Light illumination intensity dependence of photovoltaic parameter in polymer solar cells with ammonium heptamolybdate as hole extraction layer

Light illumination intensity dependence of photovoltaic parameter in polymer solar cells with ammonium heptamolybdate as hole extraction layer

Accepted Manuscript Light Illumination Intensity Dependence of Photovoltaic Parameter in Polymer Solar Cells with Ammonium Heptamolybdate as Hole Extr...

920KB Sizes 0 Downloads 15 Views

Accepted Manuscript Light Illumination Intensity Dependence of Photovoltaic Parameter in Polymer Solar Cells with Ammonium Heptamolybdate as Hole Extraction Layer Zhiyong Liu, Shengli Niu, Ning Wang PII: DOI: Reference:

S0021-9797(17)31024-X http://dx.doi.org/10.1016/j.jcis.2017.09.010 YJCIS 22761

To appear in:

Journal of Colloid and Interface Science

Received Date: Revised Date: Accepted Date:

6 July 2017 30 August 2017 1 September 2017

Please cite this article as: Z. Liu, S. Niu, N. Wang, Light Illumination Intensity Dependence of Photovoltaic Parameter in Polymer Solar Cells with Ammonium Heptamolybdate as Hole Extraction Layer, Journal of Colloid and Interface Science (2017), doi: http://dx.doi.org/10.1016/j.jcis.2017.09.010

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Light Illumination Intensity Dependence of Photovoltaic Parameter in Polymer Solar Cells with Ammonium Heptamolybdate as Hole Extraction Layer Zhiyong Liub, Shengli Niua* and Ning Wangc* a

Key Laboratory of Zoonosis of Liaoning Province, School of Animal Science and Veterinary

Medicine, Shenyang Agricultural University, Shenyang 110866, People’s Republic of China b

College of Science, Shenyang Agricultural University, Shenyang 110866, People’s Republic of

China c

Key Laboratory of Physics and Technology for Advanced Batteries (Ministry of Education),

College of Physics, Jilin University, Changchun 130012, China *

Corresponding authors: ([email protected] (S.L Niu); [email protected] (N. Wang))

1

Abstract: A low-temperature, solution-processed molybdenum oxide (MoOX) layer and a facile method for polymer solar cells (PSCs) is developed. The PSCs based on a MoOX layer as the hole extraction layer (HEL) is a significant advance for achieving higher photovoltaic performance, especially under weaker light illumination intensity. Fourier transform infrared spectroscopy (FTIR) and X-ray photoelectron spectroscopy (XPS) measurements show that the (NH4)6Mo7O24 molecule decomposes and forms the molybdenum oxide (MoOX) molecule when undergoing thermal annealing treatment. In this study, PSCs with the MoOX layer as the HEL exhibited better photovoltaic performance, especially under weak light illumination intensity (from 100 to 10 mW cm−2) compared to poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS)-based PSCs. Analysis of the current density-voltage (J-V) characteristics at various light intensities provides information on the different recombination mechanisms in the PSCs with a MoOX and PEDOT:PSS layer as the HEL. That the slopes of the open-circuit voltage (VOC) versus light illumination intensity plots are close to 1 unity (kT/q) reveals that bimolecular recombination is the dominant and weaker monomolecular recombination mechanism in open-circuit conditions. That the slopes of the short-circuit current density (JSC) versus light illumination intensity plots are close to 1 reveals that the effective charge carrier transport and collection mechanism of the MoOX/indium tin oxide (ITO) anode is the weaker bimolecular recombination in short-circuit conditions. Our results indicate that MoOX is an alternative candidate for high-performance PSCs, especially under weak light illumination intensity. 2

Keywords: hole extraction layer; Ohmic contact; bimolecular recombination.

3

1. Introduction Polymer solar cells (PSCs) have attracted increasing attention as a renewable energy source in recent years [1-3]. Significant efforts are now focused on further enhancing their power conversion efficiency (PCE), which has recently exceeded 12% for ternary PSCs [4, 5]. Regular PSCs have a sandwich structure with corrosive and hygroscopic

poly(3,4-ethylenedioxylenethiophene):poly(styrenesulfonic

acid)

(PEDOT:PSS) as the hole extraction layer (HEL), Ca as the electron extraction layer (EEL), and a low-work-function metal such as Al as the cathode [6, 7]. Ca is active and can be easily oxidized by oxygen and water [8, 9], which is detrimental to device lifetime and limits its applications. Owing to their low cost and stable nature, metal oxides (V2O5 [10], MoO3 [11], and WO3 [12]) have been used as the HEL to replace PEDOT:PSS. Because of its non-acidic nature and deeper energy level of highest occupied molecular orbital (EHOMO), MoO3 is one of the most promising materials for fabricating PSCs. However, the required vacuum deposition method has limitations in large-scale production [11, 13]. Thus, the design and synthesis of solution-processed MoO3 as an anode buffer layer in PSC devices has attracted the interest of many research groups [14-16]. Yang has reported that the (NH4)6Mo7O24-4H2O molecule decomposes and forms the molybdenum oxide (MoOX) molecule when subjected to lower-temperature annealing treatment (optimized thermal annealing temperature and times equal to or greater than 100 °C and 10 min, respectively) [17]. In the work, we fabricated a MoOX layer by spin-coating (NH4)6Mo7O24-4H2O solution on the top of indium tin 4

