapsulated PSCs fabricated with or without g-C

₃

N

₄ under constant 1 sun illumination. Reproduced with permission from ^[72], Copyright 2019, The Royal Society of Chemistry.

 under constant 1 sun illumination. Reproduced with permission from [96], Copyright 2019, The Royal Society of Chemistry.

Similarly, in a previous work Jiang et al. ^[69] reported that the addition of 0.6 wt% pure sheets of g-C

Similarly, in a previous work Jiang et al. [56] reported that the addition of 0.6 wt% pure sheets of g-C

₃

N

₄

 produced by standard pyrolysis of urea into MAPbI

₃

 layer (

Figure 6

a,b) increased the crystallization and grain sizes of the perovskite film, proved by SEM characterization, as well as, its carrier mobility (

Figure 6

c). Thus, conductive AFM measurements confirmed the improvement of the corresponding perovskite absorber/spiro HTM interface. In parallel, the defects density was reduced ((

Figure 6

d) which consequently reduced hysteresis (

Figure 6

e) and increased the PCE from 16.22% to 19.49%. In this work the authors examined thoroughly the impact of different solvents, which were used to dissolve g-C

₃

N

₄

, on the perovskite’s crystal growth and managed to control the crystallization and reduce the charge recombination (

Figure 6f). Among the employed solvents (DMF, DMSO, ethanol), DMF did not show to significantly affect the cell parameters. In a similar vein, Wei et al. ^[73] prepared ultrafine E-g-C

f). Among the employed solvents (DMF, DMSO, ethanol), DMF did not show to significantly affect the cell parameters. In a similar vein, Wei et al. [97] prepared ultrafine E-g-C

₃

N

₄

 nanoparticles, using melamine precursor and annealing at 550 °C followed by intercalation in H

₂

SO

₄

 and fast stripping in NH

₃

 aqueous solution. The obtained nanomaterials, self-located at the MAPbI

₃

 grain boundaries via hydrogen bonding interaction (

Figure 7

), worked as prohibitory of electron-hole recombination. Thus, they were successfully incorporated into PSCs improving the device performance by 35%.

Figure 6.

 (

a

) Architecture of MAPbI

₃

:g-C

₃

N

₄

 based n-i-p PSC and (

b

) corresponding cross-sectional SEM image, (

c

) Hall mobility of MAPbI

₃

 and MAPbI

₃

:g-C

₃

N

₄

 in various solvents films, (

d

) current density–voltage (J–V) curves of hole-dominated devices with a structure of FTO/PEDOT:PSS/MAPbI

₃

 (with and without g-C

₃

N

₄

)/MoO

₃

/Ag, (

e

) J–V curves of MAPbI

₃

 and MAPbI

₃

:g-C

₃

N

₄

 based devices by forward and reverse scan, and (

f

) photoluminescence (PL) spectra of MAPbI

₃

 and MAPbI

₃

:g-C

₃

N

₄

 (DMF) films deposited on TiO

₂/FTO. Reproduced with permission from ^[69], Copyright 2017, Wiley-VCH GmbH.

/FTO. Reproduced with permission from [56], Copyright 2017, Wiley-VCH GmbH.

Figure 7.

 (

a

) Schematic illustration of the self-location process of E-g-C

₃

N

₄

 at MAPbI

₃

 grain boundaries by hydrogen bonding interaction. (

b

,

c

) Top-view FESEM images of MAPbI

₃

 film and E-g-C

₃

N

₄

 modified MAPbI

₃ film, respectively. Reproduced with permission from ^[73], Copyright 2019, Elsevier.

 film, respectively. Reproduced with permission from [97], Copyright 2019, Elsevier.

The presence of integrated graphite carbon nitride is shown by FESEM top-view images (

Figure 7

b,c) of the modified MAPbI

₃

 films demonstrating that most of the ultrafine E-g-C

₃

N

₄

 nanoparticles are located on the grain boundaries. During the crystallization process often a large number of electron trap sites is created, leading to electron loss and charge recombination. The ultrafine exfoliated graphitic carbon nitride (E-g-C

₃

N

₄

) nanoparticles are covered with N-H or O-H groups and, thus, they are easily adsorbed by forming hydrogen bonds with N-H bonds present on the perovskite grain boundaries. These nanoparticles were applied as intermediate materials between the perovskite absorber and the HTM) and resulted in increased PCE by passivizing the defects at grain boundaries.

