Characterization of NiO x nanoparticles
X-ray diffraction was used to study the phase composition of as synthesized nanoparticles with different precursor concentrations (NP-1, NP-2, NP-3, and NP-4) and the patterns recorded are shown in Fig. 2(a). According to XRD patterns, all the NiOx nanoparticles demonstrated five major diffraction peaks at 2θ values of 37.2°, 43.2°, 62.8°, 75.4°, and 79.4° corresponding to the (111), (200), (220), (311), and (222) crystallographic planes of cubic NiO [JCPDS, No. 04-0835], respectively. No other peaks were found on the diffractogram.
The average crystallite size (D) of NiOx nanoparticles was estimated using Debye Scherrer formula along the (200) diffraction plane50 using Eq. (6):
$$\:D=\:\frac{0.9\:\lambda\:}{\beta\:\text{cos}\theta\:}$$
6
where λ is the X-ray wavelength, β is the full width half maximum and θ is the diffraction angle of the concerned peak. The nanoparticles synthesized at concentrations of 0.2 M, 0.5 M and 0.8 M (NP-1, NP-2, and NP-3) show similar crystallite size of 6 nm whereas nanoparticles grown at higher concentration of 1.2 M (NP-4) depict increased crystallite size of 9 nm. This phenomenon can be easily understood by the fact that with the increase in the concentration of solution number of precursor ions per unit of time and volume increases that leads to increase in growth rate yielding larger particles. The similar behaviour was reported by some other groups for different material systems synthesized by chemical precipitation method51,52.
Figure 2(b) shows the FTIR spectra of NiOx nanoparticles (NP-1, NP-2, NP-3, and NP-4) synthesized at different precursor concentration. For all the samples, a wide band at around 3400–3600 cm− 1 has been detected which could be assigned to O-H stretching vibrations of free H2O or OH group53. The peak at around 1630 cm− 1 probably arises from the O-H bending mode of H2O that is adsorbed on the material surface. Sharp peak at ca 420 cm− 1 is characteristic of Ni-O vibration54,55. A strong absorption band in the FTIR spectra of all the samples appears in the region of 1200–1515 cm− 1 with the peak positioned at around 1350 cm− 1, whereat this band is the strongest in the sample NP-4 prepared from the most concentrated solution. Strong IR bands in this spectral region were also recorded for Ni(OH)2 phase synthesized from NiNO353. Namely, vibrations at ~ 1390 cm− 1 and ~ 1490 cm− 1 were assigned to O-H bend and lattice OH in Ni(OH)2, respectively, and vibrations at around 1280 cm− 1, 1310 cm− 1 and 1340 cm− 1 were assigned NO3− ions53. IR bands positioned at ~ 1350 cm− 1 and ~ 830 cm− 1 were recorded in spectra of layered nickel hydroxynitrate phases and assigned to the nitrate stretch and bend modes, respectively55.
Thus, according to the IR spectra, the synthesized NiOx nanoparticles still contain an intermediate phase of synthesis, Ni(OH)2, as well as the traces of the Ni(NO3)2 precursor, and their concentration is higher in nanoparticles synthesized from more concentrated solutions although only NiO phase was detected by XRD.
UV–Visible absorption spectra of semiconductor materials have been regarded as a significant tool to analyse fundamental information related to its optical properties. The band gap (Eg) of different nanomaterials has been calculated using absorption spectra and following Tauc plots as shown in Eq. (7):
$$\:{\left(\text{A}\text{h}{\nu\:}\right)}^{\text{n}}=\text{B}(\text{h}{\nu\:}-\text{E}\text{g})$$
7
where hυ is the energy of photon; A is absorbance, B is the constant related to the material; and n indicates either 2 or 1/2 for direct transition and indirect transition, respectively56. Therefore, the optical band gap of the material can be obtained by drawing tangent to the linear portion of the (Ahυ)n – hυ curve and calculating intercept on hυ axis.
