Bidentate coordination-induced trap passivation and phase stability in perovskite solar cells via ionic liquid engineering

Abstract

Ionic liquid (IL) engineering has emerged as a promising strategy to improve the performance and stability of perovskite solar cells (PSCs), especially under ambient processing conditions. In this work, we investigate the role of 1-(2-ethoxyethyl)-1-methylpyrrolidinium dicyanamide (Pyr-DCA) as an additive for perovskite precursor solutions and compare its passivation effects with those of the widely used thiocyanate (SCN⁻)-based IL. Density functional theory (DFT) simulations reveal that DCA⁻ exhibits stronger binding affinity to undercoordinated Pb²⁺ ions due to its bidentate nitrogen coordination, effectively passivating deep-level trap states. Incorporation of Pyr-DCA into the perovskite film leads to increased grain size, improved crystallinity, and lower trap density, resulting in enhanced charge carrier lifetimes and reduced nonradiative recombination. Devices treated with Pyr-DCA show improved power conversion efficiency (PCE), moisture resistance, and long-term operational stability. In situ GIWAXS measurements performed under 1 Sun illumination and electrical bias confirm that DCA⁻ suppresses the formation of degradation-associated δ-phase and PbI₂, maintaining the structural integrity of the perovskite α-phase. This work highlights the dual chemical and structural stabilization offered by DCA⁻ and demonstrates its promise for enabling scalable and stable PSC fabrication under ambient conditions.

---

Jinseck Kim

Dr Jinseck Kim received his BS in nano science and technology from Pusan National University in 2009 and his MS in chemical engineering from POSTECH in 2011. He then joined LG Chem as a research scientist, working on development of organic solar cell and soluble OLED materials. From 2019 to 2023, he received his PhD at KAIST through an industrial research program focusing on organic solar cells, perovskites, and electrochromic materials. He is currently an Associate Professor in the Department of Polymer–Nano Science and Technology at Jeonbuk National University. His research centers on conjugated molecules and polymers for organic electronics and energy applications.

1 Introduction

Organic–inorganic halide perovskites have emerged as leading materials in the field of photovoltaics due to their exceptional optoelectronic properties, including high absorption coefficients, long carrier diffusion lengths, and tunable bandgaps.^1–3 Since their first application in solar cells, power conversion efficiencies (PCEs) of perovskite solar cells (PSCs) have rapidly risen, now surpassing 26%,⁴ approaching those of traditional silicon-based photovoltaics. In addition to solar energy harvesting, their versatile properties make them suitable for use in light-emitting diodes (LEDs),^5–8 photodetectors,^9–13 and chemical sensors.^14–17 However, the presence of intrinsic defects such as halide vacancies, undercoordinated metal cations, and surface/interface trap states critically undermines device stability and long-term operational performance, particularly under ambient environmental conditions.^18,19

To address these challenges, a wide range of defect passivation strategies have been developed. Additive engineering, introducing small molecules, polymers, halide salts, and ionic liquids (ILs), has proven particularly effective for tailoring the perovskite microstructure and suppressing trap-assisted recombination. Among these, ionic liquids are of notable interest due to their dual capability to modulate interfacial energetics and improve environmental tolerance. ILs consist of bulky organic cations and diverse functional anions, offering a modular platform for surface engineering.²⁰ Anions such as thiocyanate (SCN⁻), dicyanamide (DCA⁻), bis(trifluoromethylsulfonyl)imide (TFSI⁻), and tetrafluoroborate (BF₄⁻) have been used to interact with undercoordinated Pb²⁺ or halide vacancies, effectively passivating deep trap states.^21–27 Cations bearing functional groups such as pyrrolidinium or imidazolium rings further enhance charge transport and environmental stability by suppressing moisture infiltration and reinforcing film morphology.^28–31

Despite extensive exploration of SCN⁻-based ILs, the comparative effectiveness of other anions like DCA⁻ has not been fully elucidated. Dicyanamide, in particular, features multiple nitrogen donor atoms capable of coordinating with undercoordinated Pb²⁺, offering potentially stronger passivation than SCN⁻. In this study, we systematically compare the surface passivation behavior and device performance of PSCs treated with DCA⁻versus SCN⁻-based ILs. Employing pyrrolidinium-based cations (Pyr⁺) to ensure hydrophobicity and moisture resistance, we analyze the effect of these ILs on defect suppression, charge carrier dynamics, and long-term device stability. Notably, all devices were fabricated under ambient atmospheric conditions, underscoring the potential of DCA⁻-based passivation strategies for scalable, low-cost PSC manufacturing.

2 Results & discussion

Perovskite solar cells were fabricated in a conventional n–i–p architecture with the device structure: ITO/SnO₂/perovskite/Spiro-OMeTAD/Au. The ionic liquids 1-(2-ethoxyethyl)-1-methylpyrrolidinium thiocyanate (Pyr-SCN) and 1-(2-ethoxyethyl)-1-methylpyrrolidinium dicyanamide (Pyr-DCA) were introduced by direct mixing into the perovskite precursor solution (Fig. 1a). The films were deposited via a one-step spin-coating method under ambient atmospheric conditions. Further details of the synthesis and fabrication procedures are described in the Experimental section.

Fig. 1 (a) Schematic diagram of PSCs fabrication with IL (Pyr-SCN, Pyr-DCA). Top-view SEM image of (b) pristine, (c) Pyr-SCN and (d) Pyr-DCA. AFM image of (e) pristine, (f) Pyr-SCN and (g) Pyr-DCA.