Yang et al. in 2019 ^[74] implemented g-C

Yang et al. in 2019 [98] implemented g-C

₃

N

₄

 into carbon-based PSCs and reported that the addition of g-C

₃

N

₄

 into the precursor solution of the perovskite can result in improved surface morphology and larger grain size, and, thus, obtain high crystal quality of perovskite film. Specifically, the addition of 0.5wt% g-C

₃

N

₄

 reduced the root-mean-square roughness (RMS) from 15.3 nm, for the pristine perovskite, to 11.5 nm, and increased the average grain size from 150 nm, for the pristine one, to 270 nm determined by SEM and AFM measurements. The PCE of carbon based PSCs was significantly improved from 10.5% to 12.8% as a result of crystal quality improvement. In addition to this, in order to reduce the charge recombination, they spin-coated an Al

₂

O

₃

 insulating layer on the ETL surface. This modification at the interface between ETL and perovskite reduced the charge carrier recombination and resulted in an increased PCE of 14.3%.

Li et al. in 2019 ^[75] employed various passivation agents as perovskite additives including sulfonic, amino, nitrato, and hydroxy organic groups incorporated in g-C

Li et al. in 2019 [99] employed various passivation agents as perovskite additives including sulfonic, amino, nitrato, and hydroxy organic groups incorporated in g-C

₃

N

_4,

 prepared using melamine as precursor following a heat treatment process and further cure in sulfuric acid. The authors investigated the materials morphology with top-view SEM, their optoelectronic properties with steady state (PL), time-resolved (TRPL) photoluminescence and UV–vis absorption spectra, as well as the photovoltaic performance of the resulting PSCs with current density–voltage (J–V) curves. Improved nucleation and crystallinity were observed following successful g-C

₃

N

₄

 functionalization of the perovskite expressed by grain size increase, energy disorder decrease and efficient passivation of the absorber. The devices based on perovskites functionalized by graphitic carbon nitride present PCE values as high as 20.08%, clearly outperforming those without passivation. This was understood in terms of greater charge mobility and was further supported by Nyquist plots (

Figure 8a) confirming higher values of charge recombination resistance. Furthermore, the doping of iodine in graphite carbon nitride and its incorporation to the triple cation perovskite film was investigated by Cao et al. ^[76], who synthesized g-CNI by direct mixing dicyandiamide with ammonium iodide and subsequent heat treatment at 550 °C. As a result, they obtained perovskite films with high crystallinity and low trap states density. Improved photovoltaic performance of the corresponding PSCs was reported, attributed to the interaction of the iodide of the g-CNI group with the under-coordinated Pb ions on both the surface and at the grain boundaries of perovskite absorber (

a) confirming higher values of charge recombination resistance. Furthermore, the doping of iodine in graphite carbon nitride and its incorporation to the triple cation perovskite film was investigated by Cao et al. [100], who synthesized g-CNI by direct mixing dicyandiamide with ammonium iodide and subsequent heat treatment at 550 °C. As a result, they obtained perovskite films with high crystallinity and low trap states density. Improved photovoltaic performance of the corresponding PSCs was reported, attributed to the interaction of the iodide of the g-CNI group with the under-coordinated Pb ions on both the surface and at the grain boundaries of perovskite absorber (

Figure 8

b). Following optimization of the g-CNI concentration, the modified devices showed PCE as high as 18.28%, significantly higher than that of the reference cell. Based on XRD measurements showing that the lattice diffraction peaks (in shape and position) of the two perovskites (modified and reference) are very similar, the authors suggested that g-CNI was not inserted in the perovskite lattice but is located on the material surface or at the grain boundaries.

Figure 8.

 (

a

) Nyquist plot of control and devices passivated by functionalized C

₃

N

₄. Reproduced with permission from ^[75], Copyright 2019, Wiley-VCH GmbH. (

. Reproduced with permission from [99], Copyright 2019, Wiley-VCH GmbH. (

b) Schematic diagram of the mechanism of g-CNI modified PSCs. Reproduced with permission from ^[76], Copyright 2019, The Royal Society of Chemistry. (

) Schematic diagram of the mechanism of g-CNI modified PSCs. Reproduced with permission from [100], Copyright 2019, The Royal Society of Chemistry. (

c

) Schematic illustration of charge transfer between CsPb(Br

_x

I

_1−x

)

₃

 nanocrystals and g-C

₃

N

₄. Reproduced with permission from ^[77], Copyright 2020, Wiley-VCH GmbH.

