Free Access
Issue
Eur. Phys. J. Appl. Phys.
Volume 89, Number 3, March 2020
Article Number 30201
Number of page(s) 6
Section Physics of Organic Materials and Devices
DOI https://doi.org/10.1051/epjap/2020190322
Published online 06 May 2020

© EDP Sciences, 2020

1 Introduction

Organic solar cells (OSCs) have attracted tremendous attention during the past three decades [117], since they have been widely considered as a sort of low-cost, flexible, environmentally friendly, and renewable energy source. The best power conversion efficiencies (PCE) of OSCs have achieved 16–17% [3,4], well matching the level of their counterpart amorphous silicon solar cells. Nevertheless, in order to surpass the commercial benchmark of 20%, more efforts are needed to develop novel organic electron donors and acceptors and to better understand the device working mechanisms, e.g., precisely determining the exciton diffusion length (LD) of organic electron donors under the operating condition of device.

The optical absorption of organic donor contributes to the coverage of OSCs over the solar spectrum [3,4]. In order to maximize the light-utilizing efficiency, most of the excitons photogenerated in organic donor layer should diffuse towards the donor/acceptor interface and then get dissociated into electrons and holes there prior to the geminate recombination. As a result, the exciton diffusion in organic donor, characterized by LD, is a crucial process in governing the performance of OSCs. In planar-heterojunction OSCs, the LD of organic donor sets the bottom limit to the optimal thickness of organic donor layer. The longer LD, the greater optimal thickness of organic donor layer is allowed, increasing optical absorption and thereby making higher PCE. In bulk-heterojunction OSCs, the LD of organic donor lays down the top limit on the size of separated organic donor phase. The longer LD, the larger phase separation of donor and acceptor is favored, resulting in more efficient charge transport and thereby higher PCE. Thus, the acquisition of long-LD organic donors is of great importance to achieve high-efficiency OSCs.

To date, there are various methods employed to determine the LD of organic materials, for example, the spectrally resolved photoluminescence quenching, only suitable to organic materials with appreciable photoluminescence [5,18], uniting theoretical calculation with experimentally measured external quantum efficiencies to determine LD, varying organic layer thickness and then utilizing some specific models to extract LD [57,11,19,20], etc. The relatively easy and direct approach suggested by Stübinger et al. is to monitor the photocurrent intensity as a function of photoactive layer thickness in a planar-heterojunction OSCs and then extract LD from the optimal thickness of photoactive layer at the strongest photocurrent response [19]. However, the complicated estimation of the optical interference effects is very often involved in the approach, leading to some uncertainty in determining LD [20]. Increasing the intensity of the light illumination is found to enlarge LD [14].

Here, the exciton diffusion of copper phthalocyanine (CuPc) is investigated in planar-heterojunction OSCs with a novel device geometry containing a 2 nm layer of bathocuproine (BCP) introduced into the CuPc layer. By combining simple optical simulation with the photovoltaic performance of devices, the LD for CuPc is determined to be 12.5–15 nm, comparable to the reported value of 10 ± 3 nm [5].

2 Experimental

Indium tin oxide (ITO) coated glass substrates were commercially purchased with a sheet resistance of 10 Ω per square. The CuPc and BCP were bought from Jilin OLED Materials Company Ltd and C60 was obtained from Sigma-Aldrich Company. All organic materials were used as received. ITO glasses were treated in UV-ozone for 15 min prior to being loaded into the vacuum chamber.

The background pressure of the device fabrication was 3 × 10−4 Pa. A series of devices with structure of ITO/ CuPc 20-D nm/BCP 2 nm/CuPc D nm/C60 40 nm/BCP 10 nm/Al (the S devices) were prepared, where D = 7.5, 10, 12.5, and 15 nm, respectively. A control device with structure of ITO/CuPc 20 nm/C60 40 nm/BCP 10 nm/Al (the C device) was fabricated at the same time. The active area of solar cell was controlled to be around 8.0 mm2 by patterning cathode deposition. The current density–voltage (J–V) characteristics of devices were recorded using a programmable Keithley 2400 sourcemeter under simulated 1 sun irradiation of air mass 1.5G (CXE-350 arc Xe lamp) in the air.

In order to explain the experimental results, the optical modeling for the S and C devices was carried out [13].

thumbnail Fig. 1

The J–V characteristics of the S devices with structure of ITO/CuPc 20-D nm/BCP 2 nm/CuPc D nm/C60 40 nm/BCP 10 nm/Al and the C device with structure of ITO/ CuPc 20 nm/C60 40 nm/BCP 10 nm/Al under the illumination.

3 Results and discussion

3.1 The influences of the thin BCP layer on device performance

The J–V characteristics of the S and C devices under the illumination are shown in Figure 1. The photovoltaic parameters of the S devices are compared in Figure 2.