Figure 2(c) depicts the estimation of band gap of different nanoparticles synthesized by varying concentration using Tauc plots. Band gap values for NP-1, NP-2 and NP-3 are around 3.70 eV, but NP-4 shows a little smaller band gap of 3.65 eV. Band gap values of NiOx nanoparticles are close to that measured for the ultrasonically sprayed NiOx thin films57,58 The smaller band gap of NP-4 could be explained by slightly larger crystallite size of 9 nm compared to 6 nm for NP-1 – NP-3. In addition, the higher amount of Ni(OH)2 in NP-4, as shown by FTIR study, may influence the NiOx band gap as the Eg of Ni(OH)2 is 3–3.5 eV59.
Characterization of Sb 2 S 3 solar cell
After examining the physical and structural properties of nanoparticles, it is equally significant to study the properties of NiOx HTM deposited on Sb2S3 absorber layer before introducing gold contact. The structural properties of the glass/FTO/TiO2/Sb2S3/NiOx structure were studied by XRD, corresponding XRD patterns are shown in Fig. 3(a). The XRD pattern shows the peaks corresponding to Sb2S3, FTO and anatase TiO2 layers. The peaks at 2θ values of 15.9°, 17.8°, 29.2° and 30.5° observed in diffractogram, are characteristic of the orthorhombic stibnite structure of Sb2S3 (ICDD PDF 01-075-4013)60.There are no peaks observed related to NiOx suggesting formation of an extremely thin coating over glass/FTO/TiO2/Sb2S3 stack. The transmission spectra of glass/FTO/TiO2/Sb2S3 with and without NiOx in the stacked structure were recorded by UV–VIS spectrometry as shown on Fig. 3(b). It can be observed from the figure that the transmission spectrum of glass/FTO/TiO2/Sb2S3/NiOx is similar to that of glass/FTO/TiO2/Sb2S3 owing to transparent nature of thin NiOx film. The average visible transmittance (AVT %) was calculated and found to be around 22 % The transmission spectra clearly demonstrate semi-transparent behaviour of Sb2S3 solar cells in the visible range of the spectrum.
Figure 4(a), (b) and (c) illustrate surface morphology of NS-1, NS-2, NS-4 devices, respectively and demonstrate the formation of NiOx porous layer over Sb2S3 absorber layer. The morphological studies reveal uniform coverage of NiOx nanoparticles over Sb2S3 absorber layer in all samples. The cross-sectional image of NS-2 device (Fig. 4(d)) shows the thicknesses of TiO2 and Sb2S3 layers of approximately 40 nm and 90 nm, respectively, while NiOx layer is a noticeably thinner, ca 20 nm.
Device performance
The current density–voltage (J–V) characteristics of all the champion devices with varied NiOx concentrations in HTM layers measured under one sun illumination (AM 1.5G) are depicted in Fig. 5(a), while solar cell output parameters along with shunt and series resistance are summarized in Table 2. It could be seen from Fig. 5(a) that the NiOx precursor concentration of 0.5 M (NS-2) results in the most efficient solar cell compared to solar cells with other NiOx concentrations. The external quantum efficiency (EQE) curve and the integrated current density curve of the best performing cell, NS-2, is presented in Fig. 5(b). To understand the performance of these devices and obtain information about reproducibility, statistical box plots of Voc, Jsc, FF and efficiency are shown in Fig. 6. Figure 6(a) shows the effect of NiOx HTM precursor concentration on the Voc of Sb2S3 solar cells. The figure clearly shows that with increase in the precursor concentration, the Voc value spirals up from ca 360 mV, characteristic of the cell without an HTM, to ca 560 mV, and stabilizes around this value. Thus, application of NiOx HTM layer significantly reduces the recombination of charge carriers at the back contact by transferring holes and blocking electrons at the interface, and hence resulting in higher value of Voc.
