Efficient Ambient‑Air Fabrication of Mesoporous Perovskite Solar Cells Using N‑Butyl‑Amine‑Enhanced PbI₂ Precursors

Background

Since the first demonstration of PSCs in 2009, the industry‑leading efficiencies have climbed from 3.8 % to 22.7 % [1,2,3], driven by the exceptional optical absorption, low exciton binding energy, long carrier diffusion lengths, and high mobilities of hybrid perovskites [4–12]. However, the hygroscopic organic cations render these materials highly sensitive to moisture, posing a critical barrier to large‑scale production. Traditionally, PSCs are fabricated inside nitrogen glove boxes or under tightly controlled humidity to mitigate degradation [13–15]. Recent work has shown that even mild moisture can promote the crystallization of perovskite films, yet excessive humidity leads to poor morphology and reduced performance [16–19]. Despite numerous attempts, PSCs produced in ambient air rarely exceed 16 % PCE, and the precise role of moisture in film formation remains contentious. This study addresses this gap by revealing how controlled humidity and a BTA additive can be leveraged to fabricate high‑efficiency PSCs directly in ambient air.

Methods

Device Architecture and Layer Deposition

The mesoporous PSC structure (FTO/c‑TiO₂/mp‑TiO₂/MAPbI₃/Spiro‑OMeTAD/Ag) is illustrated in Fig. 1a. Fluorine‑doped SnO₂‑coated glass (sheet resistance 7 Ω sq⁻¹) was sequentially cleaned with acetone, ethanol, isopropanol, DI water, and isopropanol. A compact TiO₂ layer was spin‑coated at 3000 rpm for 30 s, repeated twice, then annealed at 150 °C for 15 min after each coat, and finally at 500 °C for 30 min in air. The mesoporous TiO₂ layer was deposited by spin‑coating TiO₂ paste (18NRD) diluted 1:7 (EtOH) at 5000 rpm for 45 s, dried at 80 °C for 40 min, sintered at 500 °C for 30 min, and post‑treated with aqueous TiCl₄ (70 °C, 30 min), rinsed, and annealed again at 500 °C for 30 min. The active area was 0.1 cm².

Perovskite Film Preparation

PbI₂ (1 M in DMF) was spin‑coated onto mp‑TiO₂ at 3000 rpm for 30 s, then annealed at 70 °C for 15 min. When employing BTA, a small volume (≈0.5 % w/v) was added to the PbI₂ solution. After cooling to room temperature, a methylammonium iodide (MAI) solution was spin‑coated at 4000 rpm for 45 s, followed by annealing at 100 °C for 30 min to form MAPbI₃. The hole‑transport layer (Spiro‑OMeTAD) was spin‑coated at 2000 rpm for 45 s from a 80 mg mL⁻¹ chlorobenzene solution containing 28.8 µL TBP and 17.7 µL Li‑TFSI (520 mg mL⁻¹ acetonitrile). Finally, silver electrodes were thermally evaporated.

a Schematic of the mesoporous PSC stack. b Cross‑sectional SEM showing FTO/c‑TiO₂/mp‑TiO₂/MAPbI₃/Spiro‑OMeTAD/Ag.

For each substrate, four devices were fabricated. Devices with at least three cells showing a maximum PCE deviation <3 % were included in the statistical analysis.

Characterization Techniques

Current–voltage (J‑V) curves were recorded with a Keithley 2400 under a calibrated AM 1.5G solar simulator (100 mW cm⁻²). Scans were performed in reverse bias from 1.2 V to –0.2 V at 100 mV s⁻¹. SEM imaging (Hitachi S‑4800, 15–60 kV) and X‑ray diffraction (Cu‑Kα, 10°–70°) characterized film morphology and crystallinity. UV–Vis absorption spectra were obtained on a Cary 5000 spectrophotometer (200–1200 nm, 1 nm steps). All measurements were conducted in ambient air without humidity control.

Results and Discussion

To isolate the effect of humidity, two‑step spin‑coating experiments were carried out at 30 °C under 0–45 % RH without BTA. As RH increased from 0 to 15 %, open‑circuit voltage (V_OC), short‑circuit current density (J_SC), fill factor (FF), and PCE improved markedly, confirming that a modest moisture level facilitates ion diffusion and crystalline ordering [13,23]. However, further increase to 45 % RH caused a steep drop in all parameters, with PCE falling to 5.02 % (V_OC = 1.00 V, J_SC = 9.84 mA cm⁻², FF = 51.02 %). Excess moisture induces rough, porous films and accelerates perovskite decomposition, undermining J_SC [18]. Thus, the optimal ambient humidity for BTA‑free devices is 15 % RH, yielding a maximum PCE of 13.21 % (average 12.48 %).

