Previous theoretical calculations show that azetidinium has the right radial size to form a 3D perovskite with lead halides [G. Kieslich et al., Chem. Sci. 5, 4712 (2014)] and has been shown to impart, as the A-site cation of the ABX3 unit, beneficial properties to ferroelectric perovskites [B. Zhou et al., Angew. Chem., Int. Ed. 50, 11441 (2011)]. However, there has been very limited research into its use as the cation in lead halide perovskites to date. In this communication, we report the synthesis and characterization of azetidinium-based lead mixed halide perovskite colloidal nanocrystals. The mixed halide system is iodine and chlorine unlike other reported nanocrystals in the literature, where the halide systems are either iodine/bromine or bromine/chlorine. UV-visible absorbance data, complemented with photoluminescence spectroscopy, reveal an indirect-bandgap of about 2.018 eV for our nanocrystals. Structural characterization using transmission electron microscopy shows two distinct interatomic distances (2.98 Å ± 0.15 Å and 3.43 Å ± 0.16 Å) and non-orthogonal lattice angles (≈112°) intrinsic to the nanocrystals with a probable triclinic structure revealed by X-ray diffraction. The presence of chlorine and iodine within the nanocrystals is confirmed by energy dispersive X-ray spectroscopy. Finally, light-induced electron paramagnetic resonance spectroscopy with PCBM confirms the photoinduced charge transfer capabilities of the nanocrystals. The formation of such semiconducting lead mixed halide perovskites using azetidinium as the cation suggests a promising subclass of hybrid perovskites holding potential for optoelectronic applications such as in solar cells and photodetectors.

Organic/inorganic lead halide perovskites (LHPs), using earth-abundant elements, have shown tremendous potential in achieving lab-scale efficiencies of solar cells approaching that of the crystalline silicon solar cells.1 The current high-performing LHPs are composed of methylammonium (MA), formamidinium (FA), and caesium (Cs) as the A-site cations in ABX3, which form a 3D structure. This dimensionality, i.e., 3D, 2D, or 1D, estimated in terms of the Goldschmidt tolerance factor, is dependent on the size of the cation with all others remaining the same.

In this context, the extended tolerance factor approach developed by Kieslich et al. has shown azetidinium [Az; Fig. 1(a)] as a prospective A-site cation with a tolerance factor of 0.98 with lead iodide and radial size MA < Az < FA.2,3 Further inspiration to explore Az for optoelectronic perovskites comes from its utilization as the A-site cation in ABX3 type ferroelectric perovskites and the consequent very distinctive properties of the single crystals attributed to Az. For example, in these perovskites, large anomalies were observed in the temperature-dependent measurements of Raman modes and the alternating current (AC) relative permittivity attributed to the ring puckering motion of an oriented azetidine during phase transitions;4–6 the latter (low frequency) was shown to reach up to ∼106 in [(CH2)3NH2][Cu(HCOO)3].4 Recent work has shown the formation of a lead iodide perovskite with Az, albeit with low efficiency in a solar cell, and a mixed cation perovskite with MA, which showed negligible hysteresis compared to the MAPbI3 perovskite, but without any detrimental impact on the performance.7 Moreover, as thin films and single crystals, AzPbI3 was shown to result in much higher stability toward water compared to MAPbI3.7,8 These initial studies suggest that Az could be a very good candidate in the search for a stable perovskite through either replacement of or partial mixing with the existing cations.

FIG. 1.

(a) Molecular structure of azetidinium in 2D and 3D (drawn in MolView12). The 3D structure highlights the restricted rotational motion of the molecule unlike with MA and FA which possess 3D rotational degrees of freedom due to the unrestricted motion of the atoms. (b) Synthesis of AzPbI2Cl: addition of oleic acid (OA) and oleyl amine (OLA) to a mixture of PbI2 (yellow) and (CH2)3NH2Cl (white), and mixing results in the formation of an orange-colored compound. (c) Left: the resultant product (CS) in toluene and right: filtrate from the CS filtered with a 0.2 μm PTFE filter, both luminescing in a UV-box (365 nm).

FIG. 1.

(a) Molecular structure of azetidinium in 2D and 3D (drawn in MolView12). The 3D structure highlights the restricted rotational motion of the molecule unlike with MA and FA which possess 3D rotational degrees of freedom due to the unrestricted motion of the atoms. (b) Synthesis of AzPbI2Cl: addition of oleic acid (OA) and oleyl amine (OLA) to a mixture of PbI2 (yellow) and (CH2)3NH2Cl (white), and mixing results in the formation of an orange-colored compound. (c) Left: the resultant product (CS) in toluene and right: filtrate from the CS filtered with a 0.2 μm PTFE filter, both luminescing in a UV-box (365 nm).

Close modal

Given the ideal size of Az to form a 3D perovskite with lead halides, research into utilizing this molecule as the cation in LHPs has been, surprisingly, very limited. In this regard, we explore the utilization of Az in the synthesis of perovskite colloidal nanocrystals (NCs). NCs, owing to their facile synthesis, offer excellent handle over size control and, importantly, are intrinsically defect-tolerant.9 Additionally, the option of color-tunability achieved through mixed halides and the corresponding bright photoluminescence (PL) spanning the visible spectrum renders NCs as excellent low-cost materials for the fabrication of LEDs and photodetectors.10 Here, we report the synthesis of colloidal NCs using azetidine hydrochloride and lead iodide and characterize the optoelectronic and structural properties. We determine an indirect-bandgap and an atypical structure within the NCs with two distinctive interatomic distances and non-orthogonal lattice angles. Electronic transport is observed in the form of charge transfer to phenyl-C61-butyric acid methyl ester (PCBM) characterized using light-induced electron paramagnetic resonance (LEPR) spectroscopy.

