ARTICLESPUBLISHED ONLINE: 5 OCTOBER 2014 | DOI: 10.1038/NMAT4097

Energy harvesting of non-emissive triplet excitonsin tetracene by emissive PbS nanocrystalsNicholas J. Thompson1†, MarkW. B. Wilson1,2†, Daniel N. Congreve1†, Patrick R. Brown1,Jennifer M. Scherer2, Thomas S. Bischof1,2, Mengfei Wu1, Nadav Geva1, MatthewWelborn1,Troy Van Voorhis1, Vladimir Bulović1, Moungi G. Bawendi1,2* and Marc A. Baldo1*

Triplet excitons are ubiquitous in organic optoelectronics, but they are often an undesirable energy sink because they are spin-forbidden from emitting light and their high binding energy hinders the generation of free electron–hole pairs. Harvestingtheir energy is consequently an important technological challenge. Here, we demonstrate direct excitonic energy transferfrom ‘dark’ triplets in the organic semiconductor tetracene to colloidal PbS nanocrystals, thereby successfully harnessingmolecular triplet excitons in the near infrared. Steady-state excitation spectra, supported by transient photoluminescencestudies, demonstrate that the transfer e�ciency is at least (90± 13)%. Themechanism is a Dexter hopping process consistingof the simultaneous exchange of two electrons. Triplet exciton transfer to nanocrystals is expected to be broadly applicable insolar and near-infrared light-emitting applications, where e�ective molecular phosphors are lacking at present. In particular,this route to ‘brighten’ low-energy molecular triplet excitons may permit singlet exciton fission sensitization of conventionalsilicon solar cells.

Bound electron–hole pairs, known as excitons, dominate theoptoelectronic properties of disordered and low-dimensionalsemiconductors1,2. In disordered materials, such as organic

semiconductors, the excitons are localized and exhibit distinctspin-0 singlet and spin-1 triplet states. Molecular singlet excitonsmediate light absorption and emission because the ground state isusually a singlet1. In contrast, triplet excitons are typically dark, butthey can be the dominant excitation insidemolecular optoelectronicdevices as they exhibit long lifetimes, are often the lowest energyexcitation in organic photovoltaic blends3, and comprise threequarters of the excitons formed from free carriers4,5. Triplets arealso the products of singlet exciton fission, a phenomenon in someorganic semiconductors which can efficiently convert a high-energysinglet into a pair of lower-energy triplets6–9.

Here, we demonstrate that triplet excitons can be efficientlytransferred from molecules to inorganic nanocrystals even whenelectric-dipole coupling is negligible. The mechanism is the short-range Dexter process—a simultaneous, correlated transfer of twoelectrons that depends on the wavefunction overlap between donorand acceptor2,10. Indeed, the same mechanism is used in visible-spectrum organic light-emitting diodes (OLEDs) to transfer tripletexcitons formed in a non-emissive host to guest molecules withstrong spin–orbit coupling and efficient phosphorescence fromtheir triplet states11. However, molecular acceptors have provedineffective in the infrared (hν < 1.5 eV) because of strong, non-radiative dissipation12.

We show that inorganic nanocrystals are an especially effectivemeans of harvesting non-emissive triplets with energies in theinfrared. This is in contrast to prior, unsuccessful, attempts totransfer triplets to bulk inorganic semiconductors13–15, as well asreported work with colloidal nanocrystals, which has relied onFörster transfer from emissive molecular donor states such assinglet excitons16 or spin-mixed triplet excitons of a phosphorescent

molecule17. In sum, Dexter energy transfer-based strategies for‘brightening’ triplet excitons may provide a general route forcircumventing the intrinsic unsuitability of spin-1 states in opticaland optoelectronic applications.

