1.

Introduction

Sustainable Energy Research initiatives have reached an important tipping point globally.
The damaging impact of the human-induced climate has been well established and
transitioning away from fossil fuel-based generation, towards renewables, is no longer an
'option'; it is mandatory. Among the many renewable energy sources available, solar
energy has emerged as one of the most compelling for meeting the dual challenge of
global climate change and energy security because; like all renewable energy sources,
no greenhouse gases are produced directly during the generation of electricity; and after
the installation of solar power systems, the negative environmental impact is limited (1).
Developing efficient, stable and cost-effective photovoltaic (PV) technologies is a critical
factor in achieving the transition to sustainable energy systems globally (2).

1.1 Physics of Perovskite Solar Cells

1.1.1 Fundamental Working Principle

A Perovskite Solar Cell (PSC) operates on the photovoltaic effect, converting photon
energy into electrical energy through a semiconductor absorber (the perovskite layer)
which is placed between an Electron Transport Layer (ETL) and a Hole Transport Layer
(HTL) (3). The process consists of three stages:

Photogeneration: When sunlight hits the semiconductor, it produces free charge pairs
(electrons 𝑒 − and holes ℎ+ ) due to the absorption of photons with energies greater than
the bandgap (Eg) which excite electrons within the semiconductor from Valence Band
(VB) to Conduction Band (CB). The low binding energy of excitons (<50 MeV) in the
perovskites allows thermal dissociation into free carriers at room temperature.

Charge Separation: The work function difference between the n-type contact ETL and p-
type contact HTL establishes a built-in electric field (Vin) which is essential to separate
charge carries before recombination. This electric field, via drift-diffusion transport, drifts
the electrons towards ETL and holes towards HTL.

Charge Collection: The ETL and HTL perform complementary functions of selectively
extracting electrons and holes respectively, while blocking the other charge carrier. The
extracted carriers are subsequently collected at FTO anode and metal cathode
electrodes, which generate current through the external circuit.
 Fig. 1. (a) Device structure of PSC. (b) Schematic representation of working and
charge transportation mechanism in electron transport layers of PSCs devices
(37)

1.1.2 The Illuminated Diode Equation

Mathematically, the electrical behaviour of illuminated solar cells is given by the

Illuminated Diode Equation, which superimposes light-generated current with the
reverse saturation current of the p-n junction:
𝑞𝑉
𝐽 = 𝐽𝑝ℎ − 𝐽0 (𝑒 𝑛𝑘𝐵 𝑇 −1)

Physical Interpretation of Terms:

𝐽 (Total Current Density): The net current density flowing out of the solar cell under load.

𝐽𝑝ℎ (Photocurrent Density): The current density generated by light absorption, which
represents the maximum available current from photoexcitation and carrier collection. It
depends on the absorption spectrum, quantum efficiency, and transport properties.

𝐽0 (Saturation Current Density): This represents recombination losses through radiative

and non-radiative pathways. A low 𝐽0 indicates superior device quality with fewer defects
and reduced non-radiative recombination via trap-assisted Shockley-Read-Hall (SRH)
processes.
𝑞𝑉
(𝑒 𝑛𝑘𝐵 𝑇 ): This describes the diode junction’s forward-bias diffusion current. As applied
voltage 𝑉 increases, this term grows exponentially, which represents the reverse current
opposing the photocurrent.

𝑞: Elementary charge (1.602 × 10⁻¹⁹ C); 𝑉: Applied voltage across the cell terminals.

𝑛(Ideality Factor): 𝑛 = 1 indicates ideal band-to-band recombination, while 𝑛 ≈

2 indicates trap-assisted recombination via mid-gap defect states which is caused by
point defects or grain boundaries. Perovskites typically exhibit 𝑛 between 1 and 1.5 in
high-quality samples.
𝑘𝐵 𝑇: Thermal energy (Boltzmann constant × Temperature). At 300 K, 𝑘𝐵 𝑇/𝑞 ≈ 0.0259 V.

1.1.3 Key Performance Parameters

Solar cells are evaluated using four standard metrics extracted from the J–V curve:

Short-Circuit Current (JSC )

Definition: The maximum current density when the cell operates at zero voltage (𝑉 = 0).

Physical Measure: At short circuit, the exponential term vanishes (e0 – 1 = 0), giving:

𝐽𝑆𝐶 ≈ 𝐽𝑝ℎ

Significance: 𝐽𝑆𝐶 quantifies the light-harvesting efficiency and carrier collection

capability. It is limited by the bandgap energy (narrower bandgaps absorb more photons)
and the external quantum efficiency (EQE), which is the fraction of incident photons
converted to collected electrons. Modern perovskites achieve 𝐽𝑆𝐶 values of 23–26 mA
cm⁻², subject to the AM1.5G solar spectrum and thermalization loss of hot carriers.