The surface morphology and microstructure of the resulting perovskite films were examined using scanning electron microscopy (SEM), as shown in Fig. 1b–d. In all samples, bright particles attributed to residual PbI₂ were observed. In the pristine sample, PbI₂ appeared as isolated grains, while in the Pyr-SCN and Pyr-DCA treated films, it manifested as thin plate-like features located along the grain boundaries. Previous studies have suggested that such lamellar PbI₂ passivates intergranular defects more effectively than particulate PbI₂, potentially contributing to enhanced device stability.^32–34 Grain size analysis revealed an average crystal size of approximately 600 nm for the pristine film, which increased to 760 nm and 822 nm in Pyr-SCN and Pyr-DCA treated films, respectively. The Pyr-DCA film exhibited the largest and most uniformly distributed grains, as corroborated by the inset images.

Atomic force microscopy (AFM) images were obtained to assess the surface topography of the films (Fig. 1e–g). The measured root-mean-square (RMS) roughness values were 41.1 nm (pristine), 39.8 nm (Pyr-SCN), and 41.9 nm (Pyr-DCA). A slight increase in surface roughness was observed in the Pyr-DCA sample, which can be attributed to the formation of PbI₂ platelets aligned along the plane of the perovskite grains, contributing to topographic undulations, as well as the enlargement of grain size.³⁵

Fig. 2a–c presents grazing-incidence wide-angle X-ray scattering (GIWAXS) images of the perovskite films. In the out-of-plane direction (q_xy = 0), distinct diffraction features are observed at q_z ≈ 0.78, 0.84, and 0.93 Å⁻¹, corresponding to the δ-phase of degraded perovskite, residual PbI₂, and the α-phase of perovskite, respectively. To gain quantitative insights, the intensity profiles along q_z at q_xy = 0 were extracted for each sample (Fig. 2d–f). The pristine film displays a dominant PbI₂ peak with a relatively weak α-phase signal, suggesting incomplete perovskite crystallization under ambient processing conditions. In contrast, IL-treated films exhibit reduced PbI₂ intensity and significantly enhanced α-phase peaks, indicating improved phase purity and crystallization.

Fig. 2 GIWAXS image of (a) pristine, (b) Pyr-SCN and (c) Pyr-DCA. Out-of plane (q_xy = 0 Å⁻¹) of (d) pristine, (e) Pyr-SCN and (f) Pyr-DCA. FWHM at q_z ≈ 0.92 Å⁻¹ (100 phase) of each perovskite film. XPS data of (g) Pb 4f and (h) N 1s of each perovskite film. (i) PL data of each perovskite, which structure was perovskite on glass. PL mapping images of (j) pristine, (k) Pyr-SCN and (l) Pyr-DCA. (m–o) OM image corresponding to the PL mapping area.

To assess the degree of crystallinity, the full-width at half-maximum (FWHM) values of the α-phase peaks were compared (Fig. 2d–f). A narrowing trend from pristine to Pyr-SCN and further to Pyr-DCA is observed, consistent with an increase in perovskite crystallite size. This trend aligns well with the SEM-derived grain size analysis and reinforces the role of IL additives, particularly Pyr-DCA, in promoting grain growth and phase stabilization.

Furthermore, a pronounced diffraction peak at 0.84 Å⁻¹ was observed in the pristine film, corresponding to PbI₂, indicating the formation of large crystalline PbI₂ domains. In contrast, this peak was significantly suppressed in the IL-treated films, suggesting more dispersed PbI₂ formation. Notably, the α-phase perovskite diffraction peak near 0.93 Å⁻¹ exhibited the highest intensity in the Pyr-DCA sample, indicative of enhanced crystallinity. This result aligns with the observed improvements in grain size and uniformity, supporting the role of DCA⁻ in promoting more ordered crystal growth during film formation.

X-ray photoelectron spectroscopy (XPS) was employed to analyze elemental composition and electronic environments within the films. The Pb 4f spectra (Fig. 2g) reveal a weak shoulder feature in the pristine sample, attributed to metallic Pb⁰. This signal originates from incomplete incorporation of Pb into the perovskite lattice, often associated with suboptimal crystallization. Notably, this Pb⁰ feature is significantly suppressed in both Pyr-SCN and Pyr-DCA samples, indicating improved film formation and reduced metallic lead byproduct.

Further insights into the chemical interactions were gained from N 1s XPS spectra (Fig. 2h). The Pyr-DCA-treated film shows a distinct peak corresponding to the C≡N group of the DCA⁻ anion, confirming its incorporation. Additionally, the C–N peak attributed to the FA⁺ cation shifts to lower binding energy in Pyr-DCA compared to Pyr-SCN, suggesting stronger electronic interaction or charge transfer between the DCA⁻ anion and the perovskite lattice. This supports the hypothesis that DCA⁻ contributes more effectively to defect passivation via coordination with undercoordinated Pb²⁺ sites.

Steady-state photoluminescence (PL) spectra of the films coated on glass substrates are displayed in Fig. 2i. The IL-treated films exhibit substantially higher PL intensity compared to the pristine sample, indicating reduced nonradiative recombination. This enhancement is attributed to effective passivation of trap states by IL anions, which likely fill halide vacancies and coordinate undercoordinated Pb²⁺ centers. A slight blue shift in the PL peak position is also observed, consistent with UV-vis absorption data and bandgap changes.

To further probe spatial distribution of nonradiative defects, PL mapping was performed on film regions containing cracks (Fig. 2j–l). The crack sites refer to the presence of surface wrinkles. The optical images of PL mapping sites were shown in Fig. 2m–o. The pristine film exhibits almost no PL emission at crack sites, suggesting localized defect accumulation and severe nonradiative recombination. In contrast, both IL-treated films, particularly Pyr-DCA, show significantly higher PL intensity in these areas, implying effective suppression of trap states even in structurally disrupted regions. These observations underscore the role of IL additives in passivating surface and grain boundary defects, thereby improving optoelectronic quality and photostability of perovskite films processed under ambient conditions.

