Article https://doi.org/10.

1038/s41467-024-48877-y

Low-temperature strain-free encapsulation

for perovskite solar cells and modules
passing multifaceted accelerated
ageing tests

Received: 19 September 2023 Paolo Mariani1,8, Miguel Ángel Molina-García2,8, Jessica Barichello1,
Marilena Isabella Zappia2, Erica Magliano 1, Luigi Angelo Castriotta 1,
Accepted: 15 May 2024
Luca Gabatel 2,3, Sanjay Balkrishna Thorat2, Antonio Esaú Del Rio Castillo2,
1234567890():,;
1234567890():,;

Filippo Drago 4, Enrico Leonardi5, Sara Pescetelli 1, Luigi Vesce1,

Francesco Di Giacomo1, Fabio Matteocci 1, Antonio Agresti 1, Nicole De Giorgi2,
Check for updates Sebastiano Bellani2,8 , Aldo Di Carlo1,6 & Francesco Bonaccorso 2,7

Perovskite solar cells promise to be part of the future portfolio of photovoltaic

technologies, but their instability is slow down their commercialization. Major
stability assessments have been recently achieved but reliable accelerated
ageing tests on beyond small-area cells are still poor. Here, we report an
industrial encapsulation process based on the lamination of highly viscoelastic
semi-solid/highly viscous liquid adhesive atop the perovskite solar cells and
modules. Our encapsulant reduces the thermomechanical stresses at the
encapsulant/rear electrode interface. The addition of thermally conductive
two-dimensional hexagonal boron nitride into the polymeric matrix improves
the barrier and thermal management properties of the encapsulant. Without
any edge sealant, encapsulated devices withstood multifaceted accelerated
ageing tests, retaining >80% of their initial efficiency. Our encapsulation is
applicable to the most established cell configurations (direct/inverted,
mesoscopic/planar), even with temperature-sensitive materials, and extended
to semi-transparent cells for building-integrated photovoltaics and Internet of
Things systems.

Perovskite solar cells (PSCs) promise to revolutionize the photovoltaic and high-throughput solution-manufacturing processes5,6. Never-
(PV) industry thanks to power conversion efficiencies (PCEs) up to theless, to reach a Levelized Cost of Energy (LCoE) competing with
26.1% and 33.9% in single-junction and tandem configurations1–4, those of market-dominating crystalline Si (c-Si) solar cells (>USD$0.05/
respectively, as well as their cost-effectiveness in terms of materials kWh)7, long-term stability is still a challenge for PSCs8, especially once

1
CHOSE—Centre for Hybrid and Organic Solar Energy, University of Rome Tor Vergata, Via del Politecnico 1, 00133 Rome, Italy. 2BeDimensional S.p.A., Via
Lungotorrente Secca 30R, 16163 Genova, Italy. 3Department of Mechanical, Energy, Management and Transport Engineering (DIME), Università di Genova,
Genova, Italy. 4Nanochemistry Department, Istituto Italiano di Tecnologia, Via Morego 30, 16163 Genova, Italy. 5GreatCell Solar Italia SRL, Rome, Italy. 6ISM-
CNR, Istitute of Structure of Matter, Consiglio Nazionale delle Ricerche, Rome, Italy. 7Graphene Labs, Istituto Italiano di Tecnologia, Via Morego 30, 16163
Genova, Italy. 8These authors contributed equally: Paolo Mariani, Miguel Ángel Molina-García, Sebastiano Bellani. e-mail: [email protected];
[email protected]; [email protected]

Nature Communications | (2024)15:4552 1

Article https://doi.org/10.1038/s41467-024-48877-y

assembled at module level, in which additional failure mechanisms, polyolefins (POEs), polyurethanes (PUs) and thermoplastic poly-
e.g., potential induced degradation and reverse bias effects, must be urethanes (TPUs). Also, edge sealants made of PIB-based butyl rubber
considered8,9. In general, the lifetime of a PSC is determined by both its tapes, UV-curable polymers, epoxy resins, silicones, and glass frits are
intrinsic (e.g., polymorphism, defects, lattice strains and ion migra- commonly used to realize stable devices41,42. In particular, early reports
tion) and extrinsic (e.g., moisture, oxygen, heat, UV-light and reverse demonstrated small-area PSCs passing accelerated ageing tests in
bias) factors10,11. Main degradation pathways include structural transi- compliance with international standards (i.e., international Electro-
tions and phase segregation of perovskite films or charge-transporting technical Commission -IEC- 61215) and international Summit on
layers (CTLs)12,13, often accompanied by morphological alterations11,14. Organic PV Stability (ISOS) protocols10,35, e.g., damp heat (≥1000 h at
Such mechanisms are frequently initiated by the migration of ions15 85 °C, relative humidity -RH- =85% for IEC 61215, ambient temperature
and outgassing of volatile molecular species at perovskite grain for ISOS-D-243; PCE retention >80%) and thermal cycling (≥200 tem-
boundaries and material interfaces16. These effects are commonly perature cycles between −40 °C and 85 °C for IEC 61215, minimum
exacerbated in the presence of lattice defects17/strains18 in the per- temperature >−40 °C for other lab procedures; PCE retention
ovskite, as well as interfacial stresses resulting from the mismatches of >80%)33,36,40,44–51, as well as humidity freeze test40. However, such results
lattices and thermal expansion coefficients (TECs) of the cell have not yet been fully validated in perovskite solar modules (PSMs),
materials19. Several strategies, including compositional20 and remarking a commercialization gap between science and technology52.
dimensional21,22 engineering, defect passivation23,24, grain boundary A recent study reported an effective encapsulation strategy based on a
modification25,26 and material interface engineering27,28, have been self-crosslinked fluorosilicone polymer gel, achieving non-destructive
proposed, and holistically combined to improve the intrinsic stability encapsulation at room temperature of both PSCs and a 25 cm2 (active
of PSCs10,11,14,29. Moisture and oxygen can react with perovskite absor- area = 15.8 cm2)-PSM. By using an unspecified epoxy edge sealant, the
bers, which finally decompose. By forming hydrogen bonds, water encapsulated PSMs passed the IEC 61215 damp heat test (1000 h, PCE
generates deprotonated organic cations, thus weakening the bond retention = 98%), but were not subjected to other relevant accelerated
between the organic cation and the Pb-halide octahedral30. The proton ageing tests36. Also, the corresponding encapsulated PSCs remarkably
can be then transferred to halide ions (e.g., I−) via water molecules, passed the IEC 61215 damp heat test (1000 h, PCE retention = 98%) and
producing volatile species (e.g., CH3NH2, HI and PbI2). Meanwhile, thermal cycling (220 cycles, PCE retention = 95%), even though con-
oxygen can diffuse into the perovskite by occupying halide vacancies, tinuous light soaking at 55 ± 5 °C for 1000 h led to a PCE drop of
and charged superoxides can form upon the photoexcitation of per- almost 20%. Overall, the encapsulation of PSMs by industrially com-
ovskites. These processes induce acid-base reactions with organic patible (high-throughput) methodologies has been rarely
cations, which entail the formation of volatile species31,32. Importantly, described36,53, whereas ageing tests have not been systematically per-
the interplay between intrinsic and extrinsic factors ultimately deter- formed beyond direct outdoor performance evaluation36,53. For PSMs,
mines the overall PSC stability10. In this context, elevate temperature/ the encapsulant materials must be compatible with a high throughput
temperature variation33, illumination and reverse bias34 aggravate (minute-time scale) and cost-effective process36,54. Compared to tra-
intrinsic degradation effects, being the light-sensitive perovskite ditional PV encapsulants, those for PSMs must also consider thermal
absorbers subjected to photodissociation at elevated temperatures15. management functionalities due to the low thermal conductivity of
Encapsulation strategies have been consensually recognized as key in perovskite absorbers36,55. In addition, they must also have Pb-
the realization of PSCs and corresponding modules lasting at least 20 sequestrating abilities due to the risk of Pb release (~40 μg/kWh)56
years in outdoor conditions10,35–37. However, well-established encap- into the environment caused by the high solubility product constant
sulation strategies reported for commercial PV technology cannot (Ksp) of well-established perovskite by-products (i.e., 4.4 × 10−9 M for
meet the distinctive requirements of PSCs, whose encapsulation con- PbI2, which is 11 orders of magnitude higher than those of PbS and
cepts are still premature10,35,38. The main requirements that an encap- PbSe)57,58. Moreover, low-Young modulus encapsulants are typically
sulant system should accomplish are10,35,39: 1) chemical inertness and recommended for PSCs and PSMs to avoid delamination issues caused
chemical compatibility with underlying cell materials (e.g., no release by the mismatch of materials TECs44. Lastly, the encapsulant must not
of degrading chemicals, such as acetic acid and methacrylic acid for introduce cell-to-module losses caused by the large area of the edge
the case of ethylene vinyl acetate -EVA- and Surlyn, respectively); 2) low seals, whose width around the glass edge should be less than 1 cm for
water vapor transmission rate (WVTR) (≤10−4 g m−2 day−1) and oxygen square meter-sized solar panels. So far, most of the stability results
transmission rate (OTR, ≤10−3 cm3 m−2 day−1 atm−1) to hinder the access have been demonstrated with encapsulant areas even larger than the
of moisture and oxygen, while constraining the outgassing of volatile photoactive area38,40,44,45. This means that such results must be still
species40; 3) resistance to degradation processes (i.e., yellowing and validated on practical device configurations ensuring market-
release of degrading products for PSC materials) induced by UV attracting LCoEs7.
radiation; 4) thermal stability up to 85 °C and low temperature In this work, we address the multifaceted challenges of encapsu-
(≤120 °C) processability to ensure compatibility with the thermal sta- lants for PSMs by proposing an industrially compatible solvent- and
bility of perovskite and common CTLs; 5) optical transparency (i.e., strain-free encapsulation strategy based on a viscoelastic (semi-solid)/
transmittance ≥90% from 400 to 1100 nm) for front side encapsulants; highly viscous (liquid) polyolefin, namely homopolymer PIB (not
6) electrically insulating properties (i.e., resistivity >1013 Ω cm and high incorporating additives commonly used in PIB-based tapes, including
dielectric constant) to prevent leakage current and, hence, alleviate butyl rubber edge sealants). By selecting a proper molecular weight of
potential-induced degradation; and 7) mechanical properties, such as homopolymer PIB, the latter can exhibit a (highly viscoelastic) semi-
flexibility (i.e., low Young modules, preferably <20 MPa at 25 °C) and solid-to-(highly viscous) liquid transition increasing the temperature
adhesivity (i.e., adhesion strength >0.1 MPa) to withstand thermo- from −40 °C to 85 °C, as those used to age PV devices through stan-
mechanical stresses originated from daily temperature variation36, as dardized tests. Polysibutylene is often reported as a common encap-
simulated by thermal cycling/shock ageing tests. So far, the combi- sulant material for PSCs. However, common PIB-based encapsulants
nation of glass/pressure-tight polymer/glass encapsulation, in which contain several additives (e.g., isobutylene-isoprene co-polymer,
the solar cell/module is sandwiched between two glass sheets using an silanes, tackifiers such as glycerol rosin ester, lamellar minerals such as
encapsulant adhesive atop the PSC (blanket-cover approach), has been talc and kaolin, metal oxides, carbon black and even molecular sieve
successfully applied for the realization of highly stable PSCs10,35,36. desiccants) that enable the crosslinking of PIB chemically bond to
Among the adhesive encapsulant candidates, noteworthy examples surfaces, improve anti-ageing and impermeability properties, adjust
are EVA, ionomers (Surlyn, Bynel and Jurasol), polysibutylene (PIB), the rheological/mechanical properties, and also control the esthetic

