Effects of Hydroiodic Acid Concentration on the Properties of CsPbI3 Perovskite Solar Cells.

Inorganic cesium lead triiodide (CsPbI3) perovskite materials are becoming increasingly attractive for use in perovskite/silicon tandem solar cells, due to their almost ideal band gap energy (E g) of about 1.7 eV. To be useful as photovoltaic absorbers, the CsPbI3 must form the cubic or black phase (α-CsPbI3). To do so at relatively low temperatures, hydroiodic acid (HI) is required as a solution additive. This paper demonstrates CsPbI3 perovskite solar cells with an efficiency of 6.44%, formed using a HI concentration of 36 μL/mL. This value is higher than the previous most commonly used HI additive concentration. Herein, by undertaking a systematic study of the HI concentration, we demonstrate that the structural, morphological, optical, and electrical properties of CsPbI3 solar cells, processed with this HI additive concentration, are superior.

Introduction

Solar cells based on perovskite materials have recently received considerable attention in the photovoltaics community.[1−8] Due to the high efficiency and simple, low-cost fabrication technique, it is considered as the most exciting among all of the emerging photovoltaic technologies.[9−11] Several attractive features possessed by the perovskite materials include: high absorption coefficient,[12] long carrier diffusion length,[13] high electron mobility,[14] and tuneable band gap.[15] From the initial reported power conversion efficiency (PCE) of 3.8% since 2009,[16] the efficiency of perovskite-based solar cells have recently reached above 22%.[17] Perovskite materials can be expressed by the generic formula ABX3, where A typically stands for an organic cation (usually methyl ammonium [CH3NH3+] or formamidinium [CHNH3+]), B stands for divalent metal ion (typically Pb2+ or Sn2+), and X is a single or mixed halide ion (Cl–, Br–, or I–).[18−21] Despite showing several demonstrated advantages, perovskite solar cells are yet to reach the commercialization stage, due mainly to the problem with stability.[22] The hygroscopic organic cation is considered as the main reason behind the poor stability.[23] Therefore, researchers are currently trying to develop perovskites by replacing the organic cation by an inorganic counterpart.[24] Several reports have demonstrated different all-inorganic perovskite approaches employing CsPbBr3,[25] CsPbI2Br,[26] CsPbIBr2,[27] and CsPbI3.[28] To address the ever-increasing energy demand, researchers are also looking to go beyond the theoretical efficiency of single-junction solar cells by implementing multijunction configuration known as tandem solar cells.[29] Holman and co-workers have shown that the maximum efficiency of a tandem solar cell can reach up to 43% under 1 sun illumination considering a Si-wafer-based bottom cell band gap of 1.12 eV and a top cell band gap of 1.7 eV.[30,31]

Among the available inorganic perovskites, cesium lead triiodide (CsPbI3) has a band gap of 1.7 eV, which makes it a potential candidate to be used as a top cell absorber layer for a tandem structure.[32] CsPbI3 has been gaining attention as a novel photovoltaic absorber under the perovskite family.[33−36] In 2014, Kim et al. reported CsPbI3 absorber-based solar cells for the first time with an efficiency of 0.09%.[37] Afterward, Snaith et al. demonstrated CsPbI3-based solar cells with an efficiency of 2.9%.[28] Taima and co-workers employed sequential vapor deposition to obtain an efficiency of 5.71%.[38] CsPbI3 has two phases, i.e., (i) orthorhombic or yellow phase (δ-CsPbI3) and (ii) cubic or black phase (α-CsPbI3).[32] The black phase is the desired phase to be used in the solar cell absorber, however, CsPbI3 is only stable in the yellow phase at room temperature.[36,39] High temperature (>310 °C) is required to convert the yellow phase into black phase.[28,36] Snaith and co-workers used 33 μL/mL HI as an additive in the CsPbI3 perovskite precursor solution to enable low-temperature (≈100 °C) phase transition.[28] It is believed that the strained crystal lattice resulting from the addition of HI in the precursor is responsible for low-temperature phase transition. Different groups have reported CsPbI3 solar cells using this HI additive route to obtain low-temperature phase transition.[28,32,35,36] However, no detailed study has been reported on the effects of varying HI additive concentrations on the properties of the CsPbI3 perovskite layer as well as on the entire photovoltaic device.