The impact of NiOx type on the device current density (Jsc), is shown in Fig. 6(b). Current density first increases from ca 10 mA·cm− 2 (NS-0) to ca 12 mA·cm− 2 (NS-2, NiOx precursor solution 0.5 M) and decreases thereupon to ca 7 mA·cm− 2 (NS-4). The external quantum efficiency (EQE) curve and the integrated current density curve of the best performing cell, NS-2, are presented in Fig. 5(b). The integrated photocurrent Jsc of 12.35 mA·cm− 2, as measured for the NS-2 device, corresponds well to that obtained from the J-V curve, 12.4 mA·cm− 2 (Table 2). It should be noted that EQE curves do not exhibit a dip in spectral region of 500–650 nm, characteristic of solar cells that use P3HT as an HTM49,61. Consequently, NiOx HTM exhibits good perspective to be used in semi-transparent devices.
Fill factor (FF) values depending on the NiOx precursor concentration are presented in Fig. 6(c) and show downtrend, from 40% (NS-0, without HTM) to 30% (NS-4). A possible explanation for the downtrend trend in the FF value in Fig. 6(c) may be related to a large series resistance, changing the boundary condition for the predominant recombination path in the device, under illumination condition. A higher NiOx precursor concentration induces a high series resistance at the back interface (Table 2) and this effect is further enhanced under illumination conditions where changes from low to high injection generation occurs. This effect generates enhanced recombination at the absorber/NiOx interface due to a light modulated potential barrier. An enhanced back surface recombination mechanism predominant around the maximum power point results in a low FF. In Fig. 5(a), the shapes of the light J–V characteristics of solar cells with high NiOx precursor concentration indicate the presence of kink anomaly phenomenon. This effect seems to be more pronounced in the fourth quadrant, implying that a large series resistance (Table 2) together with a low doping in the Sb2S3 absorber promotes accumulation of charge carrier at the Sb2S3/NiOx back interface. The later effect changes the boundary conditions for the charge recombination under light conditions – from interface and bulk recombination to intense recombination at the Sb2S3/NiOx back interface, resulting in the degradation of all PV parameters and final PCE of the solar cells.
So far, the highest efficiency of 2.65% is recorded for NS-2 device with NiOx from 0.5 M solution, which is 60% higher than the Sb2S3 solar cells without HTM while efficiencies of devices using NiOx from more concentrated solutions demonstrate lower performance (Fig. 6(d) and Table 2) mainly due to drop in Jsc and FF. Thus, more concentrated solutions are not recommended for the production of NiOx nanoparticles, as these may contain higher quantity of residues from the precursor and intermediates, as confirmed by FTIR study.
Table 2
Photovoltaic parameters of NiOx HTM based Sb2S3 champion solar cells with variation in NiOx HTM precursor concentration during synthesis.
Solar cell | Concentration of precursor (M) | Voc (mV) | Jsc (mA·cm− 2) | FF (%) | Efficiency (%) | Series resistance (Ωcm2) | Shunt resistance (Ωcm2) |
NS-0 | 0 | 356 | 11.2 | 42 | 1.66 | 5 | 162 |
NS-1 | 0.2 | 500 | 6.5 | 30 | 0.96 | 37 | 96 |
NS-2 | 0.5 | 555 | 12.4 | 39 | 2.65 | 12 | 175 |
NS-3 | 0.8 | 563 | 9.9 | 37 | 2.08 | 26 | 169 |
NS-4 | 1.2 | 545 | 8.7 | 32 | 1.51 | 19 | 103 |
Results obtained in this study show that an inorganic HTM composed of NiOx nanoparticles synthesized by a chemical precipitation method is an effective material to decrease the recombination at the solar cell back interface with no reducing the optical transparency of the device. We also showed that the nanoparticles preparation parameters viz, precursor solution concentration has strong impact on the properties of nanoparticles and solar cell output performance and should be carefully optimized for highly efficient devices.