Report

External Quantum Efficiency Above 100% in a Singlet-Exciton-Fission–Based Organic Photovoltaic Cell

See allHide authors and affiliations

Science  19 Apr 2013:
Vol. 340, Issue 6130, pp. 334-337
DOI: 10.1126/science.1232994

Splitting Singlets

Solar cell efficiency is limited because light at wavelengths shorter than the cell's absorption threshold does not channel any of its excess energy into the generated electricity. Congreve et al. (p. 334) have developed a method to harvest the excess energy in carbon-based absorbers through a process termed “singlet fission.” In this process, high-energy photons propel two current carriers, rather than just one, by populating a singlet state that spontaneously divides into a pair of triplet states. Although it works in a functioning organic solar cell, the efficiency needs improving.

Abstract

Singlet exciton fission transforms a molecular singlet excited state into two triplet states, each with half the energy of the original singlet. In solar cells, it could potentially double the photocurrent from high-energy photons. We demonstrate organic solar cells that exploit singlet exciton fission in pentacene to generate more than one electron per incident photon in a portion of the visible spectrum. Using a fullerene acceptor, a poly(3-hexylthiophene) exciton confinement layer, and a conventional optical trapping scheme, we show a peak external quantum efficiency of (109 ± 1)% at wavelength λ = 670 nanometers for a 15-nanometer-thick pentacene film. The corresponding internal quantum efficiency is (160 ± 10)%. Analysis of the magnetic field effect on photocurrent suggests that the triplet yield approaches 200% for pentacene films thicker than 5 nanometers.

Conventional solar cells generate one electron for each photon that is absorbed. The output voltage is defined by the bandgap, and solar cells waste any excess photon energy as heat. Summing the thermal loss over the solar spectrum yields the Shockley-Queisser efficiency limit of 34% for solar cells containing a single, optimized semiconductor junction (1).

Splitting excited states, or excitons, generated after the absorption of high-energy photons presents one pathway beyond the single junction efficiency limit. Instead of harvesting a single electron, several charges can be obtained by dissociating the child excitons. For example, so-called multiple exciton generation mechanisms have been used to produce an average of more than one electron from an ultraviolet photon with energy four times the bandgap (2).

Singlet exciton fission is a type of multiple exciton generation mechanism found in organic semiconductors (3, 4). It is notable because spin conservation disallows the usual competing loss process: thermal relaxation of the high-energy exciton into a single low-energy exciton. In fission, the low-energy exciton is a dark state, inaccessible by a direct transition from either the high-energy exciton or the ground state. Only the evolution of the high-energy state into two dark excitons is spin-allowed. Consequently, prior studies have suggested that singlet fission can be efficient even in the visible spectrum, harnessing photons of just twice the energy of the child excitons (510).

There is a side effect of spin in singlet fission, however. The dark exciton controls the electrical properties of the cell. These are decoupled from the optical absorption, which is controlled by the bright, high-energy exciton. Thus, fission does not itself increase the power efficiency of a solar cell. It potentially doubles the photocurrent at the cost of losing at least half the open circuit voltage. To overcome the Shockley-Queisser limit, solar cells could combine fission with a conventional material that fills in the absorption spectrum above the dark exciton (1115). First, however, singlet fission must demonstrate that it can break the conventional barrier of one electron per photon.

The best understood fission material to date is pentacene, an acene with five rings. Its dynamics are illustrated in Fig. 1. Optical excitation generates a delocalized spin 0, or singlet, exciton. Within about 80 fs (510), the pentacene singlet exciton splits into a pair of spin 1, or triplet, excitons. A pair of triplet excitons can be combined in nine different spin states. As some triplet-pair states have singlet character (total spin of 0), singlet fission is spin-allowed. Under an applied magnetic field, the singlet character of the triplet-pair states is redistributed, changing the number of states with singlet character. Thus, the rate of singlet fission is dependent on the local magnetic field, offering a unique probe of fission dynamics in thin films and devices (4, 11, 12, 16, 17). Although triplet excitons are dark states, energy may be extracted from them if they are dissociated into charge. This is possible at a junction between pentacene and the fullerene C60 when the pentacene is oriented approximately perpendicular to the interface (12).

Fig. 1 Singlet fission dynamics in pentacene.

Calculations of singlet and triplet excitons and charge transfer states at the pentacene/fullerene interface are shown, with the purple (orange) density indicating where less (more) electron density is found in the excited state. The delocalized singlet exciton and two localized triplet excitons are circled in red. The loss pathway for singlet excitons is direct dissociation into charge before singlet exciton fission.