All procedures are carried out in air under ambient conditions in a fume-hood. All chemicals were used as received without further purification. The standard hot-injection method for synthesis could not be carried out because AzCl did not dissolve in solvents used for PbI2 even at high temperatures (up to 180 °C) and with the addition of hydroiodic acid. Hence, the synthesis was carried out by mixing and thoroughly grinding in a mortar 1 mmol of lead iodide (PbI2; 99.99%; Sigma-Aldrich) and 1 mmol of azetidine hydrochloride [(CH2)3NH2Cl; 97%; Sigma-Aldrich]. Oleylamine (0.5 ml; 70%; Sigma-Aldrich) and oleic acid (1 ml; 99%; Sigma-Aldrich) were added as ligands to this mixture and mixed further thoroughly. The color starts changing within a minute from yellow into orange upon addition of the ligands, as shown in Fig. 1(b). Further mixing was carried out to ensure the complete conversion of PbI2. Toluene (anhydrous; Sigma-Aldrich) was added to this mixture, and the solution was transferred to a vial and stirred overnight. This solution, hereinafter referred to as the colloid solution (CS), was used for subsequent optoelectronic and structural characterization. Considering the exact stoichiometry of the compound remains unknown, we will refer to the same in the future as the “AzPbI2Cl” perovskite assuming that the structural unit is of the form ABX3, at least based on the initial molar ratios used.

Structural characterization was carried out using transmission electron microscope (TEM) imaging of the NCs. The filtrate from the CS was washed twice in methyl acetate (anhydrous; Sigma-Aldrich) and drop-casted onto a carbon TEM grid and allowed to dry under ambient conditions. NC imaging and energy dispersive X-ray spectroscopy (EDS) were performed using a JEOL ARM200CF microscope operated at 200 kV accelerating voltage equipped with a 0.98Sr JEOL EDS detector. For the imaging, a ∼5.5 pA STEM probe with ∼23 mrad convergence semi-angle was used and the annular dark field (ADF) detector collection angles spanned ∼50–242 mrad. Several images were recorded at a 4 µs pixel dwell time and 512 × 512 pixels to observe the structural change induced by the electron beam. From this image series, initial images with minimal structural changes were realigned and averaged [Fig. 2(a)] using the Smart-Align software.11 For EDS, it was not possible to obtain sufficient X-ray yield from a single NC to determine the elemental composition. Therefore, a wide field of view area featuring a NC ensemble or cluster (Fig. S1 of the supplementary material) was chosen and scanned with an increased probe current of ∼400 pA. By increasing the probe current, significant structural damages occurred to the NCs; however, the elemental composition of the cluster was verified to contain Pb, I, and Cl, as discussed below.

FIG. 2.

(a) Transmission electron microscope (TEM) image of the nanocrystals (NCs). (b) Zoomed-in image of the top NC in (a). The red and green boxes show two examples of an array of atoms whose center positions were determined by interpolating the data using cubic splines, as shown in plot (c); (c) the profile here corresponds to the red box in (b) while the peak maxima, taken as the atomic centers, are circled in blue. Maxima of the border peaks, which correspond to the edges of the NCs, were not considered in the estimation of spacings to exclude blurring and boundary effects. Two distinct interatomic distances were obtained from such analysis: 2.98 Å ± 0.15 Å (d-1) and 3.43 Å ± 0.16 Å (d-2). (d) TEM image of another NC. (e) FFT image of (d). FFT d-1 and FFT d-2 are line profiles taken to estimate the peak positions [black circles in (f)] corresponding to the d-spacing in the NCs. FFT d-1 is averaged over 2 pixels as the intensity is spread over 2 pixels, while FFT d-2 is only 1 pixel. (f) Bar plots of the line profiles FFT d-1 and FFT d-2. (g) Boxplot of the interatomic distances within the nanocrystals comparing the results obtained from real space analysis (d-1 and d-2), as shown in (b), and reciprocal space analysis (FFT d-1 and FFT d-2), as in (f). The estimated values between real space and reciprocal space are similar, i.e., d-1 and FFT d-1; d-2 and FFT d-2, and can be grouped into at least 2 distinct spacings. The diamond symbols are outliers of the values obtained in the estimation of the interatomic distances.

FIG. 2.

(a) Transmission electron microscope (TEM) image of the nanocrystals (NCs). (b) Zoomed-in image of the top NC in (a). The red and green boxes show two examples of an array of atoms whose center positions were determined by interpolating the data using cubic splines, as shown in plot (c); (c) the profile here corresponds to the red box in (b) while the peak maxima, taken as the atomic centers, are circled in blue. Maxima of the border peaks, which correspond to the edges of the NCs, were not considered in the estimation of spacings to exclude blurring and boundary effects. Two distinct interatomic distances were obtained from such analysis: 2.98 Å ± 0.15 Å (d-1) and 3.43 Å ± 0.16 Å (d-2). (d) TEM image of another NC. (e) FFT image of (d). FFT d-1 and FFT d-2 are line profiles taken to estimate the peak positions [black circles in (f)] corresponding to the d-spacing in the NCs. FFT d-1 is averaged over 2 pixels as the intensity is spread over 2 pixels, while FFT d-2 is only 1 pixel. (f) Bar plots of the line profiles FFT d-1 and FFT d-2. (g) Boxplot of the interatomic distances within the nanocrystals comparing the results obtained from real space analysis (d-1 and d-2), as shown in (b), and reciprocal space analysis (FFT d-1 and FFT d-2), as in (f). The estimated values between real space and reciprocal space are similar, i.e., d-1 and FFT d-1; d-2 and FFT d-2, and can be grouped into at least 2 distinct spacings. The diamond symbols are outliers of the values obtained in the estimation of the interatomic distances.