oxide (ITO) glass, followed by thermal annealing treatment. A blend of thieno[3,4-b]thiophene/benzodithiophene (PTB7) and [6,6]-phenyl C71-butyric acid methyl ester (PC71BM) acts as a photoactive layer. In order to avoid the oxidizing reaction of Ca by oxygen and water in the air, the conjugated polyelectrolyte poly[9,9-bis(3′-(N,N-dimethylamino)-propyl-2,7-fluorene)-alt-2,7-(9,9-dioctylfluoren e)] (PFN) layer is fabricated as an EEL by spin-coating PFN solution on top of the photoactive layers to replace Ca, and the composite layer of PFN/Al acts as a cathode. The PSC structure is ITO/MoOX/PTB7:PC71BM/PFN/Al. Over the course of an entire day, the light illumination intensity during the time the PSC was illuminated by the most light was lower than 100 mW cm−2. Thus, photovoltaic performance as a function of light illumination intensity is important [18, 19]. At the same time, the open-circuit voltage (VOC) and short-circuit current density (JSC) as a function of light illumination intensity was measured to gain deeper insight into the charge carrier recombination kinetics among the photoactive layer, ITO modified layer and ITO anode. The photovoltaic parameters [VOC, JSC, fill factor (FF), and PCE)] as functions of light illumination intensity were measured under illumination using an AM 1.5G solar simulator, and the light illumination intensity changed between 10 and 100 mW cm−2. The control PSC device with a PEDOT:PSS layer as the HEL had the same configuration. We also studied the recombination dynamics of PSCs by analyzing the change of photovoltaic performance parameter dependence on the light illumination intensity. The PSC configuration and energy-band diagram are shown in Figures 1(a) and 1(b); the energy levels of the highest occupied molecular orbitals (EHOMO) and 5

those of the lowest unoccupied molecular orbitals (ELUMO) of all material were obtained from the literature [20, 21].

2. Experiment PTB7 was purchased from One Material, Inc.; PC71BM from Nano.C; and PEDOT:PSS, (NH4)6Mo7O24-4H2O, o-dichlor-obenzene (ODCB, anhydrous, 99%), isopropanol solvent, and 1, 8-Diiodooctane (DIO) from Sigma-Aldrich Co. PFN was purchased from Luminescence Technology Corp. and Al from Alfa Aesar Co. (NH4)6Mo7O24-4H2O 0.4 g was added to the blend of isopropanol solvent and deionized water (10 mL, isopropanol: water ratio of 4:1), and used to prepare the solution of (NH4)6Mo7O24 because of the better contact ability of isopropanol solvent with a glass substrate. PTB7:PC71BM (1:1.5, wt.%) blends were dissolved in ODCB solution overnight with a PTB7 concentration of 10 mg ml−1. Approximately 3% (1,8-diiodooctane (DIO)/1,2-dichlorobenzene(DCB), v/v) of DIO added to the PTB7:PC71BM blend is helpful in obtaining better photovoltaic results. The PFN solution was dissolved in a methanol solution (0.2 mg ml−1 of methanol) in the presence of a small amount of acetic acid (2 l ml−1). ITO-coated glass (R=12–14 Ω sq−1) was cut into 3×3-cm2 chips and used as substrates. Each ITO substrate was ultrasonicated, followed by washing with acetone, isopropyl alcohol, and deionized water for 20 min. Subsequently, the ITO glass substrates were dried under a stream of nitrogen and heated on a hot stage. Finally, an UV-ozone treatment of the ITO glasses was performed for 15 min [22]. The (NH4)6Mo7O24 and PEDOT:PSS solutions were 6

spin-coated onto an ITO electrode, which was then thermally treated at 130 °C for 15 min. The thickness of the MoOX layer and PEDOT:PSS layer is 15 and 20 nm, respectively. The blended solution of PTB7:PC71BM was then spin-coated on top of the modified ITO layer and thermally annealed at 130 °C in a glove box. The PTB7:PC71BM photoactive layer has a nominal thickness of ~100 nm (with a variation of ~10 nm over the entire film). Next, the PFN solutions were spin-coated on top of PTB7:PC71BM layer in a thickness of 10 nm. Finally, the Al cathode was vacuum-evaporated through a shadow mask to define the active area of the devices (3 ×3 mm2) under a pressure of approximately 4×10−5 Pa. The control devices were fabricated with a PEDOT:PSS layer acting as the HEL using the spin-coating method with the same device structure. The current-voltage curves were obtained using a standard source measurement unit (Keithley 2400). The J-V characteristics of the PSCs were performed in a glove box under illumination at 100 mW cm−2 using an AM 1.5G solar simulator. The average photovoltaic parameter values were obtained from five devices fabricated in parallel.