In this work, we present a systematic investigation of the influence of the HI concentration (30, 33, 36, 39, and 42 μL) on the structural, optical, and morphological properties of the perovskite layer. Also, the perovskite layers formed with different HI additive concentrations were incorporated in the inverted device structure indium tin oxide (ITO)/poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS)/perovskite/[6,6]-phenyl C71 butyric acid methyl ester (PC71BM)/Ag. The average efficiency was found to be highest for devices containing the perovskite layers formed with the 36 μL HI additive.

Results and Discussion

X-ray diffraction (XRD) was carried out to characterize the structural properties of the CsPbI3 thin films with various HI concentrations. Figure shows the XRD patterns of the CsPbI3 thin films with 30, 33, 36, 39, and 42 μL/mL HI additive concentration. For all of the films, the presence of (100), (110), (200), (211), and (220) diffraction peaks indicate the formation of cubic or black phase CsPbI3, which is consistent with previous reports.[28,35] The lattice constant “a” for the cubic phase structure [hkl] was calculated by the Bragg’s law and Vegard’s law.[40−43]where d corresponds to the interplanar spacing value, λ is the X-ray wavelength (0.15406 nm), n is the order number, θ is the Bragg’s angle, and acubic is the lattice constant. The calculated d and acubic values are represented in Table S1.

Figure 1

X-ray diffraction (XRD) patterns of CsPbI3 layers with different HI additive concentrations in the precursor solution. The patterns are offset for clarity.

Changing the HI concentration does not change the position of the peaks. In terms of peak intensity, films formed with 36 μL HI have shown maximum intensity for the (100) and (200) peaks, which typically corresponds to the black phase of CsPbI3.[36] This observation clearly suggests that using 36 μL HI instead of 33 μL leads to increased crystallinity. From the XRD patterns, the peak splitting of the (110) peak and the small shoulder observed for the (200) peak are also consistent to the literature suggesting the presence of strain in the crystal.[28] Different research groups have already reported that strain is capable of inducing phase transitions in a crystal structure and can completely shift the phase diagram for a material.[44,45] Moreover, the addition of HI in the perovskite precursor is believed to enhance the solubility of PbI2 by the formation of an intermediate compound HPbI3.[46,47] Hence, it is assumed that the presence of strain due to the incorporation of HI additive is the main reason for allowing lower-temperature phase transition for attaining CsPbI3 in the black phase.

To study the effect of HI concentration on different structural parameters, the mean crystallite sizes (D), microstrain values (ε), and dislocation densities (δ) were calculated using the following equations, respectively.[48,49]where β is the full width at half-maximum of the diffraction peak located at 2θ and n is a factor which almost equals to unity for minimum dislocation density. The calculated values of the crystallite size, microstrain, and dislocation densities are represented in Table S1.

Figure depicts the normalized peak intensity and the crystallite size. From the normalized peak intensities, it is clearly visible that all of the films have shown strong orientation along the (100) and (200) planes. For all of the variations of HI, the two dominant peaks (100) and (200), corresponding to the black phase of CsPbI3, have shown maximum crystallite size for 36 μL HI additive. For 30 μL HI, the corresponding crystallite sizes for all of the peaks were much smaller. The trend in crystallite size shows the optimum value is 36 μL, moving to either higher or lower HI concentrations causes a reduction. Figure represents the microstrain and dislocation density on the films deposited with various HI concentrations. The presence of microstrain is visible from the peak splitting of the (110) peak. From Figure , it is evident that the microstrain is maximum on the (110) peak compared to all of the other peaks. However, for 36 μL HI concentration, both the microstrain and dislocation density are found to reduce drastically. Dislocation density indicates imperfection in a crystal. The overall dislocation density values corresponding to the dominant peaks of (100) and (200) are minimum for the HI concentration of 36 μL.

Figure 2

Normalized peak intensity and crystallite size of CsPbI3 perovskite layers with various HI concentrations. The normalized peak intensities are shown in the red bars and the crystallite sizes are shown in the green bars.

Figure 3

Calculated microstrain and dislocation density of CsPbI3 perovskite layers with various HI concentrations. Microstrain and dislocation density values are shown in the red and green colored bars, respectively.