Close modal

The shape of the NCs (the boundaries are blurred by the surface-coated ligands that remained after washing) appears to be spherical but can also be partly discerned to be hexagonal in contrast to the clear cubic shape in the case of, e.g., MAPbI3 and CsPbBr3.9,13,14 This could be ascribed to the kinetics of the NC formation, which is dependent on the temperature of the reaction, as shown by Hassan et al.13 and Protesescu et al.,15 where controlled NC synthesis at high temperatures produced such cubic shaped structures. We expect the same in our case, where the limiting factors in the formation of a unit cell could possibly be not only the incorporation of Az into the lattice but also the much larger difference in the sizes of the halides, i.e., Cl–I compared to Cl–Br and Br–I; the kinetics is uncontrolled in our case due to the primitive synthesis method employed. Analysis of the size of the NCs reveals the average size to be ≈4.0 nm with σ ≈ 0.3 nm (approximate because of blurred boundaries).

The real and reciprocal images of the NCs were analyzed to derive the structural parameters, as shown in Fig. 2. In the real-space images, a row of atoms was selected, an example shown in Fig. 2(b), and the corresponding line profile (averaged over the box width) was plotted as a function of distance. These data were interpolated by fitting with a cubic spline function, shown in Fig. 2(c), to determine the locations of the peak maxima (blue circles in this figure), which correspond to the atomic centers; cubic splines gave the best fit compared to lower order splines and Gaussian functions. With the centers determined, the interatomic distances were calculated as the difference between the centers. Peaks close to and at the border of the rows were not considered in the estimation to remove uncertainties from blurring and boundary effects. This methodology was carried out for multiple rows in two different NCs in Fig. 2(a). The analysis (28 atomic centers in total from 7 rows) shows that the interatomic distances can be categorized into two distinct values (7 and 21 counts) given by mean and standard deviations as 2.98 Å ± 0.15 Å (d-1) and 3.43 Å ± 0.16 Å (d-2), respectively, as shown in the boxplot of Fig. 2(g).

This is further corroborated by the analysis of the Fast Fourier Transform (FFT) image of a third NC [Figs. 2(d) and 2(e)], which is less resolved compared to Fig. 2(a). Line profiles (FFT d-1, green, and FFT d-2, red, corresponding to 2 different peaks) in the FFT image are shown in the bar plot in Fig. 2(f). The peaks along the red line on either side of the center can be seen as a faint signal spread over only 1 pixel in contrast to the strong signal along the green profile which is averaged over 2 pixels along its width; histogram analysis is detailed in the supplementary material. Final values of the interatomic distances are shown in the box plot, Fig. 2(g), along with the real space analysis, and have a mean and standard deviation, 3.09 Å ± 0.20 Å (FFT d-1) and 3.47 Å ± 0.16 Å (FFT d-2), in close agreement with d-1 and d-2, respectively [outliers, shown as diamond markers in Fig. 2(g), have been excluded from the calculation of mean and standard deviations]. The box plot shows correlation between the real and reciprocal space analysis, albeit with a relatively wide spread in the estimated values of the interatomic distances and the occurrence of outliers due to the low counts in the images, but confirms the presence of at least 2 distinct interatomic distances within the NCs. Further optimization of the TEM imaging process so that the NCs can withstand the electron beam for longer duration would help achieve images of high signal-to-noise ratios for high quality structural characterization.

Angles inherent to the NCs were derived by measuring the angle between two rows of atoms, as shown with an example in Fig. 2(b). The angles are non-orthogonal and approximately 112° or 180° − 112° = 68°. The non-orthogonality of the angles can also be clearly seen in the FFT image [Fig. 2(e)], where the other radial peak is located at a non-orthogonal angle with respect to the first. Thus, from the structural characterization, we can see that the structure is dissimilar to that of the MA/FA/Cs LHPs, where cubic or orthorhombic structures are prevalent in the photoactive phases,16,17 and could be attributed to both the new cation, Az, and the much larger difference in the properties of the halides Cl and I.

X-ray diffraction (XRD) measurements were carried out to further characterize the structure. The CS was washed once in methylacetate in a centrifuge (8000 rpm for 6 min), and the precipitate was dried and collected for XRD measurements. Powder XRD data were recorded at room temperature on a Rigaku SmartLab diffractometer equipped with a Cu–Kα X-ray source with λ = 1.54 Å. Data processing, indexing, and subsequent refinement were carried out in the Rigaku PDXL software. For the refinement, including the background, the whole-powder-pattern-decomposition (WPPD) method was used. The data and the fits are plotted in Fig. 3. The structure obtained from the analysis is triclinic (space group P-1) with the lattice parameters a = 8.775 Å, b = 10.621 Å, c = 8.707 Å, α = 116.01°, β = 115.57°, and γ = 71.87°. The absence of the PbI2 peak at 12.8° (peak positions in the indexing information text file in the supplementary material) indicates that the synthesis has resulted in a well-structured perovskite without any PbI2 impurities in the crystals.18 The initial indexing search before refinement resulted in many different results, all with only one single phase; however, the above indexing gave one of the lowest fit metrics during refinement (from 5° to 90°) with Rwp = 2.29% and Rp = 1.70%. The reason for choosing this indexing is that the derived lattice angles α and β are in close agreement with the angle obtained from TEM analysis. Hence, we believe the derived triclinic structure to be the probable unit cell structure in our NCs. For a concrete and complete elucidation of the structure, including the type of octahedral connection, characterization of single crystals is essential, but, as stated in Sec. II, because it was not possible to dissolve AzCl in solvents for PbI2, single crystals could not be grown. However, structural analysis of azetidine lead halides from XRD appears to be a major challenge due to twinning and distortions reported in the case of AzPbI38 and speculated to be between 2D and 3D.7 To circumvent the solubility issues, alternate synthesis routes will be explored in the future.

FIG. 3.

XRD data of the synthesized perovskite colloids and the corresponding fits of the synthesized perovskite colloids. The peak positions corresponding to the diffraction planes are given in the supplementary material.

FIG. 3.

XRD data of the synthesized perovskite colloids and the corresponding fits of the synthesized perovskite colloids. The peak positions corresponding to the diffraction planes are given in the supplementary material.