3. Results and Discussion In this work, the decomposition of the (NH4)6Mo7O24 molecule is studied using FTIR and XPS. In Figure 2(a), the FTIR spectrum of the (NH 4)6Mo7O24 layer with and without thermal annealing treatment is shown. The peaks at 3405 and 1633 cm−1 can be assigned to the H–O–H bending vibration of crystalline water; the broad peak at 3215 cm−1 and the narrow peak at 1395 cm−1 can be related to the N–H stretching 7

vibration and flexural vibration, respectively, which are associated with NH 4+ groups. The three characteristic peaks at 909, 658, and 577 cm−1 imply the presence of the Keggin structure of the (NH4)6Mo7O24 molecule and the Mo–O–ion bond [23, 24]. However, after the thermal annealing treatment of the (NH 4)6Mo7O24 layer, the strength of the absorption peaks centered at 3215 and 1395 cm−1 decrease obviously, while those at 3405 and 1633 cm−1 completely disappear. This implies that the crystalline water has been wiped off, and that many NH4+ groups have decomposed, while only a small quantity of NH4+ groups have. Moreover, the Mo–O ion bond and Keggin structure show little change. The comparison between the FTIR spectra of the (NH4)6Mo7O24 layer with and without thermal annealing might imply that the vast majority of NH4+ groups were decomposed and that the crystalline water was completely displaced by the thermal annealing treatment, and, furthermore, that the Keggin structure of the (NH4)6Mo7O24 molecules and Mo–O ion bond is stable [23, 24]. Figure 2(b) shows the XPS curve of the MoOX layer. The Mo 3d peak includes two doublets peaks (233.8 and 236.8 eV corresponding to the Mo 3d3/2 and Mo 3d5/2 bands, respectively) in the form of a Gaussian function. The two major contributing peaks at approximately 233.8 and 236.8 eV are the typical XPS peaks of the Mo 6+ ion. The minor ones centered at approximately 417.5 and 400.2 eV have been identified as the Mo 3p1/2 and Mo 3p3/2 doublet peaks of the Mo 5+ ion [25, 26]. The MoOX layer exhibits a mixture of a majority of Mo 6+ ions and a small amount of Mo 5+ ions. Figure 3(a) and Table 1 show the current density-voltage (J-V) and photovoltaic performance parameters [VOC, JSC, FF, PCE, series resistance (RS), and shunt 8

resistance (RSH)] of MoOX-based and PEDOT:PSS-based PSCs. The external quantum efficiency (EQE) curve of the corresponding PSCs are shown in Figure 3(b). When a MoOX layer replaces the PEDOT:PSS layer as the HEL, the photovoltaic performance slightly increased from a PCE of 8.91% (the other parameters VOC, JSC, and FF are 0.76 V, 17.9 mA cm−2 and 65.1% respectively) to a PCE of 9.12% (with VOC is 0.76 V, JSC is 18.3 mA cm−2 and FF is 65.6%), which is attributed to the enhancement of JSC and FF and to the fact that VOC did not change. In addition, the MoOX-based PSC showed lower RS and higher RSH values compared to the PEDOT:PSS-based PSC. The difference between the JSC value measured under simulated AM 1.5G light as a standard source and that calculated using the EQE curve is less than 5%. The EQE results agree with the higher JSC values of the PSCs mentioned above. In previous work [17, 20], it was reported that the energy level of the MoOX layer is slightly shallower than the EHOMO value of the PEDOT:PSS layer, which generated a lower potential barrier between EHOMO of the PTB7 layer and the MoOX/ITO anode [17]. In addition, the metallic nature of the MoOX layer and the Ohmic contact between both the MoOX layer and ITO glass favors charge carrier transport and collection. This result is demonstrated by the higher JSC and FF values of MoOX-based PSCs. However, the potential barrier difference between EHOMO of PTB7 and PEDOT:PSS and the non-Ohmic contact of both PEDOT:PSS and ITO glass does not favor charge carrier transport and collection [27, 28]. This leads to lower FF and JSC values for the PEDOT:PSS-based PSCs. Figure 4 shows the dark current of MoOX-based and PEDOT:PSS-based PSCs with varying magnitudes of 9

leakage current. The leakage current in PSCs can be considered as undesirable current that is injected from the electrodes prior to voltage turn-on [29]. In the operating regime (from 0 V to VOC), the leakage current flows in the opposite direction of the photocurrent, and therefore reduces JSC and the FF. However, a higher leakage current is exhibited in the reverse bias voltage range (applied voltage lower than 0 V) of a MoOX-based PSC, which is attributed to the failure of the deeper ECB and EVB values of the MoOX layer (the ECB and EVB values are 5.1 and 8.4 eV, respectively) to block electron transport between the photoactive layer and ITO anode (the ELUMO values of PTB7 and PC71BM are 3.3 and 4.0 eV, respectively). Owing to “quantum effects” of the semiconductor, many electrons will hop from the ELUMO of the photoactive layer to the conduction band of transition-metal oxides (TMOs) and recombine with holes at the HEL/anode interface [30-32]. However, the MoOX-based PSC is shown to have the higher current density in the forward-bias voltage range, which indicates the better hole transport and collection abilities of the MoOX/ITO composite anode. At the same time, for the devices with PEDOT:PSS and MoOX as the HEL the current ratios of forward-bias voltage at 1 V and reverse bias voltage at −1 V are 6.13103 and 6.95103, respectively. Higher current ratios for the MoOX-based PSCs might be the result of the strong charge carrier transport and collection ability compared with PEDOT:PSS-based PSCs [33]. We further investigated the light intensity dependence of the photovoltaic parameters and analyzed the effects of charge recombination in PSCs containing a MoOX layer and a PEDOT:PSS layer as the HEL. The current density-voltage (J-V) characteristics at various illumination 10