The proposed triplet energy transfer process is depictedschematically in Fig. 1a. We employ lead (II) sulphide (PbS)colloidal nanocrystals as the inorganic semiconductor acceptorfor a non-emissive triplet exciton. Synthesis of the nanocrystalsis described in the Supplementary Information. Triplet excitonsare generated in tetracene via singlet fission, which producesdark18 triplet excitons rapidly (τ <200 ps; refs 19–22) and at yieldsapproaching 200% in neat films23,24. The respective energy levelsof tetracene and the nanocrystals as determined by ultravioletphotoelectron spectroscopy (UPS) are shown in Fig. 1b. Theinterface between tetracene and the nanocrystals is a type Iheterojunction. Further, excitonic energy transfer is energeticallypreferred over two consecutive charge transfers since both ofthe possible charge transfer intermediates are expected to havehigher energies than the initial tetracene triplet energy (1.25 eV;refs 25–27). Indeed, we confirm the absence of charge formation atthese interfaces in the Supplementary Information using magnetic-field-dependent studies. The morphology of the sample is shownin Fig. 1c. It consists of several monolayers of PbS nanocrystalsthat are coated with a thermally evaporated layer of tetracene. Thetetracene layer is rough and consistent with Stranski–Krastanovgrowth of an initial inhomogeneous coating of tetracene followedby island formation.

To prove transfer of triplet excitons from tetracene to PbSnanocrystals, we proceed as follows: first, we measure the excitationspectrum of the tetracene/PbS bilayers and observe that opticalexcitation of tetracene results in emission from PbS. Because thisonly demonstrates the existence of energy transfer from tetraceneto PbS, we determine the yield of excitons transferred to PbS per

1Energy Frontier Research Center for Excitonics, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA, 2Department ofChemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA. †These authors contributed equally to this work.*e-mail: mgb@mit.edu; baldo@mit.edu

NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials 1© 2014 Macmillan Publishers Limited. All rights reserved.

ARTICLES NATUREMATERIALS DOI: 10.1038/NMAT4097

500 nm

a b c

Singletfission En

ergy 3.9 eV

1.25 eV

4.9 eV

PbSNC Tc

5.3 eV

0 20 40Height (nm)

60 81.7

3.0 eV

3 Å Ø4.8 nm

Triplettransfer

Figure 1 | Schematic of triplet exciton transfer from tetracene to a PbS nanocrystal with decanoic acid ligands. a, Singlet excitons in tetracene firstundergo singlet fission then triplet transfer to the nanocrystal. The tetracene simulations are from ref. 23. Note that the figure is not to scale; the tetracenemolecules have been expanded relative to the nanocrystal to show the excitonic states. b, Energy level diagram of the tetracene/PbS nanocrystal interface,with the tetracene highest occupied molecular orbital (HOMO, red) energy from UPS measurements35 and a lower bound for the lowest unoccupiedmolecular orbital (LUMO, blue) energy obtained by adding the singlet energy from the emission peak20,24. This method does not account for Coulombicstabilization. Nanocrystal valence (red) and conduction (blue) band energy levels are from UPS measurements (Supplementary Information), and areaccurate to within 0.1 eV. The tetracene triplet exciton is schematically represented by a bound electron (white dot) and hole (blue dot) within the energygap of tetracene. The tetracene triplet exciton energy is 1.25 eV (refs 25–27). c, AFM micrograph of the tetracene layer on PbS nanocrystals.

photon absorbed by tetracene. We find that the yield exceeds one,proving the involvement of triplet excitons generated by singletexciton fission in tetracene. We find that the efficiency of energytransfer is consistent with a Dexter mechanism and exponentiallydependent on the length of the ligand spacers on the surface ofthe nanocrystals. The dominance of triplet, rather than singletenergy transfer is confirmed using an external magnetic field tovary the rate of singlet exciton fission. Reducing the fission ratelessens the energy transferred to the nanocrystal. We conclude bycharacterizing the rate of triplet energy transfer using transientphotoluminescence.