Nature Communications | (2024)15:4552 2

Article https://doi.org/10.1038/s41467-024-48877-y

features (e.g., color)54,59,60. In this work, differently from commercially leading to atomic impermeability75. Furthermore, 2D h-BN flakes
available PIB-based encapsulants commonly used in literature for exhibit a high thermal conductivity (e.g., >700 W m−1 K−1 for monolayer
PSCs40,61,62, we propose low-molecular weight homopolymer PIB as a h-BN and >100 W m−1 K−1 for few-/multi-layer h-BN)76,77, thus improving
transparent (viscoelastic) semi-solid/(highly viscous) liquid proces- the thermal management properties of polymers when used as
sable in form of laminable films. The latter, herein deposited on glass additives36. Inspired by our previous activities on anticorrosive coat-
substrates, can be used as primary encapsulant for PSCs via an ings based on solid PIB with high molecular weight (800,000)65, the
industrially compatible solvent- and strain-free lamination protocols, barrier properties of the low-molecular weight semi-solid/liquid PIB
aiming at solving limitations of current approaches based on solid proposed in this work were first tested through electrochemical
encapsulants. In addition, we show that the adhesion, barrier and methods. Figure 1a shows the potentiodynamic anodic polarization
thermal management properties of our homopolymer PIB encapsulant measurements and their corresponding Tafel analysis for representa-
can be improved by the addition of two-dimensional (2D) inorganic tive PIB- and PIB:h-BN-coated steels (bare steel results are also shown
fillers, namely few-layer hexagonal boron nitride (h-BN) (nano)flakes for comparative purposes) immersed in a corrosive environment (3.5
produced at industrial scale through a patented wet-jet milling (WJM) wt.% NaCl water solution). These experiments were performed fol-
exfoliation process of the native bulk powder63–65. After encapsulation lowing the ASTM G5-14 standard (see Methods sections). The results
(without extra edge sealing), our PSCs (either in mesoscopic and pla- evidence the barrier properties of PIB and PIB:h-BN films, which
nar n-i-p configurations or inverted p-i-n configurations) and PSMs decrease the corrosion rate from 7.3 ×10−1 mm year−1 for structural steel
(mesoscopic n-i-p configurations), based on a perovskite chemistry to average values of 1.5 × 10−1 mm year−1 and 1.7 × 10−4 mm year−1 for the
well-established in large-area devices and at module level PIB- and PIB:h-BN-coated steels, respectively (Fig. 1b). Noteworthy, the
(Cs0.08FA0.80MA0.12Pb(I0.88 Br0.12)3, retained more than 80% of the addition of 2D h-BN flakes into PIB matrix improved the reproducibility
initial PCE after accelerated tests. These tests include thermal stress of the anticorrosion performance of PIB films, even though both PIB
(ISOS-D-2 at 85 °C, >1000 h), light soaking (ISOS-L-1, >1000 h), custo- and PIB:h-BN films lead to minimum corrosion rate as low as 5.6 ×
mized thermal shock test (200 cycles between −40/ + 85 °C with 10−4 mm year−1 and 1.2 × 10−5 mm year−1. In general, the data proved the
abrupt temperature changes) and modified humidity freeze test (10 superior barrier properties of PIB:h-BN films (average corrosion inhi-
cycles with abrupt temperature changes between +85 °C and −40 °C bition efficiency = 99.97%) compared to homopolymer PIB (average
and including a water immersion step before device freezing). Note- corrosion inhibition efficiency ranging = 79.53%). The same trend was
worthy, our customized/modified accelerated thermal shock and observed for homopolymer and composite films produced with solid
humidity freeze tests served as rapid ageing protocols to assess the (high-molecular weight) PIB films, as shown in Fig. S1 and ref. 65. This
moisture and temperature variation sensitivities of module effect is attributed to the superior hydrophobicity resulting from the
designs66–68, as targeted by more time-consuming IEC 61215 damp heat presence of 2D h-BN flakes (water contact angle of 88.3° ± 0.4° and
(1000 h) and thermal cycling (30-50 days) protocols, without the need 97.9° ± 0.4° for solid PIB and PIB:h-BN films, respectively, Fig. S2a–c),
to rely on expensive laboratory equipment (i.e., closed climatic as shown in ref. 65. The adhesive properties of solid PIB and PIB:h-BN
chambers with RH control). Despite the superior encapsulant prop- films were measured through pull-off tests following the ASTM D4541-
erties of PIB:h-BN encapsulants, homopolymer PIB was used as trans- 02 standard, showing that the incorporation of 2D h-BN flakes into PIB
parent encapsulant for semi-transparent PSCs, reaching a PCE of 6.8% increases the adhesive strength of the homopolymer PIB by 25%
and a bifaciality factor (defined as the ratio of PCEs measured with (Fig. S2d). Notably, both water contact angle and pull-off measure-
front and rear illuminations) as high as 89% after encapsulation. ments were carried out on solid PIB and PIB:h-BN films since semi-
Overall, our results indicate that semi-solid/liquid encapsulation con- solid/liquid films do not permit reliable measurements with these
cepts efficiently mitigate either thermal and thermomechanical stres- techniques. The WVTR of the proposed encapsulants was measured in
ses during encapsulant application, while providing excellent barrier a glass/pressure-tight polymer/glass system through calcium corro-
performance for the realization of long-term stable PSCs and PSMs, sion test (Ca test) (Fig. 1c, d), which analyzes the Ca corrosion through
aiming at tackling the competition with Si-based PVs. in-situ resistance measurements78,79. This sample configuration simu-
lates the glass/pressure-tight polymer/glass encapsulation concept
Results used for PSCs and PSMs in this work (blanket-cover approach), in
Encapsulants characterization which the moisture cannot pass through the glass (thus, its entrance
Two types of encapsulants were prepared for the encapsulation of occurs laterally from the device edge through the encapsulants)80. The
PSCs and PSMs, as described in the Methods section. Specifically, the calculated WVTR was ca. 2×10−5 g m−2 d−1 for both systems based on
first encapsulant is based on a room temperature highly viscous liquid semi-solid/liquid PIB and PIB:h-BN encapsulants (Fig. 1e). The UV-Vis
transparent PIB with low-molecular weight (95,000), while the second transmittance spectra of the encapsulated Ca films remained almost
one is an opaque composite of the same PIB and 2D h-BN (nano)flakes unchanged over 15 days, thus excluding the Ca oxidation and, conse-
(hereafter named PIB:h-BN), being the latter produced by WJM exfo- quently, the ingress of moisture (Fig. 1f). The effects of 2D h-BN flakes
liation of bulk h-BN crystals63,65,69. Despite its amorphous and semi- on the thermal management properties were evaluated through
solid/liquid nature (in the temperature range of −40/ + 85 °C), the infrared thermal imageing of glass/PIB/glass and glass/PIB:h-BN/glass
homopolymer PIB herein selected exhibits a marked packing of its systems (area = 5.6 cm×5.6 cm), which were realized through a lami-
molecular chains, leading to high barrier properties65,70. Importantly, nation protocol resembling the one used for the encapsulation of PSCs
PIB has a resistivity on the order of 1016 Ω cm, which is superior to that and PSMs (see Methods section for details). The samples were heated
of EVA (ranging between 1013 and 1015 Ω cm)35, thus leading to potential up to 90 °C and then quickly transferred onto an Al platform at 25 °C.
induced degradation-suppressing properties35,71. In addition, previous By means of an infrared camera, the maximum temperature of the
studies have proved that the incorporation of 2D h-BN flakes into the samples was monitored over time. As shown in Fig. S3, the presence of
PIB matrix is an effective strategy to enhance the barrier properties of 2D h-BN flakes improves the heat dissipation ability of the system
pristine polymer against the permeation of water (and, thus, moisture) compared to that based on bare PIB, reducing by 11.2% the time to
and other corrosive species65,72. The barrier properties of 2D h-BN are reach 30 °C. Despite it was not possible to perform reliable thermal
generally ascribed to its morphology with high-specific surface-area conductivity measurements of (viscoelastic) semi-solid/(highly vis-
(1488 m2 g−1 for monolayer h-BN)73 and hydrophobic nature74. More- cous) liquid PIB because of the impossibility of realizing self-standing
over, the delocalized dense cloud of overlapping π-orbitals of h-BN bulk objects with suitable thickness, a 2D h-BN flakes content of 5 wt%
represents a physical barrier against molecules or ions penetration, in other more solid polymeric matrix (as used in our PIB:h-BN) typically

Nature Communications | (2024)15:4552 3

Article https://doi.org/10.1038/s41467-024-48877-y

Fig. 1 | Characterization of the barrier properties of the encapsulants. a Anodic encapsulants. c, d Schematics (top-view and cross-section, respectively) of the
polarization curves (Tafel plots) of steel protected by PIB (low-molecular weight) sample configuration used for the Ca test. e WVTR measured through the encap-
and PIB:h-BN encapsulants (data acquired for the samples showing the highest sulants in the glass/pressure-tight polymer/glass encapsulation systems through
corrosion rate among different replicas). The Tafel plot measured for bare steel is the Ca test. f UV-Vis transmittance spectra of the samples measured through the Ca
also shown for comparison. b Statistical analysis of the corrosion rates of the film at different times of environmental exposure.
investigated systems and average corrosion inhibition efficiencies of the

improves significantly the thermal conductivity of the pristine polymer the encapsulant, while using two flat metallic ribbons (commercial
(e.g., by more than 80% in epoxy systems, as measured through Hot tape-like charge collectors) to bring the electrical contacts externally.
Disk measurements following ISO 22007-2 standard). The encapsulants were applied through an industrially compatible,
high-throughput (total duration <45 min) lamination protocol (see
Preliminary encapsulants assessment in PSCs details in the Methods section). Figure 2c shows a photograph of a
The encapsulants were first tested in mesoscopic n-i-p PSC config- representative device encapsulated with PIB:h-BN. Figure 2d shows the
urations (active area = 1 cm2) based on Cs0.08FA0.80MA0.12Pb(I0.88 JV curves (reverse voltage scan) of a representative PSC, before and
Br0.12)3 perovskite and poly[bis(4-phenyl)(2,4,6-trimethylphenyl) after encapsulation with PIB:h-BN. The JV curve recorded after 240 h-
amine] (PTAA) hole-transporting layer (HTL), following ISOS-D-2 (at ISOS-D-1 test, performed before the ISOS-D-2 test, is also reported.
85 °C) and ISOS-L-1 protocols (after 240 h-ISOS-D-1 preconditioning)43. Tables S1 and S2 list the PV parameters of the cells tested through ISOS-
Noteworthy, PTAA was selected because of its superior thermal sta- D-2 and ISOS-L-1 protocols, as extrapolated from their JV curves (for
bility compared to other well-established HTLs, e.g., 2,2′,7,7′-tetra- both reverse and forward voltage scan modes). The as-fabricated
kis[N,N-di(4-methoxyphenyl)amino]−9,9′-spirobifluorene (spiro- mesoscopic n-i-p PSCs based on PTAA HTLs show PCEs up to ~18.8%.
OMeTAD), used for state-of-the-art performance PSCs81. Instead, the Despite far from record certified PCE achieved by on small area active
choice of this perovskite chemistry and the overall cell configuration area (26.1% on 0.057127 cm2)3,4, our PCEs are significant for 1 cm2-active
relies on its high PCE proved at module/farm level by our group53. area PSCs, whose record certified PCE of 21.6% remain unchallenged
Prospectively, other perovskite chemistries, e.g., MA-free ones show- since 201988. A recent article reported a record PCE of 24.35% for a
ing superior stability compared to MA-containing perovskites40, could 1.007 cm2 cell by the NUS/SERIS group4, but details are not disclosed in
be also considered to further improve the stability results presented in any accessible report88. As shown hereafter, either mesoscopic n-i-p
this work. The cell structure was FTO/c-TiO2/graphene-incorporating configurations based on spiro-OMeTAD HTL and planar n-i-p config-
m-TiO2/Cs0.08FA0.80MA0.12Pb(I0.88 Br0.12)3/PEAI/PTAA/Au (Fig. 2a) (see urations can improve further the PCE of our mesoscopic n-i-p PSCs
Methods section for the definition of the acronyms). Graphene was based on PTAA HTLs, reaching maximum value of ~20.2% (for spiro-
incorporated into the m-TiO2 to improve the electron extraction effi- OMeTAD-based mesoscopic n-i-p PSCs), approaching further record
ciency of the mesoscopic electron- transporting layer (ETL)53,82–84, as certified PCE on large-area PSCs. For the encapsulated devices, the
well as to improve the stability of MA-based perovskites85. In addition, data indicate that the lamination process marginally affects the overall
PEAI served as an ultrathin perovskite-passivating layer, as reported in cell performances (absolute PCE drop <1%, regardless of the type of the
previous studies86,87. Fig. 2b sketches the cell layout, which was encapsulant). During the ISOS-D-1 (i.e., dark storage at ambient tem-
designed to entirely cover the non-compact layers of the device with perature and ambient RH) test lasting 240 h the cells retained their