intensities of PSCs with MoOX as the HEL and PEDOT:PSS as the HEL are displayed in Figures 5(a) and 5(b), respectively. In the operating regime, the photovoltaic parameters of both PSCs showed a stronger dependence on light intensity. However, the better photovoltaic performance at lower incident light intensity was obtained for the MoOX-based PSCs compared to the PEDOT:PSS-based PSCs. The device photovoltaic parameters as functions of the incident light intensity are shown in Figure 6 (Figures 6a–6d show the VOC, FF, JSC, and PCE values, respectively). The photovoltaic parameter dependences on illumination light intensity of the two PSCs exhibit the same change trend but different change ratios. Figure 6(a) shows that the VOC value varies logarithmically [ln(I)] with light intensity. The VOC value clearly decreased at lower incident light intensities when a PEDOT:PSS layer was used as the HEL; the highest VOC value of 0.76 V and lowest VOC value of 0.65 V were obtain at 100 mW cm−2 and 10 mW cm−2 respectively. On the other hand, the PSCs with a MoOX layer as the HEL showed the same VOC value of 0.76 V at 100 mW cm−2, while the slightly decreasing ratio and higher VOC value of 0.67 V was obtained at 10 mW cm−2 compared with PEDOT:PSS-based PSCs. The MoOX-based PSCs exhibited the more stable VOC dependence on light illumination intensity, and the VOC value difference is 0.09 V at both the highest and lowest light illumination intensities, while the VOC value difference of PEDOT:PSS-based PSCs is 0.11 V at the same light illumination intensity. The VOC value, which is determined by the energy difference between the lowest unoccupied molecular orbital (LUMO) of the acceptor and the highest occupied molecular orbital (HOMO) of the donor, can be found using 11

the formula [34, 35] VOC 

Egap e



kT (1  PD ) NC2 ln( ), q PDG

where k is Boltzmann’s constant, q is the elementary charge, T is the temperature (in K), Egap is the energy difference between the HOMO of the donor and the LUMO of the acceptor, PD is the dissociation probability of the electron-hole pairs, δ is the Langevin recombination constant, NC the effective density of states, and G the generation rate of bound electron-hole pairs. The maximum VOC value is governed by the difference between the LUMO of the acceptor and the HOMO of the donor. At open-circuit conditions, where most of all of the photogenerated charge carriers will be recombined in the photoactive layers, the recombination mechanism can be evaluated using the formula [36]

VOC  n

kT ln( I ). . q

Monomolecular recombination is dominant at open-circuit conditions when the slope has strong higher than one unity (kT/q). In the case of bimolecular recombination, the slope value (n) is close to one unity (kT/q). The slope for the device with a MoOX layer and a PEDOT:PSS layer as the HEL is 1.03kT/q and 1.32kT/q, respectively. According to the Shockley-Read-Hall (SRH) mechanism, electrons and holes prefer to recombine through trap states or recombination centers, resulting from interfacial defects and/or impurities in materials, which is called monomolecular recombination [37]. When the slope value (n) is close to the value of kT/q, the bimolecular recombination is dominant among the photoactive layer, ITO modified layer and ITO 12

anode. When the slope value is greater than a unit kT/q, monomolecular recombination is dominant at open-circuit conditions. When the additional mechanism of SRH recombination is involved, however, the SRH mechanism competes with bimolecular recombination, and a stronger dependence of VOC on light intensity with a slope greater than kT/q is observed [38]. The slope value of MoOX-based PSCs reaches 1.03kT/q and is close to unity, indicating that the VOC losses are likely to be dominated by bimolecular recombination. The stronger dependence of VOC on light intensity and the higher slope value of PEDOT:PSS-based PSCs implies that recombination at open-circuit conditions is a combination of monomolecular and bimolecular recombination. While the role of PEDOT:PSS in suppressing SRH recombination is not clear, we believe that PEDOT:PSS might reduce the density of interfacial traps between the photoactive layer and HEL [38]. In addition, the weaker potential barrier between the photoactive layer and MoOX, compared to that of the PEDOT:PSS layer, increases the built-in voltage (Vbi) across the devices. Since the internal voltage (Vint) is the difference between Vbi and the applied voltage in operating PSCs, a larger Vbi leads to a larger Vint at any applied voltage. An increase in Vint also might facilitate the charge carriers’ escape from shallow traps and reduce monomolecular recombination [38]. However, other factors also play important roles in determining the VOC value, such as the potential barrier, charge carrier recombination, built-in field, etc. [44,45]. The lower potential barrier between the EHOMO of the PTB7 and of the MoOX/ITO electrode compared to the PEDOT:PSS/ITO electrode decreases the probability of 13