As noted above, energy transfer from tetracene is probed bymeasuring the near-infrared emission from the nanocrystals whilepreferentially exciting the tetracene layer. In Fig. 2a we measure theexcitation spectrum of a thin film of PbS nanocrystals coated with a20-nm-thick film of tetracene. The absorption spectra of tetraceneand the nanocrystals are distinct, with PbS absorption dominantat longer wavelengths (λ> 550 nm), and PbS and tetracene bothabsorbing at shorter wavelengths (λ< 550 nm); see Fig. 2b. Theabsorption of tetracene is structured, with three characteristic peaksin the blue–green spectrum1,20. Thus, the appearance of tetracenepeaks in the excitation spectrum of PbS nanocrystal luminescenceconfirms energy transfer from tetracene to the nanocrystals.

To control the yield of exciton harvesting, we note that therate of energy transfer generally increases with decreasing distancebetween the donor and acceptor2,10. Our PbS nanocrystals aresynthesized with insulating oleic acid (OA) ligands, which passivatesurface traps and provide solubility in organic solvents. In the solidstate, however, the ligand acts as a spacer between the nanocrystalsand the tetracene. Using solid-state ligand exchange, the OA ligandcan be replaced with a number of shorter alkyl carboxylic acidderivatives28. We highlight two examples in Fig. 2, decanoic (capric)acid (DA) and caprylic (octanoic) acid (CA). As shown in Fig. 2a,both ligands result in improved energy transfer, demonstratedby the increased excitation efficiency in the regions of tetraceneabsorption. However, decreasing the ligand length reduces thequantum yield of the nanocrystals. The typical quantum yield ofour films of OA-functionalized nanocrystals is approximately 9%,decreasing to 0.5% for the shorter CA-functionalized nanocrystals.

Correcting the excitation spectra in Fig. 2a for the absorptionspectra shown in Fig. 2b gives the relative photoluminescencequantum yield shown in Fig. 2c. It is evident that the intrinsic

quantum yield of emission from PbS nanocrystals is largelyindependent of incident wavelength in our measurement range.The tetracene-coated samples, however, exhibit either dips or peaksin their blue–green spectrum depending on the ligand length.These features result from the competition between ‘shadowing’—the absorption of photons in tetracene that would otherwisehave been absorbed by the nanocrystals—and efficient energytransfer from tetracene to the PbS nanocrystals. The wavelength-dependent quantum yield of photons from the nanocrystal can beexpressed as:

QY(λ)=QYNC

(ABSNC(λ)+ηfisηETABSTc(λ)

ABSNC(λ)+ABSTc(λ)

)where ABSNC is the absorption of the nanocrystal, QYNC is theintrinsic quantum yield of the nanocrystal, ABSTc is the absorptionof tetracene, ηfis is the yield of excitons in tetracene after singletexciton fission and ηET is the exciton transfer efficiency fromtetracene to the nanocrystal.

Analysis of the photoluminescence quantum yield shows thatthe efficiency of energy transfer is notably improved when thenanocrystal ligand is short. In Fig. 2c we find that nanocrystalstreated with OA exhibit dips in the quantum yield for regions ofstrong tetracene absorption while CA-treated nanocrystals havepeaks. Using the measured absorption spectra for neat films oftetracene and PbS nanocrystals we can determine the product ηfisηETby fitting, as described in detail in the Supplementary Information.An example fit is plotted in Fig. 2c, where ηfisηET= 2.0. Averagingthirteen different samples with caprylic acid ligands, we determineηfisηET=1.80±0.26.

The critical observation that ηfisηET>1 provides definitive proofthat the direct transfer of triplet excitons is the primary methodof energy transfer in our device, because only the efficient transferof triplets generated in pairs by fission could result in moreexcitons transferred to the nanocrystals than photons absorbedin the tetracene. Specifically, assuming ηfis = 2, the lower boundfor the efficiency of triplet transfer to CA-treated nanocrystals isηET=0.90±0.13. Furthermore, as shown in Fig. 3, we find that thetransfer efficiency has an exponential dependence on the number ofcarbon–carbon bonds in the alkyl carboxylic acid ligand, which isconsistent with Dexter transfer. The fit follows exp(−2βnn), whereβn = 0.098 (C–C bonds)−1 and n is the number of C–C bonds.