Open-Circuit Voltage (𝑉𝑂𝐶 )

Definition: The voltage at which the net current is zero (𝐽 = 0).

Physical Measure: Solving the diode equation for 𝑉 when 𝐽 = 0 yields:

𝑛𝑘𝐵 𝑇 𝐽𝑝ℎ
𝑉𝑂𝐶 = ln⁡ ( +1)
𝑞 𝐽0

Significance: 𝑉𝑂𝐶 represents the maximum voltage the cell can generate and is
fundamental to determining power output. It increases logarithmically with
photocurrent-to-saturation-current ratio and highly sensitive to recombination losses.
Defects in the perovskite layer increase 𝐽0 because of trap-assisted recombination,
causing 𝑉𝑂𝐶 to decrease logarithmically. This connection emphasizes how crucial defect
passivation is to maximizing device performance.

Fill Factor (FF)

Definition: A measure of the "squareness" of the J–V curve, which quantifies the ideality
of power extraction.

Formula:
𝑉𝑚𝑝 × 𝐽𝑚𝑝
𝐹𝐹 =
𝑉𝑂𝐶 × 𝐽𝑆𝐶

where 𝑉𝑚𝑝 and 𝐽𝑚𝑝 are the voltage and current density at the maximum power point.

Physical Measure: The performance of a perovskite solar cell is negatively affected by

series resistance (𝑅𝑠 ) arising from contact resistance, ETL/HTL conductivity limitations,
and metal electrode impedance, and also by shunt resistance (𝑅𝑠ℎ ) resulting from
leakage pathways such as pinholes and grain-boundary defects. When optimally
produced, laboratory-made perovskite solar cells have a FF of greater than 80%, with FF
values nearing 85%.

Power Conversion Efficiency (PCE)

Definition: The fraction of incident solar power converted to electrical power.

Formula:
𝑉𝑂𝐶 × 𝐽𝑆𝐶 × 𝐹𝐹
𝑃𝐶𝐸 =
𝑃𝑖𝑛

where 𝑃𝑖𝑛 = 1000 W m⁻² under standard AM1.5G test conditions.

Significance: The maximum efficiency for photovoltaic devices is found using the PCE.
Optimization of PCE requires simultaneous optimization of three competing factors: (i)
𝐽𝑆𝐶 must be maximized through improved light absorption and carrier collection
efficiency, (ii) 𝑉𝑂𝐶 must be maximized by reducing defects and providing optimal band
alignment at ETL/absorber and absorber/HTL interfaces, and (iii) FF must be maximized
by supressing resistive losses and leakage currents.

1.2 Silicon-based Solar Cells

Most of the energy produced in solar power is made up of silicon solar cells (more than
90% of installed capacity), indicating that they comprise a significant portion of the total
installed capacity of photovoltaic technology around the world (4). Despite this market
dominance, efficiency of silicon solar cells is limited by physical and thermodynamic
laws, governing radiative recombination in semiconductors , known as the Shockley-
Queisser (S-Q) limit, which sets a theoretical maximum PCE of approximately 29.4% for
single-junction devices operating under standard terrestrial illumination conditions (AM
1.5G spectrum). This represents an insurmountable barrier for conventional silicon
photovoltaic architectures. At present, commercially available monocrystalline silicon
solar cells typically have efficiencies between 25 to 27%, while cutting-edge R&D
prototypes have achieved maximum efficiencies of 26.81% in the laboratory (5).
Consequently, because commercial silicon solar cells already operate at or near the Q-
S limit, further improvements in single-junction silicon solar cell efficiencies are unlikely.

Silicon-based solar cells, beyond the efficiency constraint, present significant

environmental and cost-based challenges due to the manufacturing process The
process of purifying silicon to meet the rigorous requirements for solar cell production is
highly energy-intensive and requires extensive amounts of electricity to perform high-
temperature reduction of the silicon dioxide (quartz) material in electric arc furnaces. For
every metric tonne of metallurgical-grade silicon produced directly from quartz,
approximately 4.7 to 5 tonnes of CO2 (carbon dioxide) equivalents are released into the
atmosphere (6).