Nature Communications | (2024)15:4552 4

Article https://doi.org/10.1038/s41467-024-48877-y

Fig. 2 | Characterization of PSCs: ISOS-D-1/D-2 and ISOS-L-1 stability tests. representative mesoscopic PSC encapsulated with PIB:h-BN. d JV curves (reverse
a Sketch of the structure of the mesoscopic n-i-p PSCs based on Cs0.08 voltage scan) measured for a representative mesoscopic PSC before and after
FA0.80MA0.12Pb(I0.88 Br0.12)3 perovskites and PTAA HTLs. b Schematic of the cell encapsulation with PIB:h-BN (before and after 240 h-ISOS-D1). e, f PV parameters of
layout (active area = 1 cm2), in which the non-compact layers of the device are fully the mesoscopic PSCs without encapsulation and with PIB and PIB:h-BN encapsu-
covered by the encapsulant. Two flat metallic ribbons are connected to the cell lants, acquired over >1000 h of ISOS-D-2 and ISOS-L-1 tests.
terminals to bring the electrical contacts externally. c Photograph of a

performances, confirming their shelf-life stability (absolute PCE drop the proposed lamination protocol is compatible with more
<1%). Afterwards, the cells underwent ISOS-D-2 and ISOS-L-1 tests, temperature-sensitive HTLs compared to PTAA. As shown in Fig. S4,
sampling the cell PV parameters over >1000 h. As shown in Fig. 2e, f, Table S3, the lamination of PIB:h-BN encapsulant does not significantly
the unencapsulated cells quickly degraded during both the ageing affect the cell performances (absolute PCE drop <1%), confirming the
tests, showing a T80 (defined as the time at which the PCE drops to 80% results proved for PTAA-based mesoscopic n-i-p PSCs (see Fig. 2d).
of its initial magnitude; values estimated from a multi-order poly- Afterwards, 240 h-ISOS-D-1 test proved the shelf-life stability of the
nomial fitting of the PCE data with R2 > 0.999) < 70 h and <5 h for the investigated cells. Lastly, during the ISOS-D-2 test the unencapsulated
ISOS-D-2 and ISOS-L-1 tests, respectively. Contrary, the encapsulated cell degraded, showing a T80 < 240 h, while the encapsulated cells have
cells have shown T80 > 1000 h, regardless of the type of the encapsu- shown T80 > 1000 h. In general, the hygroscopicity of lithium bis(tri-
lant. Noteworthy, no edge sealants were used during these experi- fluoromethanesulfonyl) imide (LiTFSI) and the evaporation of 4-tert-
ments, confirming the excellent barrier properties of our primary butylpyridine (tBP), used as spiro-OMeTAD dopants, promote moist-
encapsulants under PV operating conditions. Moreover, the stability of ure entry into the cell structure and morphological changes of both
the encapsulated cells was achieved in the presence of interconnection perovskite and spiro-OMeTAD HTL81. These effects are accelerated
ribbons, confirming the reliability of our encapsulants for practical PV with increasing temperature, leading to the formation of pinholes that
panels composed of ribbon-interconnected solar cells. Furthermore, accelerate iodine migration to iodine-sensitive cellular components
for the small-area cells investigated in this work, interconnect ribbons (e.g., metal electrodes) and even cause connections between CTLs
may have negative effects on the overall lamination process. However, (shunting pathways), leading to PCE losses81. In this context, our results
the latter was successful thanks to the ability of the adopted highly demonstrate that the combination of advanced PIB-based encapsu-
viscous liquid encapsulants to dissipate thermomechanical stresses, lants, which block moisture entry into PSCs, and ultrathin perovskite
leading to a strain-free encapsulation approach. passivation layers, e.g., PEAI, is a promising strategy to stabilize spiro-
As mentioned above, our encapsulation approach was also pro- OMeTAD-based PSCs operated at high temperature. In particular, PEAI
bed on mesoscopic n-i-p PSCs based on spiro-OMeTAD, proving that effectively passivate defects and trap states at the perovskite/spiro-

Nature Communications | (2024)15:4552 5

Article https://doi.org/10.1038/s41467-024-48877-y

OMeTAD interface89, helping our spiro-OMeTAD-based cell to with- tests (Fig. 3e, f), showing T80 < 100 h and <3 h, respectively, resembling
stand the lamination process at 90 °C for 10 min. However, our data the instability observed for unencapsulated PSCs. Contrary, the
indicate that proper encapsulants that effectively protects against the encapsulated PSMs have shown T80 > 1000 h.
air/moisture ingress into the cell structure are crucial to avoid doped To further assess the reliability of our encapsulants, as well as to
spiro-OMeTAD degradation through oxidation90. Our findings are well specifically evaluate the barrier and thermal management functional-
aligned with existing studies reporting excellent thermal stability of ities of 2D h-BN flakes as encapsulant additives, encapsulated PSMs
spiro-OMeTAD-based PSCs91–93. The effectiveness of PIB:h-BN encap- were subjected to two customized accelerated ageing stress, i.e., a
sulants was further demonstrated on 1 cm2-active area planar n-i-p thermal shock test (between -40/ + 85 °C) and, subsequently, a mod-
PSCs based on low-temperature processed SnO2 as the ETL. As shown ified humidity freeze test. Figure 4a, b show the temperature profiles
in Fig. S5 and Table S4, as fabricated cells exhibited a maximum PCE as and environmental conditions (e.g., water immersion and air expo-
high as 19.0%. After encapsulation with PIB:h-BN, they exhibited a T80 sure) for our thermal shock and humidity freeze tests, while their
of ca. 2000 h during ISOS-D-2 test, whereas the unencapsulated ones comparison with IEC 61215 thermal cycling and humidity freeze tests is
have shown a lower T80 of 700 h. Even though the use of SnO2 ETL in depicted in Fig. S7. Importantly, the cycle times of our tests (20 min for
planar PSCs may eliminate photo-induced degradation associated to the thermal shock test and 25 min for the humidity freeze test) are
TiO2-based ETLs in mesoscopic structures10,11, the reproducibility of significantly reduced compared to IEC 61215 tests minimum cycle
planar PSCs at module level is still lower compared to mesoscopic times (>2.5 h for thermal cycling; >22 h for humidity freeze test). Thus,
configurations. The latter have been recently assembled with PTAA in our accelerated ageing stresses, abrupt temperature changes are
HTLs into a 4.5 m2 stand-alone solar farm infrastructure53,94 and, supposed to induce severe thermomechanical stresses caused by the
therefore, selected for the realization of the PSMs reported in this work thermal expansion and contraction of materials with different TECs,
hereafter. The universality of our encapsulation approach was also critically jeopardizing the adhesion between various layers and the
tested on 1 cm2-active area inverted p-i-n configurations based on reliability of the electrical interconnections98. Previous studies in Si
PTAA as the HTL and [6,6]-phenyl-C61-butyric acid methyl ester PERC cells proved that accelerated thermal cycling can represent an
(PCBM) as the ETL. Long chain alkylammonium salt phenethyl effective tool to rapidly prototype novel PV materials and module
ammonium chloride (PEACl) was used for perovskite surface treat- configurations while triggering degradation pathways that may not
ment to simultaneously passivate the grain boundaries and the per- occur during traditional IEC 61215 thermal cycling68. Also, in our
ovskite/PCBM interface95. As shown in Fig. S6 and Table S5, the cells humidity freeze test, the water immersion step after heating the PSM
retain their performance after encapsulation with PIB:h-BN, resulting to +85 °C and before freezing the PSM to -40 °C aims at replacing the
in T80 > 1000 h during ISOS-D-2 test, whereas the unencapsulated ones long (>20 h) high-temperature (+85 °C) step at 85% RH required by IEC
have shown T80 < 360 h. 61215 protocol to determine the ability of the module to withstand
humidity penetration. Table S8 lists the PV parameters of the modules
Encapsulant validation in PSMs tested through customized thermal shock and humidity freeze tests, as
The PIB and PIB:h-BN encapsulants were subsequently assessed in extrapolated from their JV curves (for both reverse and forward vol-
mesoscopic n-i-p PSMs (based on PTAA HTLs) consisting of 5 in series- tage scan modes). As shown in Fig. 4c, the PSM encapsulated with
connected cells with an active area of 2 cm2 (total active area = 10 cm2), PIB:h-BN withstood 200 thermal shock cycles, retaining 84.5% of the
as reported in previous studies96. Fig. 3a shows the layout of the starting PCE. With the PIB encapsulant, the PSM retained 82.1% of the
encapsulated mesoscopic n-i-p PSMs. The encapsulants were applied initial PCE after 200 cycles. These data indicate that the use of 2D h-BN
following the same lamination protocols reported for PSCs, paying flakes as thermally conductive additives in encapsulants is an effective
attention to entirely covering the porous layers of the module struc- strategy to improve the overall thermal management properties of
ture. As for the case of the PSCs, no edge sealants were used in com- PSMs, integrating passive cooling abilities into the encapsulant system.
bination with our primary encapsulants. The PSMs were designed to This is consistent with the thermal properties measured for our
avoid the need for ribbons to contact the positive and negative module encapsulants (Fig. S3). After the thermal shock test, the PSMs encap-
terminals. Specifically, we printed two silver busbars along the edges of sulated with PIB and PIB:h-BN were stressed further through the
the modules exceeding the encapsulant. Thus, contrary to the case of humidity freeze test (Fig. S8), retaining 72.1% and 86.0% of their PCE
PSCs, the thermomechanical stresses associated with the presence of (before starting this test) after 10 cycles, respectively (Fig. 4d). Overall,
ribbons (both during the encapsulation and during the operation of PIB:h-BN slightly outperformed homopolymer PIB during thermal
the PSM) have been eliminated. Figure 3b, c show the photographs of a shock and humidity freeze tests, as expected by its distinctive barrier
representative PSM before (front and rear sides) and after encapsula- and thermal management properties (Fig. 1a, b, Figs. S2 and S3).
tion (rear side) with PIB:h-BN, respectively. Figure 3d shows the JV The effectiveness of the PIB:h-BN encapsulant to protect the PSMs
curves (reverse voltage scan) of two representative PSMs, before and from extrinsic factors was also assessed by measuring the Pb leakage of
after encapsulation with PIB and PIB:h-BN, respectively. The JV curves the encapsulated PSM immersed in water through inductively coupled
recorded after a 240 h-ISOS-D-1 test performed before the subsequent plasma optical emission spectroscopy (ICP-OES) (Fig. S9). After water
ageing tests are also reported. Tables S6 and S7 list the PV parameters immersion the unencapsulated PSM rapidly degraded, showing yel-
of the modules tested through ISOS-D-2 and ISOS-L-1 tests, as extra- lowing associated with the decomposition of the perovskite to PbI2.
polated from their JV curves (for both reverse and forward voltage scan Because of its high solubility (340 mg L−1, solubility product constant =
modes). Our PSMs reached a maximum PCE of 17%, which is relevant 4.4 × 10−9 M)57,58 PbI2 rapidly dissolved in water, causing cracking of the
for module configurations proved at solar farm level53 where the Au rear electrode. The detected Pb leakage (>60 µg cm−2 after 24 h) is
scalability and batch-to-batch reproducibility of the materials must be consistent with the Pb content in perovskite, typically between 0.1 and
ensured together with high manufacturing yields. Prospectively, our 1 g m−256. Contrary to unencapsulated devices, the perovskite in the
encapsulant approach may also be assessed on more efficient PSMs encapsulated PSMs retained its starting color, preserving the per-
configurations, now reaching record PCE up to 19.9% on 10 cm2 active ovskite phase. Consequently, the Pb leakage was drastically inhibited
area97. As for the case of the cells, the encapsulation process marginally to values lower than 1 µg cm−2 after 24 h (low Pb water contamination is
affected the overall module performances (absolute PCE drop <1%), likely associated with perovskite residuals nearby the encapsulant
which were also retained during 240 h-ISOS-D-1 tests (absolute PCE edges and not with the degradation of perovskite over the PSM active
drop <1%) regardless of the type of the encapsulant. Afterwards, the area). Similar Pb-leakage inhibition was observed for a PIB encapsulant
unencapsulated PSMs quickly degraded during ISOS-D-2 and ISOS-L-1 protecting simple perovskite films.