2 NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials© 2014 Macmillan Publishers Limited. All rights reserved.

NATUREMATERIALS DOI: 10.1038/NMAT4097 ARTICLES

450

Exci

tatio

n (n

orm

.)A

bsor

ptio

nRe

lativ

e qu

antu

m y

ield

2

0.3

0.1

1.6

1.4

1.2

1.0

0.8

4

6

8a

b

c

500 550Wavelength (nm)

600 650 700

450 500 550Wavelength (nm)

600 650 700

450 500 550Wavelength (nm)

600 650 700

CA

DA

OA

OOH

OOH

OOH

Figure 2 | Steady-state observations of triplet transfer. a, Excitationspectra for tetracene-capped films of 0.95 eV PbS nanocrystals with OA,DA and CA ligands. The spectra are normalized by the mean valuebetween 630 and 670 nm. Nanocrystal emission was measured for allwavelengths greater than 950 nm. b, Visible absorption versus wavelengthof each of the films in a. Also plotted is the absorption of tetracene(orange). c, Relative photoluminescence quantum yield of each of the filmsin a and a neat film of nanocrystals (light blue) normalized by the meanvalue between 630 and 670 nm. The fit to the QY for the nanocrystals witha CA ligand is shown as a dashed black line. The fit calculated ηfisηET=2.0.The tetracene layer is 20 nm thick. Ligand chemical structures are shown inthe inset of a.

Assuming a radial ligand morphology, this extracted value of βn isnearly one order of magnitude greater than typically observed incharge transfer studies29. We note, however, that this measurementis a steady-state characterization of the yield of energy transfer,rather than a direct measurement of its rate. Indeed, the spatialdependence of the rate cannot be determined in the presence ofcompeting channels that may also depend on the length of thenanocrystal ligands.

Triplet exciton transfer from tetracene to the PbS nanocrystalsis confirmed and further distinguished from singlet excitontransfer by the steady-state magnetic field dependence of singletexciton fission19. A magnetic field B> 0.4 T slows the effectiverate of triplet generation, shifting the balance from tripletstowards singlets1,19,30. Magnetic field studies are especially usefulin devices, since they can isolate the contribution of fission tothe overall performance. Indeed, the magnetic field dependence ofphotocurrent in pentacene26,31 and tetracene23 solar cells has been

6 8 10

100

50

20

1012

Ligand length (number of bonds)

Tran

sfer

effi

cien

cy (%

)

14 16 18 20

Figure 3 | Energy transfer e�ciency versus the number of carbon–carbonsingle bonds in the nanocrystal ligand. 0.95 eV PbS nanocrystals wereused and we assumed a singlet fission e�ciency of ηfis=2. The solid line isan exponential fit. The vertical error bars give the fractional uncertainty inthe transfer e�ciency, which we estimated as the standard error of themean value from 13 nominally identical CA-exchanged bilayer devices.

used previously to measure singlet fission yields and demonstratethat triplet excitons are converted to photocurrent. In Fig. 4a weplot the steady-state change in fluorescence from the tetracenecoating and fluorescence from the tetracene-coated nanocrystalfilm as a function of magnetic field. When exciting both tetraceneand the PbS nanocrystals with continuous-wave light (λ=460 nm),we observe that the magnetic field dependence of each materialis characteristic of singlet exciton fission, and closely matchesprevious photocurrent- and emission-based measurements19,23,26,31.Further, the dependences have opposite sign, which occurs becausethe inhibition of fission preserves singlets, leading to additionalphotoluminescence from the tetracene, while decreasing thereservoir of triplets, so that the nanocrystals emit less light. Weemphasize that the field dependence is due to singlet fission—asshown in Fig. 4a, the nanocrystal fluorescence is unaffected bythe magnetic field when the bilayer film is excited at wavelengthswhere tetracene does not absorb. This is consistent with theoreticalpredications that a B= 0.5 T field is insufficient to significantlyperturb the excitonic states of colloidal nanocrystals at roomtemperature32. We also observed no significant change in linewidthor position of the emission spectrum under a B= 0.5 T magneticfield (Supplementary Fig. 1).