1.3 Third-Generation Solar Technologies and Perovskite Emergence

The photovoltaic research community has shifted much of its focus away from traditional
solar technologies towards developing "third generation" forms of solar energy systems
as a direct result of the limitations placed upon these systems by silicon (7). Third
generation solar technologies consist of a variety of new and innovative materials and
structures that are being explored as potential solutions for generating solar power (7).
Third-generation solar cells include organic photovoltaics (OPVs), quantum dot solar
cells (QDSCs), dye-sensitized solar cells (DSSCs), and, most prominently, metal-halide
perovskite solar cells (PSCs) (8).

The efficiency of perovskite solar cells has been on an unparalleled improvement track
compared to previous work on solar cells in the past few decades. Efficiency has
increased from approximately 3.8% in 2009 to over 26% at the time of 2025 for laboratory
single-junction solar cells, as well as nearly 34% for laboratory perovskite-silicon tandem
solar cells (9). Beyond their rapid efficiency gains, several properties of perovskites
position them as unrestrained alternatives:

- Low Manufacturing Intensity: Compared to silicon’s energy-intensive process of growing

and purifying its crystalline structure, perovskite solar cells can be manufactured using
low-energy methods of solution-based deposition at low temperatures (≤150°C) (10).

- Flexible Device Architecture: By changing the composition and the cation of the halides
in perovskite absorber layers, it is possible to tune the perovskite absorber layer to absorb
different wavelengths of light. Because of this tunability, perovskite cells can be
manufactured in tandem with other cells to form tandem arrays. Laboratory tandem
arrays of perovskite-silicon have achieved efficiencies exceeding 33%, exceeding the
theoretical efficiency of silicon alone (10).

- Thin-Film Technology: Because perovskite materials exhibit extremely high optical

absorption coefficients, the active layers of perovskite cells may be only 300–500 nm
thick compared to the 100+ micrometres of active layer material required for silicon,
therefore using much less cost and material to manufacture perovskite devices (11).

1.4 Lead Substitution Strategies in Perovskites

Perovskite crystals are made up of cations labelled A and B, and halide anions labelled X
in the (ABX3) stoichiometry (12). All perovskite crystalline materials will have similar and
unique physical characteristics which make them suitable for utilization in photovoltaic
applications through several ways. These include high optical absorption coefficients,
tunable visible and near-infrared spectrum bandgaps, long carrier diffusion lengths, and
high intrinsic defect tolerance compared to many traditional semiconductor structures
[13][14].

Although conventional lead-based perovskites are highly efficient at converting sunlight

into electricity, there are two major barriers to their widespread commercialization:
lead's (Pb) toxicity presents significant environmental and health hazards, and
conventional lead-based perovskites are unstable when subjected to environmental
stressors such as moisture, heat, and UV radiation [15][16]. When exposed to moisture,
water molecules can adsorb onto the perovskite's surface, thus altering interatomic
distances and creating irreversible pathways for chemical degradation [15].
Furthermore, exposure to UV and O2 will result in increased rates of oxidative degradation
[17].

Lead toxicity has prompted researchers to employ a variety of substitution methods to

mitigate the effects of lead. One of the simplest methods is to replace lead with other
isovalent elements from Group 14 on the periodic table. Tin (Sn) and Germanium (Ge) are
particularly attractive as direct lead substitutes from a Group 14 perspective because of
their chemical properties [18][19]. However, these elements also suffer from significant
drawbacks related to stability, as Sn²⁺ and Ge²⁺ ions are very unstable and readily oxidizes
to +4 oxidation states in air (under ambient temperature and humidity conditions).
[18][20][21]. The oxidation of Sn²⁺ and Ge²⁺ ions creates large numbers of p-type dopant
defects, which contribute to a rapid decline in device performance [18][21].

The double perovskite architecture is a more robust strategy than either partial or total
replacement of a lead ion through an element of the same valence. With this type of
heterovalent substitution, two lead cations (Pb²⁺) are replaced with two different types of
metal cations, one that is monovalent (B⁺), and another that is trivalent (B³⁺), or,
alternatively, with divalent and tetravalent (B²⁺ and B⁴⁺) metals [22][18]. As a result of
maintaining charge neutrality during the above-described cation substitutions, this
approach to heterovalent substitution permits the incorporation of elements that would
otherwise be intrinsically non-toxic, as described in references [22][18][23]. Examples of
successful double perovskites include Cs₂AgBiBr₆ and related Cs₂AgB'X₆ compounds
(where B' = Bi, Sb, In) , all of which possess excellent chemical stability, low toxicity, and
long carrier lifetimes [22][23].