Close modal

EDS carried out on an ensemble of the NCs during TEM imaging verified the presence of Cl along with Pb and I (in Fig. 4). Calculated elemental compositions of the ensemble from the spectra are shown in Table S1 in atomic percent along with the % error arising from X-ray counting statistics.19 The compositions were calculated using inbuilt k-factor routines of the Thermo NSS software. We are aware that such k-factor routines could have large systematic errors up to 20%.20,21 Additionally, we utilized a combination of K + L lines for the k-factor quantification in order to reduce the counting statistics error. Considering that the ionization mechanism for each line is different, combining the lines could further add to the uncertainty in the estimation of the atomic ratios. Thus, although we were able to confirm the incorporation of Cl in the NCs, the true composition, and hence the stoichiometry, remains inconclusive. To determine the accurate elemental composition, a combination of EDS, energy electron loss spectroscopy (EELS), and known elemental standard comparison would be required,19,20,22–24 which will be explored in the future.

FIG. 4.

EDS spectra of an ensemble of the nanocrystals (Fig. S1). The spectra confirm the incorporation of chlorine into the perovskite. Estimated atomic percentages with corresponding counting statistics errors are shown in Table S1.

FIG. 4.

EDS spectra of an ensemble of the nanocrystals (Fig. S1). The spectra confirm the incorporation of chlorine into the perovskite. Estimated atomic percentages with corresponding counting statistics errors are shown in Table S1.

Close modal

Figure 1(c) shows the CS luminescing as bright orange color in a UV-box (365 nm). For UV-visible absorbance measurements, the CS was filtered using a 0.2 μm PTFE filter and collected in a vial [Fig. 1(c)]; unlike the CS which was turbid, the filtrate was a clear solution with dispersed colloids or NCs. 500 μl of the filtrate was added to a cuvette, with a path length of 1 cm, and toluene was added until the volume was 3 ml. This was utilized to record steady-state photoluminescence (PL) and UV-visible absorbance spectra, as shown in Fig. 5. PL was measured with an automated spectrofluorometer (Fluorolog, Horiba Jobin-Yvon) equipped with a 450 W xenon lamp excitation source and a photomultiplier tube detector; the excitation wavelength was 450 nm (2.755 eV). The absorbance measurement was recorded using a commercial Varian Cary 60 spectrophotometer.

FIG. 5.

(a) PL and UV-visible absorbance of the nanocrystals in solution. The bandgap is 2.018 eV and is of indirect nature derived from the Tauc plot in Fig. S3(a). (b) A single Gaussian function fit (dashed line) to the PL data (black; left y-axis) and the first derivative of (αhν)1/2 with respect to hν (red; right y-axis) to obtain the PL emission peak and the indirect bandgap, respectively.

FIG. 5.

(a) PL and UV-visible absorbance of the nanocrystals in solution. The bandgap is 2.018 eV and is of indirect nature derived from the Tauc plot in Fig. S3(a). (b) A single Gaussian function fit (dashed line) to the PL data (black; left y-axis) and the first derivative of (αhν)1/2 with respect to hν (red; right y-axis) to obtain the PL emission peak and the indirect bandgap, respectively.

Close modal

PL data in Fig. 5(a) show counts per second (CPS)/μA vs energy in eV with an emission peak at 1.953 eV obtained by fitting a single Gaussian function to the data, as shown in Fig. 5(b). The full width at half maximum (FWHM) is about 200 meV, higher than that observed with Cs/FA/MA LHP NCs: 90–110 meV corresponding to emission peaks ranging from 1.65 eV to 3.10 eV.9,10,25 However, since the NCs are coated with ligands, further optimization of washing steps or the use of different ligands10 could possibly help achieve a narrower bandwidth. Comparing the PL of the diluted NCs [Fig. 5(a)] with that of the unfiltered solid colloids (Fig. S2), we see that the emission peak is at a similar energy and the dilution has not affected the optical properties of the NCs significantly in contrast to the observations reported in the work of Tong et al.26 However, the shape of the PL differs between the larger colloids and the NCs indicating structural effects on the optical properties. It is possible that further dilution could have an effect on the structure and, consequently, the PL emission.

Looking at the onset of the absorbance, at the first instance and then from the Tauc plot analysis [Fig. S3(a)], AzPbI2Cl can be classified as an indirect bandgap semiconductor. The bandgap is obtained by, first, calculating the first derivative of (αhν)1/2 with respect to hν, seen in Eq. S1 and plotted in Fig. S3(b), and then fitting a Gaussian function to the peak corresponding to the onset of the absorption, which comes out to be 2.018 eV; the Gaussian fits are shown in Fig. 5(b). This yields a Stokes shift of 65 meV.

MA and FA form LHP semiconductors of direct bandgap nature (primarily, expected to also have a weak indirect band27–29), and thus, based on the extended tolerance factor analysis by Kieslich et al.,2,3 similar results should have been expected using Az as the cation as the molecule’s radial size lies between MA and FA. However, our results have shown that this is not the case. The electronic band structure in LHPs is determined by the overlap of Pb and halide orbitals, which, in turn, is determined by the corresponding bond angles and bond distances in the unit cell and affected by the spin–orbit coupling at the band edges.30–38 Theoretical calculations have shown that the rotation of the cations MA and FA in the perovskite cage also affects the bond angles and distances in the unit cell, thereby affecting the band structure leading to a dynamic direct/indirect nature.27,28 Furthermore, it has been shown that the type of octahedron connectivity mode, i.e., corner, edge, or face-sharing resulting from a different dimensionality (1D, 2D, or 3D), can affect the dispersity of the band structure, thus determining the bandgap character, i.e., direct or indirect.31,32,39–43 In this regard, first-principle calculations have shown that the high-performing MAPbI3 in its 3D cubic phase with corner-sharing octahedral units has a positive enthalpy of formation and thus is thermodynamically metastable.44 However, its hexagonal polymorph, albeit 1D, with face-sharing octahedral units has a negative enthalpy of formation and an indirect bandgap of 2.6 eV. It is possible that the combination of Az orientation and the large difference in the halides, i.e., Cl and I, would have affected the connectivity modes leading to an indirect bandgap in our NCs. Thus, understanding the complete structure should provide insights into the origin of the indirect bandgap in our material. Unfortunately, because of the lack of complete structural information in our case, we could not determine the connectivity modes and so cannot definitively conclude on the origin of the indirect-bandgap in our NCs.