Ultraviolet photoelectron spectroscopy (UPS) was conducted to determine the work function (WF) and valence band maximum (VBM) of the films (Fig. 3a–c). These were calculated using the equations:

WF = hν − E_cutoff, VBM = − (WF + E_initial) (1)

where hν = 21.2 eV. The calculated WF values are −3.88 eV (pristine), −4.11 eV (Pyr-SCN), and −4.03 eV (Pyr-DCA). Corresponding VBM values are approximately −5.40 eV, −5.61 eV, and −5.55 eV, respectively. This is attributed to the passivation of traps induced by iodine vacancies by SCN⁻ and DCA⁻ anions, as well as their low electrostatic potential, which leads to a reduction in the energy level. The downward shift in energy levels for IL-treated samples suggests favorable energy alignment at the electron transport layer interface, potentially improving electron extraction and reducing interfacial recombination losses.

Fig. 3 (a–c) UPS spectra of the processed perovskite films with Pyr-SCN and Pyr-DCA. (d) Absorption spectra of the processed perovskite films with Pyr-SCN and Pyr-DCA. (e) Energy level diagram of each layer forming perovskite solar cells.

Optical absorption spectra (Fig. 3d) show a consistent blue shift upon IL treatment, implying a slight widening of the bandgap. Tauc plot analysis estimates the bandgaps to be 1.526 eV (pristine) and 1.541 eV for both Pyr-SCN and Pyr-DCA. An increased bandgap can be advantageous for improving the open-circuit voltage (V_oc) of solar cells, contributing to higher overall efficiency.

A schematic energy band diagram constructed from UPS and optical bandgap data is shown in Fig. 3e. The IL-treated films exhibit lower-lying conduction bands relative to the pristine sample, improving band alignment with the electron transport layer (ETL) and promoting more efficient electron extractions.

To gain deeper insight into the chemical interactions between the IL anions and the perovskite surface, density functional theory (DFT) calculations were performed to simulate the binding energy between each anion and an undercoordinated Pb²⁺ site on the perovskite surface (Fig. 4a). The binding energy of the SCN⁻ ion with Pb²⁺ was calculated to be −1.85 eV, whereas the DCA⁻ ion showed a stronger interaction with a binding energy of −2.10 eV. This enhanced binding affinity suggests that DCA⁻ can more effectively coordinate with undercoordinated Pb²⁺, leading to superior passivation of surface defects and enhanced crystal stability. These results support the observed improvements in film morphology, crystallinity, and defect suppression in Pyr-DCA-treated samples.

Fig. 4 (a) Binding energy between SCN⁻ ion and Pb²⁺ ion, and DCA⁻ ion and Pb²⁺ ion, which were calculated with DFT. (b) Light intensity dependence of V_oc and ideality factor. (c) Dark J–V curves of each PSCs. (d) Mott–Schottky and (e) Nyquist plot of PSCs with and without IL. (f–h) SCLC of perovskite electron only devices. Device was constructed of ITO/SnO₂/perovskite/C₆₀/Ag. (i) Cross-sectional SEM of PSC. (j) TRPL of perovskite film on glass. (k) EQE and integrated J_sc of each PSC.

The ideality factor (n_id) of the PSCs was extracted from the slope of the V_OCversus light intensity curve using the equation:

where I is the light intensity, k is the Boltzmann constant, T is the temperature, and q is the elementary charge (Fig. 4b). The ideality factor values were found to be 1.38 for the pristine device, 1.25 for Pyr-SCN, and 1.30 for Pyr-DCA. Since values closer to 1 are indicative of reduced trap-assisted recombination, these results further confirm that the incorporation of ILs leads to more ideal diode behavior due to improved defect passivation. The short-circuit current density (J_sc) was also analyzed as a function of light intensity according to the power law I ∝ J_sc^α. The extracted α values were 0.990 (pristine), 0.995 (Pyr-SCN), and 0.990 (Pyr-DCA), indicating near-ideal photogenerated current behavior across all samples, with minimal bimolecular recombination.

Dark current–voltage measurements revealed that incorporation of ILs led to a marked suppression in dark current over the entire bias range, indicating reduced leakage and improved diode rectification behavior (Fig. 4c). Linear extrapolation in the semi-log region allowed estimation of the reverse saturation current, which was found to be lowest in Pyr-DCA devices, consistent with superior trap passivation.

To examine charge accumulation and built-in potential, Mott–Schottky analysis was conducted (Fig. 4d). The built-in potential (V_bi) values increased from 0.979 V (pristine) to 1.00 V (Pyr-SCN) and 1.045 V (Pyr-DCA), suggesting improved electric field formation across the junction in IL-treated devices. Complementary impedance measurements were performed using electrochemical impedance spectroscopy (EIS), with Nyquist plots presented in Fig. 4e. The high-frequency intercept, indicative of series resistance, decreased from 19.1 Ω (pristine) to 8.3 Ω (Pyr-SCN) and 13.2 Ω (Pyr-DCA), which supports improved charge transport across interfaces. The combined effects of increased V_bi and reduced series resistance are likely to contribute to enhanced V_oc in IL-treated devices.

Trap density (n_t) was quantified via space-charge-limited current (SCLC) measurements using electron-only devices (ITO/SnO₂/perovskite/C₆₀/Ag) (Fig. 4f–h). The trap-filled limit voltage (V_TFL) was identified from the inflection point in the J–V curve and found to be 0.24 V (pristine), 0.13 V (Pyr-SCN), and 0.12 V (Pyr-DCA). Using the relationship:

where ε is the relative dielectric constant of the perovskite (46.9),³⁶ε₀ is the vacuum permittivity (8.85 × 10⁻¹² F m⁻¹), q is the elementary charge, and L is the film thickness (540 nm, Fig. 4i), the trap densities were calculated as 4.27 × 10¹⁵ cm⁻³ (pristine), 2.31 × 10¹⁵ cm⁻³ (Pyr-SCN), and 2.13 × 10¹⁵ cm⁻³ (Pyr-DCA). The significantly reduced trap densities affirm that both SCN⁻ and DCA⁻ passivate electronic defects, with DCA⁻ offering marginally better suppression.

