Chemical Physics Letters 557 (2013) 88–91

Contents lists available at SciVerse ScienceDirect

Chemical Physics Letters

journal homepage: www.elsevier .com/locate /cplet t

Cascade organic solar cells with energy-level-matched threephoton-harvesting layers

Myungsun Sim, Jong Soo Kim, Chiyeoung Shim, Kilwon Cho ⇑Department of Chemical Engineering, Polymer Research Institute, Pohang University of Science and Technology, Pohang 790-784, Republic of Korea

a r t i c l e i n f o

Article history:Received 30 August 2012In final form 28 November 2012Available online 7 December 2012

0009-2614/$ - see front matter � 2012 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.cplett.2012.11.087

⇑ Corresponding author. Fax: +82 54 279 8269.E-mail address: kwcho@postech.ac.kr (K. Cho).

a b s t r a c t

The performances of organic photovoltaic cells were improved by matching the energy levels of threephoton-harvesting layers with distinct light absorption spectra. Pentacene, phthalocyanine, and C70,which have shallow, intermediate, and deep energy levels, respectively, were introduced as photon har-vesting layers to form a cascade structure and secure appropriate band offsets at all interfaces. This archi-tecture yielded higher values of the open-circuit voltage and short-circuit current density than a standardbilayer structure based on pentacene and C70. A device prepared with metal-free phthalocyanine per-formed better than a device with Cu phthalocyanine because the energy levels were more appropriatelytuned.

� 2012 Elsevier B.V. All rights reserved.

1. Introduction

Recently, organic photovoltaic (OPV) cells have been exten-sively studied as future energy sources because of their low cost,light weight, and mechanical flexibility [1]. The power conversionefficiencies (PCEs) of OPV cells have been steadily improved byapplying several strategies [2–13]. The inherent tradeoff betweenthe light absorption and the exciton diffusion efficiency has hin-dered improvements in PCE values because optical absorptionlengths (LA � 100 nm) are typically much longer than exciton diffu-sion lengths (LD � 10 nm) [4,14]. One approach to addressing thisproblem is to use multilayered photoactive materials with differ-ent light absorption spectra. The multilayered structures compris-ing several phtotoactive materials can increase the LA and broadenthe span of the solar spectrum that the photoactive layer can ab-sorb, thereby increasing the absorption efficiency. The multilay-ered structures also do not reduce the efficiency of excitondiffusion by providing multiple junctions where exciton dissocia-tion occurs. Several examples of multilayered structures have beenreported including interfacial layer-inserted structures [15–19]and multiple-device stacked structures [10–13]. The device perfor-mances of these structures were enhanced by higher open-circuitvoltages (VOC); however, the short-circuit current densities (JSC) de-creased slightly. The JSC of the interfacial layer-inserted structureswas lower than that of simple donor/acceptor bilayer structuresdue to the reduced charge separation efficiency. In multiple-devicestacked structures, the JSC was limited to that of either subcell dueto charge carrier recombination at the intermediate electrode. Theenhancement of VOC and limitation of JSC in these multilayered

ll rights reserved.

structures are closely related to the energy levels of the layers.However, not many studies have been reported on the control ofthe energy levels for simultaneous enhancement of both VOC andJSC.

Here, we enhanced both the VOC and JSC of the OPV cells bymatching the energy levels of three photon-harvesting layers withdifferent absorption spectra so that the highest occupied molecularorbital (HOMO) and the lowest unoccupied molecular orbital(LUMO) levels of the three materials formed a cascade energy bandstructure. The band offsets at all interfaces in the photoactive layerwere designed to be sufficiently large to permit exciton dissocia-tion. To demonstrate this concept, we employed pentacene, phtha-locyanine dyes, and C70 as photon-harvesting materials. Pentaceneis a p-type material and exhibits a high hole-mobility and shallowHOMO and LUMO levels. Compared to C60, which is a representa-tive n-type material, C70 displays better light absorption propertiesand deeper HOMO and LUMO levels. The large band offset betweenthe pentacene and C70 could produce an excellent of interlayermaterial with optimal energy levels. Several phthalocyanine dyematerials were tested as interlayer materials such that their energylevels formed a cascade energy band structure and broadened thespan of the solar spectrum absorbed by the photoactive layer.