The double perovskite crystal structure is defined by [A₂B'B''X₆], where the B'- and B''-site
cations coordinate with six (X) halide (anions) to form octahedral coordination
surrounding the cations as (B'X₆)⁵- and (B''X₆)³- [22]. The octahedra and the resulting three-
dimensional network are connected by corner and edge sharing of their (X) halide anions,
while the A-site cations occupy the voids of this octahedral framework [22]. The
preservation of the octahedral three-dimensional coordination geometry enables double
perovskites to exhibit some of the same advantageous optoelectronic characteristics as
traditional perovskites.
2. SCAPS-1D Simulation Methodology

The SCAPS-1D (Solar Cell Capacitance Simulator), developed by the Department of

Electronics and Information Systems (ELIS) at the University of Gent, Belgium, serves as
the main research tool in this study. This software provides a numerical simulation tool
that has all of the features needed for simulating thin film solar cells by providing the
ability to define up to 7 layers of semiconductors with properties such as bandgap,
electron affinity, dielectric permittivity, and defect density distribution.

The simulation solves three coupled differential equations self-consistently:

1. Poisson's Equation: Relating electrostatic potential ($\psi$) to charge distribution.

2. Continuity Equation for Electrons:

3. Continuity Equation for Holes:

Standard AM1.5G Illumination (1000 W/m2) and an initial temperature of 300K were used
for the simulations. The frequency of the AC analysis for the C-V measurements was set
to 106 Hz. The configuration of the device was the standard n-i-p architecture: glass / FTO
/ ETL / Absorber (LNMO/ENMO/DNMO) / HTL / Metal Contact.

3. Simulation Design

The numerical study was designed to systematically explore the combined influence of
transport layers and absorber composition on device performance within a SCAPS-1D
framework. For each of the three double perovskite absorbers, La₂NiMnO₆ (LNMO),
Eu₂NiMnO₆ (ENMO), and Dy₂NiMnO₆ (DNMO), a full combinatorial sweep was defined
over electron transport layers (ETLs) and hole transport layers (HTLs):

• 7 candidate ETLs

• 10 candidate HTLs

• 3 absorbers (LNMO, ENMO, DNMO)

ETL C60 PCBM Zn TiO SnO2 WS2 ZnS
O 2

Thickness 50 50 30 30 100 100 50

(nm)
Bandgap 1.7 2 3.3 3.2 3.6 1.8 3.6
(eV)
Electron 3.9 3.9 4.1 3.9 4 3.95 3.9
Affinity
(eV)
Dielectric 4.2 3.9 9 32 9 13.6 9
Permittivi
ty
Nc(cm-3) 8*101 2.5*102 101 101 2.2*10 1018 2.2*10
9 1 9 9 18 18

Nv(cm-3) 8*101 2.5*102 101 101 1.8*10 2.4*10 1.8*10

9 1 9 9 19 19 19

Ve (cm/s) 107 107 107 107 107 107

Vh (cm/s) 107 107 107 107 107 107
μe 8*10-2 0.2 20 20 100 100 100
(cm2/Vs)
μh(cm2/Vs 3.5*1 0.2 10 10 25 100 12.5
) 0-2
ND(cm-3) 1017 2.93*10 101 101 1017 1018 1.1*10
17 7 7 18

NA(cm-3)
Defect 1015 1015 101 101 1015 1015 1015
Density(c 4 4

m-3)
Referenc 23, 23, 24, 23, 27, 26 23, 26, 30
es 24, 25 25 26, 28, 25
24 26,
24,
29
HTL Cu P3 PE NiO C CF Cu Spiro CBT CuO
SN HT DO uI TS 2O _Me S
T:P OTA
SS D
Thick 50 50 50 100 20 10 20 150 100 50
ness 0 0 0
(nm)
Bandg 3.6 1.7 1.6 3.8 3. 2.1 2.1 2.9 1.9 2.17
ap 1 7 7
(eV)
Electr 1.7 3.5 3.4 1.46 2. 3.2 3.2 2.2 3.6 3.2
on 1
Affinit
y (eV)
Diele 10 3 3 10.7 6. 9 7.1 3 5.4 7.11
ctric 5 1
Permi
ttivity
Nc(cm 2.2 2*1 2.2 2.8*1 2. 2.2 2.0 2.2* 2.02*101
-3
) *10 021 *10 019 2* *10 2*1 1018 7
19 18
10 18 017
19