The PL quantum yield turned out to be less than 0.5% and remained similar even after about 12 months after synthesis, indicating a stable structure; however, the NCs aggregated with time. Such low quantum yield can be attributed not only to the indirect bandgap45 nature but also to non-radiative recombination due to factors such as surface defects—also causing the broadening of the PL peak in Fig. 5. As mentioned earlier, further optimization should improve the optical properties of the NCs.

A good electronic interaction of the NCs with other semiconducting or charge transport materials is of pivotal importance for achieving efficient organic and hybrid optoelectronic devices. An effective way to probe whether NCs can “communicate” efficiently with other molecules is the investigation of their electron transfer capabilities. To get a direct view of the electron transfer process, we applied light-induced EPR (LEPR) spectroscopy. LEPR spectroscopy is a well-suited technique to detect long-living paramagnetic species (such as radical anions and cations) generated after visible light absorption and electron transfer processes.46–50 

The LEPR spectra were recorded at 80 K and 280 K on a Bruker Elexsys E580 X-band spectrometer equipped with a nitrogen gas-flow cryostat for sample temperature control. The sample temperature was maintained with an Oxford Instruments CF9350 cryostat and controlled with an Oxford Instruments ITC503. LEPR spectra were acquired using white light illumination from a xenon lamp, IR filtered, and focused onto a quartz optical fiber ending in the optical window of the EPR spectrometer cavity. The system delivered about 50 mW cm−2 of light irradiance to the sample. The LEPR experimental parameters were a modulation amplitude of 1 G and a microwave power of 0.2 mW. LEPR measurements were performed on thick films of neat PCBM, as control, and the PCBM:NC blend. As for the neat PCBM (99%; Solenne B.V.), a 2 mg ml−1 solution of PCBM in toluene was prepared. As for the PCBM:NC blend, a PCBM solution in toluene (8 mg ml−1) was mixed with an equal volume of NC solution (from the filtered CS) in toluene. The final solutions were poured inside a quartz EPR tube (an inner diameter of 3 mm) and evaporated under vacuum, leaving a thick film on the inner tube wall. Subsequently, EPR tubes were sealed under vacuum. We chose PCBM since it is the most common n-type material used in perovskites and often used as an electron-acceptor in many optoelectronic applications.51 The LEPR measurements are performed at 80 K to slow down the recombination dynamics of the photogenerated charges and have a stronger EPR signal.

The LEPR spectra (light ON minus dark) recorded at 80 K in the neat PCBM film and the PCBM:NC blend are reported in Fig. 6. Before light illumination (dark), no EPR signal is detected in any sample, underlining that no paramagnetic species are present in the ground state. After visible light irradiation of the sample, a sharp EPR signal is detected in the PCBM:NC blend. The signal is characterized by a g-factor (g = 1.9998 ± 0.0005), a linewidth (ΔBpp ≈ 1 G), and an anisotropic line shape typical of the long-living radical anion localized on PCBM;52 to confirm that the electron transfer was to the PCBM and not to any other entity such as ligands or solvents, best-fit spectral simulation was carried out and described in Fig. S5 of the supplementary material. The generation of the PCBM radical anion after visible light illumination of the PCBM/perovskite NC blend was already revealed in literature in the case of MAPbI3, and it is usually attributed to the photoinduced electron transfer process.52,53 The same mechanism can take place in our PCBM:NC blend. After light absorption, an electron transfer between NCs (electron-donor) and the PCBM (electron-acceptor) occurs leaving a radical anion localized on PCBM and a hole localized on the NC. Bearing that S = 1/2, the hole should be detectable through EPR spectroscopy. Nevertheless, its EPR spectrum is not detected probably because of a severe line broadening due to fast magnetic relaxation times favored by the spin–orbit coupling interaction of the hole with the heavy nuclei of our NCs.53 Conversely to that observed in our PCBM:NC blend, the neat PCBM film does not show any photoinduced signal corroborating that the electron transfer does not occur between PCBM molecules but only between NCs and PCBM. The same EPR analysis was also performed at 280 K to rule out the effect of temperature and phase transitions on the electron transfer process. The spectrum reported in Fig. S4 shows the same results of the 80 K analysis. The lower EPR intensity at 280 K can be rationalized considering different factors. First, since LEPR spectroscopy is a steady-state technique, the signal intensity is given by a balance between the generation and recombination of the charge carriers. At higher temperatures, charge recombination is faster, and therefore, the LEPR signal is lower.48 Second, the EPR signal is proportional to the paramagnetic susceptibility which follows the Curie law and so is inversely proportional to the temperature.54 Thus, at higher temperatures, the intensity is lower. Finally, spin relaxation times also influence EPR intensity, and being faster at higher temperatures results in a decrease in the signal intensity.48 

FIG. 6.

LEPR spectra of PCBM (brown) and PCBM:NC blend (red) films acquired at 80 K. The spectra shown are the difference between the spectra acquired under visible light (light ON) and before (dark) sample illumination.

FIG. 6.

LEPR spectra of PCBM (brown) and PCBM:NC blend (red) films acquired at 80 K. The spectra shown are the difference between the spectra acquired under visible light (light ON) and before (dark) sample illumination.

