- Email: [email protected]
An insight on oxide interlayer in organic solar cells: From light absorption and charge collection perspectives
Organic Electronics 31 (2016) 266e272
Contents lists available at ScienceDirect
Organic Electronics journal homepage: www.elsevier.com/locate/orgel
An insight on oxide interlayer in organic solar cells: From light absorption and charge collection perspectives Zhenghui Wu, Bo Wu, Hoi Lam Tam, Furong Zhu* Department of Physics, Institute of Advanced Materials, and Institute of Research and Continuing Education (Shenzhen), Hong Kong Baptist University, Kowloon Tong, NT, Hong Kong
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 December 2015 Received in revised form 24 January 2016 Accepted 25 January 2016 Available online xxx
A comprehensive study of the effect of oxide interlayer on the performance of bulk-heterojunction organic solar cells (OSCs), based on poly[[4,8-bis[(2-ethylhexyl)oxy] benzo [1,2-b:4,5-b'] dithiophene2,6- diyl] [3-fluoro-2-[(2-ethylhexyl)carbonyl] thieno [3,4-b] thiophenediyl]] (PTB7): [6,6]-phenyl C71 butyric acid methyl ester (PC70BM) blend system, is carried out by optical simulation, interfacial exciton dissociation and charge collection analyses. It is found that a PTB7:PC70BM blend layer thickness optimized for maximum light absorption in OSCs does not generally give rise to the highest power conversion efficiency (PCE). OSCs, e.g., based on PTB7:PC70BM blend system, can benefit from the oxide interlayer in two ways, (1) to enhance the built-in potential for reducing recombination loss of the photogenerated charges, and (2) to improve charge collection by removal of unfavorable interfacial exciton dissociation. The combined effects result in ~20% improvement in PCE over an optimized control cell, having an identical layer configuration without an oxide interlayer. © 2016 Elsevier B.V. All rights reserved.
Keywords: Transient photocurrent Oxide/organic hetero-interface Charge collection efficiency Interfacial exciton dissociation Absorption enhancement
1. Introduction Power conversion efficiency (PCE) of >10% for both single junction and tandem organic solar cells (OSCs) has been demonstrated recently [1,2] through continued progresses made in both new material development and device innovation. However, the performance of OSCs still faces a major challenge to overcome the mismatch between optical absorption length and charge transport scale, caused by the low charge mobility in organic materials. High performing OSCs require a balance between charge transport and optical absorption in selecting the most suitable active layer in a narrow thickness range from 75 to 300 nm. It has been revealed that the unbalanced charge mobility in the photoactive layer, due to oxygen-induced charge traps, is one of the degradation mechanisms [3]. The degradation caused by the interfacial passivation could be avoided by the removal of the low work function cathode [4]. Different approaches have been proposed to enhance the performance of OSCs, e.g., optimizing the nanoscale morphology of organic blend layer to benefit the charge transport [5], incorporating photonic structures in the active layer to achieve
* Corresponding author. E-mail address: [email protected] (F.R. Zhu). http://dx.doi.org/10.1016/j.orgel.2016.01.040 1566-1199/© 2016 Elsevier B.V. All rights reserved.
broadband absorption enhancement via the optical coupling effect [6], and interposing an oxide optical spacer between the active layer and the electrode [7e10]. The application of metal oxides in OSCs and the impact of the properties of metal oxide/organic hetero-interfaces on cell performance have attracted a lot of attentions. For example, the use of a titanium suboxide [8] or ZnO interlayer [9] between the active layer and the reflective electrode in OSCs has been shown to increase the absorption in the active layer. The enhancement in the photocurrent generation due to the ZnO spacer is found to be more beneficial in OSCs with a thin active layer (<60 nm), but has less effect in device with a thicker layer (>90 nm) [10]. The ZnO layer also serves as a hole blocking layer to reduce the recombination rate [11]. Despite numerous reports about the improvement on OSCs due to the incorporation of the oxide spacer, mechanisms of improving cell performance due to oxide interlayer in OSCs are still inconclusive and controversial. €s et al. [12] reported that no optical beneficial effect can be Ingana expected by interposing an optical spacer layer in OSCs with an already optimized active layer thickness. In a recent work, we found that the insertion of a ZnO interlayer between the active layer and the metal contact enables to eliminate the unfavorable exciton dissociation that would otherwise occur at the organic/ metal interface, thereby improving the charge collection efficiency [13]. This suggests that the removal of the unfavorable interfacial
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
exciton dissociation is a perquisite for high performing OSCs. In addition to the absorption enhancement, it is also important to understand the effect of oxide interlayer on charge collection properties at the organic/cathode interface. The organic/oxide hetero-interfaces have a crucial impact on interfacial exciton dynamics, interfacial charge trapping behavior and the overall device performance. The complexity of physics operating in these OSCs offers a major driving force for new developments. In this work, a systematic study on the effect of ZnO interlayer on OSCs, based on poly[[4,8-bis[(2-ethylhexyl)oxy] benzo [1,2-b:4,5-b'] dithiophene2,6- diyl] [3-fluoro-2-[(2-ethylhexyl)carbonyl] thieno [3,4-b] thiophenediyl]] (PTB7): [6,6]-phenyl C71 butyric acid methyl ester (PC70BM) blend system, was carried out. The absorption enhancement in OSCs due to the presence of a ZnO interlayer between the organic active layer and metal contact is analyzed using optical admittance analysis. Effect of metal oxide/organic hetero-interfaces on the charge recombination and exciton dissociation dynamics at organic/electrode interface in OSCs was analyzed using light intensity dependent current densityevoltage (JeV) characteristics and transient photocurrent (TPC) measurements. Our results showed a 20% power conversion efficiency (PCE) improvement in OSCs, realizing through a simultaneous enhancement in light absorption and effective charge collection.