The activation energy dependence of the triplet energy transferprocess is shown in Fig. 4b. Dexter energy transfer favours acceptormaterials with excited states equal or less in energy than theexcited state in the donor. In PbS nanocrystals, this energy isstrongly dependent on the size of the particles due to quantumconfinement25,26,33,34. Thus, through the synthesis of different sizednanocrystals (Supplementary Information), we can examine themagnetic field dependence of fluorescence for nanocrystals witha range of exciton energies; see Fig. 4b. Nanocrystals withexciton energy greater than the tetracene triplet energy of 1.25 eV(refs 25–27) exhibit a positive change in fluorescence under largemagnetic field, indicating energy transfer in these systems isprimarily from singlet excitons. The ‘threshold’ transition to anegative magnetic field effect when the nanocrystal’s optical gapis less than 1.25 eV is difficult to reconcile with efficient singlettransfer, and instead indicates that singlet exciton fission and tripletexciton transfer overwhelms Förster transfer in this systemwhen thetransfer is exothermic.

NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials 3© 2014 Macmillan Publishers Limited. All rights reserved.

ARTICLES NATUREMATERIALS DOI: 10.1038/NMAT4097

3 12

8

4

0

0 0.1 0.2 0.3 0.4 0.5

−4

−8

−12

Chan

ge in

nan

ocry

stal

fluo

resc

ence

(%)

Chan

ge in

nan

ocry

stal

em

issi

on (%

)

Magnetic field (T)

0 0.1

1.09 eV

0.94 eV

1.23 eV

1.33 eV

0.2 0.3 0.4 0.5Magnetic field (T)

Change in tetracene fluorescence (%)

2

1

0

−1

1

0

−1

−3

−2

−4

−2

−3

ex = 460 nmλ

ex = 660 nmλ

ex = 460 nmλ

ex = 460 nmλ

a

b

Figure 4 | Monitoring triplet transfer via singlet fission. a, Change inemission from 0.95 eV (λ= 1,308 nm) PbS–OA nanocrystal thin filmscoated with 40 nm of tetracene (circles) and fluorescence from thetetracene coating (squares) as a function of external magnetic field, when abilayer film is excited with λ=460 nm light (absorbed by both layers) andλ=660 nm light (absorbed by nanocrystals only). b, Change in PbS–OAnanocrystal thin film emission for various nanocrystal energiescharacterized by the energy of the first absorption peak in solution.Samples were coated with 40 nm of tetracene. The change in fluorescenceis positive for nanocrystal films with energy greater than the triplet energyof tetracene. All data points indicate the energy of the first absorption peak.The bilayer films are excited with λ=460 nm light.

Finally, the dynamics of triplet exciton transfer are obtained usingtransient near-infrared photoluminescence spectroscopy. In Fig. 5awe observe that the photoluminescence decay of the nanocrystallayer is multi-exponential, with a longest lifetime of nearly 1 µs. Theintrinsic nanocrystal decay dynamics do not change significantly oncoating the nanocrystals with tetracene. However, when tetraceneis optically excited, the resulting nanocrystal transient showsenhanced emission at all times following the initial excitation. Inlight of our previous findings, we presume this delayed emission toresult from the influx of long-lived triplet excitons from tetracene.To obtain the time-dependent flux of excitons from the tetraceneto the nanocrystal layer and distinguish between direct opticalpumping of the nanocrystals and triplet transfer from tetracene,we deconvolve the photoluminescent impulse response of thenanocrystals from the total transient response (SupplementaryInformation). The resulting exciton flux is plotted in Fig. 5b. Itappears to be predominantly diffusion-limited, but with a rapidinitial component due to energy transfer from tetracene moleculesimmediately adjacent to the nanocrystals. In the inset of Fig. 5b,we show that the initial time constant of energy transfer to

100

Nan

ocry

stal

PL

(a.u

.)Ex

cito

n flu

x (a

.u.)