Nv(cm 1.8 2*1 1.8 1019 1. 1.8 101 1.8* 1.1*1019

-3
) *10 021 *10 8* *10 9 1018
18 18
10 19
19

Ve 10 107 107 107

(cm/s 7

)
Vh 10 107 107 107
(cm/s 7

)
μe 10 1.8 4.5 12 10 21. 20 10-4 30 10
(cm2/ 0 *10 *10 0 98 0
Vs) -3 -2

μh(cm 25 1.8 4.5 2.8 43 21. 80 10-4 10 0.1

2
/Vs) 6*1 *10 .9 98
0-2 -2
ND(c
m-3)
NA(cm 101 101 101 1018 10 101 101 1018 1018 1018
-3
) 8 8 8 18 8 8

Defec 101 101 101 1015 10 101 101 1015 1015 1015
t 5 5 5 14 5 5

Densi
ty
(cm-3)
Refer 23, 23 23 23 30 23, 30, 27, 26 30
enes 27, , 25 27 28,
30 27 31
,
23
,
24
,
29

LAYERS ENMO DNMO LNMO FTO ITO

Thickness 350 350 350 500 500
(nm)
Bandgap (eV) 1.06 1.03 1.05 3.5 3.5
 Electron 3.52 3.52 3.52 4 4
Affinity (eV)
Dielectric 9 13 3.5 9 9
Permittivity
Nc(cm-3) 1018 1018 1018 2.2*1018 2.2*1018
Nv(cm-3) 1018 1018 1018 1.8*1019 1.8*1019
Ve (cm/s) 107 107 107 107
Vh (cm/s) 107 107 107 107
μe (cm2/Vs) 22 22 22 20 20
μh(cm2/Vs) 22 22 22 10 10
ND(cm-3) 1019 1021
NA(cm-3) 7*1016 7*1016 7*1016
Defect 1014 1014 1014 1014 1015
Density (cm-
3
)
References 29 29 29, 23, 24 29, 24, 28, 23, 26
27, 30

This yields a total of

7 × 10 × 3 = 210
distinct device configurations to be evaluated in the full optimization space.

3.1 Baseline Reproduction and Model Validation

The initial step in validating the simulation environment was to duplicate the prototype
device from Kumar et al. (29), which consists of the FTO/TiO₂/X/CuI/Au (where X=LNMO,
DNMO, ENMO) structure. The exact structure number, layer thicknesses and material
characteristics (band energy level gaps, electron affinities, density of defects, mobility,
and doping) were obtained from the reference study and associated supplementary data.
The goal of this step was to confirm the numerical implementation by demonstrating that
the simulated values of power conversion efficiency (PCE), short-circuit current density
(Jsc), open circuit voltage (Voc), and fill factor (FF) were comparable to those reported by
Kumar et al.
3.2 Optimization Strategy

To avoid an exhaustive 210-point sweep, a staged optimization protocol was adopted.

Phase 1: HTL Screening at Fixed ETL

During the first part of the experiment, we fixed one type of ETL material (TiO₂) and kept
the LNMO as our photoactive layer or active region. Next, we tested different types of
materials to serve as our HTLs with all other variables being equal. For each type of our
10 HTL candidates:

• Fixed: Absorber = LNMO; ETL = TiO₂

• Variable: HTL (10 candidates)

Steady-state J–V characteristics under AM1.5G illumination were computed and we

selected the configuration having the highest PCE as the champion HTL, with 10 total
simulations performed in this first phase.

Phase 2: ETL Screening at Fixed HTL

After identifying champion HTL, this material was fixed while the ETL set was
systematically screened. TiO₂ was the reference material for the ETLs and has already
been evaluated in Phase 1, leaving six ETLs to be simulated in combination with LNMO
and champion HTL remaining:

• Fixed: Absorber = LNMO; HTL = champion HTL

• Variable: ETL (6 candidates, excluding the TiO₂ baseline)

Each configuration's PCE, Jsc, Voc, and FF were evaluated. The ETL that produced the
highest PCE was identified in combination with champion HTL; therefore, the
ETL/HTL/absorber stack for the ETL that produced the highest PCE with champion HTL
was established as the optimal device configuration for LNMO and claimed as champion
ETL/HTL. Therefore, this phase required six simulations.

Phase 3: Transferability to Other Absorbers

In order to evaluate whether the optimal ETL/HTL pair for LNMO can also be utilized for
the other two double perovskites ENMO and DNMO, the champion ETL/HTL
combinations obtained from the previous study of LNMO were directly replicated on
ENMO and DNMO. Two simulations were performed:

• Champion ETL/HTL + ENMO

• Champion ETL/HTL + DNMO

This stage evaluates the robustness of the optimization approach against absorber
replacements and besides checking if there is still a performance increase on LNMO
when changing the rare earth cation used. Overall, this hierarchical technique decreased
the practical amount of optimizations to complete; looking through and analyzing 210
theoretical configurations before arriving at just 1 (validation) + 10 (HTL tests) + 6 (ETL
tests) + 2 (transferability tests) = 19 targeted simulations; while still sampling the most
relevant region of the design space.