mechanisms of ZnO-induced performance enhancement of OSCs. From optical point of view, a flat OSC can be considered as a thin film system consisting of m layers. The optical electric field distribution in the cells can be optimized optically with an appropriate choice of the active layer thickness or by insertion of an oxide interlayer between the bulk heterojunction and the metal contact. Under normal incidence, the absorption spectrum of the ith layer, for example the active layer Ai(l), in an m layered thin film solar cell can be calculated by optical admittance method [15,16]:
Ai ðlÞ ¼ ½1 RðlÞ½1 ji ðlÞ
ITO/glass substrates with a sheet resistance of 10 U/square were used for the OSC fabrication. They were cleaned in ultrasonic bath sequentially with detergent, deionized water, acetone and isopropanol for 20 min each, followed by the in situ oxygen plasma treatment prior to the cell fabrication. The PTB7 (1 Material):PC70BM (Nano C) blended in a weight ratio of 1:1.5 was fully dissolved in chlorobenzene (CB) (SigmaeAldrich, 99.8%) with 3% 1, 8-Diiodooctane (DIO) (SigmaeAldrich) at 60OC before use. A set of structurally identical OSCs with different active layer thicknesses ranging from 60 to 130 nm was then fabricated on PEDOT:PSScovered ITO/glass substrates by spin-coating inside a N2-purged glove-box with O2 and H2O levels <0.1 ppm. ZnO NPs were synthesized following the procedure described by Hermann-Jens [14]. A 10 nm thick ZnO interlayer was deposited on the active layer by spin-coating inside the glove-box. The samples were then transferred to the adjacent vacuum system with a base pressure of <5.0 104 Pa for forming a 100 nm thick Al top contact, evaporated at a rate of 1 Å/s. OSCs were transferred back to the glove-box for JV characteristic measurement under an air mass (AM) 1.5G irradiation of 100 mW/cm2. The intensity of the solar simulator was calibrated using a silicon reference cell with a KG-5 filter. The transient photocurrent in the devices, e.g., ITO/PEDOT:PSS/ PC70BM (400 nm)/ZnO(10 nm)Al (30 nm) and a control sample of ITO/PEDOT:PSS/PC70BM (400 nm)/Al (30 nm), was generated using a pulsed Nd:YAG laser with the wavelength of 355 nm and pulse duration of <5 ns, illuminated from a semitransparent top Al cathode side. The devices were connected to the 1 MU input terminal of an oscilloscope (Agilent DSO8064A Infiniium Digital Oscilloscope) to measure the transient photovoltage. The corresponding transient photocurrent of the devices was obtained by converting the transient photovoltage to the current using the internal resistor of the functional generator (50 U) at different biases. 3. Results and discussions It is known that the performance of OSCs is dependent on the intensity and incident angle of solar irradiation, varied due to sunrise to sunset during the day. In the simulation, we considered incoming light at normal to the cell to unraveling the fundamental
i1 Y
jj ðlÞ;
(1)
j¼1
where jj is the ratio of the time averaged numerical magnitude of the Poynting's vector at the jth and (j-1)th boundaries, R(l) is the wavelength-dependent reflectance of the cell. Using the flux of the incident solar radiation F(l), e.g., AM1.5G measured in the units of W/m2,nm, the integrated absorbance of the active layer, defined as Ai which is proportional to the short circuit current density (JSC) of the cell, can be calculated as [15,16]:
Z Ai ¼
2. Experimental section
267
Ai ðlÞ FðlÞ dl Z : FðlÞ dl
(2)
Equation (2) allows evaluating the optimal active layer, e.g., PTB7:PC70BM, thickness in OSCs through maximizing its integrated absorptance. Essentially, the optical phenomena in thin film OSCs have a profound impact on light absorption in the active layer. This procedure takes into account the interference effects and allows us to optimize light absorption enhancement in the active layer, thereby improving JSC and PCE. Light absorption in different types of OSCs under normal incidence was calculated in the wavelength range from 380 to 780 nm. The thickness of the organic photoactive layer was varied over the range from 40 to 240 nm. This region is comparable in thickness to the active layer in OSCs made with different organic photoactive materials. To understand the effect of the ZnO interlayer on absorption enhancement in the OSCs, the integrated absorbance of the active layer in OSCs of ITO/PEDOT:PSS (40 nm)/PTB7:PC70BM/Al (100 nm) (control cell) and that of the structurally identical cells with a ZnO interlayer inserted between the active layer and the metal contact: ITO/PEDOT:PSS (40 nm)/PTB7:PC70BM/ZnO/Al (100 nm), was calculated using eqs. (1) and (2). The wavelength dependent refractive indices of each layer in the OSC system were measured using variable angle spectroscopic ellipsometer [17]. The integrated absorptance of the PTB7:PC70BM layer, calculated for the control OSC and the structurally identical cells with different ZnO interlayer thicknesses (10 nm, 20 nm and 30 nm), as a function of its layer thickness ranging from 40 to 240 nm is shown in Fig. 1. It is clear that the integrated absorbance displays an oscillation behavior with increase in the active layer thickness, having a relative absorption maximum occurred at the blend layer thickness of ~105 nm for a control OSC. However, the thickness of the PTB7:PC70BM layer corresponding to the absorption maximum in the integrated absorptance becomes thinner, e.g., ~90 nm for OSC with a 10 nm thick ZnO interlayer. A slight decrease in the integrated absorbance is observed when the thickness of the ZnO interlayer further increases from 10 to 30 nm. This suggests that the insertion of an oxide interlayer between the organic active layer and the metal contact does not have an obvious contribution to the absorption enhancement in active layer as compared to that in an optically optimized control cell. Although a thin oxide interlayer alone cannot account for light
268
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
absorption enhancement in the already optically optimized control cell, OSCs with an optimized ZnO interlayer is still beneficial for efficient device operation. The results in Fig. 1 reveal that OSCs with an appropriate ZnO interlayer (~10 nm thick) permits a thinner active layer (~90 nm) to achieve the similar maximum light absorption as compared to an optimized thicker active layer (~105 nm) used in the control cell. The inset in Fig. 1 is the crosssectional view of the OSC. It is known that light absorption in OSCs is limited due to the presence of a mismatch between optical absorption length and charge transport scale, caused by the low charge mobility in conjugated polymers. The use of a thinner photoactive layer without an obvious loss in light absorption, enabled in OSCs with a ZnO interlayer, as shown in Fig. 1, helps to create a higher internal electric field across the active layer, thereby reducing the exciton recombination losses and improving the charge transport. It shows that ZnO interlayer is advantageous for the efficient operation of the PTB7:PC70BM-based OSCs. The presence of the ZnO interlayer allows using a thinner active layer without moderating the absorption in the optimized control cells with a thicker active layer. A combination of the efficient absorption and higher internal potential favors the efficient charge transport, and thereby improving the cell performance. The relevance of the above discussion in the absorption in the active layer can be seen by measuring the performance of the control cell and the OSCs with a thin ZnO interlayer. Two sets of structurally identical OSCs, ITO/PEDOT:PSS (30 nm)/PTB7:PC70BM/ Al(100 nm) and ITO/PEDOT:PSS (30 nm)/PTB7:PC70BM/ZnO (~10 nm)/Al(100 nm), with different active thicknesses ranging from 55 to 130 nm, were made for comparison studies. The short circuit current density of both sets of the OSCs as a function of the active layer thickness was measured and normalized for comparison and then plotted in Fig. 2. The correlation between the JSC and the active layer thickness, measured for both types of the OSCs, is very similar to the simulation results. For the control cells, JSC reaches to the relative maximum value at about 105 nm thick active layer, while for OSCs with a 10 nm thick ZnO interlayer, JSC reaches to the relative maximum value at a thinner active layer of ~90 nm, following with a graduate decay, which is consistent with the simulation results. The optical admittance analysis agrees well with the experimental results in showing that OSCs having an optimal ZnO interlayer interposed between the 90 nm thick active layer and
Fig. 1. Calculated integrated absorbance of the active layer in control cell and OSCs with different ZnO interlayer thicknesses of 10 nm, 20 nm and 30 nm. Inset in Fig. 1 shows the cross sectional view of the OSC with a ZnO interlayer added between the active layer and the metal contact.
the Al contact corresponded to the best cell performance, as presented in Fig. 2. It is found that the highest PCE of the control cells was obtained for the cell having a ~90 nm active layer. It is clear that a PTB7:PC70BM blend layer thickness optimized for maximum light absorption (~105 nm) in the control OSCs, e.g., achieving highest JSC, does not give rise to the highest PCE. Such a deviation in the optical optimization and device performance essentially reflects the mismatch between the optical absorption length and charge transport scale in the organic semiconductors. Two types of OSCs were fabricated: one was the control cell optimized without the use of an oxide interlayer, the other was the cell optimized with a ZnO (10 nm) interlayer. JeV characteristics of both cells are plotted in Fig. 3a, the open-circuit voltage (VOC), JSC and fill factor (FF) of the OSC with a ZnO interlayer between the active layer and the Al cathode increased from 0.73 to 0.74 V, 13.7e14.5 mA/cm2 and 0.65 to 0.73, respectively, leading to an overall 20% increase in PCE from 6.51% to 7.82%. The steady improvements in JSC and VOC, seen in Fig. 3a, are the factors contributing to the enhancement in overall cell performance. This suggests that light absorption and transport of the photo-generated carriers are more favorable in OSCs with a ZnO interlayer, although a slight thinner photoactive layer thickness was used in the cells. A summary of device parameters obtained for both types of the OSCs is listed in Table 1. Incident photon to current efficiency (IPCE) spectra measured for a control cell and the OSC with a 10 nm thick ZnO interlayer between the active layer and the Al contact are plotted in Fig. 3b. IPCE characteristics provide the information of light absorption in the active region contributing to exciton generation and the overall external quantum efficiency of the OSCs. Light absorption in the
Fig. 2. Comparison of normalized PCE and JSC, measured for a set of (a) control cells and (b) structurally identical OSCs with a ~10 nm thick ZnO interlayer, as a function of the active layer thickness over the thickness range from 40 to 130 nm.
