Nanoplasmonic sensing of CH₃NH₃PbI₃ perovskite formation in mimic of solar cell photoelectrodes

Inorganic-organic hybrid halide perovskites have shown great potential for a breakthrough of next generation solar devices.^1–7 Lead-based and lead-free perovskites have been studied in terms of crystal structure, thin film deposition and device performance by theoretical calculations and experimental studies.⁸ Methylammonium lead halides exhibit several advantages of light harvesting and transport properties.^9–13 Mixed-halide perovskites have shown significant improvement in tuning the bandgap by adjusting the stoichiometry.¹⁴ Formation dynamics of Methylammonium lead iodide (CH₃NH₃PbI₃) has been found to be processing specific.¹⁵ Perovskite films have been deposited either via thermal evaporation,^16,17 or spin coating techniques.¹⁸ The two common pathways often applied during film formation are one-step method in which the reactants are thoroughly mixed prior to deposition^18,19 and two-step method where the precursors are sequentially deposited.^1,9 Perovskite morphology and charge carrier density was controlled using the sequential deposition at low temperature.^12,13,20,21 Defect-states recombination in three-dimensional perovskites have been overcome using perovskite multi-quantum-wells.²² While conversion of polycrystalline perovskite is preferred for solar cells, growth of single crystals into multi-dimensional shapes has gained much attention for various applications.^18,23,24

In a preceding work, Rajab et al²⁵ used nanoplasmonic sensing (NPS) to detect CH₃NH₃PbI₃ perovskite formation at the interface of thick TiO₂ films with Au nanodisks, where complete conversion of perovskite formation was characterized by slow NPS red shifts while incomplete reactions were characterized by fast methylammonium iodide (CH₃NH₃I) crystallization. When complete reaction activation energies were reduced, perovskite formation was characterized by relatively fast NPS red shifts.

In this work, Rajab et al. demonstrate the capability to use the fast, minute and varying NPS changes to detect the formation kinetics of CH₃NH₃PbI₃ perovskite at the lower interface of thin mesoporous TiO₂ films. The time-resolved spectral shifts of the Localized Surface Plasmon Resonance (LSPR) are used to calculate the CH₃NH₃PbI₃ perovskite reaction rate constants at the different conditions. We assess the perovskite structures according to their formation kinetics and analytical results obtained by the characterization techniques.

Insplorion sensors with a dense TiO₂ coating were purchased from Insplorion AB (Gothenburg, Sweden). TiO₂ paste (The average particle size 20 nm) was acquired from Dyesol. Lead iodide (PbI₂) and methyl ammonium iodide (CH₃NH₃I: MAI) were purchased from Solaronix. Dimethylformamide (DMF), propanol and ethanol were purchased from Sigma Aldrich.

Standard films comprising of fused silica coated with Au nanodisks (100 nm diameter and 20 nm height) and 10 nm dense layer of compact TiO₂ as a dielectric spacer layer were used to analyze the nanoplasmonic sensing of Au nanosensors, similar to the early studied NPS experimental system arrangements.^26–29 These films were then spin-coated with mesoporous TiO₂ films prepared by mixing a commercial TiO₂ paste from dyesol and ethanol (2:7 wt %) at 5000 rpm for 30 s. The films are calcinated at 500 °C for 30 min. The thickness of mesoporous TiO₂ films was confirmed using VASE Ellipsometer VB-400. Then, 1M (462 mg/mL) of lead iodide (PbI₂) in dimethylformamide (DMF) was prepared under stirring at 70 °C. The mesoporous TiO₂ films were consequently infiltrated with PbI₂ by spin coating at 6000 rpm for 90 s and dried at 70 °C for 30 min. Different concentrations of MAI in 2-propanol ranging from 5 to 15 mg/mL at various temperatures from 25 to 53°C were prepared for in situ monitoring of perovskite formation.