charge carrier recombination at trap sites and enhances the charge carrier collection efficiency. Other researchers reported that the reduction of the potential barrier between the EHOMO of the donor and the ITO buffer layer caused the reduced degree of charge carrier loss during the charge carrier transport and extraction processes [46,47]. Monomolecular recombination can occur between trapped charges and oppositely charged carriers at the interface between the ITO anode buffer layer and the photoactive layer during the carrier collection processes [48,49]. At the same time, the built-in field between the ITO and Al electrodes has tremendous effects on the VOC value [50]. At lower incident light intensities, the built-in field within the PSCs is reduced, hindering the hole transport across the potential barrier between the PTB7 layer and MoOX/ITO anode [51,52], as confirmed by the obvious decrease of the VOC value in PEDOT:PSS-based PSCs at low light intensities. As shown in Figure 6(b), both devices exhibit a similar FF change trend as a function of light intensity (from 10 to 100 mW cm−2). The MoOX-based and PEDOT:PSS-based PSCs showed the highest FF values, 78.2% and 77.9%, respectively, at the lowest light illumination intensity. Along with the increase of incident light intensity, the two PSCs showed a decrease in FF, achieving the lowest values, 65.8% and 64.9%, respectively, under 100 mW cm−2 incident light intensity, which is attributed to the higher bimolecular recombination of the photoactive layer [39, 40]. Bimolecular recombination is a significant loss mechanism in most semiconductor conjugated polymers [41, 42]. As a result, the decrease in FF at higher light intensities revealed that the FF values were heavily influenced by bimolecular 14

recombination at higher light intensities [43, 44]. However, a similar decrease in the ratio of the two PSCs has been shown, which indicates the weak effect of charge carrier loss and the same charge carrier recombination mechanism for the two interfaces of the HEL/photoactive layer at the maximum power point condition. At strong incident light intensities of the PSCs, the bimolecular recombination is increasingly prominent and the FF values decrease for the two PSCs. JSC values are closely linked to the charge carrier collection and transport between the photoactive layer and the modified electrode layer. The analysis of JSC as a function of the incident light intensity provides deeper insight into the charge recombination kinetics of PSC devices [35, 45, 46]. The power-law dependence of JSC on incident light intensity can be determined from J SC  I 

where I is the incident light intensity and  is obtained by fitting the data. As shown in Figure 6(c), JSC appears to scale linearly with light intensity for the two PSCs. Figure 7 shows the optical transmittance spectra and visible optical transmittance ratios of the MoOX and PEDOT:PSS layers. The MoOX layer exhibits a slight enhancement in optical transmittance ability, with an optical transmittance ratio of 91.1% compared to that of the PEDOT:PSS layer (90.2%). A power-law dependence of ~1 was found (=0.937 and 0.914 for MoOX-based and PEDOT:PSS-based PSCs, respectively), indicating a recombination-limited photocurrent [47, 48]. For the two PSCs, similar JSC values are exhibited at a lower light illumination intensity, specifically 2.3 and 2.2 mA cm−2 for MoOX-based and PEDOT:PSS-based PSCs, 15

respectively. However, an obvious enhancement of the difference in the JSC values of the two PSCs was obtained along with an increase of the light illumination intensity. The slightly higher visible light transmissivity of the MoO X layer compared to the PEDOT:PSS layer cannot explain this result. At the same time, the different change trends of the JSC value for both PSCs is again illustrated by the different interface contacts engineered among the photoactive layer, HEL and ITO anode. The metallic nature of the MoOX layer leads to the Ohmic contact between the MoOX layer and ITO anode. However, the semiconductor conjugated nature of PEDOT:PSS forms the non-Ohmic contact between the PEDOT:PSS layer and ITO anode, which is not conducive to charge carrier transport and collection [27, 28]. The nearly linear dependence of JSC is consistent with sweep-out at short-circuit conditions, but also indicates that bimolecular recombination is relatively weak. The deviation from =1 is typically attributed to bimolecular recombination [38]. It is apparent that MoOX-based PSCs exhibit a linear relationship in the photocurrent density dependence on light intensity on a double-logarithmic scale with  values of 0.937. Therefore, bimolecular recombination in the MoOX-based PSCs should be rather low at short-circuit conditions, indicating that the photogenerated charge carriers can be effectively collected by the individual electrodes [37]. For the PEDOT:PSS-based PSCs, the value is significant deviation of 1 is typically attributed to the strong bimolecular recombination at the short circuit condition. An ideal contact should satisfy two essential requirements: (1) a zero injection potential barrier, and (2) an infinite surface recombination velocity. A large injection barrier or finite surface 16

recombination velocity will block and accumulate charges at the electrodes and thus modify the built-in electrostatic field, which significantly degrades the electrical performance of PSCs [49]. In Figure 6(d), the PCE of MoOX-based and PEDOT:PSS-based PSCs are plotted as functions of the light intensity. The best PCE values of the two PSCs were obtained under the weakest incident light intensity; specifically, 12.2% and 11.3% for MoOX-based and PEDOT:PSS-based PSCs, respectively. The PCE value decreased to 9.2% at