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
experimentally optimized control cell with a ~90 nm thick active layer is lower than that in the optimized cell with the same active layer thickness having a ~10 nm thick ZnO interlayer. It is apparent that cells with a ZnO interlayer apparently possess a better interfacial contact at the organic/cathode interface, therefore an efficient charge collection and a higher drift current are expected. The change in the IPCE spectra at the wavelength <500 nm and the wavelength region from 500 to 750 nm, as shown in Fig. 3b, is due to the interference effect in the OSCs [17]. There is a slight decrease in the IPCE measured for the OSCs with a ZnO interlayer at the wavelength <500 nm, and an obvious increase in IPCE over a spectral region from 500 to 750 nm. PTB7 does not have a strong absorption below 500 nm. Incorporation of a thin ZnO interlayer in the cells still benefit from an overall absorption enhancement over the wavelength region from 500 to 750 nm, attaining a ~6% increase in JSC as compared to the control cell. JSC calculated using IPCE spectra of the control cell and the OSCs with a ZnO interlayer also agrees with measured JeV characteristics, showing that the improvement in performance of the OSCs with a ZnO interlayer contributes the improvement in JSC, as well as VOC and FF illustrated in Fig. 4. The results in Fig. 4 reveal that VOC and FF decrease with increase in the thickness of the active layer in the cells. The behavior of the active layer thickness-dependent VOC and FF in OSCs originates from the limited charge mobility in the organic
Fig. 3. (a) The JV characteristics and (b) IPCE spectra measured of control cell and OSC with a 10 nm thick ZnO interlayer between the active layer and the Al contact.
269
Table 1 The characteristics of optimized devices with/without the ZnO interlayer. The values of JSC and PCE calibrated using IPCE measurements are also given in the parentheses for comparison. OSCs
VOC (V)
JSC (mA/cm2)
FF (%)
PCE (%)
Control cell Cells with a ZnO interlayer
0.73 0.74
13.3 (13.7) 14.0 (14.5)
65.1 72.9
6.32 (6.51) 7.55 (7.82)
semiconductors. It is shown clearly that that VOC and FF become less sensitive to the change in the active layer thickness when a ZnO interlayer was inserted between the active layer and Al contact. The charge collection properties at the organic/cathode interface, caused by the insertion of an oxide interlayer between the active layer and the metal contact, also play an important role in determining the performance of the cells. To further understand the origin of the enhancement in the performance of the OSCs with a ZnO interlayer as compared to a control cell, charge recombination characteristics in the OSCs were analyzed. Fig. 5 shows the double logarithmic plot of photocurrent Jph (Jph ¼ JL e JD, where JL is the current density measured under AM1.5G irradiation of 100 mW/cm2 and JD is the dark current) as a function of the effective bias Veff (Veff ¼ V0 eVb, where V0 is the built-in potential measured at Jph ¼ 0, and Vb is the applied bias), measured for the OSCs with a 10 nm thick ZnO interlayer and a control cell. In this work, V0 of both types of OSCs is approximately 0.8 V. The saturated photocurrents Jsat of both types of the OSCs are also shown in Fig. 5. Jsat in the OSCs depends only on the charge generation as Jph saturates at high Veff (>1.0 V), giving rise to almost 100% collection of
Fig. 4. Comparison of (a) VOC and (b) FF measured for control cells and OSCs having a ~10 nm thick ZnO buffer as a function of the active layer thickness.
270
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
the photo-generated charges. The ratio of Jph to Jsat reflects essentially a measure of an overall loss in the photo-excited charges, e.g., due to bimolecular recombination, trap-assisted defect states, occurred during transport processes prior to the collection by the electrodes [18]. At low Veff region, the number of photo-generated carriers that can be extracted decreases due to the increase in the charge recombination, leading to a poor charge collection. As shown in Fig. 5, Jph obtained for a control cell decreases much faster than that in the OSC with a ZnO interlayer at Veff <0.5 V. The loss due to the recombination of photo-generated carrier at the organic/cathode interface is obviously much higher in the control cell as compared to that in the OSCs with a ZnO interlayer at Veff <0.5 V. Both types of OSCs have an identical anode contact and active layer configuration, the change in the JphVeff characteristics at low Veff region is apparently associated with the change in the interfacial properties at the organic/cathode contact. This reveals that presence of a thin ZnO interlayer between the active layer and the Al cathode augments the photocurrent in this Veff region, revealing improvement in charge collection efficiency at the cathode contact. Dissociation of the photo-excited excitons occurs at the interface between the two dissimilar materials, e.g., organic/electrode and donor/acceptor interfaces, the later process dominates the photocurrent generation in the OSCs. The charges generated due to the exciton dissociation at the donor/acceptor interface are drifted under the internal built-in potential, forming the field-dependent drift current in the cells. The transient photocurrent in OSCs comprises (1) the drift current formed in the active layer, and (2) the contribution from the interfacial exciton dissociations occurred at the organic/electrode interfaces [13]. The drift current is a function of the internal electric field, while the current generated via interfacial exciton dissociation is independent on the internal electric field. Contributions to the transient photocurrent due to the drift current and the interfacial exciton dissociation cannot be decoupled in the JV characteristics of the cells measured under the steady state condition. However, the behavior of exciton dissociation at the organic/electrode interfaces and its contribution to the transient current can be analyzed by applying an opposite external bias to suppress the drift current in the transient photocurrent measurements. It then becomes possible to exam the interfacial exciton dissociation in the absence of internal built-in potential. Therefore, the correlation between the interfacial exciton dynamics and charge collection at organic/electrode interface can be obtained by measuring transient current for devices at different biases. In this work, charge collection and the interfacial
Fig. 5. Photocurrent density Jph as a function of the effective bias Veff across the active layer measured for a control cell and the OSC with a 10 nm thick ZnO interlayer.