. Reproduced with permission from [101], Copyright 2020, Wiley-VCH GmbH.

In order to further study g-C

₃

N

₄ and its effect on energy/charge transfer, Sheng et al. ^[77] prepared heterostructures of CsPb(Br

 and its effect on energy/charge transfer, Sheng et al. [101] prepared heterostructures of CsPb(Br

_x

I

_1−x

)

₃

 and g-C

₃

N

₄

 nanosheets with the standard procedure from melamine and reported an enhancement of charge separation and transportation due to special band alignments (

Figure 8

c) revealed by stead-state photoluminescence (PL) spectra, time-resolved PL spectra, and photodegradation- and temperature-dependent PL results. Compared to other heterostructures, the heterostructures formed by perovskite nanocrystals (PNCs) and g-C

₃

N

₄

 nanosheets (CN) not only broaden the absorption spectra, but also enhance the absorption capacity. The efficient electron–hole separation by charge transfer lead to a significant charge transfer efficiency up to 98.16%. In general, the utilization of g-C

₃

N

₄

 as additive or in combination with functionalized groups as a passivation strategy improves significantly the overall photovoltaic performance by suppressing charge recombination.

4.2. g-C

₃

N

₄

 as Charge Carrier Transport Material

The most significant research efforts have focused on the addition of g-C

₃

N

₄

 into the perovskite absorber, however, lately a few studies came out using it as electron transport layer (ETL) or/and hole transport layer (HTL). As the number of investigations dealing with the application of g-C

₃

N

₄ in the ETL of PSCs increases, more promising results are expected by this emerging, innovative and most of all promising technological approach. In 2019 for the first time, to the author’s knowledge, Chen et al. ^[78] developed a hybrid ETL nanocomposite consisting of SnO

 in the ETL of PSCs increases, more promising results are expected by this emerging, innovative and most of all promising technological approach. In 2019 for the first time, to the author’s knowledge, Chen et al. [51] developed a hybrid ETL nanocomposite consisting of SnO

₂

 with g-C

₃

N

₄

 quantum dots (QDs). The synthesis includes treatment with H

₂

SO

₄

 and HNO

₃

 acids as well hydrothermal cure in autoclave with NH

₃

·H

₂

O in order to form porous graphitic carbon nitride quantum dots (g-CNQDs). In this way multiple benefits were achieved. Firstly, the intrinsic interface crystal defects of SnO

₂

 were reduced.

In fact, the lone-pair electrons of nitrogen can interact with the under-coordinated Sn to effectively passivate the defects of SnO

₂

 related to oxygen-vacancies (

Figure 9

a). This resulted in a PCE exceeding 22%, which was attributed to the ability of g-C

₃

N

₄

 for trap-healing, its high resistance to corrosion and its significant chemical stability. In addition, devices based on the G-SnO

₂

 electron transport layer show exceptional long-term stability, retaining 90% of their initial performance when they are stored for 1500 h in a humid environment (with about 60% humidity). In conclusion, the utilization of g-C

₃

N

₄ as ETL in perovskite solar cells enhances significantly both the performance and the long-term stability. In a following study, Liu et al. ^[79] also used graphitic carbon nitride quantum dots (g-CNQDs) as modifying reagent at the SnO

 as ETL in perovskite solar cells enhances significantly both the performance and the long-term stability. In a following study, Liu et al. [102] also used graphitic carbon nitride quantum dots (g-CNQDs) as modifying reagent at the SnO

₂

/perovskite interface (

Figure 9

b). The g-CNQDs were synthesized by a heat treatment of urea and sodium citrate. This led to the decrease of the SnO

₂

 surface roughness, less grain boundaries and the facilitation of the crystal growth of perovskite absorber (

Figure 9

c,d) as detected by SEM measurements. The best PSC exhibited a PCE of 21.23% with negligible hysteresis. Furthermore, the long-term stability of the device was evaluated after being stored in ambient air under room light with a humidity of ≈30% upon 30 days’ exposure. Impressively, the PCE of g-CNQD modified PSC kept over 90% of initial performance after 30 days. Another group innovated by modifying the interfaces of ETL/perovskite and perovskite/HTL with the addition of g-C

₃

N

₄ derived from urea. Liu et al. ^[80] addressed that the g-C

 derived from urea. Liu et al. [103] addressed that the g-C

₃

N

₄

 modification does not exhibit desirable band alignments among ETL, HTL, and MAPbI

₃

 perovskite absorber at the two interfaces; the authors observed an outstanding PCE increase of 19.69% for the dual incorporated PSC over the 18.03% of the reference device. The performance improvement was assigned to the reduction of the trap density proved by TEM and SEM evaluation at both ETL/perovskite and perovskite/HTL interfaces. Overall, the addition of g-C

₃

N

₄

 either as additive in the perovskite and/or as a modification at the ETL/HTM transport layer leads to higher efficiencies (

Table 1

) and long-term stability of the perovskite solar cells.