100 mW cm−2

for MoOX-based PSCs, whereas the PCE of

PEDOT:PSS-based PSCs exhibited a lower PCE under the same light illumination intensity with an efficiency of 8.8%. As shown in Figure 6(d), the PCE dependence was largely determined by the trend in VOC, with FF and JSC playing slight roles. For the two PSCs, the PCE increase ratio between 10 and 100 mW cm−2 for MoOX-based and PEDOT:PSS-based PSCs was 32.4% and 28.1%, respectively. VOC played a primary role in the entire light illumination process, as shown in Figure 6(a). Notably, for the MoOX-based PSCs, the VOC was slightly affected by the light intensity compared to PEDOT:PSS-based PSCs. This is attributed to the slight potential barrier between the EHOMO value of PTB7 and the MoOX/ITO anode and to the fact that the hole can cross the potential barrier even considering the weaker built-in field under lower light illumination intensity, which favors the result of slightly higher JSC and FF values compared to PEDOT:PSS-based PSCs as a function of light illumination intensity [50, 51]. The higher decrease ratio of PEDOT:PSS-based PSCs indicates the strong monomolecular recombination at open-circuit conditions, the obvious 17

bimolecular recombination at short-circuit conditions, and the non-Ohmic contact between PEDOT:PSS and the ITO anode at different incident light intensities [38, 52].

4. Conclusions In this work, the dependence of J-V characteristics on light illumination intensity was investigated to characterize the charge carrier recombination mechanisms in MoOX-based PSCs. After thermal annealing treatment of the (NH 4)6Mo7O24 layer, the (NH4)6Mo7O24 molecule decomposed and formed the MoOX layer. MoOX-based PSCs were shown to exhibit better photovoltaic performance, especially at lower incident light intensities. For the MoOX-based PSCs, the dependence of VOC on the natural logarithm of light intensity with a slope of 1.03kT/q indicates that bimolecular recombination is the dominant mechanism and the weaker monomolecular recombination at open-circuit conditions. The nearly linear dependence of JSC on light intensity with  value is close to 1, which indicates efficient charge carrier transport and collection and weak bimolecular recombination at short-circuit conditions. At the same time, the Ohmic contact between the MoOX layer and ITO anode, and the approximately zero injection potential barrier between the photoactive layer and MoOX layer, favors hole transport and collection. Thus, MoOX-based PSCs show better photovoltaic performance, particularly at lower light intensities.

5. Acknowledgements This work was financially supported by the General project of scientific research 18

of the Education Department of Liaoning Province (Grant No.L2015480). Scientific Research Foundation of Shenyang Agricultural University (Grant No. 880415039). We thank LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript.

19

Reference [1] T.D. Nielsen, C. Cruickshank, S. Foged, J. Thorsen, F.C. Krebs, Solar Energy Materials and Solar Cells 94 (2010) 1553. [2] G. Dennler, M.C. Scharber, C.J. Brabec, Advanced Materials 21 (2009) 1323. [3] P.S. Mbule, H.C. Swart, O.M. Ntwaeaborwa, Journal of Colloid and Interface Science 436 (2014) 9. [4] W. Zhao, S. Li, S. Zhang, X. Liu, J. Hou, Adv Mater 29 (2016) 1604059. [5] S. Li, L. Ye, W. Zhao, S. Zhang, S. Mukherjee, H. Ade, J. Hou, Adv Mater 28 (2016) 9423. [6] M. Jorgensen, K. Norrman, S.A. Gevorgyan, T. Tromholt, B. Andreasen, F.C. Krebs, Adv Mater 24 (2012) 580. [7] M. Jørgensen, K. Norrman, F.C. Krebs, Solar Energy Materials and Solar Cells 92 (2008) 686. [8] A.K. Pandey, M. Aljada, M. Velusamy, P.L. Burn, P. Meredith, Advanced Materials 24 (2012) 1055. [9] J.Y. Kim, S.H. Kim, H.H. Lee, K. Lee, W. Ma, X. Gong, A.J. Heeger, Advanced Materials 18 (2006) 572. [10] G. Terán-Escobar, J. Pampel, J.M. Caicedo, M. Lira-Cantú, Energy & Environmental Science 6 (2013) 3088. [11] S. Shao, J. Liu, J. Bergqvist, S. Shi, C. Veit, U. Würfel, Z. Xie, F. Zhang, Advanced Energy Materials 3 (2012) 349. [12] H. Choi, B. Kim, M.J. Ko, D.-K. Lee, H. Kim, S.H. Kim, K. Kim, Organic Electronics 13 (2012) 959. [13] I. Hancox, K. Chauhan, P. Sullivan, R.A. Hatton, A. Moshar, C. Mulcahy, T. Jones, Energy & Environmental Science 3 (2010) 107. [14] A.A. Mane, M.P. Suryawanshi, J.H. Kim, A.V. Moholkar, Journal of Colloid and Interface Science 483 (2016) 220. [15] C.-C. Chueh, C.-Z. Li, A.K.Y. Jen, Energy & Environmental Science 8 (2015) 1160. [16] B. Xiao, H.B. Wu, Y. Cao, Mater. Today 18 (2015) 385. [17] S. Murase, Y. Yang, Advanced Materials 24 (2012) 2459. [18] M. Corazza, F.C. Krebs, S.A. Gevorgyan, Solar Energy Materials and Solar Cells 143 (2015) 467. [19] G. del Pozo, B. Arredondo, B. Romero, G. Susanna, F. Brunetti, Solar Energy 144 (2017) 105. [20] Z. Liu, S. Hui, N. Wang, Journal of Luminescence 181 (2017) 310. [21] H.M. Thirimanne, K.D.G.I. Jayawardena, J.V. Anguita, V. Stolojan, C.A. Mills, S.R.P. Silva, Advanced Electronic Materials 2 (2016). [22] J.-C. Ke, Y.-H. Wang, K.-L. Chen, C.-J. Huang, Journal of Colloid and Interface Science 465 (2016) 311. [23] T. Zhaoyi, H. Zhaoya, Z. Dong, W. Xiaolin, Journal of Radioanalytical and Nuclear Chemistry 299 (2014) 1165. [24] J. Wienold, R.E. Jentoft, T. Ressler, European Journal of Inorganic Chemistry 2003 (2003) 1058. [25] J. Xiong, Z. He, S. Zhan, B. Yang, X. Zhang, P. Cai, C. Xu, X. Xue, J. Zhang, Current Applied Physics 17 (2017) 1021. [26] J. Liu, X.K. Wu, S.J. Chen, X.D. Shi, J. Wang, S.J. Huang, X.J. Guo, G.F. He, Journal of Materials Chemistry C 2 (2014) 158. [27] M.E. Thompson, A.N. Bartynski, C. Trinh, A. Panda, B. Lassiter, K. Bergemann, J.D. Zimmerman, S.R. Forrest, Nano Lett (2013). [28] Y.-S. Kim, T. Kim, B. Kim, D.-K. Lee, H. Kim, B.-K. Ju, K. Kim, Organic Electronics 14 (2013) 1749. [29] C.M. Proctor, T.Q. Nguyen, Applied Physics Letters 106 (2015) 083301. [30] J.-J. Chao, S.-C. Shiu, C.-F. Lin, Sol Energ Mat Sol C 105 (2012) 40. 20