exciton dissociation processes at the organic/cathode interface in the cells were investigated using TPC measurements. TPC measurement is a technique to study the transient dynamics of the photo-generated carriers with a time scale of ~10 ns in electronic devices. The details of the TPC measurement are described in our previous work [13]. Before the exposure of the devices to light, a dark transient current (ID) was recorded as the background current. When the active layer was excited by the laser, the transient current (IL) was recorded again. Then transient photocurrent (ITPC) of the device can be calculated by:
ITPC ¼ IL ID :
(3)
To analysis the result of the interfacial exciton dissociation on charge collection at the organic/Al cathode interface, two devices with layered structures of ITO/PEDOT:PSS/PC70BM(400 nm)/ Al(30 nm) (control device) and ITO/PEDOT:PSS/PC70BM(400 nm)/ ZnO(~10 nm)/Al (30 nm) were made for TPC measurements. The effect of ZnO/Al bilayer on exciton dissociation at the oxide/electrode interface can be unraveled by probing its contribution to the transient photocurrent, measured by suppression of the fielddrifted current in the bulk, e.g., applying a reverse bias to compensate the built-in potential in the device. As ZnO has a lower work function compared to that of the Al cathode, the use of the ZnO/Al bilayer contact results in an increase in the built-in potential across the active region of the devices with a ZnO interlayer, e.g., increasing from ~0.5 V (Al contact) to ~0.8 V (ZnO/Al contact). Therefore, a slight higher compensation voltage of ~0.9 V was used in the transient photocurrent measurement, to suppress the fielddependent photocurrent generated in the samples with a ZnO interlayer. The transient photocurrents measured for both devices at different biases are plotted in Fig. 6a and b. The positive transient photocurrent decays were observed for both devices without the external bias, suggesting that the electric field drifted currents are dominated in both devices. However, when the TPC measurements were performed for both devices with an opposite external bias, very different transient photocurrent behaviors were observed. A clear rapid negative transient photocurrent signal for the control device was observed at a reverse bias of 0.4 V, shown in Fig. 6a. The prompt negative transient photocurrent, originated from the fast interfacial exciton dissociation at the PC70BM/Al interface, is apparently unfavorable for the electron collection. In comparison with the flow of the transient photocurrent generated at the PC70BM/Al interface shown in Fig. 6a, there was an obvious positive transient photocurrent seen in the device with a presence of a 10 nm thick ZnO interlayer between active layer and Al contact at a reverse bias of 0.9 V, shown in Fig. 6b. It becomes clear that the interfacial exciton dissociation at the organic/Al interface in the control device impedes the electron collection. In contrast, the presence of a 10 nm thick ZnO interlayer between the active layer and the metal cathode in the OSCs promotes an efficient charge collection through elimination of the adverse interfacial exciton dissociation process occurred in the ones without a ZnO interlayer. The TPC results agree well with the analysis based on the JpheVeff characteristics of the cells shown in Fig. 5. It is clear that the enhanced Jph, observed in the OSCs with a ZnO interlayer at the low Veff, arises from the favorable interfacial exciton dissociation behavior that inhibits the interfacial bimolecular recombination. The difference in charge recombination and extraction in both types of the OSCs, caused by the adverse exciton dissociation processes at the organic/cathode interface, can be further examined by analyzing the JV characteristics of the OSCs under different light intensities. The empirical relationship between Jph and the intensity of light (I) is [19,20]:
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
271
variation in the exponent of JphI characteristics reflects the difference in charge accumulation at the organic/cathode interface in both cells. The charge accumulation at the organic/cathode in the control cell becomes more notable as compared to the OSC having a ZnO interlayer. A relatively high charge recombination in the control cell is then expected. It is seen that charge accumulation at the organic/cathode is much stronger in the control cell than that in the OSC with a ZnO interlayer. A structurally identical OSC with a thin ZnO interlayer apparently favors an efficient charge collection process, thereby yielding a high PCE of 7.82%, which is about 20% higher than that of a control cell (6.51%). 4. Conclusion Functionalized metal oxide/organic hetero-interfaces have a crucial impact on the optical, interfacial exciton dynamics, interfacial charge trapping behavior of the overall performance of organic solar cells. The presence of a thin ZnO interlayer layer between the photoactive layer and the cathode is favorable for efficient operation of OSCs. The ZnO interlayer serves a dual purpose to enable a thinner active layer without moderating the absorption via interference effect as well as to augment charge collection efficiency via favorable interfacial exciton dissociation at the organic/ cathode interface. This simple yet effective implementation can improve the device performance with the PCE of PTB7:PC70BMbased OSCs by ~20%.
Fig. 6. Transient photocurrent signals, due to the interfacial exciton dissociation at (a) PC70BM/Al and (b) PC70BM/ZnO/Al interfaces.
Jph fI a ;
(4)
where the exponent a in eq. (4) is usually less than one. Recombination of the photo-generated charges in the device generally results in a deviating from one [18]. With increase in the intensity of light, charge accumulation in the active layer of an OSC will be quickly saturated due to the limited trap states. Therefore, less “bright” light irradiation can be used to analyze the charge accumulation processes in the OSCs. The double logarithmic plot of JphI characteristics, measured for the OSC with a 10 nm thick ZnO interlayer and the control cell, are shown in Fig. 7a and b. The power law dependence of Jph on light intensity is clearly seen. The power exponents of JphI characteristics measured for both cells approach to one at Veff of >0.25 V. This suggests that most photo-generated charges can be collected at the organic/electrode contacts without charge accumulation in the cell at high Veff. As the charge mobility in organic materials is field dependent, less efficient charge transport and collection in the OSCs would occur at the low Veff, leading to the build-up of space charges that gives rise to higher bimolecular recombination, and thereby a lower power exponent in the JphI characteristics. As shown in Fig. 7a, the exponent of JphI characteristics observed for the control cell decreases to 0.67 at Veff ¼ 0.09 V. A higher exponent value of 0.87 is obtained for OSC with a ZnO interlayer under the same effective voltage. As both cells have the same anode/organic contact and blend system, it can be considered that the charge accumulation behaviors at the anode contact and in the active layer of both cells are very similar. The
Fig. 7. Double logarithmic plot of photocurrent density as a function of light intensity for (a) a control cell and (b) the OSC with a 10 nm thick ZnO interlayer, measured under different effective voltages.
272
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
Acknowledgments This work was supported by Research Grants Council of Hong Kong Special Administrative Region, China, Project No. T23-713/11, GRF12303114, National Natural Science Foundation of China (No. 61275037), Shenzhen Peacock Project, KQTD20140630110339343 and Hong Kong Baptist University FRG2/14-15/081. References [1] C.C. Chen, W.H. Chang, K. Yoshimura, K. Ohya, J. You, J. Gao, Z. Hong, Y. Yang, Adv. Mater 26 (2014) 5670. [2] Y. Liu, J. Zhao, Z. Li, C. Mu, W. Ma, H. Hu, K. Jiang, H. Lin, H. Ade, H. Yan, Nat. Commun. 5 (2014) 5293. [3] C. Vijila, G.M. Ng, Mein Jin Tan, W.P. Goh, F.R. Zhu, Appl. Phys. Lett. 95 (2009) 263305. [4] X.Z. Wang, X.X. Zhao, G. Xu, Z.K. Chen, F.R. Zhu, Sol. Energy Mater. Sol. Cells 104 (2012) 1. [5] W.L. Xu, B. Wu, F. Zheng, H.B. Wang, Y.Z. Wang, F.G. Bian, X.T. Hao, F.R. Zhu, Org. Electron 25 (2015) 266. [6] W.X. Lan, Y.X. Cui, Q.Y. Yang, M.F. Lo, C.S. Lee, F.R. Zhu, AIP Adv. 5 (2015) 057164. [7] J.Y. Kim, S.H. Kim, H.H. Lee, K. Lee, W. Ma, X. Gong, A.J. Heeger, Adv. Mater 18
(2006) 572. [8] A. Roy, S.H. Park, S. Cowan, M.H. Tong, S. Cho, K. Lee, A.J. Heeger, Appl. Phys. Lett. 95 (2009) 013302. [9] J. Gilot, I. Barbu, M.M. Wienk, R.A.J. Janssen, Appl. Phys. Lett. 91 (2007) 113520. [10] Y. Zhang, A.K. Pandey, C. Tao, Y. Fang, H. Jin, P.L. Burn, P. Meredith, Appl. Phys. Lett. 102 (2013) 013302. [11] A.K.K. Kyaw, D.H. Wang, D. Wynands, J. Zhang, T.Q. Nguyen, G.C. Bazan, A.J. Heeger, Nano Lett. 13 (2013) 3796. , O. Ingana €s, Appl. Phys. Lett. 94 (2009) [12] B.V. Andersson, D.M. Huang, A.J. Moule 043302. [13] B. Wu, Z.H. Wu, H.L. Tam, F.R. Zhu, Appl. Phys. Lett. 105 (2014) 103302. [14] W. Hermann-Jens, United States Patents 2004, 6710091. [15] F.R. Zhu, P. Jennings, J. Cornish, G. Hefter, K. Luczak, Sol. Energy Mater. Sol. Cells 49 (1997) 163. [16] F.R. Zhu, T. Fuyuki, H. Matsunami, J. Singh, Sol. Energy Mater. Sol. Cells 39 (1995) 1. [17] H.X. Liu, Z.H. Wu, J.Q. Hu, Q.L. Song, B. Wu, H.L. Tam, Q.Y. Yang, W.H. Choi, F.R. Zhu, Appl. Phys. Lett. 103 (2013) 043309. [18] V.D. Mihailetchi, L.J.A. Koster, J.C. Hummelen, P.W.M. Blom, Phys. Rev. Lett. 93 (2004) 216601. [19] D. Gebeyehu, C. Brabec, F. Padinger, T. Fromherz, J. Hummelen, D. Badt, H. Schindlerd, N.S. Sariciftci, Synth. Met. 118 (2001) 1. [20] D. Gebeyehu, M. Pfeiffer, B. Maennig, J. Drechsel, A. Werner, K. Leo, Thin Solid Films 451e452 (2004) 29.