3.2.1 Rationale for Sequential Optimization Across Absorbers

In general, changing the absorber layer in a heterojunction device can strongly perturb
the band alignment at both ETL/absorber and absorber/HTL interfaces, since the
conduction band minimum (CBM) and valence band maximum (VBM) positions depend
on band gap and electron affinity (or ionization potential). Under such circumstances, an
ETL that is optimal for one absorber may introduce an energetic spike or cliff for another,
degrading charge extraction or enhancing interfacial recombination. For the present
material set, however, the underlying energy‑level landscape justifies the use of a
sequential, absorber‑agnostic optimization strategy. (32, 33)

In the set of parameters used for the three absorbers, the electron affinity (which
determines the conduction band minimum (CBM) in reference to the vacuum) is the
same for all three. The values are ENMO: 3.52 eV, DNMO: 3.52 eV, and LNMO: 3.52 eV.
Because the CBM determines the energy levels of the junction at ETL/absorber, this
means that all three absorbers have the same conduction band offset when placed
against a given ETL (32). Although the band gaps of the absorbers are slightly different —
ENMO has a gap of 1.06 eV, DNMO has a gap of 1.03 eV, and LNMO has a gap of 1.05 eV
— these small differences in band gap will result in only a small difference in the valence
band maxima (VBM) of the absorbers, provided the CBM is equal (as determined by the
common electron affinity). As a result, the relative alignments of the three absorbers
when placed against a given hole transport layer (HTL) will also be very similar. Thus,
LNMO, ENMO, and DNMO are electronically “near‑degenerate twins” in the context of the
initial simulation phase, and a transport‑layer stack optimized for LNMO is expected to
remain close to optimal for ENMO and DNMO as well (32).

3.2.2 Justification for Separating ETL and HTL Optimization

When performing sequential optimization, one of the possible issues is that there are
likely to be significant interactions between ETLs and HTLs so that optimizing them
separately may lead to missing significant interactions between them. The current
strategy is still valid due to three primary reasons:

(i) “Sandwich” principle and physical separation

In the models of perovskite-based solar cells, it has been observed that for thicknesses
of between 300 and 500 nm corresponding to experimentally observed thickness and
models of oxide double perovskites and hybrid perovskites, the ETLs and HTLs are
physically spaced by a bulk absorber layer at all times between them, therefore, the local
electrostatics at each interface are predominantly determined by the adjoining two
layers:

• An ETL interacts with the conduction band of an absorber (via a conduction

band coupling during operation)
• An HTL interacts with the valence band of an absorber (via a valence band
coupling during operation)

As a result, the ETL interacts only with the ETL/absorber interface while the HTL interacts
only with the absorber/HTL interface. Hence, if any HTL provides a poor valence‑band
alignment (e.g., excessive barrier for hole extraction or a deep recombination‑active cliff)
then all devices with this fault will perform poorly irrespective of which type of ETL is
placed on the HTL side (33,34).

(ii) Locality of band alignment

Band alignment is a intrinsically local interfacial property.

• At the ETL/absorber interface, the performance at this interface is primarily

affected by the conduction band offset, which is determined by the electron
affinity difference of the ETL and the absorber.
• At the absorber/HTL interface, the performance at this interface is primarily
affected by the valence band offset, which is related to the ionization potential
difference of the absorber and the HTL.

Because the two interfaces are physically separated by a comparatively thick absorber
layer, optimization of the ETL/absorber interface for favourable conduction‑band
alignment does not alter the optimization of the absorber/HTL interface’s valence‑band
alignment, and vice versa. This locality supports the validity of independently screening
HTLs at fixed ETL and then screening ETLs at fixed HTL (32).

(iii) Consistency with established SCAPS methodology

In SCAPS-1D modelling of perovskite, researchers typically fix a widely used transport

layer (such as TiO₂ as ETL) while screening several HTLs and then revisit the ETL once an
ideal HTL is found. This sequential optimization protocol is in line with standard practice.
Similar to this work, the reference double perovskite study optimizes absorber
parameters first and states clearly that ETL, HTL, and defect properties can be tuned later
to achieve further improvements (29).