Close modal

The LEPR analysis provided a direct confirmation that a strong electronic interaction, resulting in charge transfer and charge separation, occurs between NCs and PCBM suggesting the potential of our NCs in applications relevant for optoelectronics. From these observations and the high low frequency dielectric constants imparted by Az to the ferroelectric perovskites,4 we speculate the charge generation to be efficient. The high dielectric constants in the ferroelectric perovskites have been attributed to the strong polarization caused by a highly oriented Az4–6 in contrast to the dynamic orientational disorder of linear molecules MA and FA.27,28,55 Theoretical calculations on MAPbI3 show a low barrier for the rotation of the MA cation leading to dynamic orientational disorder and thus no long-range ordering of the dipoles for the ferroelectric effect under the electric field.30,55 While MA and FA are linear molecules possessing 3D degrees of rotational freedom,27,28,55 azetidine, however, is a four-membered ring with bond angles much more constricted compared to higher order rings [Fig. 1(a)]. The ring structure restricts torsional motion of the atoms, thereby reducing the rotational degrees of freedom. When constrained in a cage-like structure as in a perovskite and ionically bound to the halides, this is even more the case. Consequently, the abnormally high dielectric constants due to high polarization reported in Ref. 4 and attributed to the ring-puckering of Az can only occur if Az is highly oriented in the perovskite structure. Hence, presuming the case to be similar in our NCs, this constraint should induce a relatively highly oriented dipole compared to MA and FA. Moreover, the ability of the dipole to orient in a stable way could also be affected by the distribution of the different halides surrounding the cation in terms of a preferential electrostatic attraction and also hydrogen bonding between the H of the N of the cation and the halide, which, in this case, is preferentially Cl because of its high electronegativity compared to I.

The result of such a highly oriented dipole is a high polarization in the material in the presence of an electric field and the consequent high dielectric constants.56 Such values are desired in optoelectronic materials as high dielectric constants, such as in GaAs, lead to efficient charge generation as a result of a strong screening effect.

Our work has shown the suitability of azetidine in forming perovskite-type NCs with lead iodide and revealing properties with the potential for optoelectronic applications. Interestingly, the synthesized NCs result from the chlorine-adduct of azetidine rather than iodine or bromine. TEM analysis of the NCs showed two distinctive interatomic distances, 2.98 Å and 3.43 Å, and a lattice angle of ≈112°. XRD characterization revealed a probable triclinic unit cell with the lattice angles in agreement with the TEM analysis but needs further confirmation through characterization of a single crystal. LEPR measurements verified the electron transfer capabilities of the NCs through photoinduced charge transfer to PCBM. While EDS of the NCs confirmed the presence of chlorine, accurate composition could not be verified. Nevertheless, this points toward the formation of lead mixed halide perovskites with chlorine and iodine observed for the first time in NCs, and not seen before with Cs, MA, or FA cations, with azetidinium as the cation which seems to play a role in driving such formation, implying that not just the size (determining the tolerance factor) but also the geometry of the molecule, affecting the rotational degrees of freedom, could play a role in perovskite formation. The indirect bandgap of 2.018 eV displayed in AzPbI2Cl unlike the primarily direct bandgap nature of Cs/MA/FA lead halide perovskites may entail the need for thick films for applications in solar cells. However, mixed cation combinations with Cs/MA/FA can be explored7 to develop systems with excellent photovoltaic performance (owing to Cs/MA/FA) and high stability (negligible hysteresis and water resistant owing to azetidine).7,8 Along with the latter, systematic studies, including developing azetidine analogs with inorganic elements, open doors to understand and subsequently harness the high AC dielectric constants endowed by azetidine (attributed to its ring-puckering motion4–6 due to its geometry) to the ferroelectric perovskites in lead halide perovskites potentially leading to a paradigm-shift in the pursuit of stable, high efficiency perovskite optoelectronic technologies.

See the supplementary material for the TEM image of a NC cluster; histogram analysis of the FFT line profiles; elemental compositions calculated from the EDS spectra of the NCs; PL spectra of an unfiltered colloid solution; Tauc plot; the equation for the first derivative of the Tauc plot; LEPR spectra of the NCs measured at 280 K; and best-fit EPR spectral simulation. Indexing information—the text file contains the peak positions and the corresponding intensities obtained from the XRD data analysis.

S.V.K. conceived and directed the study with M.K.R. and H.J.S. Y.H., worked with S.V.K. on the synthesis of NCs, optical and XRD measurements, and prepared the TEM samples. A.P. measured and analyzed the EPR data, carried out the best-fit spectral simulations, and prepared the EPR part of the manuscript. A.V. carried out the TEM imaging and EDS and analyzed the EDS spectra. Analyses of the optical, TEM, and XRD data were carried out by S.V.K. The manuscript was mainly prepared by S.V.K. and all authors participated in the preparation and review.

S.V.K. would like to thank Dr. Amir-Abbas Haghighirad and Dr. Dharmalingam Prabharakaran for assistance with testing alternative synthesis routes and corresponding XRD. We thank the Centre for Advanced Electron Spin Resonance (CAESR), Department of Chemistry, University of Oxford for EPR measurements. The work at CAESR was supported by the EPSRC (Grant No. EP/L011972/1). A.P. would also like to thank Dr. William Myers, CAESR facility, for his kind assistance with EPR measurements. TEM/EDX work was performed on the South of England electron microscope funded through EPSRC (Grant No. EP/K040375/1). S.V.K. expresses gratitude to EPSRC (WAFT, Grant No. EP/M015173/1) and UKRI (START, Grant No. ST/R002754/1) and A.P. to the European Union’s Horizon 2020 research and innovation programme (SEPOMO, Marie Sklodowska Curie, Grant Agreement No. 722651) for funding. The research materials supporting this publication can be publicly accessed on the Oxford University Research Archive via the following persistent identifier: https://doi.org/10.5287/bodleian:5RBk00g9e. The research materials are available under a CC BY license.

The authors declare the following competing financial interest: HJS is a founding member of Oxford Photovoltaics Ltd. and Helio Display Materials Ltd., which seek to commercialize perovskite solar cells and LEDs.