Nature Communications | (2024)15:4552 6

Article https://doi.org/10.1038/s41467-024-48877-y

Fig. 3 | Characterization of PSMs: ISOS-D-1/D-2 and ISOS-L-1 stability tests. encapsulation (rear side) with PIB:h-BN. d JV curves (reverse voltage scan) mea-
a Schematic of the mesoscopic n-i-p PSM layout (cell active area = 2 cm2; total active sured for the as-fabricated mesoscopic n-i-p PSMs before and after encapsulation
area = 10 cm2), in which the non-compact layers are entirely covered by the with PIB (top panel) and PIB:h-BN (bottom panel) (before and after 240 h-ISOS-D1).
encapsulant. b Photograph of a representative mesoscopic n-i-p PSM as fabricated e, f PV parameters of the PSMs without encapsulation and with PIB and PIB:h-BN
(front and read sides: top and bottom picture, respectively) and c after encapsulants acquired over >1000 h of the ISOS-D-2 and ISOS-L-1 tests.

Proof-of-concept semi-transparent PSCs based on PIB of applications, our PIB encapsulants were applied to semi-transparent
encapsulants PSCs based on a wide-bandgap (~2.3 eV)106 FaPbBr3 perovskite. Pre-
Even though PIB:h-BN encapsulants outperformed PIB ones during the viously, the deposition of FaPbBr3 perovskite (1.4 M) in a complete
most aggressive ageing tests (i.e., those involving abrupt temperature semi-transparent stack has been optimized by our group with both
changes) reported for our PSMs, homopolymer PIB still has shown spin and blade coating technique reaching a maximum average visible
satisfactory performances, rarely achieved in literature without addi- transmittance (AVT) of 52% and a maximum bifaciality factor of
tional edge sealants. Consequently, transparent PIB encapsulants may 86.5%107. Here, we fabricated the FTO/c-TiO2/FaPbBr3/PTAA/ITO
find applications for high-PCE perovskite-based tandem systems82,99,100 structure with a FaPbBr3 solution of 1 M allowing to achieve an AVT
and building-integrated PVs –PIBVs– (e.g., smart windows, façades and value exceeding 60%. Figure 5a shows the UV-Vis transmittance spec-
agrivoltaics)101. Also, semi-transparent PV architectures, especially tra of a representative semi-transparent PSC before and after encap-
solution-processed ones (e.g., PSCs and organic solar cells), have sulation with PIB. Interestingly, the AVT increased from 58.1% to 62.7%
attracted significant interest for indoor applications to power portable after encapsulation. Based on the reflectance spectra of the samples,
electronics and photonic devices for the Internet of Things (IoT), e.g., this behavior is attributed to the decrease in reflection losses (i.e.,
distributed sensors, remote actuators, and communication improved matching of the refractive indices of the interface materials)
devices102–105. Here, indoor PVs may be subjected to less aggressive after device encapsulation. The increase of the transmittance after PIB
environmental conditions compared to outdoor PVs and, thus, PIB encapsulation is also observed for bare FTO, supporting our conclu-
encapsulants may still be a suitable choice. To investigate these types sion. The antireflective properties of PIB have been also confirmed by

Nature Communications | (2024)15:4552 7

Article https://doi.org/10.1038/s41467-024-48877-y

Fig. 4 | Accelerated ageing tests for PSMs: thermal shock and humidity the mesoscopic n-i-p PSMs encapsulated with PIB and PIB:h-BN acquired over >200
freeze tests. a Temperature profile of the thermal shock test performed on the cycles of the thermal shock test. d PV parameters of the mesoscopic n-i-p PSMs
mesoscopic n-i-p PSMs encapsulated with PIB and PIB:h-BN. b Temperature profile encapsulated with PIB and PIB:h-BN acquired over >10 cycles of the customized
and environmental exposure conditions of the humidity freeze test performed on humidity freeze test.
the mesoscopic n-i-p PSMs encapsulated with PIB and PIB:h-BN. c PV parameters of

Fig. 5 | Characterization of semi-transparent PSCs. a UV-Vis transmittance shown (right y-axis). The photograph of the semi-transparent PSCs is also shown.
spectra of a semi-transparent PSC before and after encapsulation with PIB (samples b JV curves measured for a representative semi-transparent PSC before and after
named PSC and PSC/PIB, respectively), bare FTO and FTO/PIB/glass (sample named encapsulation with PIB, for both front and rear side illuminations.
FTO/PIB) (left y-axis). The reflectance spectra of ITO and ITO/PIB samples are also

Nature Communications | (2024)15:4552 8

Article https://doi.org/10.1038/s41467-024-48877-y

the reflectance spectra measured for ITO and PIB-coated ITO (ITO/ tert-butylpyridine)cobalt(III) tri[bis(trifluoromethane)sulfonimide]
PIB), also shown in Fig. 5a. Thus, PIB can acts as a kind of antireflective (FK209 Co(III) TFSI), bathocuproine (BCP) and copper beads (Cu beads),
coating and, prospectively, future optical modeling and simulations ethanol (EtOH) (anhydrous, ≥99.8 %), acetone (≥99.5%), acetonitrile
(beyond the scope of this work) could be used to further reduce (ACN) (≥99.8%), dimethylformamide (DMF) ≥ 99%), dimethyl sulfoxide
reflection losses by controlling the PIB thickness after the lamination (DMSO) (>99%), chlorobenzene (CB) (99.8%), 1,2-dichlorobenzene
process. Figure 5b shows the JV curves measured for representative (DCB) (99%), toluene (>99.7%), N‐methyl‐2‐pyrrolidone (NMP) (>97%),
semi-transparent PSCs before and after encapsulation for both front 2-propanol (IPA) (anhydrous, 99.5%), ethyl acetate (anhydrous, 99.8%),
and rear side illuminations. Table S9 reports the PV parameters tBP and LiTFSI were purchased from Sigma-Aldrich. The SnO2 disper-
extracted from the JV curves, showing that the encapsulation pro- sion in water (15%) was purchased from Alfa Aesar. Poly[bis(4-phenyl)
cesses almost retain the PCE of the unencapsulated devices. Sig- (2,4,6-trimethylphenyl)amine] SOL2426M (average molecular weight,
nificantly, the encapsulated cell exhibited a bifaciality factor of 89%, 105 kDa) was purchased from Solaris Chem. [6,6]-phenyl-C61-butyric acid
which is similar to the one measured before the encapsulation (92%). methyl ester was purchased from Solenne. Graphene dispersion in EtOH
(0.9 mg mL−1) was supplied by BeDimensional S.p.A. Room temperature
Discussion highly viscous (Brookfield viscosity >100000 mPas at 10 rpm at tem-
In summary, we report a blanket-cover encapsulation approach for perature <120 °C) PIB (LMW-80, average molecular weight 95,000) was
PSCs and PSMs based on the lamination of (highly viscoelastic) semi- provided by TER Chemicals. Solid PIB (Oppanol® N80, average mole-
solid/(highly viscous) liquid PIB-based adhesives atop the mesoporous cular weight 800,000) was purchased from BASF. Bulk powder of h-BN
cell layers. The viscoelasticity of PIB intrinsically limits the thermo- was supplied by Alfa Aesar. All the chemicals were used as received
mechanical stresses caused by both the encapsulation process and the unless specified otherwise. Fluorine Tin Oxide (FTO)-coated glasses
temperature gradients occurring during accelerated ageing stresses. (sheet resistance RSH = 7 Ω sq−1) were purchased from NSG-Pilkington.
The incorporation of 2D h-BN flakes into the PIB matrix improves the Indium-tin oxide (ITO) coated-glasses (RSH = 7 Ω sq−1) were purchased
barrier and thermal management properties of homopolymer PIB, from Kintec. Silver (Ag) paste 7713 was purchased from Dupont. Indium
which, however, has demonstrated to be an optimal transparent Tin Oxide (ITO) target for sputtering was purchased from
encapsulant by itself for the realization of devices operating with TestBourne Ltd.
reduced temperature fluctuation (e.g., indoor applications and specific
BIPVs) compared to those taking place outdoor for conventional PVs. Encapsulants preparation
Without using any edge sealants the PSCs and PSMs encapsulated with To produce homopolymer PIB encapsulants, PIB was first dissolved in
our PIB-based encapsulants withstood multiple accelerated ageing toluene with a PIB:toluene weight ratio of 1:1.5 and vigorously stirred
tests, including ISOS-D1 preconditioning (240 h), ISOS-D2 (85 °C, (500 rpm) for 12 h at 800 rpm and 80 °C until a homogeneous solution
>1000 h), ISOS-L1 (light soaking, >1000 h), as well as a customized was obtained. For the production of PIB:h-BN encapsulants, few-layer
thermal shock test (200 cycles) and modified humidity freeze test (10 h-BN flakes were produced utilizing the patented WJM method by
cycles), retaining more than 80% of their initial (at the beginning of the BeDimensional S.p.A63,69. Experimentally, a mixture of NMP and bulk h-
test) PCE. Noteworthy, these results have been achieved with an MA- BN with an NMP:h-BN weight ratio of 98:2 was pressurized into two jet
based perovskite, namely Cs0.08FA0.80MA0.12Pb(I0.88 Br0.12)3, and streams, which collided in a nozzle (to produce the shear forces
may, therefore, be further improved by using more stable MA-free responsible for the exfoliation mechanism)63,69. The WJM-produced h-
perovskite chemistries. The combination of our semi-solid/liquid PIB- BN flakes dispersion was dried using a customized drier (BeDimen-
based encapsulants with advanced edge sealants can also be beneficial sional S.p.A), ensuring solvent residuals less than 1 wt%, as detected by
for the long-term stability and robustness of practical PSMs. Despite means of thermal gravimetric analysis (TGA) in N2 atmosphere from 25
the superior encapsulant properties of PIB:h-BN compared to homo- to 800 °C at a heating rate of 10 °C min−1 using TGA Q500 (TA Instru-
polymer PIB, the latter was used for the realization of proof-of-concept ments) thermogravimetric analyzer. Eventually, WJM-produced h-BN
semi-transparent PSCs based on wide-bandgap FaPbBr3. The encap- flakes were added into the PIB solution followed by mixing in a pla-
sulated semi-transparent PSCs reached a PCE of 6.8% with a bifaciality netary centrifugal mixer (Thinky ARE-250 Mixing and Degassing
factor as high as 89%, which is similar to the one measured before the Machine) at 1000 rpm for 5 min to produce h-BN/PIB composite resins
encapsulation (92%). Prospectively, semi-transparent PSMs encapsu- with a h-BN weight percentage (wt.%) of 5% (excluding the solvent).
lated with homopolymer PIB can be used for indoor applications The h-BN content was previously optimized through electrochemical
(which entail conditions less harsh than outdoor ones), as well as for tests compliant with ASTM G5-14, ASTM G59-97, ASTM G61-86, and
the front-side encapsulation of high-PCE perovskite-based tandem ASTM G106-89 standards to maximize the barrier properties of the h-
systems and BIPVs, possibly in combination with an edge sealant. The BN-incorporating encapsulants65. The electrochemical characteriza-
results reported in this study represent a breakthrough towards the tions were carried out on samples produced either using (highly vis-
realization of long-term stable PSMs through low-temperature and coelastic) semi-solid/(highly viscous) liquid PIB or solid (high-
cost-effective semi-solid/liquid encapsulants combining high- molecular weight) PIB. To produce the encapsulants, the PIB and PIB:h-
throughput processability, perovskite/CTL compatibility, barrier BN resins were deposited by doctor blading onto 1 mm-thick glass
properties, adhesivity, thermal and light-soaking stability, without the substrates to be used for the glass/pressure-tight polymer/glass
need to consider advanced perovskite chemistries with high internal encapsulation of PSCs and PSMs. The resulting films were dried at
stability or specific edge sealants. room temperature for 1 h, followed by 15 h at 60 °C to evaporate the
residual solvent.
Methods
Materials Encapsulants characterization
TiO2 (titanium dioxide) paste (30 NR-D), formamidinium iodide (FAI), The thickness of the resulting homopolymer PIB or PIB:h-BN films was
methylammonium bromide (MABr), methylammonium chloride (MACl) between 600 and 700 µm, as measured with a Trotec BB20 thickness
and PEACl were purchased from GreatCell Solar. Lead(II) iodide (PbI2), measurement system (for these measurements, the encapsulant films
lead(II) bromide (PbBr2), and cesium iodide (CsI) were purchased from were deposited on metallic substrates following the same procedure
TCI. Cesium bromide beads (CsBr), titanium(IV) isopropoxide (TTIP), and parameters used for the glass coating). The water contact angle of
diisopropoxytitanium bis(acetylacetonate) (Ti(AcAc)2), acetyl acetone the PIB and PIB:h-BN films was measured with an OSSILA L2004A1
(AcAc), phenethylammonium iodide (PEAI), tris(2-(1H-pyrazol-1-yl)-4- contact angle goniometer, imageing a 10 μL water drop deposited on