Time (ns)

Time (ns)

NC

PL (a

.u.)

10−1

10−2

10−3 1

0

0.1

20 40

10−4

10−3

10−4

10−5

0 500 1,000 1,500 2,000 2,500

Time (ns)

0 500 1,000 1,500 2,000 2,500

With tetracene

No tetracene

λex = 635 nm

b

a

Figure 5 | Monitoring triplet transfer in the time domain. a, Near-infrarednanocrystal fluorescence as a function of time for tetracene-coated 0.95 eV(λ= 1,308 nm) PbS–CA nanocrystal thin films pumped within the tetraceneabsorption band at λ=532 nm (red) and beyond the tetracene absorptionband at λ=635 nm (dark blue). We also plot the transient response of thepure nanocrystal film with no tetracene coating to pulsed excitation atλ=532 nm (light blue). b, The exciton flux from tetracene to thenanocrystal film obtained by deconvolving the photoluminescence impulseresponse of the nanocrystals without tetracene from thetetracene-sensitized transient response. Inset: The nanocrystal PL (NC PL)transient with direct optical pumping of the nanocrystals removed (orange),compared against the ‘with tetracene’ and ‘no tetracene’ transients from a.Emission stemming from transferred triplets dominates after 10 ns.

the nanocrystals is <10 ns. Although we cannot rule out thepossibility of some singlet exciton transfer at these early times,we note that transient measurements of tetracene fluorescenceshown in Supplementary Fig. 3 show little evidence of quenchingby nanocrystals within the first 20 ns, confirming the absence ofsignificant singlet exciton energy transfer. Beyond 20 ns, we observesignificant quenching of the emission from tetracene only underexcitation conditions which ensure a large population of free tripletsthat interact bimolecularly.

In conclusion, we find that careful control of thetetracene/nanocrystal interface results in near-unity transferefficiency of molecular triplet excitons with negligible oscillatorstrength. This demonstration of triplet energy transfer to colloidalnanocrystals suggests a clear path towards applications. Forexample, because semiconductor nanocrystals are capable ofefficient Förster energy transfer to bulk silicon, they could serveas a ‘relay’ to shunt excitonic energy from organic materials intoconventional solar cells. This could enhance power conversion

4 NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials© 2014 Macmillan Publishers Limited. All rights reserved.

NATUREMATERIALS DOI: 10.1038/NMAT4097 ARTICLESefficiency, because singlet fission can generate two triplets forevery absorbed photon30, potentially doubling the photocurrent atselect wavelengths6–8. Similarly, if the photoluminescence quantumyield of the colloidal nanocrystals is improved, the combinationof nanocrystals and singlet exciton fission should enable theincoherent emission of up to two photons from one absorbedphoton, with potential applications including phosphors forhigh-efficiency lighting.

Received 9 June 2014; accepted 1 September 2014;published online 5 October 2014

References1. Bardeen, C. J. The structure and dynamics of molecular excitons. Annu. Rev.

Phys. Chem. 65, 127–148 (2014).2. Köhler, A. & Bässler, H. What controls triplet exciton transfer in organic

semiconductors? J. Mater. Chem. 21, 4003–4011 (2011).3. Rao, A. et al. The role of spin in the kinetic control of recombination in organic

photovoltaics. Nature 500, 435–439 (2013).4. Baldo, M., O’Brien, D., Thompson, M. & Forrest, S. Excitonic singlet–triplet

ratio in a semiconducting organic thin film. Phys. Rev. B 60, 422–428 (1999).5. Segal, M., Baldo, M., Holmes, R., Forrest, S. & Soos, Z. Excitonic singlet–triplet

ratios in molecular and polymeric organic materials. Phys. Rev. B 68,075211 (2003).