The open-circuit voltage (VOC) of the S device varies obviously with D: it remains 0.51 V when D = 7.5 and 10 nm, and then decreases to 0.48 V at D = 12.5 nm and further to 0.45 V at D = 15 nm. All the S devices have higher VOC than the C device (0.43 V). The short-circuit current density (JSC) and PCE of the S device show similar dependences on D. With D increasing from 7.5 nm to 12.5 nm, both JSC and PCE decreased gradually: JSC decreases from 3.19 mA/cm2 at D = 7.5 nm to 2.52 mA/cm2 at D = 12.5 nm and PCE decreases from 0.51% at D = 7 nm to 0.33% at D = 12.5 nm; when D changing from 12.5 to 15 nm, JSC and PCE increase to 4.35 mA/cm2 and 0.80%, respectively. All the S devices show lower JSC and PCE than the C device (5.57 mA/cm2 and 1.14%). The S device with D = 15 nm possesses higher FF than the other three S devices, while all of them give decreased FF than the C device (47.7%).

thumbnail Fig. 2

Plots of the S devices' photovoltaic parameters, VOC (a), JSC and PCE (b), and FF (c), versus D.

thumbnail Fig. 3

Schematic of the S device used for optical simulation. The CuPc layer is divided into zones 1 and 2 by the 2 nm-thick BCP layer.

3.2 The roles of the 2 nm BCP layer in the CuPc layer

3.2.1 The effect of the 2 nm BCP on the VOC of the S device

As shown in Figure 3, the S device is equivalent to the series connection of the CuPc (zone 2)/BCP and CuPc (zone 1)/C60 heterojunctions, because the CuPc (zone 1) can function as an interconnection layer. The CuPc (zone 2)/BCP heterojunction offers some weak charge separation, since CuPC is a p-type conductor and BCP is an n-type conductor, thereby generating certain photovoltaic effect. Therefore, the VOC of the S device is higher than the C device. In general, the VOC is expressed as(1)

where n is the ideality factor, kB is Boltzmann constant, T is temperature, J0 is background recombination current. With D increasing, the JSC from the CuPc (zone 2)/BCP heterojunction decreases accordingly, due to the decreased optical absorption of CuPc (zone 2); thus, the VOC based on the CuPc (zone 2)/BCP heterojunction decreases according to equation (1), thereby explaining the variation of VOC with D as shown in Figure 2a. Note that, the weak photovoltaic performance of the BCP/CuPc (zone 1) heterojunction is prohibited by the built-in voltage and therefore has not been considered in the following discussion.

3.2.2 The effect of the 2 nm BCP layer on the hole transport and exciton diffusion in the CuPc layer

Although zones 1 and 2 are separated by the 2 nm layer of BCP with the highest occupied molecular orbital level 1.2 eV below that (–5.2 eV) of CuPc, the hole transport is unlikely interrupted, because the hole current can efficiently tunnel through this very thin BCP layer. This is supported by the observation that the JSC of the S device with D = 15 nm is about 0.83 times that of the C device. In other words, the weakly charge-separating CuPc (zone 2)/BCP heterojunction is not the major factor to limit the JSC of the S devices.

The singlet exciton diffusion in organic materials layer can be considered as a Forster-type energy transfer [21]. Thus, the singlet excitons in CuPc are likely to flow through the BCP via the long-range energy resonance, despite the large difference in the optical bandgaps between CuPc (1.7 eV) and BCP (3.5 eV). In the S device, the efficiency E of the CuPc singlet excitons diffusing through the 2 nm BCP layer can be estimated via(2)

where R0 is the Forester radius of CuPc (∼1.5 nm) [18], R equals 2 nm. As a result, the E is calculated to be about 0.15 based on equation (2), indicating that only a small part of the CuPc singlet excitons is able to go through the 2 nm BCP layer. Hence, the 2 nm BCP layer in the S device functions as a smart sieve to allow hole current to pass it through but block most of the singlet excitons photogenerated in zone 1.

3.3 The optical simulation

The modeling calculation of the optical field intensity in the S device is given in Figure 4. It can be seen that the maximum intensity of the optical field appears near the CuPc/C60 interface. Figure 5 compares the calculated exciton generation rates (G) of the S devices to that of the C device. It can be seen that each S device exhibits slight change in the exciton generation rate profile across the entire CuPc layer, compared to the C device. Obviously, the exciton generation rate profile in the S device is not sensitive to the variation of D.