3.2.3 Limits of the Decoupling Approximation

Although ETL and HTL can be considered effectively decoupled in this study, their
interaction becomes non-negligible in certain well-defined regimes:

(i) Optical interference effects

The changes in ETL thickness (or) ETL refractive index that cause the internal optical
interference patterns to be modified change the spatial generation profile and the portion
of incoming light incident on the back side of the absorber and HTL. Although, for
adjustments of nanometer scale, these effects may be minimal; however, for ultra-thin
or highly reflective stacks they may have considerable impact (35).

(ii) Ultra‑thin absorber regime

When the thickness of an absorber is lowered to the fully depleted regime (< 100 nm for
typical perovskites), the space‑charge region formed at the interface of ETL/absorber and
absorber/HTL interfaces can overlap, causing the internal electric field which is
influenced simultaneously by both contacts. This creates a situation where the ETL & HTL
cannot be optimized independently from one another, thus they must both be considered
when optimizing a perovskite PV. In the present work, however, the thickness of the
absorber (≈ 350 nm in the reproduced structure) is chosen large enough for it to not be
impacted by either one of the contacts alone, allowing for an independent optimization
strategy (34, 36).

The sequential optimization of ETL and HTL is thus both physically justified and
numerically efficient under the thickness and parameter ranges taken into consideration,
and it is completely consistent with documented best practices in SCAPS-based device
modelling.
References

[1] Yoshikawa, K., et al. (2017). Silicon heterojunction solar cell with interdigitated back
contacts for a photoconversion e ciency over 26%. Nature Energy, 2(11), 17032

[2] Polman, A., Knight, M., Garnett, E. C., Ehrler, B., & Sinke, W. C. (2016). Photovoltaic
materials – Present e ciencies and future challenges. Science, 352(6283), aad4424

[3] Green, M. A., Hishikawa, Y., Dunlop, E. D., Levi, D. H., Hohl-Ebinger, J., & Ho-Baillie, A.
W. (2019). Solar cell e ciency tables (Version 55). Progress in Photovoltaics, 27(7), 565-
575.

[4] LONGi Solar. (2023). "LONGi sets a new world record of 33.9% for the efficiency of
crystalline silicon-perovskite tandem solar cells." *News Release*.
[Link]
silicon-perovskite-tandem-solar-cells/
[5] Pan, Z., et al. (2023). "Silicon heterojunction solar cells with up to 26.81% efficiency via low-work-
function molybdenum carbide contact." *Nature Energy*, 8(5), 463–473.
[Link]

[7] Ben Dkhil, S. "What is the 'third generation' of photovoltaic?"

*Dracula Technologies*, March 2025. [Link]
[Link]/what-is-the-third-generation-of-photovoltaic/

[8] Osman, M. I., et al. (2024). "Perovskite solar cells: Progress,

challenges, and future avenues to commercialization." *Solar Energy*,
273, 112435. [Link]

[9] Lin, R., et al. (2025). "Solar cells that combine multiple perovskite
layers surpass 30% efficiency." *Nature*, 636, 527–533.
[Link]

[10] Rakštys, K., et al. (2025). "Next-generation perovskite solar cells

are closer to commercialization." *Nature Energy*.
[Link]
[Link]

[11] U.S. Department of Energy. "Perovskite Solar Cells." Solar Energy

Technologies Office. [Link]
cells

[12] Sharma, R., & Soni, A. (2024). "ABX₃ inorganic halide perovskites
for solar cells: Chemical and crystal structure stability." *Journal of
Materials Science*, 59, 5812–5843. [Link]
09537-y

[13] Manser, J. S., Christians, J. A., & Kamat, P. V. (2016). "Intriguing

Optoelectronic Properties of Metal Halide Perovskites." *Chemical
Reviews*, 116(21), 12956–13008.
[Link]

[14] Ma, F. et al. (2023). "Developments of Highly Efficient Perovskite

Solar Cells." *Accounts of Chemical Research*.
[Link]

[15] Yin, X., Song, Z., Li, Z., & Tang, W. (2022). "A critical review on
the moisture stability of halide perovskite films and solar cells."
*Journal of Energy Chemistry*, 67, 234–251.
[Link]

[16] Ahmad, S., et al. (2024). "The impact of moisture on the stability
and degradation of halide perovskite solar cells." *Materials Horizons*,
11(3), 878–920. [Link]

[17] Jiang, X., Khan, F., & Nazeeruddin, M. K. (2024). "Unraveling the
Heat- and UV-Induced Degradation of Mixed Halide Perovskite Thin Films
via Surface Analysis Techniques." *Langmuir*, 40(20), 10456–10468.
[Link]