1.
See https://www.nrel.gov/pv/ for NREL efficiency chart.
2.
G.
Kieslich
,
S.
Sun
, and
A. K.
Cheetham
,
Chem. Sci.
5
,
4712
(
2014
).
3.
G.
Kieslich
,
S.
Sun
, and
A. K.
Cheetham
,
Chem. Sci.
6
,
3430
(
2015
).
4.
B.
Zhou
,
Y.
Imai
,
A.
Kobayashi
,
Z.-M.
Wang
, and
H.
Kobayashi
,
Angew. Chem., Int. Ed.
50
,
11441
(
2011
).
5.
T.
Asaji
,
Y.
Ito
,
J.
Seliger
,
V.
Žagar
,
A.
Gradišek
, and
T.
Apih
,
J. Phys. Chem. A
116
,
12422
(
2012
).
6.
M.
Maczka
,
T. A.
da Silva
,
W.
Paraguassu
,
M.
Ptak
, and
K.
Hermanowicz
,
Inorg. Chem.
53
,
12650
(
2014
).
7.
S. R.
Pering
,
W.
Deng
,
J. R.
Troughton
,
P. S.
Kubiak
,
D.
Ghosh
,
R. G.
Niemann
,
F.
Brivio
,
F. E.
Jeffrey
,
A. B.
Walker
,
M. S.
Islam
 et al,
J. Mater. Chem. A
5
,
20658
(
2017
).
8.
R.
Panetta
,
G.
Righini
,
M.
Colapietro
,
L.
Barba
,
D.
Tedeschi
,
A.
Polimeni
,
A.
Ciccioli
, and
A.
Latini
,
J. Mater. Chem. A
6
,
10135
(
2018
).
9.
M. V.
Kovalenko
,
L.
Protesescu
, and
M. I.
Bodnarchuk
,
Science
358
,
745
(
2017
).
10.
H.
Huang
,
M. I.
Bodnarchuk
,
S. V.
Kershaw
,
M. V.
Kovalenko
, and
A. L.
Rogach
,
ACS Energy Lett.
2
,
2071
(
2017
).
11.
L.
Jones
,
H.
Yang
,
T. J.
Pennycook
,
M. S. J.
Marshall
,
S.
Van Aert
,
N. D.
Browning
,
M. R.
Castell
, and
P. D.
Nellist
,
Adv. Struct. Chem. Imaging
1
,
8
(
2015
).
12.
See molview.org for MolView.
13.
Y.
Hassan
,
O. J.
Ashton
,
J. H.
Park
,
G.
Li
,
N.
Sakai
,
B.
Wenger
,
A.-A.
Haghighirad
,
N. K.
Noel
,
M. H.
Song
,
B. R.
Lee
 et al,
J. Am. Chem. Soc.
141
,
1269
(
2019
).
14.
Z.-Y.
Zhu
,
Q.-Q.
Yang
,
L.-F.
Gao
,
L.
Zhang
,
A.-Y.
Shi
,
C.-L.
Sun
,
Q.
Wang
, and
H.-L.
Zhang
,
J. Phys. Chem. Lett.
8
,
1610
(
2017
).
15.
L.
Protesescu
,
S.
Yakunin
,
M. I.
Bodnarchuk
,
F.
Krieg
,
R.
Caputo
,
C. H.
Hendon
,
R. X.
Yang
,
A.
Walsh
, and
M. V.
Kovalenko
,
Nano Lett.
15
,
3692
(
2015
).
16.
K.
Chen
,
S.
Schünemann
,
S.
Song
, and
H.
Tüysüz
,
Chem. Soc. Rev.
47
,
7045
(
2018
).
17.
R. J.
Sutton
,
M. R.
Filip
,
A. A.
Haghighirad
,
N.
Sakai
,
B.
Wenger
,
F.
Giustino
, and
H. J.
Snaith
,
ACS Energy Lett.
3
,
1787
(
2018
).
18.
U.
Kwon
,
M. M.
Hasan
,
W.
Yin
,
D.
Kim
,
N. Y.
Ha
,
S.
Lee
,
T. K.
Ahn
, and
H. J.
Park
,
Sci. Rep.
6
,
35994
(
2016
).
19.
M.
Watanabe
and
D. B.
Williams
,
J. Microsc.
221
,
89
(
2006
).
20.
K. E.
MacArthur
,
T. J. A.
Slater
,
S. J.
Haigh
,
D.
Ozkaya
,
P. D.
Nellist
, and
S.
Lozano-Perez
,
Microsc. Microanal.
22
,
71
81
(
2016
).
21.
M.
Watanabe
,
Z.
Horita
, and
M.
Nemoto
,
Ultramicroscopy
65
,
187
(
1996
).
22.
A. J.
Craven
,
J.
Bobynko
,
B.
Sala
, and
I.
MacLaren
,
Ultramicroscopy
170
,
113
(
2016
).
23.
G.
Haberfehlner
,
A.
Orthacker
,
M.
Albu
,
J.
Li
, and
G.
Kothleitner
,
Nanoscale
6
,
14563
(
2014
).
24.
A.
Varambhia
,
L.
Jones
,
A.
London
,
D.
Ozkaya
,
P. D.
Nellist
, and
S.
Lozano-Perez
,
Micron
113
,
69
(
2018
).
25.
Y.
Hassan
,
Y.
Song
,
R. D.
Pensack
,
A. I.
Abdelrahman
,
Y.
Kobayashi
,
M. A.
Winnik
, and
G. D.
Scholes
,
Adv. Mater.
28
,
566
(
2015
).
26.
Y.
Tong
,
F.
Ehrat
,
W.
Vanderlinden
,
C.
Cardenas-Daw
,
J. K.
Stolarczyk
,
L.
Polavarapu
, and
A. S.
Urban
,
ACS Nano
10
,
10936
(
2016
).
27.
E. M.
Hutter
,
M. C.
Gélvez-Rueda
,
A.
Osherov
,
V.
Bulović
,
F. C.
Grozema
,
S. D.
Stranks
, and
T. J.
Savenije
,
Nat. Mater.
16
,
115
(
2016
).
28.
C.
Motta
,
F.
El-Mellouhi
,
S.
Kais
,