2. Experimental methods

The devices were prepared on patterned indium tin oxide (ITO)glass substrates with 150 nm thick ITO and a 15 X/sq sheet resis-tance. After cleaning the ITO substrates, poly(3,4-ethylenedioxythiophene)-doped poly(styrenesulfonate) (PEDOT:PSS) (Bytron P,from Bayer AG) was spin-coated onto the substrates at 4000 rpmfor 60 s, followed by baking at 120 �C for 30 min. High-puritypentacene, phthalocyanine dyes such as Cu-phthalocyanine (CuPc)

M. Sim et al. / Chemical Physics Letters 557 (2013) 88–91 89

and metal-free phthalocyanine (H2Pc), and C70 were provided byNano C. Bathocuproine (BCP) was purchased from Aldrich. All or-ganic layers were deposited on the ITO substrate by vacuum evap-oration under a vacuum of 10�6 Torr. A pentacene layer (50 nm)was deposited at 0.2 Å/s followed by a 10 nm thick phthalocyaineat 1.0 Å/s and a 25 nm thick C70 at 0.5 Å/s. An 8 nm BCP layerwas deposited at 1.0 Å/s as a hole blocking layer, and finally an100 nm thick Ag layer was deposited as a cathode. The deviceswere characterized by current density–voltage (J–V) characteristicsmeasured under AM 1.5 solar illumination at 100 mW/cm2 gener-ated by an Oriel 1 kW solar simulator using a programmable Keith-ley mode 4200 power source.

Figure 2. Current density–voltage characteristics under illumination for the threedevice structures: pentacenenC70, pentacenenCuPcnC70, and pentacenenH2PcnC70.

Table 1Summary of the device parameters associated with the three device structures testedhere: pentacenenC70, pentacenenCuPcnC70, and pentacenenH2PcnC70.

Device structure JSC

(mA/cm2)VOC

(V)FF n

(%)

Pentacene (50 nm)nC70 (25 nm) 4.05 0.23 0.50 0.47Pentacene (50 nm)nCuPc (10 nm)nC70 (25 nm) 4.13 0.39 0.48 0.77Pentacene (50 nm)nH2Pc (10 nm)nC70 (25 nm) 6.29 0.40 0.51 1.28

3. Results and discussion

The photovoltaic device structure and the corresponding energyband diagrams are shown in Figure 1. The structures of the photo-active layer are the bilayer structure, pentacene (50 nm)nC70

(25 nm), and the three-layered structure, either pentacene(50 nm)nH2Pc or CuPc (10 nm)nC70 (25 nm). Although the phthalo-cyanine layer was inserted between the pentacene and C70 layersin the three-layered structure, band offsets at all interfaces in thethree-layered structure were comparable to the exciton bindingenergy (0.4–1.4 eV) [20] and efficient exciton-dissociation oc-curred due to the large band offset between the energy levels ofthe pentacene and C70. The HOMO and LUMO levels of CuPc andH2Pc were positioned, respectively, between the energy levels ofthe pentacene and C70 to form a cascade energy structure withoutenergy barriers to charge transport toward the electrodes. TheLUMO level of H2Pc was slightly deeper than that of CuPc such thatthe LUMO offset at the pentacenenH2Pc interface matched the exci-ton binding energy of pentacene (0.1–0.6 eV) [21,22]. This factorwas expected when taking into account the much lower electronaffinity (EA) of Cu (1.22 eV) compared to the EA of H2Pc (5.35 eV)[23]. The addition of a Cu atom to the H2Pc molecule introducedadditional charges and decreased the overall EA of the resultingmolecule (CuPc) [23]. The energy levels of HOMO and LUMO levelsin the organic materials were consistent with the values cited inthe literature [23,24].