The structure of our pentacene-based solar cell is shown in Fig. 2A. The core of the device is a pentacene/C60 donor-acceptor junction (4, 12, 18). To minimize triplet exciton losses, we used a thin pentacene layer and also introduced an exciton blocking layer of regio-regular poly(3-hexylthiophene) (P3HT), which was placed between the pentacene and the anode. The combination of the wide energy gap and 1.5-eV triplet energy (19) of P3HT confines pentacene triplet excitons, and its highest occupied molecular orbital (HOMO) of 4.7 eV (20) helps extract holes from pentacene. To maximize light absorption, devices were fabricated with MgF2 antireflection coatings on the front surface of the glass substrate (21).

Fig. 2 Device architecture and EQE of a pentacene solar cell.

(A) Chemical structures and architecture of the solar cell with the thickness of each layer in nanometers and energy levels of the lowest unoccupied and highest occupied molecular orbitals in electron volts (12, 18, 20, 2931). The anode is composed of indium tin oxide (ITO) and poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT:PSS). The cathode employs bathocuproine (BCP) and a silver cap. (B) External quantum efficiency of devices without optical trapping (blue line), and device measured with light incident at 10° from normal with an external mirror reflecting the residual pump light (red line). Optical fits from IQE modeling are shown with dashed lines: modeled pentacene EQE (blue dashes), modeled P3HT EQE (purple dashes), and modeled device EQE (black dashes) for comparison to the measured device efficiency without optical trapping.

The external quantum efficiency (EQE) is defined as the ratio between the number of electrons flowing out of the device and the number of photons incident upon it. We measured the EQE (Fig. 2B) for a device that features a 15-nm-thick pentacene layer. The EQE at normal incidence is (82 ± 1)% at the peak pentacene absorption wavelength λ = 670 nm. Optical modeling (21) predicts that the internal quantum efficiency (IQE), which is defined as the number of electrons collected per photon absorbed, for photoexcitation of pentacene and P3HT is (160 ± 10)% and (150 ± 10)%, respectively. The IQE of pentacene in this structure is approximately double that reported previously for pentacene (18, 22), and the high IQE of P3HT is consistent with the expected sensitization of P3HT by pentacene, as singlet excitons generated in P3HT are transferred to pentacene and then split into triplets (23). The peak EQE drops to 24% when P3HT is absent. The P3HT appears to block triplet diffusion to the anode and suppress recombination by improving hole extraction; see figs. S1 to S3 and accompanying text for further discussion of both sensitization and the efficiency enhancement due to P3HT.

The 15-nm-thick film of pentacene in the solar cell microcavity absorbs only 49% of the incident light at λ = 670 nm according to optical modeling; hence, the efficiency should improve if a light-trapping scheme is employed. Therefore, we measured the EQE in configurations designed to simulate two conventional optical trapping schemes. The first scheme mounts the cell at 45° to the incident light, with a mirror that directs reflected photons back to the device. This configuration models a sawtooth geometry such that incident light bounces at least twice within the structure (24, 25). In the second scheme, the incident angle is reduced to 10° from the normal, modeling an optical collector that focuses light through a small hole in a mirror held parallel to the surface of the cell (26). The peak EQE at λ = 670 nm for the solar cell mounted at 45° is (102 ± 1)%, increasing to (109 ± 1)% for incidence at 10° from the normal. Both light-trapping schemes yield efficiencies that meet or exceed the one electron per incident photon benchmark.

The current-voltage characteristics of the planar pentacene solar cell are shown in Fig. 3. The short-circuit current measured at AM1.5 matches the integrated EQE measured at <1 mW/cm2 to within 6%, demonstrating that the fission process in pentacene is not significantly intensity dependent. As expected, the enhanced EQE does not correspond to a high power efficiency. The open circuit voltage is 0.36 V, identical to the values of previous pentacene devices (18). It is defined by the pentacene triplet energy of 0.86 eV (3, 14). With C60 as the acceptor, the device absorbs light only above the pentacene singlet energy at 1.8 eV. Consequently, the power efficiency is (1.8 ± 0.1)%.

Fig. 3 Current density–voltage characteristic of a pentacene solar cell.

The solar cell was measured under dark (black dashed line) or AM1.5G 100 mW/cm2 (red line) conditions without optical trapping. The power efficiency is (1.8 ± 0.1)%.

Independent confirmation of the high internal quantum efficiency within the cell is provided by analysis of the photocurrent under a magnetic field. The crucial rates are identified in Fig. 1. The singlet exciton can either directly dissociate into a single electron-hole pair, kS, or undergo fission resulting in the generation of two electron-hole pairs, kfis(B). In absence of a magnetic field, three out of nine triplet-pair states have singlet character. Under a high magnetic field (B > 0.2 T), the number of triplet-triplet pairs with singlet character reduces to two, reducing the singlet fission rate, kfis(B). The photocurrent yield changes if there is effective competition between fission and the dissociation of the singlet exciton. It is not possible to generate a magnetic field effect on the photocurrent yield unless there is a singlet loss mechanism that competes with the fission process.