Time-resolved photoluminescence (TRPL) decay was fitted using a bi-exponential function to distinguish radiative and non-radiative recombination pathways (Fig. 4j). The pristine film exhibited a higher proportion of the fast-decay component (A₁), corresponding to non-radiative recombination, while IL-treated films showed increased slow-decay components (A₂), indicative of enhanced radiative recombination. The average carrier lifetimes, calculated via:

were 78.40 ns (pristine), 154.76 ns (Pyr-SCN), and 181.65 ns (Pyr-DCA), again highlighting the improved optoelectronic quality of IL-treated films, particularly those treated with DCA⁻. The detailed parameters were shown in Table 1.

Table 1 TRPL parameters

	A 1	τ 1	A 2	τ 2	τ avg
Pristine	7.54	1.74	0.27	111.59	78.40
Pyr-SCN	0.51	24.32	0.34	181.07	154.76
Pyr-DCA	0.50	18.70	0.32	204.55	181.65

---

PSC devices with architecture ITO/SnO₂/perovskite/Spiro-OMeTAD/Au were fabricated to evaluate photovoltaic performance. External quantum efficiency (EQE) and J–V curves were conducted to verify the photovoltaic performance (Fig. 4k and 5a–c). Incorporation of ILs led to enhancements in both V_oc and J_sc, translating to an overall improvement in PCE (Table 2). EQE spectra confirmed the improved spectral response, particularly for Pyr-DCA, which showed higher EQE across the entire wavelength range. The integrated J_sc values, calculated from EQE data, increased from 23.62 mA cm⁻² (pristine) to 23.95 mA cm⁻² (Pyr-DCA), in agreement with the J–V measurements. Furthermore, the hysteresis index (HI) decreased from 0.0348 for the pristine device to 0.0054 for the Pyr-DCA device, indicating that ion migration was effectively suppressed by the incorporation of the Pyr-DCA. The decrease in the HI can be attributed to the reduced trap density and the decrease in series resistance. These findings indicate that the enhanced optical absorption, reduced trap density, and favorable band alignment achieved through DCA⁻ passivation directly contribute to superior photovoltaic performance under ambient processing conditions. The concentration of Pyr-DCA reached highest value at 5 mM, suggesting that this concentration yields the most significant effect (Fig. 5d). To verify reproducibility, PCE and J_sc were evaluated across multiple devices (Fig. 5e and f).

Fig. 5 (a–c) J–V curves of PSCs under 1 Sun illumination. (d) PCE of PSCs as a function of Pyr-DCA concentration. (e–f) Box plots of the PCE and J_sc distributions for individual PSCs.

Table 2 J–V curves of PSCs

	Scan direction	V oc (V)	J sc (mA cm^−2)	FF (%)	PCE (%)	HI^a
a .
Pristine	Forward	1.120	23.49	74.73	19.65	0.0348
	Reverse	1.116	23.44	72.62	18.99
Pyr-SCN	Forward	1.109	24.05	74.12	19.77	0.0373
	Reverse	1.109	24.01	71.56	19.06
Pyr-DCA	Forward	1.125	24.08	74.41	20.16	0.0054
	Reverse	1.134	24.05	74.34	20.27

---

To assess the moisture stability conferred by IL additives, perovskite films were exposed to ambient air at 85% relative humidity (RH) in the dark (Fig. 6a). The pristine perovskite film underwent rapid degradation, visibly turning from black to yellow within 2 days, and fully decomposed after 20 days of exposure. In stark contrast, both Pyr-SCN and Pyr-DCA treated films preserved their black color and structural integrity for the full 20 days period, demonstrating the ILs' ability to enhance environmental robustness.

Fig. 6 (a) Photographs of perovskite film changing at RH 85% and dark condition and water contact angle of perovskite films. (b–d) Aging AFM images of perovskite films under RH 90%, 2 h. Insets are enlarged images of perovskite grain. (e) Aging PCE and (f) J_sc of PSCs under ambient air (RH 50%, dark). (g) Aging PCE of PSCs under high humidity (RH 85%, dark). (h) MPPT of each PSC.

Water contact angle (WCA) measurements were performed to quantify surface hydrophobicity of the perovskite films (Fig. 6a). The pristine film exhibited a low contact angle of 48.9°, indicative of poor water resistance. In comparison, Pyr-SCN and Pyr-DCA treated films showed increased contact angles of 66.8° and 66.7°, respectively. The enhanced hydrophobicity is attributed to the presence of alkyl-functionalized pyrrolidinium cations, which mitigate moisture penetration and thereby contribute to improved film stability.

AFM analysis was conducted on films exposed to 90% RH for 2 h to evaluate morphological degradation (Fig. 6b–d). The pristine film surface exhibited the emergence of nanoparticulates, consistent with the formation of PbI₂, a byproduct of perovskite decomposition.³⁷ In contrast, the surfaces of Pyr-SCN and Pyr-DCA treated films remained smooth and devoid of such degradation features, corroborating their superior resistance to humidity-induced decomposition.

Device-level operational stability was also investigated under ambient storage conditions. Photovoltaic devices were maintained in the dark at 50% RH, and their performance was monitored over time (Fig. 6e). Pristine PSCs experienced a rapid decline in PCE, dropping below 40% of their initial value within 200 h. Meanwhile, the Pyr-DCA treated devices retained more than 50% of their initial PCE even after 1000 h. Notably, J_sc retention exceeded 80% over the same duration (Fig. 6f), indicating that IL incorporation effectively suppresses moisture-induced degradation and preserves carrier transport properties.