6. Lee, J. et al. Singlet exciton fission photovoltaics. Acc. Chem. Res. 46,1300–1311 (2013).

7. Wilson, M. W. B., Rao, A., Ehrler, B. & Friend, R. H. Singlet exciton fission inpolycrystalline pentacene: From photophysics toward devices. Acc. Chem. Res.46, 1330–1338 (2013).

8. Smith, M. B. & Michl, J. Recent advances in singlet fission. Annu. Rev. Phys.Chem. 64, 361–386 (2013).

9. Thompson, N. J., Congreve, D. N., Goldberg, D., Menon, V. M. & Baldo, M. A.Slow light enhanced singlet exciton fission solar cells with a 126% yield ofelectrons per photon. Appl. Phys. Lett. 103, 263302 (2013).

10. Dexter, D. L. A theory of sensitized luminescence in solids. J. Chem. Phys. 21,836–850 (1953).

11. Baldo, M. A. et al.Highly efficient phosphorescent emission from organicelectroluminescent devices. Nature 395, 151–154 (1998).

12. Semonin, O. E. et al. Absolute photoluminescence quantum yields of IR-26Dye, PbS, and PbSe quantum dots. J. Phys. Chem. Lett. 1, 2445–2450 (2010).

13. Whitmore, P. M., Alivisatos, A. P. & Harris, C. B. Distance dependence ofelectronic energy transfer to semiconductor surfaces: n3π*Pyrazine/GaAs(110). Phys. Rev. Lett. 50, 1092–1094 (1983).

14. Hayashi, T., Castner, T. G. & Boyd, R. W. Quenching of molecular fluorescencenear the surface of a semiconductor. Chem. Phys. Lett. 94, 461–466 (1983).

15. Piland, G. B. et al. Dynamics of molecular excitons near a semiconductorsurface studied by fluorescence quenching of polycrystalline tetracene onsilicon. Chem. Phys. Lett. 601, 33–38 (2014).

16. Anni, M. et al. Förster energy transfer from blue-emitting polymers to colloidalCdSe/ZnS core shell quantum dots. Appl. Phys. Lett. 85, 4169–4171 (2004).

17. Anikeeva, P. O. et al. Photoluminescence of CdSe/ZnS core/shell quantum dotsenhanced by energy transfer from a phosphorescent donor. Chem. Phys. Lett.424, 120–125 (2006).

18. Reineke, S. & Baldo, M. Room temperature triplet state spectroscopy of organicsemiconductors. Sci. Rep. 4, 3797 (2014).

19. Burdett, J. J. & Bardeen, C. J. The dynamics of singlet fission in crystallinetetracene and covalent analogs. Acc. Chem. Res. 46, 1312–1320 (2013).

20. Wilson, M. W. B. et al. Temperature-independent singlet exciton fission intetracene. J. Am. Chem. Soc. 135, 16680–15588 (2013).

21. Tayebjee, M. J. Y., Clady, R. G. C. R. & Schmidt, T. W. The exciton dynamics intetracene thin films. Phys. Chem. Chem. Phys. 15, 14797–14805 (2013).

22. Yost, S. R. et al. A transferable model for singlet-fission kinetics. Nature Chem.6, 492–497 (2014).

23. Wu, T. C. et al. Singlet fission efficiency in tetracene-based organic solar cells.Appl. Phys. Lett. 104, 193901 (2014).

24. Burdett, J. J., Gosztola, D. & Bardeen, C. J. The dependence of singlet excitonrelaxation on excitation density and temperature in polycrystalline tetracenethin films: Kinetic evidence for a dark intermediate state and implications forsinglet fission. J. Chem. Phys. 135, 214508 (2011).