Nature Communications | (2024)15:4552 9

Article https://doi.org/10.1038/s41467-024-48877-y

the sample. To evaluate the barrier properties of the encapsulants, both Perovskite solar cell and module fabrication
electrochemical measurements and Ca tests were carried out. Electro- Mesoscopic n-i-p PSCs. Mesoscopic PSCs (active area = 1 cm2) were
chemical measurements were carried out using a BioLogic VMP3 Mul- fabricated on FTO-coated glass substrates cut in 2.5 × 2.5 cm2 size. The
tichannel Potentiostat in a three-electrode 1 L electrochemical cell at FTO layer was patterned via laser etching using an Nd:YVO4-pulsed UV
room temperature in a 3.5 wt.% NaCl aqueous solution, following the laser system (BrightSolutions, Luce 40 laser), carrying out a P1 process
procedures described in the ASTM G5-14 standard. A KCl-saturated Ag/ to electrically separate the photoelectrode from the counter electrode.
AgCl radiometer Analytical REF201 Red Rod Reference Electrode (Bio- Then, the substrates were cleaned with brushing and without scratching
logic) was used as the reference electrode, whereas a graphite rod was the FTO surface, using a cleaning solution (Hellmanex) diluted with
used as the counter electrode. The standard working electrode assem- water (2:98 vol/vol). After this step, the substrates were sonicated in an
bly consisted of a cylindrical sample of structural steel substrate (S355) ultrasonic bath first with acetone and then with IPA (10 min for each
coated by PIB or PIB:h-BN films, drilled-and-tapped with a 3-48 UNC step). Afterwards, the substrates were air dried. Then, an UV/O3-treated
thread, and screwed onto the support rod. The PIB or PIB:h-BN films with a PSD Pro Series Digital UV Ozone System (Novascan) was used to
were produced by depositing the corresponding resins by doctor remove organic contaminations. The c-TiO2 layers were then deposited
blading, followed by drying at room temperature for 1 h, and at 60 °C onto the patterned FTO by the spray pyrolysis of a dispersion of 0.16 M
for 15 h to evaporate the residual solvent. The thickness of the PIB or Ti(AcAc)2 and 0.4 M AcAc in EtOH, setting the hot plate temperature at
PIB:h-BN films was ca. 60 µm, as measured with a Trotec BB20 thickness 465 °C and using air as the gas carrier at a pressure of 1.6 bar. The nozzle
magnetic induction-based measurement system. A Teflon compression angle was about 45 °C with respect to the plane of the substrate, and the
gasket ensured a leak-free seal. The open-circuit potential was mon- nozzle was moved with a serpentine path for 12-13 cycles (one every
itored for 30 min, after which the corrosion performance of the coatings 10 seconds) until reaching a thickness of 50 nm, as measured by profi-
was investigated by potentiodynamic anodic polarization measure- lometry (Deektat Veeco 150). Then, the substrates were left for 15 min at
ments and their Tafel analysis, as described in the ASTM G5-14 standard, 465 °C before slowly cooling them to room temperature. To fabricate
for the determination of a metal’s corrosion current (icorr) and the cor- the m-TiO2 layers, the 30NR-D TiO2 paste was diluted in anhydrous EtOH
rosion potential108,109. The corrosion rate of the samples was calculated (1:5 w/w). After stirring overnight (>12 h), graphene was incorporated
KW eq icorr
from icorr according to the Faraday law, i.e.: CR = D , where CR are into the diluted 30NR-D TiO2 paste by adding 1 vol% graphene disper-
−1
the corrosion rate (in mm year ), K is a constant with a value of 3.27 × sion in EtOH53. 120 µL of the graphene-incorporating paste was spin-
10−3, Weq is the equivalent weight of iron in ferrous compounds (27.9 g coated onto c-TiO2 layers at 3000 rpm for 30 s with an acceleration of
eq−1), icorr is the corrosion current density (in μA cm−2) and D is the 1500 rpm s−1. The resulting layers were converted into graphene-
density of steel (7.85 g cm−3)110. The corrosion inhibition efficiency (ηp) of incorporating m-TiO2 through a multi-step sintering program: 1) 5
the composites was calculated from icorr by the following equation: min-temperature ramp from room temperature to 150 °C, dwell time of
i0
i
ηp % = corr0 corr x100, where i0corr and icorr are the corrosion current 5 min; 2) 15 min-temperature ramp from 120 to 325 °C, dwell time of
icorr
densities in the absence and presence of inhibitors, respectively111. 5 min, 3) 5 min- temperature ramp from 325 to 375 °C, dwell time of
Ca tests were performed on samples produced by depositing a 5 min; 4) 15 min-temperature ramp from 375 °C to 490 °C, dwell time of
Ca film on etched FTO-coated glass substrates with an area of 25 mm 30 min. To improve its wettability, the graphene-incorporating m-TiO2
× 25 mm, which were then laminated on encapsulant-coated glass layers were UV-treated for 30 min with 5000-EC UV curing lamps
substrates using a heated press at ~100 °C to 150 °C in an N2-filled (Dymax). Then, the samples were transferred to the glove box. The
glove box. Figure 1c, d depict the layout of the sample configuration perovskite precursor solution was then prepared by dissolving FAI (1 M),
used for the Ca tests. The FTO coatings were etched over a central PbI2 (1.2 M), PbBr2 (0.2 M), CsI (0.1 M), MABr (0.2 M) in DMF and DMSO
strip of the substrate with an etched area of 25 mm × 5 mm. The Ca (1:4 vol/vol) to obtain a perovskite composition of Cs0.08 FA0.80 MA0.12
films were deposited by thermal evaporation with an area of 5 mm × Pb (I0.88 Br0.12)3. After stirring overnight, 90 µL of the perovskite pre-
15 mm, which covered both parts of the FTO-etched region and FTO cursor solution was spin-coated on the ETLs with a two-step protocol: 1)
regions. The latter served as electrical contacts for the Ca films. The 2000 rpm for 10 s with an acceleration of 400 rpm s−1; 2) 5000 rpm for
Ca films were covered by the encapsulants with an area of 25 mm × 20 s with an acceleration of 2000 rpm s−1. Subsequently, 100 µL of CB,
18 mm. The WVTR through the encapsulants was measured through used as antisolvent, was poured on the spinning substrates 15 s before
electrical (quantitative)78 and optical (qualitative) analyzes of Ca the end of the spinning programme to induce the perovskite nucleation.
corrosion80. The thermal management properties of the encapsu- After the perovskite deposition, the samples were annealed at 100 °C for
lants were evaluated by monitoring the maximum temperature of 45 min to complete the crystal phase growth. An ultrathin perovskite-
glass/encapsulant/glass systems using a thermal camera (A655sc, passivating PEAI layer was then deposited onto the perovskites by spin
FLIR) placed at ~50 cm from the sample surface. The samples were coating 100 µL of PEAI solution in IPA (concentration of 5 mg mL−1) at
produced through a lamination protocol resembling the one used in 2500 rpm for 30 s, with an acceleration of 1250 rpm s−186,87. Then, the
this work for the encapsulation of PSCs and PSMs (see details samples were annealed at 100 °C for 10 min. To deposit the HTL, a PTAA
hereafter). The samples were heated at 90 °C on a hot plate with a lid solution in toluene (10 mg mL−1), doped with 7 μL mL−1 of tBP and a
and then transferred to an Al platform at 25 °C. The temperature of 10 μl mL−1 of a LiTFSI stock solution in ACN (170 mg mL−1), was spin-
the system was monitored during cooling. The thermal camera was coated at 3000 rpm for 20 s with an acceleration of 1500 rpm s−1. As
controlled with FLIR’s software (Temperature FLIR ResearchIR Max alternative to PTAA, spiro-OMeTAD was deposited by spinning 90 µL of
software), which was also used to process the temperature data. The spiro-OMeTAD solution in CB (73.5 mg mL−1) doped with tBP
adhesive properties of solid PIB and solid PIB:h-BN were measured (26.8 μL mL−1), LiTFSI (16.6 μL mL−1) (from stock solution in ACN
through pull-off tests using an Instron 3365 dual-column dynam- (520 mg mL−1) and FK209 Co(III) TFSI (7.2 μL mL−1), at 4000 rpm for
ometer equipped with a 2 kN load cell and following the ASTM 1 min, with an acceleration of 1000 rpm s−1. Finally, the mesoscopic n-i-p
D4541-02 standard. The encapsulant resins were deposited on steel PSCs were completed by depositing 100 nm-thick Au back electrodes
plates, which were clamped to the bottom anvil. Afterwards, the through thermal evaporation in a high vacuum chamber (10−6 mbar).
15 mm diameter top piston was painted with cyanoacrylate adhesive
and immediately put in contact with the sample. A force of 15 N was Planar n-i-p PSCs. The planar PSCs (active area = 1 cm2) were fabri-
applied to the sample and the adhesive was let curing for 30 min. cated as the mesoscopic ones except for the ETL, which was produced
Normal displacement was then applied to the piston, with a rate of by spin coating 100 µL of SnO2 solution in water (15%) onto the FTO-
1 mm min−1 until separation. coated glasses, previously UV-treated for 30 min to improve their