To specifically evaluate the devices under high humidity stress, the PSCs were stored in the dark at 85% RH (Fig. 6g). The pristine devices exhibited catastrophic degradation, losing virtually all photovoltaic activity within 100 h. In contrast, devices containing ILs—particularly those treated with Pyr-DCA—retained measurable performance levels after prolonged exposure. These results collectively highlight the dual function of the IL additives: enhancing both the intrinsic material stability of the perovskite film and the operational durability of the device under humid conditions.

To evaluate the operational stability under realistic conditions, maximum power point tracking (MPPT) measurements were conducted under 1 Sun illumination, at a temperature of 25–28 °C, and RH 70–80%, without any encapsulation (Fig. 6h). While the pristine PSC retained only 60% of its initial efficiency after approximately 30 h, the device incorporating Pyr-DCA maintained over 80% of its initial efficiency for more than 100 h.

To elucidate the structural evolution of perovskite films under realistic operating conditions, in situ grazing-incidence wide-angle X-ray scattering (GIWAXS) measurements were conducted using a custom-designed synchrotron beamline setup (Fig. 7a–c). This system uniquely enables simultaneous exposure of perovskite solar cells to continuous 1 Sun illumination and an applied electrical bias (35 V), while recording the dynamic evolution of the crystal structure in real time. The coupling of optoelectronic operation with structural monitoring allows for a direct correlation between performance-relevant stress factors and crystallographic changes, offering valuable insights into degradation mechanisms.

Fig. 7 (a–c) In situ GIWAXS of perovskite film under ambient air, 1 Sun illumination and 35 V for 1 h. (d) Out-of-plane of in situ GIWAXS at q_xy = 0 at 0 s and 3600 s. (e) Normalized peak of α-phase/(α-phase + PbI₂).

For the pristine perovskite film, initial GIWAXS patterns revealed prominent reflections corresponding to the α-phase perovskite structure, residual PbI₂, and the δ-phase, an intermediate phase associated with perovskite decomposition (Fig. 7a and d). Over the course of 1 h, the intensities of both α-phase and δ-phase peaks decreased progressively, while the PbI₂ peak remained relatively constant. This apparent stability of the PbI₂ signal is misleading; it results from continuous conversion of degraded perovskite into PbI₂, thus maintaining its presence even as the α-phase structure collapses. These observations are indicative of an accelerated degradation pathway in the pristine sample, where the α-phase is readily destabilized under combined photoelectrical and X-ray stress, generating new PbI₂ that eventually decomposes into metallic Pb⁰.³⁸

In contrast, perovskite films incorporating IL additives (SCN and DCA) exhibited markedly improved structural resilience under identical in situ conditions (Fig. 7b–d). Notably, no δ-phase signals were detected at any time, and the α-phase peak intensity exhibited only a gradual and moderate decline. The absence of newly formed PbI₂, alongside a progressive decrease in its pre-existing peak, suggests that IL-treated films are less prone to decomposition. This implies that the IL effectively inhibits the structural transition from perovskite to degradation products by stabilizing the lattice and passivating defect sites.

To quantitatively evaluate the compositional evolution between the photoactive perovskite α-phase and its degradation product PbI₂, we extracted the ratio of α/(α + PbI₂) from the GIWAXS data over time (Fig. 7e). This metric offers a normalized and reliable indicator for tracking the stability of the α-phase under combined stress conditions, including continuous 1 Sun illumination, electrical bias, ambient atmosphere, and X-ray irradiation—thus reflecting an accelerated degradation environment. As shown in the graph, pristine perovskite films exhibit a rapid and continuous decrease in the α/(α + PbI₂) ratio, indicating significant α-phase loss and concurrent PbI₂ formation. In contrast, films incorporating Pyr-SCN and Pyr-DCA ionic liquids show remarkably suppressed PbI₂ generation, maintaining a higher proportion of the α-phase throughout the 1 h test. Notably, Pyr-DCA–treated films demonstrate the smallest variation in the α-phase ratio over time, underscoring its superior capacity to inhibit phase decomposition and stabilize the perovskite lattice under harsh operating conditions.

Importantly, the stability of the α-phase in IL-treated films under simultaneous illumination and electrical bias, conditions known to accelerate both ion migration and lattice disruption, demonstrates the critical role of ILs in suppressing phase segregation and mitigating photo-induced damage. The absence of the δ-phase throughout the experiment further supports the notion that ILs not only enhance static humidity stability but also actively protect the perovskite structure under operational stresses. This advanced in situ methodology thus validates the dynamic stabilizing effect of ILs and highlights the importance of real-time structural monitoring in evaluating material robustness for practical photovoltaic applications.

3 Conclusion

This study systematically elucidates the multifunctional role of dicyanamide (DCA⁻)-based ionic liquids in enhancing the performance and stability of perovskite solar cells processed under ambient conditions. Among the ionic liquids tested, 1-(2-ethoxyethyl)-1-methylpyrrolidinium dicyanamide (Pyr-DCA) proved most effective due to its strong coordination capability with undercoordinated Pb²⁺ ions, as confirmed by DFT-calculated binding energies. The DCA⁻ anion, featuring multiple electron-rich nitrogen sites within its linear –N(CN)₂ structure, enables robust Lewis base interactions with the perovskite surface, thereby efficiently passivating defects and suppressing nonradiative recombination. As a result, Pyr-DCA-treated films exhibited enlarged and more uniform grains, reduced trap density, enhanced charge carrier lifetime, and improved crystallinity compared to both pristine and SCN⁻-treated counterparts. These material-level improvements translated into superior device performance, with enhanced V_oc, J_sc, and long-term operational stability. Moisture resistance was also significantly improved, attributed to the hydrophobic pyrrolidinium cation and the suppression of PbI₂ formation under humid conditions. Importantly, in situ GIWAXS measurements under simultaneous illumination and electrical bias—enabled by a custom-built synchrotron beamline setup—revealed that Pyr-DCA effectively stabilizes the perovskite α-phase and prevents degradation pathways involving δ-phase and PbI₂ formation. These findings underscore the mechanistic advantage of DCA⁻ in perovskite surface engineering and highlight its potential as a strategic component in scalable, durable, and high-performance perovskite solar cells.