[31] P. Zhai, H. Lee, Y.-T. Huang, T.-C. Wei, S.-P. Feng, Journal of Power Sources 329 (2016) 502. [32] A. Ignaczak, Electrochimica Acta 53 (2008) 2619. [33] Z.a. Tan, D. Qian, W. Zhang, L. Li, Y. Ding, Q. Xu, F. Wang, Y. Li, Journal of Materials Chemistry A 1 (2013) 657. [34] F. Wang, G. Sun, C. Li, J. Liu, S. Hu, H. Zheng, Z. Tan, Y. Li, ACS Appl Mater Interfaces 6 (2014) 9458. [35] L.C.C. Elliott, J.I. Basham, K.P. Pernstich, P.R. Shrestha, L.J. Richter, D.M. DeLongchamp, D.J. Gundlach, Advanced Energy Materials 4 (2014) 1400356. [36] S.R. Cowan, A. Roy, A.J. Heeger, Physical Review B 82 (2010) 245207. [37] Q. An, F. Zhang, Q. Sun, M. Zhang, J. Zhang, W. Tang, X. Yin, Z. Deng, Nano Energy 26 (2016) 180. [38] A.K.K. Kyaw, D.H. Wang, V. Gupta, W.L. Leong, L. Ke, G.C. Bazan, A.J. Heeger, ACS nano 7 (2013) 4569. [39] Y. Zhou, T.M. Khan, J.W. Shim, A. Dindar, C. Fuentes-Hernandez, B. Kippelen, Journal of Materials Chemistry A 2 (2014) 3492. [40] C.M. Proctor, C. Kim, D. Neher, T.-Q. Nguyen, Advanced Functional Materials 23 (2013) 3584. [41] S.R. Cowan, J. Wang, J. Yi, Y.-J. Lee, D.C. Olson, J.W.P. Hsu, Journal of Applied Physics 113 (2013) 154504. [42] S. Massip, P.M. Oberhumer, G. Tu, S. Albert-Seifried, W.T.S. Huck, R.H. Friend, N.C. Greenham, The Journal of Physical Chemistry C 115 (2011) 25046. [43] S. Shoaee, M.P. Eng, E. Espildora, J.L. Delgado, B. Campo, N. Martin, D. Vanderzande, J.R. Durrant, Energy & Environmental Science 3 (2010) 971. [44] Y. Yu, X. Li, Z. Shen, X. Zhang, P. Liu, Y. Gao, T. Jiang, J. Hua, Journal of Colloid and Interface Science 490 (2017) 380. [45] S. Mori, A. Asano, The Journal of Physical Chemistry C 114 (2010) 13113. [46] Z. Wang, X. Xu, Z. Li, K. Feng, K. Li, Y. Li, Q. Peng, Advanced Electronic Materials 2 (2016). [47] Y. Zhang, X.-D. Dang, C. Kim, T.-Q. Nguyen, Advanced Energy Materials 1 (2011) 610. [48] M. Lenes, M. Morana, C.J. Brabec, P.W.M. Blom, Advanced Functional Materials 19 (2009) 1106. [49] J. Cheng, F. Xie, Y. Liu, W.E.I. Sha, X. Li, Y. Yang, W.C.H. Choy, Journal of Materials Chemistry A 3 (2015) 23955. [50] A. Maurano, R. Hamilton, C.G. Shuttle, A.M. Ballantyne, J. Nelson, B. O’Regan, W. Zhang, I. McCulloch, H. Azimi, M. Morana, C.J. Brabec, J.R. Durrant, Advanced Materials 22 (2010) 4987. [51] L.C. Palilis, P.A. Lane, G.P. Kushto, B. Purushothaman, J.E. Anthony, Z.H. Kafafi, Organic Electronics 9 (2008) 747. [52] E.T. Hoke, K. Vandewal, J.A. Bartelt, W.R. Mateker, J.D. Douglas, R. Noriega, K.R. Graham, J.M. Fréchet, A. Salleo, M.D. McGehee, Advanced Energy Materials 3 (2013) 220.