Nature Communications | (2024)15:4552 10

Article https://doi.org/10.1038/s41467-024-48877-y

wettability. The SnO2 solution was spun at 6000 rpm for 35 s with an width of P1 was 20 µm. The distance between the cells, cell width, and
acceleration of 3000 rpm s−1. Then, the samples were annealed for 1 h cell length were set to 0.5 mm, 5 mm, and 40 mm, respectively. After
at 150 °C to form a 70-80 nm-thick SnO2 layer. the cleaning step performed according to the procedure described for
the mesoscopic PSCs, two Ag busbars (width 4 mm) were screen
Inverted p-i-n PSCs. Inverted p-i-n PSCs were produced on ITO-coated printed close to the edge of the substrate (2 mm from the edge),
glass substrates patterned with a Nd:YVO4-pulsed UV laser system parallel to the longest size of the cell, using an automated screen
(BrightSolutions, Luce 40 laser) and then cut into 2.5 × 2.5 cm2 sam- printer (Baccini, Applied Materials). After the deposition, the busbars
ples. The ITO-patterned samples were cleaned in ultrasonic bath with a were dried onto a hot plate at 120 °C for 10 min. The busbars were then
cleaning solution (Hellmanex) diluted with water (2:98 vol/vol), acet- sintered during the spray pyrolysis deposition of c-TiO2. After the
one and then isopropanol (15 min for each step). Any remaining sol- deposition of the PTAA, P2 was carried out to clean the FTO inter-
vent residual was blown off using air flow. UV-ozone treatment was connection areas. The width of P2 was 160 µm. After the thermal eva-
then performed on the substrates for 15 min to remove all the residual poration of the 100 nm-thick Au back electrodes, the adjacent cell
organic contaminants, using a PSD Pro Series Digital UV Ozone System isolation was accomplished with the P3 process. The width of P3 was
(Novascan). The samples were then transferred in a N2-filled glove box 90 µm. The resolution of our laser ablation processes led to a geo-
and the PTAA (2 mg mL−1 in toluene) was spin-coated at 5000 rpm for metrical fill factor of ~91%113. Table S10 lists the parameters used for
20 s. The samples were annealed at 100 °C for 10 min. After cooling laser ablation processes (P1, P2, and P3).
down, a film of PbI2 and CsBr (with a ratio 10:1) was thermally co-
evaporated onto the substrates. FAI (0.48 M), MABr (0.09 M) and Perovskite solar cells and modules lamination
MACl (0.09 M) were dissolved in EtOH and the solution was dynami- Both PSCs and PSMs were encapsulated through a multi-step, low-
cally spin-coated on the substrates in a flow box filled with dry air temperature (90 °C), differential pressure lamination process using an
(RH < 10%). The samples were then annealed in air (RH between 30 and automatic two-chamber solar panel laminator (CORE – Model 2, Rise
40%) at 150 °C for 15 minutes. On top of the perovskite layer, PEACl Technology srl), equipped with a cooling system to guarantee a high
(1.5 mg mL−1 in EtOH) was dynamically spin-coated at 4000 rpm and reproducibility on the lamination procedure and to reduce the expo-
subsequently annealed at 100 °C for 10 minutes. Afterwards, PCBM sure time of the materials to temperature causing their degradation.
(27 mg mL−1 in CB:DCB—3:1 volume ratio) was spun at 1350 rpm for 20 s Experimentally, the entire surface of the devices was covered with the
and annealed at 100 °C for 5 min. BCP (0.5 mg mL−1 in IPA) was PIB or PIB:h-BN encapsulant-coated glass to ensure the glass/pressure-
deposited at 2300 rpm for 20 s without any further drying. Finally, a Cu tight polymer/glass encapsulation of PSCs and PSMs (blanket-cover
layer of 100 nm was thermally evaporated on top of the samples using approach). The so-assembled laminates were put inside the laminator
a shadow mask. lower chamber, which exploited the differential pressure between the
upper and lower chambers. After locking the laminator, 1) the lami-
Semi-transparent PSCs. The fabrication of semi-transparent PSCs nator chambers were evacuated at moderate vacuum (pressure of ~1
(active area = 1 cm2) started with the c-TiO2 deposition onto FTO- mbar), while the laminates were heated from room temperature to
coated glasses as for the case of mesoscopic PSCs. Before depositing 50 °C in 215 s (~7 °C min−1). Then, 2) the upper chamber started to
the perovskite layers, the samples underwent a UV-light soaking for inflate to apply a final pressure of 30 mbar on the top of the substrates.
10 min to improve their wettability. In an N2-filled atmosphere, 80 µL of The system required 500 s to stabilize such a low pressure in the
1 M FaBr and 1 M PbBr2 in DMSO solvent were spin-coated on the laminator chamber. Then, 3) the temperature of the laminate was
samples, pre-heated at 60 °C, at 4000 rpm for 20 s with an accelera- increased from 50 °C to 90 °C in 500 s (4.8 °C min−1), which was then
tion of 2000 rpm s−1. After 10 s from the spin start, 200 µL of anhydrous kept for 600 s. Afterwards, 4) the temperature of the laminate was
ethyl acetate were dropped to induce the FaPbBr3 perovskite crystal- decreased to 50 °C in 500 s (-4.8 °C min−1). Lastly, 5) the pressure of the
lization. Subsequently, the c-TiO2/FaPbBr3 samples were annealed at chambers returned to 1000 mbar in about 1 s and the laminator started
85 °C for 10 min. After cooling down the samples at room temperature, opening while releasing the laminated devices.
90 µL of a PTAA solution in toluene (10 mg mL−1), doped with 10 µL mL−1
of tBP and 5 µL mL−1 of a stock solution of LiTFSI (170 mg mL−1 in ACN) Device characterization
were spin-coated with the same parameters used for the deposition of J-V measurements of the devices were performed with a Class-A Sun
the perovskite layer. Afterwards, ITO was sputtered atop the samples Simulator (ABET 2000) equipped with an AM1.5 G filter (ABET). The
at low temperature using an industrial magnetron sputtering (KENO- sun simulator was calibrated to 1 Sun illumination condition with a Si-
SISTEC S.R.L., KS 400 In-Line) at 1.1 × 10−3 mbar and 90 W RF power and based reference cell (RR-226-O, RERA Solutions). Arkeo platform (Cicci
purging inert Ar gas in the chamber at 40 sccm. By means of a sample Research S.r.l.) was used for J-V data acquisition under forward and
holder, the samples were moved below the ITO target at 120 cm/min reverse voltage scan modes, using a voltage step of 20 mV s−1 and a
speed for 200 cycles to achieve a 200 nm-thick ITO top electrode with voltage scan rate of 200 mV s−1.
a sheet resistance of 25 Ω/sq, as measured by a four-probe unit A UV-vis spectrophotometer (Shimadzu UV-2550) equipped
installed in Arkeo Platform (Cicci Research S.r.L.). with an integrated sphere was used for the acquisition of transmit-
tance spectra of the encapsulant both for the optical characteriza-
Mesoscopic PSMs tion of the Ca test and for the characterization of semi-transparent
The fabrication of mesoscopic PSMs followed the same step as the devices. The sheet resistance of the Ca films during the calcium
mesoscopic PSCs except for the quantities of solution poured onto the test was measured by the four-probe method using a Keithley
substrates and for the overall additional three laser ablation processes 2620 source meter (Tektronix). The AVT values of semi-transparent
(P1-P2-P3, see Fig. S9) to define the series-connected layout with an PSCs were calculated according to the method reported in
active area of 10 cm2 (single cell active area = 2 cm2), as sketched in the ISOhR 9050:2003 standard i hRusing the following
i equation:
780 780
Fig. 3a112. The size of the FTO-coated substrates was 5.6 cm × 5.6 cm. AV T = 380 DðλÞ × TðλÞ × V ðλÞdλ = 380 DðλÞ × V ðλÞdλ , in which DðyÞ is
The amounts of m-TiO2 solution, perovskite precursor solution, the incident light spectral distribution, V ðyÞ is the sensitivity factor
chlorobenzene as antisolvent, PEAI solution, and PTAA solutions were of the human eye, and TðλÞ is the transmittance.
600 µL, 450 µL, 400 µL, 450 µL and 500 µL, respectively, for each The ISOS-L-1 tests of the devices were performed in air and using
module. The P1 process consisted of patterning the FTO-coated glass an Arkeo-multichannel (Cicci Research S.r.l.) station based on 32 fully
substrates to isolate 5 adjacent cells composing the final modules. The independent source meter units (+/− 10 V @ +/−250 mA) and an