The current density–voltage (J–V) characteristics of the devicesunder illumination are shown in Figure 2 and Table 1. The open-circuit voltages (VOC) of the devices with three photon-harvestinglayers were higher than those of the device with two photon-har-vesting layers. The VOC depends on the energy difference betweenthe LUMO level of the n-type material and the HOMO level of thep-type material [25]. The HOMO levels of H2Pc and CuPc were sim-ilar and deeper than the level of pentacene, so that the VOC valuesof the devices prepared with H2Pc and with CuPc were not signif-icantly different (0.40 V for H2Pc, 0.39 V for CuPc) and higher than

Figure 1. (a) Schematic diagram of the device structure. (b) Energy band diagramsof the devices with pentacene (P5) (50 nm)nC70 (25 nm), P5 (50 nm)nCuPc(10 nm)nC70 (25 nm), and P5 (50 nm)nH2Pc (10 nm)nC70 (25 nm).

the values of a device without phthalocyanine (0.23 V). The JSC va-lue of a device without phtholocyanine and a device with CuPcwere nearly identical, and the JSC of the device with H2Pc exceededthe values of the other devices. The additional exciton-generationin the phthalocyanine layer contributed to the photocurrent ofeach device with three photon-harvesting layers; however, thenegligible enhancement in JSC value for the device prepared withCuPc may suggest that the additional CuPc layer limited chargegeneration from the pentacene layer. The higher value of JSC inthe device with H2Pc suggested that charge separation in the de-vice with H2Pc was efficient, and the presence of an additionalH2Pc layer slightly hindered charge generation in the pentacenelayer. The small changes in the FF demonstrated that the insertedphthalocyanine layer did not significantly affect the series resis-tances of the devices. As a result, the power conversion efficiencies(gp) of devices with three photon-harvesting layers were higherthan that of a device with two photon-harvesting layers.

Figure 3 shows the absorption spectra and external quantumefficiency (EQE) for each device structure and explains the in-creased JSC in terms of the presence of an additional photon-har-vesting layer with an appropriate energy level. The additionalphthalocyanine layer absorbed longer wavelengths (700–800 nm)of the solar spectrum that pentacene and C70 could not absorb, asshown in Fig 3a and b. As a result, the layer increased the lightabsorption efficiency of the device. This wider light absorptionrange additionally increased the EQE over the range 700–800 nm,corresponding to the phthalocyanine absorption spectrum, asshown in Figure 3c. The additional excitons generated in thephthalocyanine layer contributed to the photocurrent. Excitonsformed in the phthalocyanine were dissociated both at the penta-cenenphthalocyanine interface and at the phthalocyaninenC70

interface; however, the LUMO level difference at the phthalocya-ninenC70 interface (1.1 eV for CuPc and 0.95 eV for H2Pc) was muchhigher than the HOMO level difference at the pentacenenphthalo-

Figure 3. (a) Absorption spectra of each homo film of pentacene, CuPc, H2Pc, andC70. (b) Absorption spectra of the three device structures, (c) EQE spectra of thethree device structures: pentacenenC70, pentacenenCuPcnC70, and pentacenenH2PcnC70.

Figure 4. Atomic force microscopy (AFM) images of (a) a pentacene (50 nm) single film, (a pentacene (50 nm) film on an ITO substrate.

90 M. Sim et al. / Chemical Physics Letters 557 (2013) 88–91

cyanine interface (0.1 eV for CuPc and 0.25 eV for H2Pc) and largerthan the exciton binding energy of the phthalocyanine (0.6 eV)[20]; therefore, the excitons in the phthalocyanine would be pre-dominantly dissociated at the phthalocyaninenC70 interface.