The localized surface plasmon resonance (LSPR) phenomenon occurs when the Au nanoparticles are illuminated with near-visible light. The nanoparticles are sensitive to dielectric changes in their proximity via locally strongly enhanced electric fields. The LSPR is shown as a peak in an optical extinction spectrum, which comes from absorption and scattering of light by the plasmonic nanoparticles at specific wavelengths. Insplorion Xnano was used to monitor nanoplasmonic peaks during CH₃NH₃PbI₃ (MAPbI₃) perovskite formation as described elsewhere.²⁵ Initially, a blank measurement was taken for a fused silica substrate for subtraction. The coated Au sensor film was placed inside the system chamber and wavelength scan shows the nanoplasmonic peak position of a sensor is located at 800 nm-900 nm depending on the Au discs size and discs spacing. The Au nanodiscs, which act as optical antennas respond to events occurring at the interface between the sensor surface and sample material. A peak-fitting method proposed by Dahlin et al.³⁰ is applied to enable monitoring the spectral shifts on the order of 1 nm or less with a 0.01 nm resolution limit.

A fresh pure solvent was injected via the system pump to flush the tubes and to take a baseline measurement as settings are set at continuous, optimum flow rate of 0.1 µL/min to avoid bulk concentration gradient. At this point, minor spectral shifts were observed in both nanoplasmonic peak position and extinction due to change in the dielectric constant of the medium near the sensors. Upon diffusion of MAI into mesoporous TiO₂ film at the different concentrations and temperatures, the refractive index at Au nanodiscs changes with film color change (dark brown), known of perovskite reaction, as major spectral shifts of resonance wavelength and extinction of the previously located, sensor-specific peak, are simultaneously, in real time collected and monitored via Insplorer software. We have evaluated the injection time in the current study from the starting point of MAI injection until the nanoplasmonic peak has plateaued, corresponding to the transition stage in Figs. 1(a), 4(a). Typical time scale ranging from 18 s to 57 s has been calculated for the thinner 350 nm TiO₂ films, depending on MAI concentration used and operating temperature. Much prolonged injection time was observed in case of thicker 650 nm films with a range from 31 s up to 120 s.²⁵ In the standard sequential deposition method, 20 s dipping time has been applied using the 10 mg/mL MAI concentration under stagnant conditions.¹ The corresponding injection time in our study for the same MAI concentration (under continuous flow conditions) was 22 s.

For comparison, fresh solvent can be pumped again for rinsing the films. A set of 5 sensors were used for each experimental condition. The sensors were removed for characterization. X-ray diffraction (XRD), Bruker AXS D4 Endeavour X diffractometer using Cu Kα_1/2, λα₁=154.060 pm, λα₂=154.439 pm radiation, was used to obtain XRD spectra on different substrate surfaces as control samples. As indicated in, Fig. S1, the control samples show peaks of tetragonal MAI at 2θ = 19.74° and 29.79° along with a hexagonal PbI₂ at 2θ = 12.8°. Average XRD spectra on different perovskite films was taken to probe perovskite phase formation as well as content of reactants and products. As complementary characterizing technique, scanning electron microscopy (SEM) (JEOL-6300F, 5 kV), was used to obtain SEM images, which were analyzed using Image J software to generate histograms of average particle size distributions.

The effect of changing MAI concentrations from 5 to 15 mg/mL at a high temperature of 41 °C on NPS response using Au sensors/compact TiO₂/350 nm mesoporous TiO₂/PbI₂ was evaluated. Fig. 1(a and b) and Fig. S1(a) show the characteristic curves of nanoplasmonic peak position shifts and peak extinction shifts using 350 nm TiO₂ films. Fig. 1(a and b) show blue, fast shifts (higher rate constants) in the nanoplasmonic peak position for all concentrations except for the intermediate MAI concentration level at 10 mg/mL, which exhibits a red, slow shift (lower rate constant). The characteristic curves in Fig. S1(a) depict the opposite behavior in nanoplasmonic peak extinction for all concentrations.