Nature Communications | (2024)15:4552 11

Article https://doi.org/10.1038/s41467-024-48877-y

ARKEO light soaker (VIS version) with low-mismatch LED-based system 13. Slotcavage, D. J., Karunadasa, H. I. & McGehee, M. D. Light-
(400-750 nm). A standard Perturb & Observe tracking algorithm was induced phase segregation in halide-perovskite absorbers. ACS
selected for tracking the maximum power point of the devices, Energy Lett. 1, 1199–1205 (2016).
acquiring a J-V scan every 4 min. The ISOS-D-2 tests were carried out in 14. Cheng, Y. & Ding, L. Pushing commercialization of perovskite solar
a Lenton WHT4/30 oven (Hope Valley). A thermal shock test was cells by improving their intrinsic stability. Energy Environ. Sci. 14,
performed by cycling the samples between −40 °C and +85 °C with 3233–3255 (2021).
abrupt temperature changes (from room temperature to +85 °C and 15. Yuan, Y. & Huang, J. Ion migration in organometal trihalide per-
vice versa, from room temperature to −40 °C and vice versa, see ovskite and its impact on photovoltaic efficiency and stability. Acc.
Fig. 4a). The modified humidity freeze test was performed after 200 Chem. Res. 49, 286–293 (2016).
thermal shock cycles and consisted of 10 thermal shock cycles 16. Conings, B. et al. Intrinsic thermal instability of methylammonium
between −40 °C and +85 °C, each cycle starting from a step of lead trihalide perovskite. Adv. Energy Mater. 5, 1500477 (2015).
immersion in water at room temperature (see Fig. 4b). For the thermal 17. Yang, S. et al. Stabilizing halide perovskite surfaces for solar cell
shock and humidity freeze tests, a Lenton WHT4/30 oven and a low- operation with wide-bandgap lead oxysalts. Science 365,
temperature home-freezer were used to set the chamber temperature 473–478 (2019).
at +85 °C and −40 °C, respectively. 18. Zhao, J. et al. Strained hybrid perovskite thin films and their impact
Inductively coupled plasma optical emission spectroscopy mea- on the intrinsic stability of perovskite solar cells. Sci. Adv. 3,
surements were carried out on a ThermoFisher iCAP 7600 DUO eaao5616 (2017).
Thermo spectrometer to measure the Pb leakage of the modules 19. Rolston, N. et al. Engineering stress in perovskite solar cells to
immersed in water, sampling the water solution at different times until improve stability. Adv. Energy Mater. 8, 1802139 (2018).
a total immersion time of 24 h. 20. Jeon, N. J. et al. Compositional engineering of perovskite materials
for high-performance solar cells. Nature 517, 476–480 (2015).
Reporting summary 21. Shao, M. et al. Over 21% efficiency stable 2D perovskite solar cells.
Further information on research design is available in the Nature Adv. Mater. 34, 2107211 (2022).
Portfolio Reporting Summary linked to this article. 22. Grancini, G. & Nazeeruddin, M. K. Dimensional tailoring of hybrid
perovskites for photovoltaics. Nat. Rev. Mater. 4, 4–22 (2019).
Data availability 23. Zheng, X. et al. Defect passivation in hybrid perovskite solar cells
The source data generated in this study are provided in the Source using quaternary ammonium halide anions and cations. Nat.
Data file. Additional data that support the findings of this work are Energy 2, 17102 (2017).
available from the corresponding authors upon request. Source data 24. Gao, F., Zhao, Y., Zhang, X. & You, J. Recent progresses on defect
are provided with this paper. passivation toward efficient perovskite solar cells. Adv. Energy
Mater. 10, 1902650 (2020).
References 25. Liu, L. et al. Grain-Boundary “Patches” by In situ conversion to
1. Kim, M. et al. Conformal quantum dot–SnO2 layers as electron enhance perovskite solar cells stability. Adv. Mater. 30, 1800544
transporters for efficient perovskite solar cells. Science 375, (2018).
302–306 (2022). 26. Niu, T. et al. Stable high-performance perovskite solar cells via
2. Min, H. et al. Perovskite solar cells with atomically coherent grain boundary passivation. Adv. Mater. 30, 1706576 (2018).
interlayers on SnO2 electrodes. Nature 598, 444–450 (2021). 27. Agresti, A. et al. Two-dimensional (2D) material interface engi-
3. Best Research-Cell Efficiency Chart, https://www.nrel.gov/pv/ neering for efficient perovskite large-area modules. ACS Energy
cell-efficiency.html (accessed on 06/04/2024). Lett. 4, 1862–1871 (2019).
4. Green, M. A. et al. Solar cell efficiency tables (Version 63). Prog. 28. Isikgor, F. H. et al. Molecular engineering of contact interfaces for
Photovolt. Res. Appl. 32, 3–13 (2024). high-performance perovskite solar cells. Nat. Rev. Mater. 8,
5. Wang, P. et al. Solution-processable perovskite solar cells toward 89–108 (2023).
commercialization: progress and challenges. Adv. Funct. Mater. 29. Masi, S., Gualdrón-Reyes, A. F. & Mora-Seró, I. Stabilization of
29, 1807661 (2019). black perovskite phase in FAPbI3 and CsPbI3. ACS Energy Lett. 5,
6. Holzhey, P., Prettl, M., Collavini, S., Chang, N. L. & Saliba, M. 1974–1985 (2020).
Toward commercialization with lightweight, flexible perovskite 30. Leguy, A. M. A. et al. Reversible hydration of CH3NH3PbI3 in films,
solar cells for residential photovoltaics. Joule 7, 257–271 (2023). single crystals, and solar cells. Chem. Mater. 27, 3397–3407
7. De Bastiani, M., Larini, V., Montecucco, R. & Grancini, G. The (2015).
levelized cost of electricity from perovskite photovoltaics. Energy 31. Bryant, D. et al. Light and oxygen induced degradation limits the
Environ. Sci. 16, 421–429 (2023). operational stability of methylammonium lead triiodide perovskite
8. Faheem, M. B. et al. Insights from scalable fabrication to opera- solar cells. Energy Environ. Sci. 9, 1655–1660 (2016).
tional stability and industrial opportunities for perovskite solar 32. Aristidou, N. et al. Fast oxygen diffusion and iodide defects med-
cells and modules. Cell Rep. Phys. Sci. 3, 100827 (2022). iate oxygen-induced degradation of perovskite solar cells. Nat.
9. Lan, D. & Green, M. A. Combatting temperature and reverse-bias Commun. 8, 15218 (2017).
challenges facing perovskite solar cells. Joule 6, 1782–1797 (2022). 33. Li, G. et al. Highly efficient p-i-n perovskite solar cells that endure
10. Ma, S. et al. Development of encapsulation strategies towards the temperature variations. Science 379, 399–403 (2023).
commercialization of perovskite solar cells. Energy Environ. Sci. 34. Ni, Z. et al. Evolution of defects during the degradation of metal
15, 13–55 (2022). halide perovskite solar cells under reverse bias and illumination.
11. Zhang, D., Li, D., Hu, Y., Mei, A. & Han, H. Degradation pathways in Nat. Energy 7, 65–73 (2022).
perovskite solar cells and how to meet international standards. 35. Wang, Y. et al. Encapsulation and stability testing of perovskite
Commun. Mater. 3, 58 (2022). solar cells for real life applications. ACS Mater. Au 2, 215–236
12. Brennan, M. C., Draguta, S., Kamat, P. V. & Kuno, M. Light-induced (2022).
anion phase segregation in mixed halide perovskites. ACS Energy 36. Wang, T. et al. Room temperature nondestructive encapsulation
Lett. 3, 204–213 (2018). via self-crosslinked fluorosilicone polymer enables damp heat-

Nature Communications | (2024)15:4552 12

Article https://doi.org/10.1038/s41467-024-48877-y

stable sustainable perovskite solar cells. Nat. Commun. 14, 57. Hailegnaw, B., Kirmayer, S., Edri, E., Hodes, G. & Cahen, D. Rain on
1342 (2023). methylammonium lead iodide based perovskites: possible envir-
37. Matteocci, F. et al. Encapsulation for long-term stability onmental effects of perovskite solar cells. J. Phys. Chem. Lett. 6,
enhancement of perovskite solar cells. Nano Energy 30, 1543–1547 (2015).
162–172 (2016). 58. Li, J. et al. Biological impact of lead from halide perovskites reveals
38. Emery, Q. et al. Encapsulation and outdoor testing of perovskite the risk of introducing a safe threshold. Nat. Commun. 11, 310
solar cells: comparing industrially relevant process with a sim- (2020).
plified lab procedure. ACS Appl. Mater. Interfaces 14, 5159–5167 59. Kempe, M. D., Dameron, A. A., Moricone, T. J. & Reese, M. O.
(2022). Evaluation and modeling of edge-seal materials for photovoltaic
39. Raman, R. K., Gurusamy Thangavelu, S. A., Venkataraj, S. & applications. in 2010 35th IEEE Photovoltaic Specialists Con-
Krishnamoorthy, A. Materials, methods and strategies for encap- ference 256–261 (2010).
sulation of perovskite solar cells: From past to present. Renew. 60. Pocius, A. V. Adhesives and sealants. Polym. Sci.: A Compr. Ref. 8,
Sustain. Energy Rev. 151, 111608 (2021). 305–324 (2012).
40. Shi, L. et al. Gas chromatography–mass spectrometry analyses of 61. Jiang, Q. et al. Towards linking lab and field lifetimes of perovskite
encapsulated stable perovskite solar cells. Science 368, eaba2412 solar cells. Nature 623, 313–318 (2023).
(2020). 62. Shi, L. et al. Accelerated lifetime testing of organic–inorganic
41. Emami, S., Martins, J., Ivanou, D. & Mendes, A. Advanced hermetic perovskite solar cells encapsulated by polyisobutylene. ACS Appl.
encapsulation of perovskite solar cells: the route to commercia- Mater. Interfaces 9, 25073–25081 (2017).
lization. J. Mater. Chem. A 8, 2654–2662 (2020). 63. Del Rio Castillo, A. E. et al. High-yield production of 2D crystals by
42. Martins, J. et al. Novel laser-assisted glass frit encapsulation for wet-jet milling. Mater. Horiz. 5, 890–904 (2018).
long-lifetime perovskite solar cells. J. Mater. Chem. A 8, 64. Garakani, M. A. et al. Scalable spray-coated graphene-based
20037–20046 (2020). electrodes for high-power electrochemical double-layer capaci-
43. Khenkin, M. V. et al. Consensus statement for stability assessment tors operating over a wide range of temperature. Energy Storage
and reporting for perovskite photovoltaics based on ISOS proce- Mater. 34, 1–11 (2021).
dures. Nat. Energy 5, 35–49 (2020). 65. Molina-Garcia, M. A. et al. Wet-jet milling exfoliated hexagonal
44. Cheacharoen, R. et al. Design and understanding of encapsulated boron nitride as industrial anticorrosive pigment for polymeric
perovskite solar cells to withstand temperature cycling. Energy coatings. J. Phys. Mater. 6, 035006 (2023).
Environ. Sci. 11, 144–150 (2018). 66. Twesme, E. N., Weiss, P. & Mcleod, D. Evaluation of the potential
45. Cheacharoen, R. et al. Encapsulating perovskite solar cells to usefulness of water immersion for a-Si thin film modules. Sol.
withstand damp heat and thermal cycling. Sustain. Energy Fuels 2, Cells 28, 115–119 (1990).
2398–2406 (2018). 67. Jeon, Y.-J., Kim, D.-S. & Shin, Y.-E. Study of characteristics of solar
46. Kim, Y. et al. Methoxy-functionalized triarylamine-based hole- cells through thermal shock and high-temperature and high-
transporting polymers for highly efficient and stable perovskite humidity testing. Int. J. Precis. Eng. Manuf. 15, 355–360 (2014).
solar cells. ACS Energy Lett. 5, 3304–3313 (2020). 68. Schiller, C. H. et al. Accelerated TC test in comparison with stan-
47. Han, F. et al. Hyperstable perovskite solar cells without ion dard TC test for PV modules with ribbon, wire and shingle inter-
migration and metal diffusion based on ZnS segregated Cubic connection. Proc. 36th EUPVSEC 995–999 (2019).
ZnTiO3 electron transport layers. Sol. RRL 5, 2000654 (2021). 69. Del Rio Castillo, A. E., Ansaldo, A., Pellegrini, V. & Bonaccorso, F.
48. Peng, J. et al. Nanoscale localized contacts for high fill factors in Exfoliation of layered materials by wet-jet milling techniques,
polymer-passivated perovskite solar cells. Science 371, Patent Nr. WO2017089987A1 (2019).
390–395 (2021). 70. Kunal, K. et al. Polyisobutylene: A most unusual polymer. J. Polym.
49. Bush, K. A. et al. 23.6%-efficient monolithic perovskite/silicon Sci. Part B Polym. Phys. 46, 1390–1399 (2008).
tandem solar cells with improved stability. Nat. Energy 2, 17009 71. Luo, W. et al. Potential-induced degradation in photovoltaic
(2017). modules: a critical review. Energy Environ. Sci. 10, 43–68 (2017).
50. He, J. et al. Influence of phase transition on stability of perovskite 72. Cui, M. et al. Anticorrosive performance of waterborne epoxy
solar cells under thermal cycling conditions. Sol. Energy 188, coatings containing water-dispersible hexagonal boron nitride (h-
312–317 (2019). BN) nanosheets. Appl. Surf. Sci. 397, 77–86 (2017).
51. Boyd, C. C. et al. Barrier design to prevent metal-induced degra- 73. Weng, Q., Wang, X., Zhi, C., Bando, Y. & Golberg, D. Boron nitride
dation and improve thermal stability in perovskite solar cells. ACS porous microbelts for hydrogen storage. ACS Nano 7, 1558–1565
Energy Lett. 3, 1772–1778 (2018). (2013).
52. Siegler, T. D. et al. The path to perovskite commercialization: a 74. Pakdel, A., Zhi, C., Bando, Y., Nakayama, T. & Golberg, D. Boron
perspective from the United States Solar Energy Technologies nitride nanosheet coatings with controllable water repellency.
Office. ACS Energy Lett. 7, 1728–1734 (2022). ACS Nano 5, 6507–6515 (2011).
53. Pescetelli, S. et al. Integration of two-dimensional materials-based 75. Yoon, S. I., Ma, K. Y., Kim, T. Y. & Shin, H. S. Proton conductivity of a
perovskite solar panels into a stand-alone solar farm. Nat. Energy hexagonal boron nitride membrane and its energy applications. J.
7, 597–607 (2022). Mater. Chem. A 8, 2898–2912 (2020).
54. Shimpi, T. M., Moffett, C., Sampath, W. S. & Barth, K. L. Materials 76. Cai, Q. et al. High thermal conductivity of high-quality monolayer
selection investigation for thin film photovoltaic module encap- boron nitride and its thermal expansion. Sci. Adv. 5, eaav0129
sulation. Sol. Energy 187, 226–232 (2019). (2019).
55. Zhang, L. et al. Thermal conductive encapsulation enables stable 77. Guerra, V., Wan, C. & McNally, T. Thermal conductivity of 2D nano-
high-power perovskite-converted light-emitting diodes. ACS structured boron nitride (BN) and its composites with polymers.
Appl. Mater. Interfaces 13, 30076–30085 (2021). Prog. Mater. Sci. 100, 170–186 (2019).
56. Fabini, D. Quantifying the potential for lead pollution from halide 78. Paetzold, R., Winnacker, A., Henseler, D., Cesari, V. & Heuser, K.
perovskite photovoltaics. J. Phys. Chem. Lett. 6, 3546–3548 Permeation rate measurements by electrical analysis of calcium
(2015). corrosion. Rev. Sci. Instrum. 74, 5147–5150 (2003).