Over the range 500–700 nm, the main absorption peaks corre-sponding to the pentacene and phthalocyanine layers overlappedand the light absorption efficiency of the devices with three pho-ton-harvesting layer exceeded that of a device with two photon-harvesting layers, as shown in Figure 3a and b; however, the higherEQE of devices with three photon-harvesting layers were not aslarge as the relative increase in light absorption. This result indi-cated that not all excitons formed in the pentacene layer contrib-uted to the photocurrent. No morphological features of thepentacene layer were observed in atomic force microscopy (AFM)images of a phthalocyanine layer on a pentacene film (Figure 4).A 10 nm thick phthalocyanine layer prepared with 1 Å/s evapora-tion rate fully covered the surface of the pentacene, and there isno contact between the pentacene and C70. This result indicatedthat excitons generated in the pentacene layer were dissociatedonly at the pentacenenphthalocyanine interface; however, thepentacenenphthalocyanine interface yielded a lower efficiency ofexciton dissociation compared to the pentacenenC70 interface dueto the relatively low LUMO level difference. As a result, some exci-tons generated in the pentacene layer were lost in the devices withthree photon-harvesting layers.

Among the devices with three photon-harvesting layers, theCuPc and H2Pc devices displayed different EQE features over therange 500–700 nm. The EQE of the device with an additional CuPclayer was lower than the values of the device with two photon-harvesting layers, and the EQE curve were comparable to theabsorbance curve of the CuPc. These results suggested that excitongenerated in the pentacene layer could not contribute to the pho-tocurrent, and most of the photocurrent resulted from excitonsformed in the CuPc layer. This result was consistent with the con-clusions of previous studies [16]. On the other hand, the EQE of adevice prepared with an additional H2Pc layer was higher thanthe EQE of a device with two photon-harvesting layers. The deviceprepared with H2Pc exhibited not only an EQE peak at 615 nm, cor-responding to the maximum absorption peak of H2Pc, but also anEQE peak at around 630 nm, corresponding to the absorption ofpentacene. The device prepared with CuPc exhibited only theEQE peak at 615 nm, corresponding to the maximum absorptionpeak of CuPc. The EQE shoulder at around 630 nm provided solidevidence that charge transfer occurred at the pentacenenH2Pcinterface. No previous studies have reported comparable EQE fea-tures in devices with interfacial layer inserted structures [15–18].

Additional insight into charge generation in the three photon-harvesting layers was sought. The discussion above outlines the

b) a CuPc (10 nm) layer on a pentacene (50 nm) film, and (c) a H2Pc (10 nm) layer on

M. Sim et al. / Chemical Physics Letters 557 (2013) 88–91 91

reasons underlying the different EQE features of the devices withCuPc or H2Pc. The EQE is the product of the following four efficien-cies: the efficiency of light absorption in the active region of the so-lar cell (gA), the efficiency of exciton diffusion to dissociation sites(gED), the efficiency of charge transfer at the exciton dissociationinterface (gCT), and the efficiency of charge collection (gCC). Acrossthe main absorption band of pentacene, the gA of the device pre-pared with H2Pc was comparable to that of the device with CuPc,as shown in Figure 3b. H2Pc and CuPc are planar molecular withsimilar Q-band peaks because the Cu(II) ion is sufficiently smallto be accommodated in the central cavity of the phthalocyaninering and still engages in weak interaction with the phthalocyaninering [26]. The difference between gED and gCC in H2Pc and CuPc de-vices was negligible because the H2Pc and CuPc layers were suffi-ciently thin (10 nm) to permit exciton diffusion and chargetransport. Thus, the EQE in the 600–700 nm region was mainlydetermined by gCT, which was governed by the band offset at thejunction. In the present devices, the LUMO offset at the penta-cenenH2Pc junction (0.35 eV) was larger than that at the penta-cenenCuPc junction (0.2 eV). The larger LUMO level differencewas more suitable for the exciton binding energy of pentacene(0.1–0.6 eV) [21,22], and increased the probability of electrontransfer from pentacene to H2Pc [24,27].

4. Conclusions

We prepared OPV devices based on three photon-harvestinglayers with different absorption spectra. Pentacene, phthalocya-nine (CuPc and H2Pc), and C70 were included, each of which hasshallow, intermediate, and deep energy levels, respectively. Theenergy levels of the three layers were matched in an effort to in-crease both VOC and JSC relative to the values obtained in the stan-dard bilayer device composed of a p-type and an n-typephotoactive material. The band offset energies critically affectedphotocurrent in the device with three photon-harvesting layers.This approach provides the basis for new device architectures cen-

tered on multilayered photon-harvesting materials that enhancethe device performance.