4 Experimental method

4.1 Materials

Methylammonium chloride (MACl, > 98%), N,N-dimethyl formamide (DMF, 99.8%), dimethyl sulfoxide (DMSO, 99.9%), chlorobenzene (CB, 99.8%), ethyl acetate (EtOAc, anhydrous, 99.8%), acetonitrile (99.8%), bis(trifluoromethane)sulfonamide lithium salt (Li-TFSI), 4–tert-butylpyridine (tBP, 98%), and Spiro-MeOTAD (99%, HPLC) were purchased from Sigma-Aldrich. Acetone (99.5%) and IPA (99.8%) were purchased from Samchun, Korea. SnO₂ (15% in H₂O colloidal dispersion) was purchased from Alfa Aesar. PbI₂ (99.99%, trace metal basis) and PbBr₂ (low water content) were purchased from TCI. Formamidinium iodide (FAI, > 99.99%), methylammonium bromide (MABr) were purchased from GreatcellSolar.

4.2 Device fabrication

ITO glass substrates (18 mm × 18 mm) were sequentially cleaned by sonication in detergent, deionized water, acetone, and IPA, followed by drying under nitrogen gas. The cleaned ITO substrates were treated with ultraviolet ozone (UVO) for 20 min, followed by spin-coating of colloidal SnO₂, which had been diluted with deionized water at a volume ratio of 1 : 4 and filtered through a 0.2 μm syringe filter. The films were then spin-coated at 3000 rpm for 30 s and annealed at 180 °C for 30 min. Subsequently, the substrates were treated with UVO for 15 min, followed by spin-coating of 100 μL of perovskite precursor solution at 4000 rpm for 30 s. During the spin-coating process, 400 μL of EtOAc was rapidly dispensed onto the substrate 15 s before the end of the spin. After spin-coating, the films were annealed at 150 °C for 15 min. The perovskite precursor solution was prepared by dissolving PbI₂ (731.11 mg), PbBr₂ (16.5 mg), FAI (250.22 mg), MACl (34 mg), and MABr (5.056 mg) in 1 mL of a mixed solvent of DMF and DMSO (v/v = 8 : 1). The ionic liquids, Pyr-SCN and Pyr-DCA, were each added to the perovskite precursor solutions at a concentration of 5 mM. And then, spiro-OMeTAD was spin-coated onto the perovskite layer at 4000 rpm for 30 s. The spiro-OMeTAD solution was prepared at a concentration of 72.3 mg mL⁻¹ in CB and doped 30 min prior to coating with 28.8 μL of tBP and 17.5 μL of a Li-TFSI solution (52 mg mL⁻¹ in acetonitrile). The resulting solution was filtered through a 0.45 μm syringe filter before use. All coating processes were carried out under ambient air conditions (25–30 °C, RH 50–60%). After completing all the coating steps, the films were stored in humidity-controlled air over 24 h. Subsequently, 55 nm of gold was thermally evaporated to form the top electrode. The active area of the resulting device was 0.12 cm².

4.3 DFT simulations

First-principles calculations based on density functional theory (DFT) were performed using the Quantum ESPRESSO package to evaluate the interaction between the dicyanamide (DCA⁻) anion and the perovskite surface. The generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) exchange–correlation functional was employed. Core–valence interactions were described using projector augmented-wave (PAW) pseudopotentials. A 2 × 2 slab model of the (001) surface of MAPbI₃, terminated with PbI₂, was constructed to mimic the experimentally relevant surface. A vacuum layer of at least 15 Å was included to avoid spurious interactions between periodic images. The Brillouin zone was sampled using a 3 × 3 × 1 Monkhorst–Pack k-point grid, and a kinetic energy cutoff of 50 Ry was used for the wavefunctions. The DCA⁻ molecule was initially placed near undercoordinated Pb²⁺ surface atoms in multiple configurations to explore favorable binding geometries. Full structural relaxation was performed until forces on each atom were below 0.01 eV Å⁻¹. The binding energy (E_bind) of DCA⁻ on the perovskite surface was calculated using the following equation:

E_bind = E_total − (E_surface + E_DCA) (5)

where E_total is the total energy of the perovskite surface with the DCA⁻ adsorbed, E_surface is the energy of the clean perovskite slab, and EDCA is the energy of the isolated DCA⁻ ion in vacuum. The optimized geometry shows bidentate coordination of the nitrogen atoms in DCA⁻ with a surface Pb²⁺ ion, resulting in a strong binding energy of approximately −1.54 eV, indicating effective chemical passivation of undercoordinated sites. In contrast, SCN⁻ displayed a weaker monodentate interaction with a binding energy of ∼ −0.92 eV under similar conditions, suggesting superior passivation capacity of DCA⁻.