The steady-state exciton concentration profile in zone 1 of the S device was simulated with an assumed LD of 13.5 nm via(3)

where P is the exciton density and τ is the lifetime (1.0 ns) [18]. Two boundary conditions for numerically solving equation (3) are adopted as follows. Firstly, P = 0 at the CuPc/C60 interface is employed, since the efficiency of the CuPc singlet exciton dissociation into hole and electron at the CuPc/C60 interface remains almost 100%. Secondly, P at the BCP/CuPc (zone 1) interface is defined as the ratio of the D to LD at D ≤ LD, and is approximated to be the ratio of (DLD) to LD when D > LD. Most of the excitons photogenerated in zone 1 tend to diffuse forward to the CuPc/C60 interface and then get dissociated into free holes and electrons there, generating efficient photocurrent; however, according to Figure 5, a certain number of excitons in zone 1 should have some tendency to diffuse backward and cause some unavoidable germination recombination, i.e., the loss of photocurrent, accounting for that planar-heterojunction OSCs are unlikely to have high internal quantum efficiencies [8]. Hence, the second assumed boundary condition is reasonable. Figure 6 shows the calculated steady-state excitons densities across zone 1 for the S devices.

thumbnail Fig. 4

The model calculation of the electric component of the optical field in the S device. The value of the optical intensity |E ∙ j(x)|2 is normalized to |E0|2. E0 is the amplitude of the incident plane wave with λ = ∼535 nm. See reference [13] for the modeling details.

thumbnail Fig. 5

The comparison of calculated exciton generation rate G(x) at λ = ∼535 nm between the S (blue curves) and control (red curves) devices. See reference [13] for the modeling details. Note that, the interruptions in the calculated exciton generation rate at the interfaces are attributed to the sudden changes of the relative permittivity between the two neighboring materials.

thumbnail Fig. 6

The calculated steady-state exciton distributions across zone 1 at the D = 7.5, 10, 12.5, and 15 nm, respectively, on the basis of the assumed LD of 13.5 nm. The CuPc/C60 interface locates at the thickness = 15 nm.

3.4 The determination of LD of CuPc

When D increases from 7.5 to 12.5 nm, the optical absorption of zone 1, which underlies the photocurrent generation in the S device, increases correspondingly; however, JSC of the S device decreases (Fig. 2b). Here, it is thought that JSC of the S device is subject to the hole-exciton scattering effect at the BCP/CuPc (zone 1) interface [13,22,23], which is dependent on the value of D.

As shown in Figure 6, with D increasing from 7.5 nm to 12.5 nm, P at the BCP/CuPc (zone 1) interface increases. As a result, the hole-exciton scattering effect becomes severer, reducing hole mean free path and thereby leading to a gradual decrease in JSC of the S device; with D rising to 15 nm, P at the BCP/ CuPc (zone 1) interface drops drastically and accordingly the hole-exciton scattering effect is much alleviated, in accordance with some recovery in JSC of the S device with D = 15 nm. Increased hole mean free path at D = 15 nm decreases the series resistance of device, thereby leading to a marked increase in FF (Fig. 2c). The variation of the simulated P at the BCP/CuPc (zone 1) interface in the S device favorably explains the experimental data. Thus, it is reasonable to conclude that the LD lies between 12.5 and 15 nm, quite close to that reported by Peumans [5].

4 Conclusions

Here, organic solar cells with a novel geometry have been fabricated to observe the exciton diffusion in organic donor CuPc under the device operating condition. By uniting the photovoltaic performance of devices with the optical modeling, the exciton diffusion length of CuPc is derived to be 12.5–15 nm. The current research provides some in-depth understanding on the working mechanism of organic solar cells.

Author contribution statement

Xi Guan fabricated and measured the devices. Shiyu Wang and Wenxing Liu performed the optical modeling. Dayan Ban supervised the optical modeling and co-wrote the manuscript. Dashan Qin inspired the work and co-wrote the manuscript.

Acknowledgments

The authors are thankful for the valuable discussion from Dr. Seyed Ghasem Razavipour.