25. Ehrler, B. et al. In situmeasurement of exciton energy in hybrid singlet-fissionsolar cells. Nature Commun. 3, 1019 (2012).

26. Jadhav, P. J. et al. Triplet exciton dissociation in singlet exciton fissionphotovoltaics. Adv. Mater. 24, 6169–6174 (2012).

27. Tomkiewicz, Y., Groff, R. P. & Avakian, P. Spectroscopic approach to energeticsof exciton fission and fusion in tetracene crystals. J. Chem. Phys. 54,4504–4507 (1971).

28. Zarghami, M. et al. p-Type PbSe and PbS quantum dot solids prepared withshort-chain acids and diacids. ACS Nano 4, 2475–2485 (2010).

29. Zhu, H., Song, N. & Lian, T. Controlling charge separation and recombinationrates in CdSe/ZnS type I core-shell quantum dots by shell thicknesses. J. Am.Chem. Soc. 132, 15038–15045 (2010).

30. Merrifield, R. E., Avakian, P., Groff, R. P., Physics, C. & June, L. Fission ofsinglet excitons into pairs of triplet excitons in tetracene crystals. Chem. Phys.Lett. 3, 386–388 (1969).

31. Congreve, D. N. et al. External quantum efficiency above 100% in asinglet-exciton-fission-based organic photovoltaic cell. Science 340,334–337 (2013).

32. Maksym, P. A. & Chakraborty, T. Quantum dots in a magnetic field: Role ofelectron–electron interactions. Phys. Rev. Lett. 65, 108–111 (1990).

33. Moreels, I. et al. Size-dependent optical properties of colloidal PbS quantumdots. ACS Nano 3, 3023–3030 (2009).

34. Jasieniak, J., Califano, M. &Watkins, S. E. Size-dependent valence andconduction band-edge energies of semiconductor nanocrystals. ACS Nano 5,5888–5902 (2011).

35. Graham, K. R. et al. Re-evaluating the role of sterics and electronic coupling indetermining the open-circuit voltage of organic solar cells. Adv. Mater. 25,6076–6082 (2013).

AcknowledgementsThis work was supported as part of the Center for Excitonics, an Energy FrontierResearch Center funded by the US Department of Energy, Office of Science, Office ofBasic Energy Sciences under Award Number DE-SC0001088 (MIT). D.N.C. was partiallysupported by the National Science Foundation Graduate Research Fellowship underGrant No. 1122374. J.M.S. was partially supported under FA9550-11-C-0028 awarded bythe Department of Defense, Air Force Office of Scientific Research, National DefenseScience and Engineering Graduate (NDSEG) Fellowship, 32 CFR 168a, and by the USArmy through the Institute for Soldier Nanotechnology (W911NF-13-D-0001). Theauthors thank O. Chen, C-H. Chuang, D. D. Grinolds, D. K. Harris and G. W. Hwang forstimulating discussions and their considerable help with nanocrystal synthesis, as well asR. Murphy for help in obtaining access to equipment.

Author contributionsN.J.T., D.N.C. and M.Wu fabricated the samples and made steady-state measurements.P.R.B. performed UPS and AFMmeasurements. M.W.B.W. and T.S.B. made transientmeasurements and M.W.B.W. prepared nanocrystal solutions for sample fabrication.J.M.S., M.W.B.W. and P.R.B. fabricated the nanocrystals. N.G. and M.Welborn simulatedthe nanocrystal structure. The project was conceived by N.J.T. and M.A.B. All authorsdiscussed the results and commented on the manuscript.

Additional informationSupplementary information is available in the online version of the paper. Reprints andpermissions information is available online at www.nature.com/reprints.Correspondence and requests for materials should be addressed to M.A.B. or M.G.B.

Competing financial interestsMIT has filed a provisional application for patent based on this technology that namesD.N.C., N.J.T., M.W.B.W., M.Wu, M.A.B., V.B. and M.G.B. as inventors.

NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials 5© 2014 Macmillan Publishers Limited. All rights reserved.