Figure 9.

 (

a

) The side view for the charge density difference of SnO

₂

 (up) and G-SnO

₂ (down) with oxygen vacancy, reproduced with permission from ^[78], Copyright 2020, Elsevier. (

 (down) with oxygen vacancy, reproduced with permission from [51], Copyright 2020, Elsevier. (

b

) Schematic illustration of the PSCs with g-CNQD modified SnO

₂

 layers, and top-view SEM images of perovskite films based on (

c

) SnO

₂

 and (

d

) SnO

₂/g-CNQD ETL. Reproduced with permission from ^[79], Copyright 2020, The Royal Society of Chemistry.

/g-CNQD ETL. Reproduced with permission from [102], Copyright 2020, The Royal Society of Chemistry.

Table 1.

 Recent development of g-C3N4 based PSC photovoltaic performance.

Structure	Jsc (mA∙cm	−2	)	Voc (V)	FF (%)	PCE (%)	Ref.	Year

FTO/compact TiO	2	/g-C	3	N	4	modified MAPbI	3	/spiro-OMeTAD/MoO	3	/Ag	24.31	1.07	74.0	19.49	[69]	[56]	2018
FTO/compact TiO	2	/g-C	3	N	4	modified MAPbI	3	/spiro-MeOTAD/Au	23.00	1.16	79.0	21.10	[72]	[96]	2019
FTO/compact TiO	2	/exfoliated g-C	3	N	4	modified MAPbI	3	/spiro-OMeTAD/Au	23.20	1.10	62.0	15.80	[73]	[97]	2019
FTO/c-TiO	2	/m-TiO	2	/Al	2	O	3	/MAPbI	3	+ 0.5 wt% g-C	3	N	4	/carbon	23.80	1.00	60.1	14.34	[74]	[98]	2019
FTO/c-TiO	2	/m-TiO	2	/MAPbI	3	+ 0.5 wt% g-C	3	N	4	/carbon	24.00	0.92	58.2	12.85
FTO/TiO	2	/g-C	3	N	4	modified CsFAMAPbI	3−x	Br	x	/spiro-OMeTAD/Au	22.63	1.06	73.0	17.53	[76]	[100]	2019
FTO/TiO	2	/iodine doped g-C	3	N	4	modified CsFAMAPbI	3−x	Br	x	/spiro-OMeTAD/Au	22.97	1.07	74.0	18.28
ITO/PTAA/CsFAMAPbI	3−x	Br	x	+ NO	3	functionalized g-C	3	N	4	/PCBM/BCP/Ag	22.84	1.11	79.20	20.08	[75]	[99]	2019
ITO/PTAA/CsFAMAPbI	3−x	Br	x	+ SO	3	functionalized g-C	3	N	4	/PCBM/BCP/Ag	22.47	1.06	75.96	18.09
ITO/PTAA/CsFAMAPbI	3−x	Br	x	+ NH	3	functionalized g-C	3	N	4	/PCBM/BCP/Ag	22.56	1.07	76.96	18.58
ITO/PTAA/CsFAMAPbI	3−x	Br	x	+ OH functionalized g-C	3	N	4	/PCBM/BCP/Ag	22.37	1.08	75.68	18.28
ITO/g-C	3	N	4	QDs modified SnO	2	/CsFAMAPbI	3−x	Br	x	/Spiro-MeOTAD/Au	24.03	1.18	78.3	22.13	[78]	[51]	2020
FTO/SnO	2	/g-C	3	N	4	/MAPbI	3	/g-C	3	N	4	/Spiro-OMeTAD/Au	21.45	1.14	80.7	19.67	[80]	[103]	2020
FTO/SnO	2	/g-C	3	N	4	QDs/(FA/MA/Cs)PbI	(3−(x+Y))	Br	(x)	Cl	(y)	/Spiro-OMeTAD/Au	23.39	1.14	79.6	21.23	[79]	[102]	2020