Acknowledgement

This work was supported by a grant (Code No. 2012-053499)from the Center for Advanced Soft Electronics under the GlobalFrontier Research Program of the Ministry of Education, Scienceand Technology, Korea.

References

[1] S.E. Shaheen, D.S. Ginley, G.E. Jabbour, MRS Bull. 30 (2005) 10.[2] S. Gunes, H. Neugebauer, N.S. Sariciftci, Chem. Rev. 107 (2007) 1324.[3] G. Dennler, M.C. Scharber, C.J. Brabec, Adv. Mater. 21 (2009) 1323.[4] P. Peumans, A. Yakimov, S.R. Forrest, J. Appl. Phys. 93 (2003) 3693.[5] X. Bulliard et al., Adv. Funct. Mater. 20 (2010) 4381.[6] J. Peet, J.Y. Kim, N.E. Coates, W.L. Ma, D. Moses, A.J. Heeger, G.C. Bazan, Nat.

Mater. 6 (2007) 497.[7] J.S. Kim, J.H. Park, J.H. Lee, J. Jo, D.Y. Kim, K. Cho, Appl. Phys. Lett. 91 (2007).[8] S.H. Park et al., Nat. Photon. 3 (2009) 297.[9] S.C. Price, A.C. Stuart, L.Q. Yang, H.X. Zhou, W. You, J. Am. Chem. Soc. 133

(2011) 4625.[10] A. Yakimov, S.R. Forrest, Appl. Phys. Lett. 80 (2002) 1667.[11] T. Ameri, G. Dennler, C. Lungenschmied, C.J. Brabec, Energy Environ. Sci. 2

(2009) 347.[12] M. Riede et al., Adv. Funct. Mater. 21 (2011) 3019.[13] L.T. Dou et al., Nat. Photon. 6 (2012) 180.[14] J.S. Kim et al., Adv. Funct. Mater. 20 (2010) 540.[15] S. Sista, Y. Yao, Y. Yang, M.L. Tang, Z.A. Bao, Appl. Phys. Lett. 91 (2007) 223508.[16] Y. Kinoshita, T. Hasobe, H. Murata, Appl. Phys. Lett. 91 (2007) 083518.[17] S.L. Lai, M.F. Lo, M.Y. Chan, C.S. Lee, S.T. Lee, Appl. Phys. Lett. 95 (2009) 153303.[18] J.S. Kim, J.H. Lee, J.H. Park, C. Shim, M. Sim, K. Cho, Adv. Funct. Mater. 21 (2011)

480.[19] Z.R. Hong, R. Lessmann, B. Maennig, Q. Huang, K. Harada, M. Riede, K. Leo, J.

Appl. Phys. 106 (2009) 064511.[20] M. Knupfer, Appl. Phys. A: Mater. Sci. Proc. 77 (2003) 623.[21] J. Lee, S.S. Kim, K. Kim, J.H. Kim, S. Im, Appl. Phys. Lett. 84 (2004) 1701.[22] P.K. Nayak, N. Periasamy, Org. Electron. 10 (2009) 1396.[23] D.R.T. Zahn, M. Gorgoi, O.D. Gordan, Sol. Energy 80 (2006) 707.[24] B.P. Rand, D.P. Burk, S.R. Forrest, Phys. Rev. B 75 (2007) 115327.[25] C.J. Brabec, A. Cravino, D. Meissner, N.S. Sariciftci, T. Fromherz, M.T. Rispens, L.

Sanchez, J.C. Hummelen, Adv. Funct. Mater. 11 (2001) 374.[26] C.G. Claessens, U. Hahn, T. Torres, Chem. Rec. 8 (2008) 75.[27] S. Sun, Z. Fan, Y. Wang, J. Haliburton, J. Mater. Sci. 40 (2005) 1429.