N.
Tabet
,
F.
Alharbi
, and
S.
Sanvito
,
Nat. Commun.
6
,
7026
(
2015
).
29.
M. A.
Green
,
A.
Ho-Baillie
, and
H. J.
Snaith
,
Nat. Photonics
8
,
506
(
2014
).
30.
F.
Brivio
,
A. B.
Walker
, and
A.
Walsh
,
APL Mater.
1
,
042111
(
2013
).
31.
L.
Mao
,
P.
Guo
,
M.
Kepenekian
,
I.
Hadar
,
C.
Katan
,
J.
Even
,
R. D.
Schaller
,
C. C.
Stoumpos
, and
M. G.
Kanatzidis
,
J. Am. Chem. Soc.
140
,
13078
(
2018
).
32.
C. C.
Stoumpos
,
L.
Frazer
,
D. J.
Clark
,
Y. S.
Kim
,
S. H.
Rhim
,
A. J.
Freeman
,
J. B.
Ketterson
,
J. I.
Jang
, and
M. G.
Kanatzidis
,
J. Am. Chem. Soc.
137
,
6804
(
2015
).
33.
P.
Azarhoosh
,
S.
McKechnie
,
J. M.
Frost
,
A.
Walsh
, and
M.
van Schilfgaarde
,
APL Mater.
4
,
091501
(
2016
).
34.
D.
Niesner
,
M.
Wilhelm
,
I.
Levchuk
,
A.
Osvet
,
S.
Shrestha
,
M.
Batentschuk
,
C.
Brabec
, and
T.
Fauster
,
Phys. Rev. Lett.
117
,
126401
(
2016
).
35.
S. D.
Stranks
and
P.
Plochocka
,
Nat. Mater.
17
,
381
(
2018
).
36.
J. L.
Knutson
,
J. D.
Martin
, and
D. B.
Mitzi
,
Inorg. Chem.
44
,
4699
(
2005
).
37.
L.
Mao
,
H.
Tsai
,
W.
Nie
,
L.
Ma
,
J.
Im
,
C. C.
Stoumpos
,
C. D.
Malliakas
,
F.
Hao
,
M. R.
Wasielewski
,
A. D.
Mohite
 et al,
Chem. Mater.
28
,
7781
(
2016
).
38.
E. R.
Dohner
,
A.
Jaffe
,
L. R.
Bradshaw
, and
H. I.
Karunadasa
,
J. Am. Chem. Soc.
136
,
13154
(
2014
).
39.
C. C.
Stoumpos
,
L.
Mao
,
C. D.
Malliakas
, and
M. G.
Kanatzidis
,
Inorg. Chem.
56
,
56
(
2017
).
40.
H.
Li
,
C. D.
Malliakas
,
F.
Han
,
D. Y.
Chung
, and
M. G.
Kanatzidis
,
Chem. Mater.
27
,
5417
(
2015
).
41.
J.
Luo
,
X.
Wang
,
S.
Li
,
J.
Liu
,
Y.
Guo
,
G.
Niu
,
L.
Yao
,
Y.
Fu
,
L.
Gao
,
Q.
Dong
 et al,
Nature
563
,
541
(
2018
).
42.
L.
Mao
,
Y.
Wu
,
C. C.
Stoumpos
,
M. R.
Wasielewski
, and
M. G.
Kanatzidis
,
J. Am. Chem. Soc.
139
,
5210
(
2017
).
43.
T. T.
Tran
,
J. R.
Panella
,
J. R.
Chamorro
,
J. R.
Morey
, and
T. M.
McQueen
,
Mater. Horiz.
4
,
688
(
2017
).
44.
A. S.
Thind
,
X.
Huang
,
J.
Sun
, and
R.
Mishra
,
Chem. Mater.
29
,
6003
(
2017
).
45.
N. B.
Mohamed
,
K.
Shinokita
,
X.
Wang
,
H. E.
Lim
,
D.
Tan
,
Y.
Miyauchi
, and
K.
Matsuda
,
Appl. Phys. Lett.
113
,
121112
(
2018
).
46.
S.
Stoll
and
A.
Schweiger
,
J. Magn. Reson.
178
,
42
(
2006
).
47.
J.
Niklas
,
K. L.
Mardis
,
B. P.
Banks
,
G. M.
Grooms
,
A.
Sperlich
,
V.
Dyakonov
,
S.
Beaupré
,
M.
Leclerc
,
T.
Xu
,
L.
Yu
 et al,
Phys. Chem. Chem. Phys.
15
,
9562
(
2013
).
48.
J.
Niklas
and
O. G.
Poluektov
,
Adv. Energy Mater.
7
,
1602226
(
2017
).
49.
C.
Krause
,
R.
Miranti
,
F.
Witt
,
J.
Neumann
,
D.
Fenske
,
J.
Parisi
, and
H.
Borchert
,
Sol. Energy Mater. Sol. Cells
124
,
241
(
2014
).
50.
V. I.
Krinichnyi
and
E. I.
Yudanova
,
AIP Adv.
1
,
022131
(
2011
).
51.
O.
Ostroverkhova
,
Chem. Rev.
116
,
13279
(
2016
).
52.
M.
Righetto
,
A.
Privitera
,
F.
Carraro
,
L.
Bolzonello
,
C.
Ferrante
,
L.
Franco
, and
R.
Bozio
,
Nanoscale
10
,
11913
(
2018
).
53.
A.
Privitera
,
M.
Righetto
,
R.
Bozio
, and
L.
Franco
,
MRS Adv.
2
,
2327
(
2017
).
54.
S. H.
Simon
,
The Oxford Solid State Basics
(
Oxford University Press
,
2013
).
55.
Z.
Fan
,
J.
Xiao
,
K.
Sun
,
L.
Chen
,
Y.
Hu
,
J.
Ouyang
,
K. P.
Ong
,
K.
Zeng
, and
J.
Wang
,
J. Phys. Chem. Lett.
6
,
1155
(
2015
).
56.
N. W.
Ashcroft
and
N. D.
Mermin
,
Solid State Physics
(
Holt-Saunders International Editions
,
1981
).

Supplementary Material