21

Figure 1. (a) Configuration and (b) energy-band diagram of the PSCs with MoOX and PEDOT:PSS as the HEL and PTB7:PC71BM as the photoactive layer.

(NH4)6Mo7O24.4H2O

a

Mo 3d5/2

H2O NH4

+

+

4000

NH4

3000 2000 Wave number (cm-1)

1000

Intensity (a.u.)

H2O

Mo-O

Transmittance (a.u.)

b

O 1s

MoOX

Mo 3d3/2 Mo 3p3/2 Mo 3p1/2 C 1s

600

500

400

300

200

Binding Energy (eV)

Figure 2. (a) FTIR spectra of (NH4)6Mo7O24-4H2O layer with and without thermal annealing. (b) XPS spectra of the MoOX layer.

22

100

MoOX

a

b

80

PEDOT:PSS

-4

-8

60

EQE (%)

-2

Current Density (mA cm )

0

-12

40

20

-16

MoOX PEDOT:PSS

0.0

0.2

0.4

0.6

0 300

0.8

400

Voltage (V)

500

600

700

800

Wavelength (nm)

Figure 3. (a) J-V characteristics of photovoltaic devices using the MoOX and PEDOT:PSS layers as the HEL. (b) External quantum efficiency (EQE) for PSCs using the MoOX and PEDOT:PSS layers as the HEL.

-2

Current Density (mA cm )

1000

MoOX

100

PEDOT:PSS

10 1 0.1 0.01 1E-3 1E-4 1E-5 1E-6 -1

0

1

Applied Bias (V)

Figure 4. J-V curves of PSCs based on PTB7:PC71BM as the photoactive layer, and the MoOX and PEDOT:PSS layers as the HEL under dark conditions.

23

0

0

b

MoOX Layer -2

Current Density (mA cm )

-2

Current Density (mA cm )

a

-4

-8

-12

-16 max power point

-20 0.0

0.2

0.4

0.6

10 20 30 40 55 70 85 100

PEDOT:PSS Layer

-4

-8

-12

-16 max power point

-20 0.0

0.8

0.2

Voltage (V)

0.4

10 20 30 40 55 70 85 100

0.6

0.8

Voltage (V)

Figure 5. J-V characteristics of PTB7:PC71BM PSCs with (a) a MoOX layer as the HEL and (b) a PEDOT:PSS layer as the HEL under various light intensities ranging from 10 to 100 mW cm−2.

80 0.76

a n=1.03 kT/q

0.72

72 n=1.32 kT/q

0.68

0.74

FF (%)

VOC (V)

0.72

68

0.70

12

0.68 0.66

8

0.64

1

4

0.64

64 10

100

20

40

60

80

100

20

12

0.937

d

0.914

11

10

PCE (%)

c

10

-2

JSC (mA cm )

0.76

16

76

b

9

10

100

20

40 -2

Incident Light Intensity (mA cm

60

80

8 100

)

Figure 6. Light intensity dependences of (a) VOC, (b) FF, (c) JSC, and (d) PCE for PSCs with PTB7:PC71BM as the photoactive layer; black squares and red circles correspond to a MoOX layer and a PEDOT:PSS layer as the HEL, respectively. 24

40

60

100

Transmittance (%)

90

80

70

60 MoOX/ITO

Transmittance Ratio 91.1

PEDOT:PSS/ITO

50 300

400

500

90.2

600

700

800

Wavelength (nm)

Figure 7. Transmittance spectra of MoOX/ITO and PEDOT:PSS/ITO; optical transmittance ratios were embedded in the optical transmittance spectrum.

25

Table 1. Photovoltaic parameters of PSCs with MoOX and PEDOT:PSS as the HEL. EEL Materials

VOC (V)

JSC (mA cm-2)

Calculated JSC

FF (%)

PCE (%)

RS ( cm-2)

RSH ( cm-2)

MoOX 0.76 PEDOT:PSS 0.76

18.3±0.3 17.9±0.2

17.6 17.2

65.6±1.5 65.1±1.4

9.12±0.31 8.91±0.32

5.44±0.13 7.12±0.34

461±12 332±16

26

80 0.76

a n=1.03 kT/q

0.72

72 n=1.32 kT/q

0.68

0.74

FF (%)

VOC (V)

0.72

68

0.70

12

0.68 0.66

8

0.64

1

4

0.64

64 10

100

20

40

60

80

100

20

12

0.937

d

0.914

11

10

9

10

100

20

40

60 -2

Incident Light Intensity (mA cm

27

)

80

8 100

PCE (%)

c

10

-2

JSC (mA cm )

0.76

16

76

b

40

60