Figure displays the top-view scanning electron microscopy (SEM) images of the CsPbI3 layer grown on top of the PEDOT:PSS layer for different additive concentrations. The difference in the morphologies observed from the SEM micrographs is due to the variation in the HI additive concentration, as all other deposition parameters and postdeposition annealing temperature were fixed. For the lowest additive concentration (HI = 30 μL/mL) from Figure a, poor grain growth and nonuniform coverage along with pores can be observed. This is attributed to a lack of HI, which hinders crystallization at low temperatures. From Figure b, uniform coverage along with improved grain growth can be observed. A few pinholes are visible from the image. The grain morphology is further improved for the CsPbI3 layer formed with 36 μL HI, which can be observed from Figure c with no visible pinholes or cracks. The grain sizes are very small, on the scale of a few nanometers. This finding correlates directly with the XRD measurements. Additionally, these results are consistent with previous reports showing smaller grain sizes for the CsPbI3 layer grown at lower temperature (≈100 °C) with the aid of HI additives.[28,35,36] Protesescu et al. demonstrated that the stability in the black phase can be enhanced by smaller nanocrystals.[50] Therefore, the smaller crystallites observed from the SEM and XRD data are beneficial for attaining the black phase. Figure d,e depicts the grain morphologies of the CsPbI3 layer deposited with 39 and 42 μL HI, respectively. The presence of numerous pinholes and cracks clearly suggests that the quality of these films is not suitable for an active layer in the photovoltaic devices. The increase in HI concentration causes an increase of H2O in the precursor solution state, which is unfavorable for perovskite formation[46] since the HI used contains water (57 wt % in H2O). Therefore, the addition of HI up to the optimized amount is beneficial for grain growth. However, if the HI concentration is further increased, the presence of water molecules is detrimental for the perovskite layer. After observing the SEM images of CsPbI3 thin-film surfaces with different HI concentrations in Figure , it can be concluded that the film deposited with 36 μL/mL HI concentration has shown better surface morphology compared to other HI concentration in the CsPbI3 precursor solution.

Figure 4

Top-view scanning electron microscopy (SEM) images of CsPbI3 layer with various HI concentrations: (a) 30 μL, (b) 33 μL, (c) 36 μL, (d) 39 μL, and (e) 42 μL in the precursor solution.

The Tauc plot of CsPbI3 films is shown in Figure . The band gap energy (Eg) of each film was calculated using the following equation[51,52]where α is the absorption coefficient, A is a constant, h is Planck’s constant, υ is the frequency of the incident photon, and n is a constant, which is equal to 1 for direct band gap semiconductors and 4 for indirect band gap semiconductors.[51] For the 30 μL HI CsPbI3 film, the calculated Eg is 1.75 eV. This value is higher than what is expected for the desired cubic phase configuration, which should be ∼1.7 eV.[28,32] Kim et al. showed that the yellow phase of CsPbI3 possesses an Eg of 2.81 eV.[36] The observation of Eg = 1.75 eV for the 30 μL film indicates that this HI concentration is too low for the full formation of the black phase. In the concentration range from 33 to 39 μL, the extracted Eg is ∼1.7 eV, representative of the fully formed CsPbI3 black phase. This finding correlates with the crystallite sizes displayed in Figure . Further increasing HI concentration to 42 μL sees the Eg increase to 1.73 eV. It is believed that the change in Eg for the higher HI concentration is due to the detrimental influence of water molecules incorporated in the HI. Deterioration in the perovskite film quality with excess HI addition has also been previously observed in the literature by Kim et al.[53] and Wang et al.[46]

Figure 5

Tauc plot for determining the bandgaps of the CsPbI3 layer deposited with different HI additive concentrations in the precursor solution. The band gap energy was obtained by extrapolating the straight-line portion of the graph to zero absorption coefficients. The intercept on the energy axis indicates the value of band gap energy.

Figure shows the absorption spectra for CsPbI3 thin films. The 30 μL HI CsPbI3 film has a shorter cut off wavelength and significantly reduced absorption in the range from 500 to 700 nm. Increasing the HI concentration leads to an increase in the cut off wavelength, and an improvement in absorption in the range from 500 to 700 nm. This absorption profile, indicative of the formation of the CsPbI3 black phase, is much more suitable for a photovoltaic absorber. Increasing the HI concentration to 42 μL leads to a reduction in the film absorption. This reduced absorption severely hinders the amount of photocurrent, which can be generated in this film.

Figure 6

Absorption spectra of CsPbI3 perovskite layer deposited with various HI additive concentrations in the precursor solution.