Contents lists available at ScienceDirect
Organic Electronics journal homepage: www.elsevier.com/locate/orgel
An insight on oxide interlayer in organic solar cells: From light absorption and charge collection perspectives Zhenghui Wu, Bo Wu, Hoi Lam Tam, Furong Zhu* Department of Physics, Institute of Advanced Materials, and Institute of Research and Continuing Education (Shenzhen), Hong Kong Baptist University, Kowloon Tong, NT, Hong Kong
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 December 2015 Received in revised form 24 January 2016 Accepted 25 January 2016 Available online xxx
A comprehensive study of the effect of oxide interlayer on the performance of bulk-heterojunction organic solar cells (OSCs), based on poly[[4,8-bis[(2-ethylhexyl)oxy] benzo [1,2-b:4,5-b'] dithiophene2,6- diyl] [3-fluoro-2-[(2-ethylhexyl)carbonyl] thieno [3,4-b] thiophenediyl]] (PTB7): [6,6]-phenyl C71 butyric acid methyl ester (PC70BM) blend system, is carried out by optical simulation, interfacial exciton dissociation and charge collection analyses. It is found that a PTB7:PC70BM blend layer thickness optimized for maximum light absorption in OSCs does not generally give rise to the highest power conversion efficiency (PCE). OSCs, e.g., based on PTB7:PC70BM blend system, can benefit from the oxide interlayer in two ways, (1) to enhance the built-in potential for reducing recombination loss of the photogenerated charges, and (2) to improve charge collection by removal of unfavorable interfacial exciton dissociation. The combined effects result in ~20% improvement in PCE over an optimized control cell, having an identical layer configuration without an oxide interlayer. © 2016 Elsevier B.V. All rights reserved.
Keywords: Transient photocurrent Oxide/organic hetero-interface Charge collection efficiency Interfacial exciton dissociation Absorption enhancement
1. Introduction Power conversion efficiency (PCE) of >10% for both single junction and tandem organic solar cells (OSCs) has been demonstrated recently [1,2] through continued progresses made in both new material development and device innovation. However, the performance of OSCs still faces a major challenge to overcome the mismatch between optical absorption length and charge transport scale, caused by the low charge mobility in organic materials. High performing OSCs require a balance between charge transport and optical absorption in selecting the most suitable active layer in a narrow thickness range from 75 to 300 nm. It has been revealed that the unbalanced charge mobility in the photoactive layer, due to oxygen-induced charge traps, is one of the degradation mechanisms [3]. The degradation caused by the interfacial passivation could be avoided by the removal of the low work function cathode [4]. Different approaches have been proposed to enhance the performance of OSCs, e.g., optimizing the nanoscale morphology of organic blend layer to benefit the charge transport [5], incorporating photonic structures in the active layer to achieve
* Corresponding author. E-mail address: [email protected] (F.R. Zhu). http://dx.doi.org/10.1016/j.orgel.2016.01.040 1566-1199/© 2016 Elsevier B.V. All rights reserved.
broadband absorption enhancement via the optical coupling effect [6], and interposing an oxide optical spacer between the active layer and the electrode [7e10]. The application of metal oxides in OSCs and the impact of the properties of metal oxide/organic hetero-interfaces on cell performance have attracted a lot of attentions. For example, the use of a titanium suboxide [8] or ZnO interlayer [9] between the active layer and the reflective electrode in OSCs has been shown to increase the absorption in the active layer. The enhancement in the photocurrent generation due to the ZnO spacer is found to be more beneficial in OSCs with a thin active layer (<60 nm), but has less effect in device with a thicker layer (>90 nm) [10]. The ZnO layer also serves as a hole blocking layer to reduce the recombination rate [11]. Despite numerous reports about the improvement on OSCs due to the incorporation of the oxide spacer, mechanisms of improving cell performance due to oxide interlayer in OSCs are still inconclusive and controversial. €s et al. [12] reported that no optical beneficial effect can be Ingana expected by interposing an optical spacer layer in OSCs with an already optimized active layer thickness. In a recent work, we found that the insertion of a ZnO interlayer between the active layer and the metal contact enables to eliminate the unfavorable exciton dissociation that would otherwise occur at the organic/ metal interface, thereby improving the charge collection efficiency [13]. This suggests that the removal of the unfavorable interfacial
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
exciton dissociation is a perquisite for high performing OSCs. In addition to the absorption enhancement, it is also important to understand the effect of oxide interlayer on charge collection properties at the organic/cathode interface. The organic/oxide hetero-interfaces have a crucial impact on interfacial exciton dynamics, interfacial charge trapping behavior and the overall device performance. The complexity of physics operating in these OSCs offers a major driving force for new developments. In this work, a systematic study on the effect of ZnO interlayer on OSCs, based on poly[[4,8-bis[(2-ethylhexyl)oxy] benzo [1,2-b:4,5-b'] dithiophene2,6- diyl] [3-fluoro-2-[(2-ethylhexyl)carbonyl] thieno [3,4-b] thiophenediyl]] (PTB7): [6,6]-phenyl C71 butyric acid methyl ester (PC70BM) blend system, was carried out. The absorption enhancement in OSCs due to the presence of a ZnO interlayer between the organic active layer and metal contact is analyzed using optical admittance analysis. Effect of metal oxide/organic hetero-interfaces on the charge recombination and exciton dissociation dynamics at organic/electrode interface in OSCs was analyzed using light intensity dependent current densityevoltage (JeV) characteristics and transient photocurrent (TPC) measurements. Our results showed a 20% power conversion efficiency (PCE) improvement in OSCs, realizing through a simultaneous enhancement in light absorption and effective charge collection.