For convenience, we write Embedded Image, where χ(B) is the modulation of the zero-field fission rate Embedded Image. The normalized change in photocurrent in steady state, δ, is then given by Eq. 1:

Embedded Image (1)

where I(B) is the photocurrent as a function of magnetic field strength. Dissociation of the singlet exciton directly into charge is only likely to compete with fission for pentacene molecules directly adjacent to the acceptor. Indeed, reductions in the singlet exciton lifetime of pentacene have been observed in very thin pentacene films (0.6 monolayer) adjacent to a C60 layer (10). Thus, we can approximately model pentacene films of varying thickness by changing the effective rate of singlet dissociation in Eq. 1.

Analytically, we can solve for χ at a given value of the magnetic field by noting that the magnitude of δ is maximized when Embedded Image. This yields Eq. 2:

Embedded Image (2)

The result for χ can be used to directly obtain the triplet yield of singlet fission from the magnetic field modulation in photocurrent (Eq. 3):

Embedded Image (3)

To obtain an independent measure of the yield of singlet fission, we fabricated multiple devices while varying the thickness of pentacene (Fig. 4). For thin layers of pentacene (d < 5 nm), we increased the optical absorption by using the multilayer photodetector architecture (4, 26). Photodetectors were measured in reverse bias to improve charge extraction. As a test of generality, both C60 and 3,4,9,10-perylene tetracarboxylic bisbenzimidazole (PTCBI) were used as acceptor molecules and found to yield similar results. Devices with thicker layers of pentacene made use of the same device architecture as in Fig. 2. The magnetic field modulation of photocurrent at 0.4 T is shown in Fig. 4A. It peaks at δmax = −(2.7 ± 0.1)% in 2-nm-thick layers of pentacene sandwiched between acceptor layers. From Eq. 2, we obtain χ = 0.85, identical to the value assumed in (4) on the basis of tetracene measurements (27).

Fig. 4 Triplet yield and IQE as a function of pentacene thickness.

(A) The magnetic field–dependent change in photocurrent measured at B = 0.4 T as a function of pentacene layer thickness. Square symbols are measured in photodetector structures, and each pentacene layer is sandwiched between C60 (blue squares) or PTCBI (red squares) acceptor films. Measurements in the solar cell architecture of Fig. 2A are circles. (B) The triplet yield from singlet exciton fission as obtained from Eq. 3. (C) A comparison of the maximum achievable quantum yield determined from the magnetic field effect (green line) with the internal quantum efficiency as determined from EQE measurements and the calculated optical absorption. The reduction in quantum efficiency observed in thin layers of pentacene is found to originate in incomplete singlet exciton fission. Gray dashed lines are a guide to the eye.

In Fig. 4B, we apply Eq. 3 to transform the magnetic field modulation data into the expected yield of triplet excitons from singlet fission. We find that singlet fission is incomplete in pentacene films with thickness d < 5 nm, accounting for the relatively low IQE in the photodetector structures. The triplet yield approaches 200% in thicker films, providing independent confirmation of the high IQE calculated for the device structure shown in Fig. 2.

The IQE, as evaluated with optical modeling (26), is shown in Fig. 4C and compared to predictions based on the magnetic field effect. The IQE is suppressed in thin layers of pentacene, increases to a maximum for d ~ 15 nm, and then is reduced in thicker films. Decreases in IQE for thicker films are presumably due to triplet exciton diffusion limitations and lower-than-unity charge collection efficiency. There are two important conclusions from this IQE comparison. First, the yield of singlet fission can be conveniently determined directly from the normalized change in photocurrent under a magnetic field. A high yield is characterized by a vanishing modulation of photocurrent under a magnetic field. Second, singlet fission in pentacene requires a relatively thick film to minimize losses due to singlet exciton dissociation. Fission is not effective in fine-grained blends of pentacene and fullerene or perylene-based acceptors.

The observation of external quantum yields exceeding 100% in the visible spectrum represents a notable advance in the application of singlet fission to solar cells. Next, fission should be paired with a low-bandgap material that harvests photons below the singlet exciton energy. This could be an organic material (11), inorganic semiconductor nanocrystal (1214), or a conventional inorganic semiconductor (15). High-quality contemporary silicon solar cells show an AM1.5 efficiency of ~25% (28); singlet fission materials such as tetracene or rubrene could be integrated with silicon cells to double the photocurrent from high-energy solar photons (λ < 550 nm), ultimately boosting the efficiency of the silicon cell to more than 30%.

Supplementary Materials

www.sciencemag.org/cgi/content/full/340/6130/334/DC1

Materials and Methods

Supplementary Text

Figs. S1 to S5

References (3237)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: This work was supported as part of the Center for Excitonics, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under award no. DE-SC0001088 (MIT). D.N.C. was partially supported by the NSF Graduate Research Fellowship under grant 1122374. MIT has filed a provisional application for patent based on this technology that names D.N.C., N.J.T., and M.A.B. as inventors.
View Abstract

Navigate This Article