4.4 Characterization

SEM images of the perovskites were obtained using field-emission SEM (S-4700, Hitachi) at the Center for University-wide Research Facilities (CURF) of Jeonbuk National University. AFM images were obtained using Multimode-8 (BRUKER) at the CURF of Jeonbuk National University. XRD was conducted using a D8 Advance (Bruker) instrument with Cu radiation as the X-ray source and Eiger2 R500K as the detector. GIWAXS and in situ GIWAXS measurements were conducted at the 3C and 9A beamlines at the Pohang Accelerator Laboratory (PAL), Republic of Korea. In situ GIWAXS samples were prepared in a lateral structure by spin-coating the perovskite layer onto a silicon oxide substrate, followed by thermal evaporation of gold electrodes with a spacing of 2 cm. XPS and UPS analyses were performed using a Nexsa XPS system (Thermo Fisher Scientific) equipped with Al Kα radiation source (1486.6 eV) as the X-ray source and He I (21.2 eV) as the UV source, respectively. XPS and UPS analyses were carried out at the Jeonju Center of the Korea Basic Science Institute (KBSI). PL data were measured using a Fluoromax plus C (Horiba) system equipped with a 150 W ozone free xenon arc lamp as the light source. TRPL data were obtained using XperRF (NANO BASE) equipped with 405 nm picosecond pulsed diode laser and OLYMPUS ×40 objective lens. PL mapping was conducted using XperRAM S Series (NANO BASE) equipped with 300 l mm⁻¹ grating at 532 nm and 100×/0.9 NA lens. UV-vis data were obtained using OPTIZEN POP (KLAB). The perovskite solar cells were measured using a solar simulator (LCS-100, Class ABB, Newport) in a glove box filled with argon gas. The light irradiance was calibrated to AM 1.5G (100 mW cm⁻²) using a Si-reference cell calibrated by MKS Instruments/Newport PV Lab before each measurement. J–V curves were measured with a 2450 SourceMeter (Keithley) using forward/reverse scan from −0.1 to 1.2 V with a step size of 0.02 V. EQE was measured using a Polaronix K3100 (Mcscience) equipped with 300 W Xe lamp as a light source. Contact angle data were measured using the Surfaceware 7 software. MPPT measurements were conducted under ambient air using a solar simulator (LumiSun-50, IEC Class A + A + A + compliant, Innovations In Optics) and K3600 Solar Cell Reliability Test System (McScience). The illumination source was calibrated to AM 1.5G (100 mW cm⁻²) using same Si-reference cell employed for the PCE measurements.

4.5 Operando GIWAXS measurements

Operando grazing-incidence wide-angle X-ray scattering (GIWAXS) measurements were performed at the 3C beamline of the Pohang Light Source II (PLS-II) in Korea to investigate the real-time structural evolution of perovskite films under operational conditions. Perovskite solar cell devices were mounted within a custom-built environmental chamber capable of temperature and humidity control, equipped with a solar simulator (AM 1.5G, 100 mW cm⁻²) for continuous illumination and a source meter (Keithley 2400) to apply electrical bias during measurement. The chamber environment was maintained at 25 °C and 30–40% relative humidity, unless otherwise stated. GIWAXS patterns were collected using a Pilatus 2 M detector, positioned at a sample-to-detector distance of approximately 250 mm, with a calibrated incident X-ray energy of 12.4 keV. A grazing incidence angle of 0.2° was used to probe the near-surface structure of the perovskite films with high sensitivity. Time-resolved GIWAXS data were recorded in situ under simultaneous illumination and bias to monitor structural changes during continuous device operation. Measurements were synchronized with the applied electrical conditions to correlate specific phase transitions with device performance parameters.

Conflicts of interest

There are no conflicts to declare.

Data availability

The datasets supporting the findings of this study are available from the corresponding author upon reasonable request. All relevant data, including raw GIWAXS images, DFT input/output files, and photovoltaic performance measurements, have been archived and can be shared in accordance with the data sharing policies of the Royal Society of Chemistry.

Acknowledgements

This work was supported by the Basic Study and Interdisciplinary R&D Foundation Fund of the University of Seoul (2025) for Min Kim. This paper was supported by research funds for newly appointed professors of Jeonbuk National University in 2024 for Jinseck Kim. This work was supported by the Korea Basic Science Institute (National research Facilities and Equipment Center) grant funded by the Korea government (MSIT) (RS-2024-00404963) for Jaewon Lee. This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT) (RS-2023-00214318) for Hye Min Oh.