mechanisms of ZnO-induced performance enhancement of OSCs. From optical point of view, a flat OSC can be considered as a thin film system consisting of m layers. The optical electric field distribution in the cells can be optimized optically with an appropriate choice of the active layer thickness or by insertion of an oxide interlayer between the bulk heterojunction and the metal contact. Under normal incidence, the absorption spectrum of the ith layer, for example the active layer Ai(l), in an m layered thin film solar cell can be calculated by optical admittance method [15,16]:
Ai ðlÞ ¼ ½1 RðlÞ½1 ji ðlÞ
ITO/glass substrates with a sheet resistance of 10 U/square were used for the OSC fabrication. They were cleaned in ultrasonic bath sequentially with detergent, deionized water, acetone and isopropanol for 20 min each, followed by the in situ oxygen plasma treatment prior to the cell fabrication. The PTB7 (1 Material):PC70BM (Nano C) blended in a weight ratio of 1:1.5 was fully dissolved in chlorobenzene (CB) (SigmaeAldrich, 99.8%) with 3% 1, 8-Diiodooctane (DIO) (SigmaeAldrich) at 60OC before use. A set of structurally identical OSCs with different active layer thicknesses ranging from 60 to 130 nm was then fabricated on PEDOT:PSScovered ITO/glass substrates by spin-coating inside a N2-purged glove-box with O2 and H2O levels <0.1 ppm. ZnO NPs were synthesized following the procedure described by Hermann-Jens [14]. A 10 nm thick ZnO interlayer was deposited on the active layer by spin-coating inside the glove-box. The samples were then transferred to the adjacent vacuum system with a base pressure of <5.0 104 Pa for forming a 100 nm thick Al top contact, evaporated at a rate of 1 Å/s. OSCs were transferred back to the glove-box for JV characteristic measurement under an air mass (AM) 1.5G irradiation of 100 mW/cm2. The intensity of the solar simulator was calibrated using a silicon reference cell with a KG-5 filter. The transient photocurrent in the devices, e.g., ITO/PEDOT:PSS/ PC70BM (400 nm)/ZnO(10 nm)Al (30 nm) and a control sample of ITO/PEDOT:PSS/PC70BM (400 nm)/Al (30 nm), was generated using a pulsed Nd:YAG laser with the wavelength of 355 nm and pulse duration of <5 ns, illuminated from a semitransparent top Al cathode side. The devices were connected to the 1 MU input terminal of an oscilloscope (Agilent DSO8064A Infiniium Digital Oscilloscope) to measure the transient photovoltage. The corresponding transient photocurrent of the devices was obtained by converting the transient photovoltage to the current using the internal resistor of the functional generator (50 U) at different biases. 3. Results and discussions It is known that the performance of OSCs is dependent on the intensity and incident angle of solar irradiation, varied due to sunrise to sunset during the day. In the simulation, we considered incoming light at normal to the cell to unraveling the fundamental
i1 Y
jj ðlÞ;
(1)
j¼1
where jj is the ratio of the time averaged numerical magnitude of the Poynting's vector at the jth and (j-1)th boundaries, R(l) is the wavelength-dependent reflectance of the cell. Using the flux of the incident solar radiation F(l), e.g., AM1.5G measured in the units of W/m2,nm, the integrated absorbance of the active layer, defined as Ai which is proportional to the short circuit current density (JSC) of the cell, can be calculated as [15,16]:
Z Ai ¼
2. Experimental section
267
Ai ðlÞ FðlÞ dl Z : FðlÞ dl
(2)
Equation (2) allows evaluating the optimal active layer, e.g., PTB7:PC70BM, thickness in OSCs through maximizing its integrated absorptance. Essentially, the optical phenomena in thin film OSCs have a profound impact on light absorption in the active layer. This procedure takes into account the interference effects and allows us to optimize light absorption enhancement in the active layer, thereby improving JSC and PCE. Light absorption in different types of OSCs under normal incidence was calculated in the wavelength range from 380 to 780 nm. The thickness of the organic photoactive layer was varied over the range from 40 to 240 nm. This region is comparable in thickness to the active layer in OSCs made with different organic photoactive materials. To understand the effect of the ZnO interlayer on absorption enhancement in the OSCs, the integrated absorbance of the active layer in OSCs of ITO/PEDOT:PSS (40 nm)/PTB7:PC70BM/Al (100 nm) (control cell) and that of the structurally identical cells with a ZnO interlayer inserted between the active layer and the metal contact: ITO/PEDOT:PSS (40 nm)/PTB7:PC70BM/ZnO/Al (100 nm), was calculated using eqs. (1) and (2). The wavelength dependent refractive indices of each layer in the OSC system were measured using variable angle spectroscopic ellipsometer [17]. The integrated absorptance of the PTB7:PC70BM layer, calculated for the control OSC and the structurally identical cells with different ZnO interlayer thicknesses (10 nm, 20 nm and 30 nm), as a function of its layer thickness ranging from 40 to 240 nm is shown in Fig. 1. It is clear that the integrated absorbance displays an oscillation behavior with increase in the active layer thickness, having a relative absorption maximum occurred at the blend layer thickness of ~105 nm for a control OSC. However, the thickness of the PTB7:PC70BM layer corresponding to the absorption maximum in the integrated absorptance becomes thinner, e.g., ~90 nm for OSC with a 10 nm thick ZnO interlayer. A slight decrease in the integrated absorbance is observed when the thickness of the ZnO interlayer further increases from 10 to 30 nm. This suggests that the insertion of an oxide interlayer between the organic active layer and the metal contact does not have an obvious contribution to the absorption enhancement in active layer as compared to that in an optically optimized control cell. Although a thin oxide interlayer alone cannot account for light
268
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
absorption enhancement in the already optically optimized control cell, OSCs with an optimized ZnO interlayer is still beneficial for efficient device operation. The results in Fig. 1 reveal that OSCs with an appropriate ZnO interlayer (~10 nm thick) permits a thinner active layer (~90 nm) to achieve the similar maximum light absorption as compared to an optimized thicker active layer (~105 nm) used in the control cell. The inset in Fig. 1 is the crosssectional view of the OSC. It is known that light absorption in OSCs is limited due to the presence of a mismatch between optical absorption length and charge transport scale, caused by the low charge mobility in conjugated polymers. The use of a thinner photoactive layer without an obvious loss in light absorption, enabled in OSCs with a ZnO interlayer, as shown in Fig. 1, helps to create a higher internal electric field across the active layer, thereby reducing the exciton recombination losses and improving the charge transport. It shows that ZnO interlayer is advantageous for the efficient operation of the PTB7:PC70BM-based OSCs. The presence of the ZnO interlayer allows using a thinner active layer without moderating the absorption in the optimized control cells with a thicker active layer. A combination of the efficient absorption and higher internal potential favors the efficient charge transport, and thereby improving the cell performance. The relevance of the above discussion in the absorption in the active layer can be seen by measuring the performance of the control cell and the OSCs with a thin ZnO interlayer. Two sets of structurally identical OSCs, ITO/PEDOT:PSS (30 nm)/PTB7:PC70BM/ Al(100 nm) and ITO/PEDOT:PSS (30 nm)/PTB7:PC70BM/ZnO (~10 nm)/Al(100 nm), with different active thicknesses ranging from 55 to 130 nm, were made for comparison studies. The short circuit current density of both sets of the OSCs as a function of the active layer thickness was measured and normalized for comparison and then plotted in Fig. 2. The correlation between the JSC and the active layer thickness, measured for both types of the OSCs, is very similar to the simulation results. For the control cells, JSC reaches to the relative maximum value at about 105 nm thick active layer, while for OSCs with a 10 nm thick ZnO interlayer, JSC reaches to the relative maximum value at a thinner active layer of ~90 nm, following with a graduate decay, which is consistent with the simulation results. The optical admittance analysis agrees well with the experimental results in showing that OSCs having an optimal ZnO interlayer interposed between the 90 nm thick active layer and
Fig. 1. Calculated integrated absorbance of the active layer in control cell and OSCs with different ZnO interlayer thicknesses of 10 nm, 20 nm and 30 nm. Inset in Fig. 1 shows the cross sectional view of the OSC with a ZnO interlayer added between the active layer and the metal contact.
the Al contact corresponded to the best cell performance, as presented in Fig. 2. It is found that the highest PCE of the control cells was obtained for the cell having a ~90 nm active layer. It is clear that a PTB7:PC70BM blend layer thickness optimized for maximum light absorption (~105 nm) in the control OSCs, e.g., achieving highest JSC, does not give rise to the highest PCE. Such a deviation in the optical optimization and device performance essentially reflects the mismatch between the optical absorption length and charge transport scale in the organic semiconductors. Two types of OSCs were fabricated: one was the control cell optimized without the use of an oxide interlayer, the other was the cell optimized with a ZnO (10 nm) interlayer. JeV characteristics of both cells are plotted in Fig. 3a, the open-circuit voltage (VOC), JSC and fill factor (FF) of the OSC with a ZnO interlayer between the active layer and the Al cathode increased from 0.73 to 0.74 V, 13.7e14.5 mA/cm2 and 0.65 to 0.73, respectively, leading to an overall 20% increase in PCE from 6.51% to 7.82%. The steady improvements in JSC and VOC, seen in Fig. 3a, are the factors contributing to the enhancement in overall cell performance. This suggests that light absorption and transport of the photo-generated carriers are more favorable in OSCs with a ZnO interlayer, although a slight thinner photoactive layer thickness was used in the cells. A summary of device parameters obtained for both types of the OSCs is listed in Table 1. Incident photon to current efficiency (IPCE) spectra measured for a control cell and the OSC with a 10 nm thick ZnO interlayer between the active layer and the Al contact are plotted in Fig. 3b. IPCE characteristics provide the information of light absorption in the active region contributing to exciton generation and the overall external quantum efficiency of the OSCs. Light absorption in the
Fig. 2. Comparison of normalized PCE and JSC, measured for a set of (a) control cells and (b) structurally identical OSCs with a ~10 nm thick ZnO interlayer, as a function of the active layer thickness over the thickness range from 40 to 130 nm.