On the basis of the above discussions, 30 μL additives display lower crystallinity, poor morphology, reduced absorption, and a higher band gap. On the other hand, 42 μL HI additive has also shown unfavorable results in terms of structural, morphological, and optical properties. In the literature, HI concentration of 33 μL has been employed widely by different groups.[28,32,35,36] It appears that the optimized amount lies between the concentrations of 33–39 μL. To validate the findings from the characterizations of single layers, planar solar cells consisting of a ITO/PEDOT:PSS/CsPbI3-perovskite/PC71BM/Ag structure were fabricated. The PCE, open circuit voltage (Voc), short circuit current density (Jsc), and fill factor (FF) of corresponding devices are shown with the box plots in Figure . The average values, calculated based on 10 devices, are presented in Table . The maximum average efficiency is found for the solar cells consisting of the CsPbI3 layer deposited with 36 μL HI additive. Increasing the concentration from 30 to 39 μL leads to an increase in Jsc, which correlates with the absorption data shown in Figure . The low Jsc observed for the 30 μL case is related to the incomplete phase transition to the black phase. The HI concentration above the 36 μL leads to a significant reduction in Voc, which governs the optimum concentration. As shown in the SEM images in Figure , these higher concentrations lead to the formation of pinholes. The poor contact due to these pinholes may cause the abrupt change in Voc. The J–V curves of the best devices fabricated with various HI concentrations are shown in Figure .

Figure 7

Plot of the (a) PCE, (b) Voc, (c) Jsc, and (d) FF as a function of HI additive concentrations. The data are average values of ten solar cells under 1 sun illumination. Error boxes represent standard deviations.

Table 1

Average J–V Parameters of Open Circuit Voltage (Voc), Short Circuit Current (Jsc), Fill-Factor (FF), and Power Conversion Efficiency (PCE) of Inverted Structure ITO/PEDOT:PSS/CsPbI3/PC71BM/Ag Devices with Different HI Additive Concentrationsa

HI conc. (μL)	PCE (%)	Voc (mV)	Jsc (mA/cm2)	FF (%)
30	5.17 ± 0.32	846 ± 0.03	10.84 ± 0.41	56.33 ± 1.94
33	5.96 ± 0.26	866 ± 0.03	12.78 ± 0.15	53.91 ± 0.57
36	6.30 ± 0.15	875 ± 0.03	12.83 ± 0.20	56.21 ± 2.93
39	5.79 ± 0.15	806 ± 0.01	13.29 ± 0.15	54.05 ± 0.86
42	4.33 ± 0.69	770 ± 0.03	11.14 ± 1.25	50.25 ± 1.36

The data are average values of 10 solar cells under 1 sun illumination (100 mW/cm2).

Figure 8

Current density–voltage (J–V) curves under 1 sun illumination (100 mW/mL) for the best performing photovoltaic cells with different HI additive concentrations in the CsPbI3 precursor solution.

The obtained solar cell parameters have shown better values corresponding to the perovskite layer with 36 μL HI additive concentration, which is consistent with the previous findings in terms of XRD, SEM, and optical characterizations. Therefore, it can be concluded that 36 μL can be used as an optimized amount for HI concentration with the CsPbI3 perovskite layer.

Electrochemical impedance spectroscopy measurements were performed to study the charge-transport characteristics of all of the devices based on various HI concentrations in the active layer. Figure shows the Nyquist plots of the CsPbI3 solar cells with different HI concentrations with an applied bias of 0.6 V. The obtained Nyquist data exhibit distinct semicircles, which were fitted with the equivalent circuit model shown in the inset of Figure . All of the measurements were performed under dark conditions where the internal resistance of the photovoltaic device consists of sheet resistance and charge-transfer resistance. In the equivalent circuit, RSH represents the sheet resistance, which basically consists of the ohmic resistance of the electrodes and the bulk resistance of the active layer.[36]RCT corresponds to the charge transfer or interfacial resistance, which is typically obtained from the diameter of the semicircle of the Nyquist plot.[54] The constant phase element (CPE) represents the nonideal capacitor element of the photovoltaic devices.[55,56] The extracted equivalent circuit parameters are given in Table . From Figure the devices with 36 μL HI have shown the smallest semicircle indicating the lowest charge-transport resistance. The lowest value of RCT (from Table ) corresponding to the devices with 36 μL HI can be attributed to the better crystallinity, morphology,and higher range of absorption confirmed from XRD, SEM, and UV–vis spectrometry, respectively. The value of RCT for the 30 μL HI case is significantly higher. This high interfacial resistance is attributed to the presence of the yellow phase at this additive concentration. Additionally, the 42 μL HI also displays much higher RCT compared to the optimum concentration. This correlates well to the SEM results in Figure , which show the presence of pinholes in this film. This poor contact may be responsible for the significantly reduced Voc for devices with this HI concentration. The RSH values are very close to each other indicating the fact that all of the devices were prepared under the same conditions. The overall values obtained for RSH, RCT, and capacitance are consistent to the literature.[36,57]

Figure 9

Nyquist plot for CsPbI3-based devices with an applied bias of 0.6 V under dark condition for different HI concentrations in the CsPbI3 precursor solution. Inset: the equivalent circuit used to fit the data.