References

  1. C.W. Tang, Appl. Phys. Lett. 48 , 183 (1986) [Google Scholar]
  2. G. Yu, J. Gao, J.C. Hummelen, F. Wudl, A.J. Heeger, Science 270, 1789 (1995) [Google Scholar]
  3. L. Meng, Y. Zhang, X. Wan, C. Li, X. Zhang, Y. Wang, X. Ke, Z. Xiao, L. Ding, R. Xia, H.L. Yip, Y. Cao, Y. Chen, Science 361, 1094 (2018) [Google Scholar]
  4. L. Hong, H. Yao, Z. Wu, Y. Cui, T. Zhang, Y. Xu, R. Yu, Q. Liao, B. Gao, K. Xian, H. Woo, Z. Ge, J. Hou, Adv. Mater. 31, 1903441 (2019) [Google Scholar]
  5. P. Peumans, A. Yakimov, S.R. Forrest, J. Appl. Phys. 93, 3693 (2003) [Google Scholar]
  6. S. Banerjee, A.P. Parhi, S.S.K. Iyer, S. Kumar, Appl. Phys. Lett. 94, 223303 (2009) [Google Scholar]
  7. D. Kurrle, J. Pflaum, Appl. Phys. Lett. 92, 133306 (2008) [Google Scholar]
  8. J.G. Xue, B.P. Rand, S. Uchida, S.R. Forrest, J. Appl. Phys. 98, 124903 (2005) [Google Scholar]
  9. O.V. Mikhnenko, P.W.M. Blom, T.Q. Nguyen, Energy Environ. Sci. 8, 1867 (2015) [Google Scholar]
  10. C. Yan, S. Barlow, Z. Wang, H. Yan, A.K.Y. Jen, S.R. Marder, X. Zhan, Nat. Rev. Mater. 3, 18003 (2018) [Google Scholar]
  11. L.A.A. Pettersson, L.S. Roman, O. Inganas, J. Appl. Phys. 86, 487 (1999) [Google Scholar]
  12. P. Peumans, S.R. Forrest, Appl. Phys. Lett. 79, 126 (2001) [Google Scholar]
  13. D. Qin, P. Gu, R.S. Dhar, S.G. Razavipour, D. Ban, Phys. Status Solidi A 208, 1967 (2011) [CrossRef] [Google Scholar]
  14. G. Xu, N. Lu, W. Wang, N. Gao, Z. Ji, L. Li, M. Liu, Org. Electron. 23, 53 (2015) [Google Scholar]
  15. P. Kumar, A. Gaur, J. Appl. Phys. 113, 094505 (2013) [Google Scholar]
  16. J. Chen, C. Cui, Y. Li, L. Zhou, Q. Ou, C. Li, Y. Li, J. Tang, Adv. Mater. 27, 1035 (2015) [CrossRef] [PubMed] [Google Scholar]
  17. N. Lu, J. Wang, D. Geng, L. Li, M. Liu, Org. Electron. 66, 163 (2019) [Google Scholar]
  18. R.R. Lunt, N.C. Giebink, A.A. Belak, J.B. Benziger, S.R. Forrest, J. Appl. Phys. 105, 053711 (2009) [Google Scholar]
  19. T. Stubinger, W. Brutting, J. Appl. Phys. 97, 3632 (2001) [Google Scholar]
  20. L.G. Yang, H.Z. Chen, M. Wang, Thin Solid Films 516, 7701 (2008) [Google Scholar]
  21. C. Madigan, V. Bulovic', Phys. Rev. Lett. 96, 046404 (2006) [CrossRef] [PubMed] [Google Scholar]
  22. V.S. Bagaev, V.V. Zaitsev, Y.V. Klevkov, V.S. Krivobok, Phys. Solid State 47, 1827 (2005) [CrossRef] [Google Scholar]
  23. Y.P. Feng, H.N. Spector, IEEE J. Quantum Electron. 24, 1656 (1998) [Google Scholar]

Cite this article as: Xi Guan, Shiyu Wang, Wenxing Liu, Dashan Qin, Dayan Ban, Determining the exciton diffusion length of copper phthalocyanine in operating planar-heterojunction organic solar cells, Eur. Phys. J. Appl. Phys. 89, 30201 (2020)

All Figures

thumbnail Fig. 1

The J–V characteristics of the S devices with structure of ITO/CuPc 20-D nm/BCP 2 nm/CuPc D nm/C60 40 nm/BCP 10 nm/Al and the C device with structure of ITO/ CuPc 20 nm/C60 40 nm/BCP 10 nm/Al under the illumination.

In the text
thumbnail Fig. 2

Plots of the S devices' photovoltaic parameters, VOC (a), JSC and PCE (b), and FF (c), versus D.

In the text
thumbnail Fig. 3

Schematic of the S device used for optical simulation. The CuPc layer is divided into zones 1 and 2 by the 2 nm-thick BCP layer.

In the text
thumbnail Fig. 4

The model calculation of the electric component of the optical field in the S device. The value of the optical intensity |E ∙ j(x)|2 is normalized to |E0|2. E0 is the amplitude of the incident plane wave with λ = ∼535 nm. See reference [13] for the modeling details.

In the text
thumbnail Fig. 5

The comparison of calculated exciton generation rate G(x) at λ = ∼535 nm between the S (blue curves) and control (red curves) devices. See reference [13] for the modeling details. Note that, the interruptions in the calculated exciton generation rate at the interfaces are attributed to the sudden changes of the relative permittivity between the two neighboring materials.

In the text
thumbnail Fig. 6

The calculated steady-state exciton distributions across zone 1 at the D = 7.5, 10, 12.5, and 15 nm, respectively, on the basis of the assumed LD of 13.5 nm. The CuPc/C60 interface locates at the thickness = 15 nm.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.