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
experimentally optimized control cell with a ~90 nm thick active layer is lower than that in the optimized cell with the same active layer thickness having a ~10 nm thick ZnO interlayer. It is apparent that cells with a ZnO interlayer apparently possess a better interfacial contact at the organic/cathode interface, therefore an efficient charge collection and a higher drift current are expected. The change in the IPCE spectra at the wavelength <500 nm and the wavelength region from 500 to 750 nm, as shown in Fig. 3b, is due to the interference effect in the OSCs [17]. There is a slight decrease in the IPCE measured for the OSCs with a ZnO interlayer at the wavelength <500 nm, and an obvious increase in IPCE over a spectral region from 500 to 750 nm. PTB7 does not have a strong absorption below 500 nm. Incorporation of a thin ZnO interlayer in the cells still benefit from an overall absorption enhancement over the wavelength region from 500 to 750 nm, attaining a ~6% increase in JSC as compared to the control cell. JSC calculated using IPCE spectra of the control cell and the OSCs with a ZnO interlayer also agrees with measured JeV characteristics, showing that the improvement in performance of the OSCs with a ZnO interlayer contributes the improvement in JSC, as well as VOC and FF illustrated in Fig. 4. The results in Fig. 4 reveal that VOC and FF decrease with increase in the thickness of the active layer in the cells. The behavior of the active layer thickness-dependent VOC and FF in OSCs originates from the limited charge mobility in the organic
Fig. 3. (a) The JV characteristics and (b) IPCE spectra measured of control cell and OSC with a 10 nm thick ZnO interlayer between the active layer and the Al contact.
269
Table 1 The characteristics of optimized devices with/without the ZnO interlayer. The values of JSC and PCE calibrated using IPCE measurements are also given in the parentheses for comparison. OSCs
VOC (V)
JSC (mA/cm2)
FF (%)
PCE (%)
Control cell Cells with a ZnO interlayer
0.73 0.74
13.3 (13.7) 14.0 (14.5)
65.1 72.9
6.32 (6.51) 7.55 (7.82)
semiconductors. It is shown clearly that that VOC and FF become less sensitive to the change in the active layer thickness when a ZnO interlayer was inserted between the active layer and Al contact. The charge collection properties at the organic/cathode interface, caused by the insertion of an oxide interlayer between the active layer and the metal contact, also play an important role in determining the performance of the cells. To further understand the origin of the enhancement in the performance of the OSCs with a ZnO interlayer as compared to a control cell, charge recombination characteristics in the OSCs were analyzed. Fig. 5 shows the double logarithmic plot of photocurrent Jph (Jph ¼ JL e JD, where JL is the current density measured under AM1.5G irradiation of 100 mW/cm2 and JD is the dark current) as a function of the effective bias Veff (Veff ¼ V0 eVb, where V0 is the built-in potential measured at Jph ¼ 0, and Vb is the applied bias), measured for the OSCs with a 10 nm thick ZnO interlayer and a control cell. In this work, V0 of both types of OSCs is approximately 0.8 V. The saturated photocurrents Jsat of both types of the OSCs are also shown in Fig. 5. Jsat in the OSCs depends only on the charge generation as Jph saturates at high Veff (>1.0 V), giving rise to almost 100% collection of
Fig. 4. Comparison of (a) VOC and (b) FF measured for control cells and OSCs having a ~10 nm thick ZnO buffer as a function of the active layer thickness.
270
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
the photo-generated charges. The ratio of Jph to Jsat reflects essentially a measure of an overall loss in the photo-excited charges, e.g., due to bimolecular recombination, trap-assisted defect states, occurred during transport processes prior to the collection by the electrodes [18]. At low Veff region, the number of photo-generated carriers that can be extracted decreases due to the increase in the charge recombination, leading to a poor charge collection. As shown in Fig. 5, Jph obtained for a control cell decreases much faster than that in the OSC with a ZnO interlayer at Veff <0.5 V. The loss due to the recombination of photo-generated carrier at the organic/cathode interface is obviously much higher in the control cell as compared to that in the OSCs with a ZnO interlayer at Veff <0.5 V. Both types of OSCs have an identical anode contact and active layer configuration, the change in the JphVeff characteristics at low Veff region is apparently associated with the change in the interfacial properties at the organic/cathode contact. This reveals that presence of a thin ZnO interlayer between the active layer and the Al cathode augments the photocurrent in this Veff region, revealing improvement in charge collection efficiency at the cathode contact. Dissociation of the photo-excited excitons occurs at the interface between the two dissimilar materials, e.g., organic/electrode and donor/acceptor interfaces, the later process dominates the photocurrent generation in the OSCs. The charges generated due to the exciton dissociation at the donor/acceptor interface are drifted under the internal built-in potential, forming the field-dependent drift current in the cells. The transient photocurrent in OSCs comprises (1) the drift current formed in the active layer, and (2) the contribution from the interfacial exciton dissociations occurred at the organic/electrode interfaces [13]. The drift current is a function of the internal electric field, while the current generated via interfacial exciton dissociation is independent on the internal electric field. Contributions to the transient photocurrent due to the drift current and the interfacial exciton dissociation cannot be decoupled in the JV characteristics of the cells measured under the steady state condition. However, the behavior of exciton dissociation at the organic/electrode interfaces and its contribution to the transient current can be analyzed by applying an opposite external bias to suppress the drift current in the transient photocurrent measurements. It then becomes possible to exam the interfacial exciton dissociation in the absence of internal built-in potential. Therefore, the correlation between the interfacial exciton dynamics and charge collection at organic/electrode interface can be obtained by measuring transient current for devices at different biases. In this work, charge collection and the interfacial
Fig. 5. Photocurrent density Jph as a function of the effective bias Veff across the active layer measured for a control cell and the OSC with a 10 nm thick ZnO interlayer.
exciton dissociation processes at the organic/cathode interface in the cells were investigated using TPC measurements. TPC measurement is a technique to study the transient dynamics of the photo-generated carriers with a time scale of ~10 ns in electronic devices. The details of the TPC measurement are described in our previous work [13]. Before the exposure of the devices to light, a dark transient current (ID) was recorded as the background current. When the active layer was excited by the laser, the transient current (IL) was recorded again. Then transient photocurrent (ITPC) of the device can be calculated by:
ITPC ¼ IL ID :
(3)
To analysis the result of the interfacial exciton dissociation on charge collection at the organic/Al cathode interface, two devices with layered structures of ITO/PEDOT:PSS/PC70BM(400 nm)/ Al(30 nm) (control device) and ITO/PEDOT:PSS/PC70BM(400 nm)/ ZnO(~10 nm)/Al (30 nm) were made for TPC measurements. The effect of ZnO/Al bilayer on exciton dissociation at the oxide/electrode interface can be unraveled by probing its contribution to the transient photocurrent, measured by suppression of the fielddrifted current in the bulk, e.g., applying a reverse bias to compensate the built-in potential in the device. As ZnO has a lower work function compared to that of the Al cathode, the use of the ZnO/Al bilayer contact results in an increase in the built-in potential across the active region of the devices with a ZnO interlayer, e.g., increasing from ~0.5 V (Al contact) to ~0.8 V (ZnO/Al contact). Therefore, a slight higher compensation voltage of ~0.9 V was used in the transient photocurrent measurement, to suppress the fielddependent photocurrent generated in the samples with a ZnO interlayer. The transient photocurrents measured for both devices at different biases are plotted in Fig. 6a and b. The positive transient photocurrent decays were observed for both devices without the external bias, suggesting that the electric field drifted currents are dominated in both devices. However, when the TPC measurements were performed for both devices with an opposite external bias, very different transient photocurrent behaviors were observed. A clear rapid negative transient photocurrent signal for the control device was observed at a reverse bias of 0.4 V, shown in Fig. 6a. The prompt negative transient photocurrent, originated from the fast interfacial exciton dissociation at the PC70BM/Al interface, is apparently unfavorable for the electron collection. In comparison with the flow of the transient photocurrent generated at the PC70BM/Al interface shown in Fig. 6a, there was an obvious positive transient photocurrent seen in the device with a presence of a 10 nm thick ZnO interlayer between active layer and Al contact at a reverse bias of 0.9 V, shown in Fig. 6b. It becomes clear that the interfacial exciton dissociation at the organic/Al interface in the control device impedes the electron collection. In contrast, the presence of a 10 nm thick ZnO interlayer between the active layer and the metal cathode in the OSCs promotes an efficient charge collection through elimination of the adverse interfacial exciton dissociation process occurred in the ones without a ZnO interlayer. The TPC results agree well with the analysis based on the JpheVeff characteristics of the cells shown in Fig. 5. It is clear that the enhanced Jph, observed in the OSCs with a ZnO interlayer at the low Veff, arises from the favorable interfacial exciton dissociation behavior that inhibits the interfacial bimolecular recombination. The difference in charge recombination and extraction in both types of the OSCs, caused by the adverse exciton dissociation processes at the organic/cathode interface, can be further examined by analyzing the JV characteristics of the OSCs under different light intensities. The empirical relationship between Jph and the intensity of light (I) is [19,20]:
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
271
variation in the exponent of JphI characteristics reflects the difference in charge accumulation at the organic/cathode interface in both cells. The charge accumulation at the organic/cathode in the control cell becomes more notable as compared to the OSC having a ZnO interlayer. A relatively high charge recombination in the control cell is then expected. It is seen that charge accumulation at the organic/cathode is much stronger in the control cell than that in the OSC with a ZnO interlayer. A structurally identical OSC with a thin ZnO interlayer apparently favors an efficient charge collection process, thereby yielding a high PCE of 7.82%, which is about 20% higher than that of a control cell (6.51%). 4. Conclusion Functionalized metal oxide/organic hetero-interfaces have a crucial impact on the optical, interfacial exciton dynamics, interfacial charge trapping behavior of the overall performance of organic solar cells. The presence of a thin ZnO interlayer layer between the photoactive layer and the cathode is favorable for efficient operation of OSCs. The ZnO interlayer serves a dual purpose to enable a thinner active layer without moderating the absorption via interference effect as well as to augment charge collection efficiency via favorable interfacial exciton dissociation at the organic/ cathode interface. This simple yet effective implementation can improve the device performance with the PCE of PTB7:PC70BMbased OSCs by ~20%.