Box charts of a V_OC, b J_SC, c FF, and d PCE for mesoporous PSCs fabricated at different RHs (0–45 %) without BTA.

Incorporating a small BTA amount into the PbI₂ precursor markedly altered film morphology. At 25 % RH, pristine PbI₂ films exhibited inhomogeneous, porous surfaces with large grains (Fig. 3a), whereas BTA‑modified films were continuous, dense, and featured smaller grains (Fig. 3b). The volatility and Lewis‑base character of BTA isolate the precursor from ambient moisture, promoting uniform spreading and slowing crystallization to yield high‑quality films. Subsequent conversion to MAPbI₃ produced dense, pinhole‑free films with large grains (Fig. 3d), while the control exhibited small grains and pinholes (Fig. 3c). X‑ray diffraction confirmed comparable crystalline phases for all samples; however, the BTA‑modified PbI₂ film showed a reduced 12.69° peak intensity, indicative of finer grains and increased peak broadening (Fig. 4a). The MAPbI₃ films displayed characteristic tetragonal perovskite peaks (14.06°, 20.00°, 23.45°, 28.42°, 31.86°, 40.59°, 43.21°) and a residual PbI₂ peak at 12.69°, suggesting incomplete conversion in the absence of BTA (Fig. 4b). The improved crystallinity and reduced PbI₂ residue in BTA‑treated films correlate with superior photovoltaic performance.

SEM images of PbI₂ films on FTO/c‑TiO₂/mp‑TiO₂ substrates without (a) and with (b) BTA; corresponding MAPbI₃ films (c, d) prepared under 25 % RH.

XRD spectra of PbI₂ films (a) and MAPbI₃ films (b) on quartz substrates, with and without BTA, under 25 % RH. Inset in (a) shows the rocking curve of the main PbI₂ peak at 12.69°.

UV–Vis absorption spectra of MAPbI₃ films (Fig. 5) confirmed the presence of the perovskite band edge (~780 nm) and a weak shoulder near 510 nm, indicative of residual PbI₂. The BTA‑treated film displayed a slightly lower absorbance due to its thinner profile, as verified by cross‑sectional SEM (insets).

UV–Vis absorption spectra of MAPbI₃ films with and without BTA. Insets: cross‑sectional SEM of the films.

Devices fabricated from BTA‑enhanced perovskite films achieved a peak PCE of 16.00 %, with J_SC = 22.29 mA cm⁻², V_OC = 1.10 V, and FF = 65.25 %—a ~40 % improvement over BTA‑free devices (max. PCE = 11.38 %). The enhanced performance arises from the smoother, larger‑grain perovskite layers that facilitate efficient charge extraction and suppress recombination.

J‑V curves of PSCs fabricated with and without BTA under 25 % RH. Inset: detailed photovoltaic parameters.

Conclusions

We demonstrated that modest ambient humidity accelerates PbI₂ crystallization, leading to rough, large‑grain films that impair device performance. However, a small moisture level (≈25 % RH) is harmless—or even beneficial—for the PbI₂‑to‑MAPbI₃ conversion when the precursor is pre‑treated with n‑butyl amine. BTA improves the uniformity and crystallinity of PbI₂, which translates into dense, high‑quality MAPbI₃ films and PSCs delivering 16.00 % PCE in ambient air. This strategy offers a scalable, glove‑box‑free route to high‑efficiency perovskite solar cells.

Abbreviations

BTA:

N‑butyl amine

c‑TiO₂:

Compact TiO₂

DMF:

N,N‑dimethylformamide

FF:

Fill factor

FTO:

Fluorine‑doped SnO₂‑coated glass

J_SC:

Short‑circuit current density

J‑V:

Current density‑voltage characteristics

Li‑TFSI:

Lithium bis(trifluoromethanesulfonyl)imide

MA:

Methylammonium

mp‑TiO₂:

Mesoporous TiO₂

PCE:

Power conversion efficiency

PSCs:

Perovskite solar cells

RH:

Relative humidity

SEM:

Scanning electron microscopy

Spiro‑OMeTAD:

2,2′,7,7′‑Tetrakis[N,N‑di(4‑methoxyphenyl)amino]‑9,9′‑spiro‑bifluorene

TBP:

4‑tert‑butylpyridine

V_OC:

Open‑circuit voltage

XRD:

X‑ray diffraction