Nature Communications | (2024)15:4552 13

Article https://doi.org/10.1038/s41467-024-48877-y

79. Castro-Hermosa, S., Top, M., Dagar, J., Fahlteich, J. & Brown, T. M. 99. Sahli, F. et al. Fully textured monolithic perovskite/silicon tandem
Quantifying performance of permeation barrier-encapsulation solar cells with 25.2% power conversion efficiency. Nat. Mater. 17,
systems for flexible and glass-based electronics and their appli- 820–826 (2018).
cation to perovskite solar cells. Adv. Electron. Mater. 5, 1800978 100. Rossi, D. et al. Design of highly efficient semitransparent per-
(2019). ovskite/organic tandem solar cells. Sol. RRL 6, 2200242 (2022).
80. Kempe, M. D., Nobles, D. L., Postak, L. & Calderon, J. A. Moisture 101. Pillai, D. S., Shabunko, V. & Krishna, A. A comprehensive review on
ingress prediction in polyisobutylene-based edge seal with building integrated photovoltaic systems: Emphasis to technolo-
molecular sieve desiccant. Prog. Photovolt. Res. Appl. 26, gical advancements, outdoor testing, and predictive main-
93–101 (2018). tenance. Renew. Sustain. Energy Rev. 156, 111946 (2022).
81. Rombach, F. M., Haque, S. A. & Macdonald, T. J. Lessons learned 102. Li, B., Hou, B. & Amaratunga, G. A. J. Indoor photovoltaics, The
from spiro-OMeTAD and PTAA in perovskite solar cells. Energy Next Big Trend in solution-processed solar cells. InfoMat 3,
Environ. Sci. 14, 5161–5190 (2021). 445–459 (2021).
82. Lamanna, E. et al. Mechanically stacked, two-terminal graphene- 103. Panidi, J., Georgiadou, D. G., Schoetz, T. & Prodromakis, T.
based perovskite/silicon tandem solar cell with efficiency over Advances in organic and perovskite photovoltaics enabling a
26%. Joule 4, 865–881 (2020). greener Internet of Things. Adv. Funct. Mater. 32, 2200694 (2022).
83. Mariani, P. et al. Low-temperature graphene-based paste for large- 104. Matteocci, F. et al. Wide bandgap halide perovskite absorbers for
area carbon perovskite solar cells. ACS Appl. Mater. Interfaces 13, semi-transparent photovoltaics: From theoretical design to mod-
22368–22380 (2021). ules. Nano Energy 101, 107560 (2022).
84. Taheri, B. et al. Graphene-engineered automated sprayed meso- 105. Singh, A. et al. Methylamine gas treatment affords improving
scopic structure for perovskite device scaling-up. 2D Mater. 5, semitransparency, efficiency, and stability of CH3NH3PbBr3-based
45034 (2018). perovskite solar cells. Sol. RRL 5, 2100277 (2021).
85. Biccari, F. et al. Graphene-based electron transport layers in per- 106. Zhang, Y. et al. Planar FAPbBr3 solar cells with power conversion
ovskite solar cells: a step-up for an efficient carrier collection. Adv. efficiency above 10. ACS Energy Lett. 3, 1808–1814 (2018).
Energy Mater. 7, 1701349 (2017). 107. Barichello, J. et al. Semi-transparent blade-coated FAPbBr3 per-
86. Lucarelli, G., De Rossi, F., Taheri, B., Brown, T. M. & Brunetti, F. ovskite solar cells: a scalable low-temperature manufacturing
Phenethylammonium iodide passivation layers for flexible planar process under ambient condition. Sol. RRL 7, 2200739 (2023).
perovskite solar cells. Energy Technol. 10, 2200314 (2022). 108. Mansfeld, F. 1973 Tafel slopes and corrosion rates from polariza-
87. Vesce, L. et al. Hysteresis-free planar perovskite solar module with tion resistance measurements. Corrosion 29, 397–402 (1973).
19.1% efficiency by interfacial defects passivation. Sol. RRL 6, 109. Flitt, H. J. & Schweinsberg, D. P. Evaluation of corrosion rate from
2101095 (2022). polarisation curves not exhibiting a Tafel region. Corros. Sci. 47,
88. Szabó, G., Park, N.-G., De Angelis, F. & Kamat, P. V. Are perovskite 3034–3052 (2005).
solar cells reaching the efficiency and voltage limits? ACS Energy 110. Nautiyal, A., Qiao, M., Cook, J. E., Zhang, X. & Huang, T. S. High
Lett. 8, 3829–3831 (2023). performance polypyrrole coating for corrosion protection and
89. Jiang, Q. et al. Surface passivation of perovskite film for efficient biocidal applications. Appl. Surf. Sci. 427, 922–930 (2018).
solar cells. Nat. Photonics 13, 460–466 (2019). 111. Ashassi-sorkhabi, H., Kazempour, A. & Frouzat, Z. Superior poten-
90. Kasparavicius, E. et al. Oxidized spiro-OMeTAD: investigation of tials of hydrazone Schiff bases for efficient corrosion protection of
stability in contact with various perovskite compositions. ACS mild steel in 1.0 M HCl. J. Adhes. Sci. Technol. 0, 1–21 (2020).
Appl. Energy Mater. 4, 13696–13705 (2021). 112. Castriotta, L. A. et al. Reducing losses in perovskite large area solar
91. Xu, R. et al. Optimizing the buried interface in flexible perovskite technology: laser design optimization for highly efficient modules
solar cells to achieve over 24% efficiency and long-term stability. and minipanels. Adv. Energy Mater. 12, 2103420 (2022).
Adv. Mater. 36, 2308039 (2024). 113. Palma, A. L. et al. Laser-patterning engineering for perovskite solar
92. Dewi, H. A. et al. Excellent intrinsic long-term thermal stability of modules with 95% aperture ratio. IEEE J. Photovolt. 7,
co-evaporated MAPbI3 solar cells at 85 °C. Adv. Funct. Mater. 31, 1674–1680 (2017).
2100557 (2021).
93. Zhang, Z., Wang, H., Jacobsson, T. J. & Luo, J. Big data driven Acknowledgements
perovskite solar cell stability analysis. Nat. Commun. 13, 7639 The work has been supported by the European Union’s Horizon 2020
(2022). research and innovation programme under grant agreement number
94. Pescetelli, S. et al. Synergic use of two-dimensional materials to 881603—GrapheneCore3, European Union’s Horizon Europe Framework
tailor interfaces in large area perovskite modules. Nano Energy 95, Programme for research and innovation under grant agreement no.
107019 (2022). 101084124—DIAMOND, European Union’s Horizon Europe Framework
95. Gharibzadeh, S. et al. Two birds with one stone: dual grain- Programme for research and innovation under grant agreement no.
boundary and interface passivation enables >22% efficient inver- 694101—2D-PRINTABLE. ADC gratefully acknowledges the support of
ted methylammonium-free perovskite solar cells. Energy Environ. the Mission Innovation grant between the Italian Ministry of Ecological
Sci. 14, 5875–5893 (2021). Transition and ENEA (agreement 21A033302 GU n. 133/5-6-2021). FDG
96. Vesce, L. et al. Ambient air blade-coating fabrication of stable acknowledges funding from the European Union’s Horizon 2020
triple-cation perovskite solar modules by green solvent quench- research and innovation program under grant agreement no.
ing. Sol. RRL 5, 2100073 (2021). 101006715—VIPERLAB. We are grateful to Prof. Liberato Manna for
97. Wang, Y. et al. Grain boundary elimination via recrystallization- supporting this research with his research team.
assisted vapor deposition for efficient and stable perovskite solar
cells and modules. Adv. Mater. 35, 2304625 (2023). Author contributions
98. Fujimori, M., Kohno, T., Tsuno, Y., Morita, K. Applicability of highly P.M. produced and characterized the PSCs and PSMSs, coordinating the
accelerated thermal cycling testing for multiple types of oly- experimental activities. E.M., S.P., L.V., F.D.G., and A.A. produced and
crystalline silicon photovoltaic modules, The 33rd European characterized the PSCs and PSMs, formulating and characterizing
Photovoltaic Solar Energy Conference and Exhibition (2017). device components. L.A.C. and F.D.G. contributed to the fabrication of

Nature Communications | (2024)15:4552 14

Article https://doi.org/10.1038/s41467-024-48877-y

PSMs, engineering the P1-P2-3 laser patterning protocol and PSM Peer review information Nature Communications thanks Jinsong
design. S.B. and M.Á.M.-G. conceived the encapsulants concepts. Huang, Xuanhua Li and the other, anonymous, reviewer(s) for their
M.Á.M.-G. and S.B. produced and characterized the encapsulants for the contribution to the peer review of this work. A peer review file is avail-
PSCs and PSMs coordinating the experimental activities. J.B. and F.M. able.
produced and characterized the semi-transparent PSCs. M.I.Z., L.G.,
S.T., A.E.D.R.C., N.D.G, S.B., and F.B. produced and characterized 2D Reprints and permissions information is available at
h-BN flakes, supporting the characterization of its polymeric compo- http://www.nature.com/reprints
sites. F.D. performed ICP-OES measurements. E.L. carried out the lami-
nation of solar cells. P.M. and S.B. wrote the first draft and revised the Publisher’s note Springer Nature remains neutral with regard to jur-
manuscript with contributions from all the authors. A.D.C. and F.B. isdictional claims in published maps and institutional affiliations.
supervised and planned the experimental activities. All authors partici-
pated in the preparation of the final manuscript. Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
Competing interests adaptation, distribution and reproduction in any medium or format, as
F.B. is a co-founder and CSO, S.B., S.T and A.E.D.R.C are a senior sci- long as you give appropriate credit to the original author(s) and the
entists, M.A.M.-G., M.I.Z, L.G. and N.G are researchers at BeDimensional source, provide a link to the Creative Commons licence, and indicate if
S.p.A., a company that is commercializing 2D materials. E.L. is a senior changes were made. The images or other third party material in this
researcher of GreatCell Solar Italia SRL, a company commercializing article are included in the article’s Creative Commons licence, unless
solar-grade materials. The remaining authors declare no competing indicated otherwise in a credit line to the material. If material is not
interests. included in the article’s Creative Commons licence and your intended
use is not permitted by statutory regulation or exceeds the permitted
Additional information use, you will need to obtain permission directly from the copyright
Supplementary information The online version contains holder. To view a copy of this licence, visit http://creativecommons.org/
supplementary material available at licenses/by/4.0/.
https://doi.org/10.1038/s41467-024-48877-y.
© The Author(s) 2024
Correspondence and requests for materials should be addressed to
Sebastiano Bellani, Aldo Di Carlo or Francesco Bonaccorso.

Nature Communications | (2024)15:4552 15