Fig. 6. Transient photocurrent signals, due to the interfacial exciton dissociation at (a) PC70BM/Al and (b) PC70BM/ZnO/Al interfaces.
Jph fI a ;
(4)
where the exponent a in eq. (4) is usually less than one. Recombination of the photo-generated charges in the device generally results in a deviating from one [18]. With increase in the intensity of light, charge accumulation in the active layer of an OSC will be quickly saturated due to the limited trap states. Therefore, less “bright” light irradiation can be used to analyze the charge accumulation processes in the OSCs. The double logarithmic plot of JphI characteristics, measured for the OSC with a 10 nm thick ZnO interlayer and the control cell, are shown in Fig. 7a and b. The power law dependence of Jph on light intensity is clearly seen. The power exponents of JphI characteristics measured for both cells approach to one at Veff of >0.25 V. This suggests that most photo-generated charges can be collected at the organic/electrode contacts without charge accumulation in the cell at high Veff. As the charge mobility in organic materials is field dependent, less efficient charge transport and collection in the OSCs would occur at the low Veff, leading to the build-up of space charges that gives rise to higher bimolecular recombination, and thereby a lower power exponent in the JphI characteristics. As shown in Fig. 7a, the exponent of JphI characteristics observed for the control cell decreases to 0.67 at Veff ¼ 0.09 V. A higher exponent value of 0.87 is obtained for OSC with a ZnO interlayer under the same effective voltage. As both cells have the same anode/organic contact and blend system, it can be considered that the charge accumulation behaviors at the anode contact and in the active layer of both cells are very similar. The
Fig. 7. Double logarithmic plot of photocurrent density as a function of light intensity for (a) a control cell and (b) the OSC with a 10 nm thick ZnO interlayer, measured under different effective voltages.
272
Z. Wu et al. / Organic Electronics 31 (2016) 266e272
Acknowledgments This work was supported by Research Grants Council of Hong Kong Special Administrative Region, China, Project No. T23-713/11, GRF12303114, National Natural Science Foundation of China (No. 61275037), Shenzhen Peacock Project, KQTD20140630110339343 and Hong Kong Baptist University FRG2/14-15/081. References [1] C.C. Chen, W.H. Chang, K. Yoshimura, K. Ohya, J. You, J. Gao, Z. Hong, Y. Yang, Adv. Mater 26 (2014) 5670. [2] Y. Liu, J. Zhao, Z. Li, C. Mu, W. Ma, H. Hu, K. Jiang, H. Lin, H. Ade, H. Yan, Nat. Commun. 5 (2014) 5293. [3] C. Vijila, G.M. Ng, Mein Jin Tan, W.P. Goh, F.R. Zhu, Appl. Phys. Lett. 95 (2009) 263305. [4] X.Z. Wang, X.X. Zhao, G. Xu, Z.K. Chen, F.R. Zhu, Sol. Energy Mater. Sol. Cells 104 (2012) 1. [5] W.L. Xu, B. Wu, F. Zheng, H.B. Wang, Y.Z. Wang, F.G. Bian, X.T. Hao, F.R. Zhu, Org. Electron 25 (2015) 266. [6] W.X. Lan, Y.X. Cui, Q.Y. Yang, M.F. Lo, C.S. Lee, F.R. Zhu, AIP Adv. 5 (2015) 057164. [7] J.Y. Kim, S.H. Kim, H.H. Lee, K. Lee, W. Ma, X. Gong, A.J. Heeger, Adv. Mater 18
(2006) 572. [8] A. Roy, S.H. Park, S. Cowan, M.H. Tong, S. Cho, K. Lee, A.J. Heeger, Appl. Phys. Lett. 95 (2009) 013302. [9] J. Gilot, I. Barbu, M.M. Wienk, R.A.J. Janssen, Appl. Phys. Lett. 91 (2007) 113520. [10] Y. Zhang, A.K. Pandey, C. Tao, Y. Fang, H. Jin, P.L. Burn, P. Meredith, Appl. Phys. Lett. 102 (2013) 013302. [11] A.K.K. Kyaw, D.H. Wang, D. Wynands, J. Zhang, T.Q. Nguyen, G.C. Bazan, A.J. Heeger, Nano Lett. 13 (2013) 3796. , O. Ingana €s, Appl. Phys. Lett. 94 (2009) [12] B.V. Andersson, D.M. Huang, A.J. Moule 043302. [13] B. Wu, Z.H. Wu, H.L. Tam, F.R. Zhu, Appl. Phys. Lett. 105 (2014) 103302. [14] W. Hermann-Jens, United States Patents 2004, 6710091. [15] F.R. Zhu, P. Jennings, J. Cornish, G. Hefter, K. Luczak, Sol. Energy Mater. Sol. Cells 49 (1997) 163. [16] F.R. Zhu, T. Fuyuki, H. Matsunami, J. Singh, Sol. Energy Mater. Sol. Cells 39 (1995) 1. [17] H.X. Liu, Z.H. Wu, J.Q. Hu, Q.L. Song, B. Wu, H.L. Tam, Q.Y. Yang, W.H. Choi, F.R. Zhu, Appl. Phys. Lett. 103 (2013) 043309. [18] V.D. Mihailetchi, L.J.A. Koster, J.C. Hummelen, P.W.M. Blom, Phys. Rev. Lett. 93 (2004) 216601. [19] D. Gebeyehu, C. Brabec, F. Padinger, T. Fromherz, J. Hummelen, D. Badt, H. Schindlerd, N.S. Sariciftci, Synth. Met. 118 (2001) 1. [20] D. Gebeyehu, M. Pfeiffer, B. Maennig, J. Drechsel, A. Werner, K. Leo, Thin Solid Films 451e452 (2004) 29.