[18] Ricerca (2023). "Lead-Free Halide Double Perovskites." *Advanced

Materials & Technologies*. Università degli Studi di Bari Aldo Moro.
[19] Pantaler, R. (2020). "Lead-free absorber materials for solar cell
applications." *Doctoral Dissertation*, University of Duisburg-Essen.
[Link]
[Link]/servlets/MCRFileNodeServlet/duepublico_derivate_00071963/Diss_Pa
[Link]

[20] Ramasamy, P., et al. (2022). "Tin perovskite solar cells with >1300
h of operational stability." *Nature Energy*, 7(6), 497–506.
[Link]

[21] Leijtens, T., et al. (2017). "Mechanism of Tin Oxidation and

Stabilization by Lead in Tin-Lead Perovskite Solar Cells." *ACS Energy
Letters*, 2(4), 970–975. [Link]

[22] Lei, J., Wang, Z., & Wu, Z. (2023). "Lead-Free Halide Double
Perovskite for High-Performance Photodetectors." *Nanomaterials*,
13(10), 1671. [Link]
[23] Sun, M., et al. (2023). "Synthesis and Optimization of Cs₂B'B''X₆
Double Perovskite Solar Cells." *Advanced Functional Materials*, 33(40),
2304591. [Link]

(23) M. K. Hossain et al., “Design insights of La₂NiMnO₆-based perovskite solar cell

employing different charge transport layers: DFT and SCAPS-1D frameworks,” Energy
Fuels, vol. 37, no. 17, pp. 13377–13396, 2023. doi: 10.1021/[Link].3c02361

(24) A. Raj, A. Anshul, V. Tuli, P. K. Singh, R. C. Singh, and M. Kumar, “Effect of appropriate
ETL on quantum efficiency of double perovskite La₂NiMnO₆ based solar cell device via
SCAPS simulation,” Mater. Today: Proc., vol. 47, pp. 1656–1659, 2021. doi:
10.1016/[Link].2021.04.581
(25) M. A. B. Siddique et al., “Performance evaluation of Eu₂NiMnO₆-based lead-free
perovskite solar cells: a SCAPS-1D study,” RSC Adv., vol. 15, pp. 35488–35509, 2025. doi:
10.1039/d5ra05366h

(26) M. K. Hossain et al., “Design and simulation of Cs₂BiAgI₆ double perovskite solar
cells with different electron transport layers for efficiency enhancement,” Energy Fuels,
vol. 37, no. 5, pp. 3957–3979, 2023. doi: 10.1021/[Link].3c00181

(27) A. Raj, M. Kumar, D. V. Singh, B. Singh, D. K. Dwivedi, and A. Anshul, “Physical

parameter optimization and band alignment approach for efficiency improvement in
Cs₂AgBiBr₆ based lead-free perovskite solar cells,” Sci. Rep., vol. 15, p. 32868, 2025. doi:
10.1038/s41598-025-02203-8

(28) A. Kumar, N. Gupta, A. Jain, A. K. Goyal, and Y. Massoud, “Numerical assessment

and optimization of highly efficient lead-free hybrid double perovskite solar cell,” Results
Opt., vol. 11, p. 100387, 2023. doi: 10.1016/[Link].2023.100387

(29) M. Kumar, A. Raj, A. Kumar, and A. Anshul, “Effect of band-gap tuning on lead-free
double perovskite heterostructure devices for photovoltaic applications via SCAPS
simulation,” Mater. Today Commun., vol. 26, p. 101851, 2021. doi:
10.1016/[Link].2020.101851

(30) R. Ranjan, N. Anand, M. N. Tripathi et al., “SCAPS study on the effect of various hole
transport layer on highly efficient 31.86% eco-friendly CZTS based solar cell,” Sci. Rep.,
vol. 13, p. 18411, 2023. doi: 10.1038/s41598-023-44845-6

(31) S. S. Hussain et al., “Numerical modeling and optimization of lead-free hybrid double
perovskite solar cell by using SCAPS-1D,” J. Renew. Energy, vol. 2021, Art. no. 6668687,
2021. doi: 10.1155/2021/6668687

(32) Sci Rep. 2020 Oct 27;10:18388. doi: 10.1038/s41598-020-73828-0

(33) J. Appl. Phys. 129, 130904 (2021) [Link]

(34) APL Energy 3, 036106 (2025) [Link]

(35) Ji, R., Zhang, Z., Hofstetter, Y.J. et al. Perovskite phase heterojunction solar cells. Nat
Energy 7, 1170–1179 (2022). [Link]

(36) [Link]

(37) [Link]