Table 2

Fitted Values of Different Electronic Parameters of Sheet Resistance (RSH), Charge-Transfer Resistance (RCT), and the Capacitance CPE from the Nyquist Plot of CsPbI3-Based Solar Cells with Different Additive Concentrations

HI conc. (μL)	RSH (Ω)	RCT (kΩ)	CPE (F)
30	200	39.1	1.27 × 10–9
33	173	13.9	8.14 × 10–10
36	225	12.8	5.14 × 10–10
39	200	19.1	1.35 × 10–9
42	182	37.7	1.44 × 10–9

Conclusions

In summary, we investigated the influence of the HI additive concentration on both the formation of CsPbI3 films and the performance of CsPbI3 perovskite photovoltaic devices. The use of a HI additive is crucial to facilitate the phase transition from the undesirable orthorhombic or yellow phase to the cubic or black phase, at temperatures amenable to low-cost production. At low HI concentrations of 30 μL/mL, absorption and XRD measurements indicated the incomplete phase transition to the black CsPbI3 phase. Correspondingly, devices fabricated with these films displayed poor photovoltaic performance. Increasing the HI concentration caused an improvement in the crystallinity of the cubic phase as well improved optical and electrical properties. The optimum HI concentration was found to be 36 μL/mL, which is higher than the commonly used concentration of 33 μL/mL. CsPbI3 perovskite photovoltaic devices fabricated using this condition achieved a maximum efficiency of 6.45%. CsPbI3 perovskite solar cells are an excellent candidate for use as the top cell in perovskite/silicon tandem solar cells. These results provide a further step toward the realization of this technology.

Experimental Section

Device Fabrication

Patterned ITO glass, purchased from Lumtec was used as the substrates. The substrates were ultrasonically cleaned with Hellmanex III soap, deionized water, acetone, and isopropanol sequence. Afterward, they were dried with high-pure N2 gas. The PEDOT:PSS (Al4083, Ossila Ltd.) solution was sonicated for 10 min and then filtered with a 0.45 μm poly(tetrafluoroethylene) filter. Afterward, the PEDOT:PSS layer was formed on the cleaned substrates via spin-coating at 4500 rpm for 30 s followed by annealing at 130 °C for 15 min. The substrates were then transferred inside a N2-filled glovebox (<0.1 ppm of O2 and H2O). Five separate 0.48 M CsPbI3 precursor solutions were made by adding cesium iodide (CsI, Sigma-Aldrich, >99.999%) and lead iodide (PbI2, Lumtec, >99.999%) in 1 mL N,N-dimethylformamide (Sigma-Aldrich, anhydrous) solvent. Prior to deposition, various concentrations (30, 33, 36, 39, and 42 μL) of hydroiodic acid (HI, 57 wt % in H2O, Sigma-Aldrich, >99.99%) were added to the previously prepared CsPbI3 precursor solutions. The perovskite layer was then formed by spin-coating at 3500 rpm for 30 s and subsequent annealing at 100 °C for 10 min inside the glovebox. For preparing the electron-transport layer, [6,6]-phenyl C71 butyric acid methyl ester (PC71BM, 1-Material, Inc., 20 mg/mL) was dissolved in cholorobenzene (Sigma-Aldrich, anhydrous). The PC71BM solution was spin-coated at 3000 rpm for 30 s on top of the perovskite layer to form the electron-transport layer. Finally, thermal evaporation was used to deposit a 100 nm Ag layer with an evaporation rate of 2.0 Å/s under a vacuum condition of 1 × 10–6 mBar.

Characterization

Current density–voltage (J–V) measurements were performed using a solar cell I–V testing system from PV Measurements, Inc. with a Keithley 2400 source meter. The light intensity was calibrated to 100 mW/cm2 intensity with an AM 1.5G solar simulator. The structural and crystallographic properties of the deposited films were studied with PANalytical Empyrean thin-film XRD machine with Cu Kα radiation. The surface morphology of the films was measured by scanning electron microscopy (FEI Nova NanoSEM450) equipment. For optical characterization (transmittance, reflectance, and absorbance) of the films, PerkinElmer Lambda 950 UV–vis–NIR spectrometer was used. The impedance analysis was conducted with an Autolab PGSTAT-30 equipped with a frequency analyzer module in the frequency range from 1 MHz to 1 Hz.