The perovskite formed at MAI concentrations of 5 mg/mL show smaller crystalline structures, Fig. 2(a), due to the fast conversion of 85 s^-1, Fig. 1(b). The MAI concentration of 5 mg/mL shows the highest intensity of XRD diffraction peaks at 2θ = 14.25°, 28.57°, Fig. 3(c–d). Hence, perovskite conversion was achieved using low MAI concentration by lowering the reaction activation energy (operating at temperature of 41 °C). The perovskite formed at other MAI concentrations of 7.5 mg/mL and 12.5 mg/mL show similar results, Fig. S3. However, the MAI concentration of 7.5 mg/mL shows dense, small crystalline morphology with highest rate constant of 125 s^-1 where XRD diffraction peaks show highest residual unreacted PbI₂ at 2θ = 12.8°, Fig. 3(b).

The perovskite formed at the 10 mg/mL MAI concentration and 41 °C is largely in cubic and nanowire form, Fig. 2(c), with slow and maximum conversion seen in NPS rate constant of 3.4 s^-1, Fig. 1(b), and highest intensity of XRD diffraction peak at 2θ = 31.95°, Fig. 3(e), respectively. The perovskite formed at MAI concentration of 15 mg/ml show similar morphology (nanowires and sheets with higher particle size distribution, Fig. S3(e)) to the 10 mg/mL with a rate constant of 50 s^-1 where XRD diffraction peaks show highest residual of MAI at 2θ = 19.74°, Fig. 3(a). It is noticeable that the slower reactions that allow growth in multi-dimensions such as those at MAI concentrations of 10 and 15 mg/mL are reflected by red or minimum blue NPS shifts.

The effect of operating temperature (from 25 to 53 °C) at a fixed high MAI concentration of 12.5 mg/mL on NPS response for Au sensors/compact TiO₂/350 nm mesoporous TiO₂/PbI₂ was evaluated. Figure 4(a and b) and Fig. S1(b) show characteristic temperature curves on nanoplasmonic peak position shifts and peak extinction shifts using compact/350 nm mesoporous TiO₂ films. The characteristic curves in Fig. 4(a and b) show red (positive) shifts of low rate constants of 3.4- 12 s^-1 in the nanoplasmonic peak position for temperatures from 25-36 °C indicative of slow conversion. They also show blue (negative) shifts of higher rate constants of 34- 87 s^-1 for temperatures from 41-53 °C indicative of fast conversion.

The conversion of PbI₂ with a dielectric constant (ε∞=6) to MAPbI₃ perovskite with much higher (ε∞∼20),³¹ Fig. 5(a–c), leads to the red shifts observed in NPS extinction at temperatures of 25-36 °C. However, the partial change in morphology from smooth (precursor PbI₂ film) to rough, sparse large crystals of MAI and MAPbI₃ perovskites, Fig. 5(d–e), reduces the contact area at the Au sensor interface leading to the blue shifts observed at temperatures of 41-53 °C.

MAI injections at varying conditions to confined PbI₂ crystals can be monitored and recorded as red or blue shifts as seen by the formation of perovskites of various geometries. The change in dielectric constants of the materials upon injection of MAI into the Au sensor-coated TiO₂/PbI₂ films provides insight on the interface interactions and subsequently the ensemble peak pattern of the gold nanosensors can result in red or blue NPS shifts indicating perovskite conversion. While red shifts indicate slow reactions with multi-dimensional shapes, blue shifts showed faster reactions with lower conversions. The effect of changing the temperature on perovskite formation at high MAI concentration shows a non-linear relationship between the rate constant and temperature. By varying operation conditions, various perovskite structures can be synthesized and predicted utilizing their NPS responses.

See supplementary material for additional NPS characterization, XRD, and SEM analysis.

Authors would like to acknowledge the support of the Ministry of Higher Education, Kingdom of Saudi Arabia for this research through a grant (PCSED-003-16) under the Promising Centre for Sensors and Electronic Devices (PCSED) at Najran University, Kingdom of Saudi Arabia.