References

1. W. Fang, L. Chen, W. Zhou, J. Wang, K. Huang, R. Zhu, J. Wu, B. Liu, Q. Fang and X. Wang, Korean J. Chem. Eng., 2025, 42, 307–322.
2. S.-B. Cho, M.-J. Kim and I.-K. Park, Korean J. Chem. Eng., 2025, 42, 81–88.
3. N. K. Elangovan, R. Kannadasan, B. Beenarani, M. H. Alsharif, M.-K. Kim and Z. H. Inamul, Energy Rep., 2024, 11, 1171–1190.
4. H. Chen, C. Liu, J. Xu, A. Maxwell, W. Zhou, Y. Yang, Q. Zhou, A. S. Bati, H. Wan and Z. Wang, Science, 2024, 384, 189–193.
5. M. Li, Y. Yang, Z. Kuang, C. Hao, S. Wang, F. Lu, Z. Liu, J. Liu, L. Zeng and Y. Cai, Nature, 2024, 630, 631–635.
6. Y. Lian, Y. Wang, Y. Yuan, Z. Ren, W. Tang, Z. Liu, S. Xing, K. Ji, B. Yuan and Y. Yang, Nature, 2025, 640, 62–68.
7. L. Kong, Y. Sun, B. Zhao, K. Ji, J. Feng, J. Dong, Y. Wang, Z. Liu, S. Maqbool and Y. Li, Nature, 2024, 631, 73–79.
8. S. Yuan, L. Dai, Y. Sun, F. Auras, Y.-H. Zhou, R.-Z. An, Y. Liu, C. Ding, C. Cassidy and X. Tang, Nat. Photonics, 2024, 18, 425–431.
9. S. I. Hussein, R. A. Ismail, N. J. Almashhadani and A. J. Addie, Opt. Mater., 2024, 115890.
10. L. Min, H. Sun, L. Guo, Y. Zhou, M. Wang, F. Cao and L. Li, Adv. Mater., 2024, 36, 2400279.
11. Y. Hu, X. Zhang, X. Mo, J. Chu, X. Fang and Z. Li, Adv. Funct. Mater., 2025, 35, 2412015.
12. L. Min, Y. Zhou, H. Sun, L. Guo, M. Wang, F. Cao, W. Tian and L. Li, Light: Sci. Appl., 2024, 13, 280.
13. S. M. Park, S. W. Park, H. Jin, D. Baek, J. Cha, W. S. Chae, T. K. Lee and M. Kim, Small, 2024, 20, 2404958.
14. D. Li, P. Zhuang and C. Sun, J. Mater. Chem. C, 2024, 12, 4544–4561.
15. D. Gao, Y. Zhuang, S. Gao, S. Huang and X. He, ACS Appl. Mater. Interfaces, 2024, 16, 6228–6238.
16. P.-C. Chen, R. Kirankumar, P.-Y. Lin, Z.-W. Chuang and S. Hsieh, J. Alloys Compd., 2024, 1002, 175320.
17. C. Zhang, Z. Wang, T. Chen, M. Tang, J. Zhang, J. Jian, Y. Wang, K. Zhou and X. Zhang, ACS Appl. Nano Mater., 2024, 7, 17329–17338.
18. L. Yan, H. Huang, P. Cui, S. Du, Z. Lan, Y. Yang, S. Qu, X. Wang, Q. Zhang and B. Liu, Nat. Energy, 2023, 8, 1158–1167.
19. J. Wang, L. Yuan, H. Luo, C. Duan, B. Zhou, Q. Wen and K. Yan, Chem. Eng. J., 2022, 446, 136968.
20. Z. Lei, B. Chen, Y.-M. Koo and D. R. MacFarlane, Chem. Rev., 2017, 117, 6633–6635.
21. S. Wang, H. Guo, J. Wu, Y. Lei, X. Li, Y. Fang, Y. Dai, W. Xiang and Y. Lin, Chem. Commun., 2022, 58, 8384–8387.
22. M.-S. Bae, Y. H. Chang, M. Q. Li, S.-W. Kim, W. Lee, J.-J. Bae, D. Kim, H.-S. Kim, J. Lim and J. Lee, Mater. Today Energy, 2024, 43, 101593.
23. M. Ebic, F. Sadegh, M. Ans, D. Prochowicz, P. Yadav, S. Satapathi and S. Akin, Small, 2024, 20, 2404190.
24. Y. Liu, C. Tang, A. Sun, R. Zhuang, Y. Zheng, C. Tian, X. Wu, Z. Li, B. Ouyang and J. Du, Mater. Horiz., 2023, 10, 5763–5774.
25. R. Panta, S. Chapagain, P. J. Armstrong, T. Druffel and C. A. Grapperhaus, ACS Appl. Energy Mater., 2024, 7, 10721–10729.
26. Y. Wang, Y. Yang, N. Li, M. Hu, S. R. Raga, Y. Jiang, C. Wang, X. L. Zhang, M. Lira-Cantu and F. Huang, Adv. Funct. Mater., 2022, 32, 2204396.
27. X. Shang, Z. He, J. Ding, M. Li, Z. Zhang, F. Meng and C. Chen, Sol. RRL, 2024, 8, 2400017.
28. K. Zhang, X. Zhang, K. G. Brooks, B. Ding, S. Kinge, Y. Ding, S. Dai and M. K. Nazeeruddin, Sol. RRL, 2023, 7, 2300115.
29. Y. Zou, J. Eichhorn, S. Rieger, Y. Zheng, S. Yuan, L. Wolz, L. V. Spanier, J. E. Heger, S. Yin and C. R. Everett, Nano Energy, 2023, 112, 108449.
30. X. Zhu, M. Du, J. Feng, H. Wang, Z. Xu, L. Wang, S. Zuo, C. Wang, Z. Wang and C. Zhang, Angew. Chem., Int. Ed., 2021, 60, 4238–4244.
31. X. Zhu, C. Wang, C. Zhang, Z. Wang, J. Feng, S. F. Liu and D. Yang, Chem. Eng. J., 2022, 434, 134759.
32. Y. Wang, Z. Shi, Y. Wang, Q. U. Khan, X. Li, L. Deng, Y. Pan, X. Zhang, Y. Yang and X. Yue, Adv. Mater., 2023, 35, 2302298.
33. C. Deng, J. Wu, Y. Yang, Y. Du, R. Li, Q. Chen, Y. Xu, W. Sun, Z. Lan and P. Gao, ACS Energy Lett., 2022, 8, 666–676.
34. Z. Ma, D. Huang, Q. Liu, G. Yan, Z. Xiao, D. Chen, J. Zhao, Y. Xiang, C. Peng and H. Li, J. Energy Chem., 2022, 66, 152–160.
35. J. Jeong, M. Kim, J. Seo, H. Lu, P. Ahlawat, A. Mishra, Y. Yang, M. A. Hope, F. T. Eickemeyer and M. Kim, Nature, 2021, 592, 381–385.
36. F. Li, X. Huang, C. Ma, J. Xue, Y. Li, D. Kim, H. S. Yang, Y. Zhang, B. R. Lee and J. Kim, Advanced Science, 2023, 10, 2301603.
37. F. Richheimer, D. Toth, B. Hailegnaw, M. A. Baker, R. A. Dorey, F. Kienberger, F. A. Castro, M. Kaltenbrunner, M. C. Scharber and G. Gramse, Nanoscale, 2022, 14, 2605–2616.
38. J. Ferrer Orri, T. A. Doherty, D. Johnstone, S. M. Collins, H. Simons, P. A. Midgley, C. Ducati and S. D. Stranks, Adv. Mater., 2022, 34, 2200383.

---

Footnote

† D. Baek, M. Q. Li, and J. Cha contributed equally to this work.

---