Effect of Diels–Alder Reaction in C60-Tetracene Photovoltaic Devices

Aug 30, 2016 - ABSTRACT: Developing organic photovoltaic materials systems requires a detailed understanding of the heterojunction interface, as it is...
0 downloads 16 Views 1020KB Size
Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Communication 60

Effect of Diels-Alder Reaction in C -Tetracene Photovoltaic Devices Andrew P Proudian, Matthew B Jaskot, Christelle Lyiza, David R. Diercks, Brian P Gorman, and Jeramy D Zimmerman Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.6b02238 • Publication Date (Web): 30 Aug 2016 Downloaded from http://pubs.acs.org on September 7, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Effect of Diels-Alder Reaction in C60-Tetracene Photovoltaic Devices Andrew P. Proudian,†,§ Matthew B. Jaskot,‡,§ Christelle Lyiza,‡ David R. Diercks,¶,‡ Brian P. Gorman,¶,‡ and Jeramy D. Zimmerman∗,†,‡ Department of Physics, Colorado School of Mines, Golden, CO 80401, Materials Science Program, Colorado School of Mines, Golden, CO 80401, and Department of Metallurgy and Materials Engineering, Colorado School of Mines, Golden, CO 80401 E-mail: [email protected]

Abstract Developing organic photovoltaic materials systems requires a detailed understanding of the heterojunction interface, as it is the foundation for photovoltaic device performance. The bilayer fullerene/acene system is one of the most studied models for testing our understanding of this interface. We demonstrate that the fullerene and acene molecules chemically react at the heterojunction interface, creating a partial monolayer of a Diels-Alder cycloadduct species. Furthermore, we show that the reaction occurs during standard deposition conditions and that thermal annealing increases the concentration of the cycloadduct. The cycloaddition reaction reduces the number of sites available at the interface for charge transfer exciton recombination and decreases the charge transfer state reorganization energy, increasing the open circuit voltage. ∗

To whom correspondence should be addressed Department of Physics, Colorado School of Mines, Golden, CO 80401 ‡ Materials Science Program, Colorado School of Mines, Golden, CO 80401 ¶ Department of Metallurgy and Materials Engineering, Colorado School of Mines, Golden, CO 80401 § Contributed equally to this work †

1

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 19

The sub-monolayer quantity of the cycloadduct renders it difficult to identify with conventional characterization techniques; we use atom probe tomography to overcome this limitation while also measuring the spatial distribution of each chemical species.

Keywords acenes; fullerenes; donor/acceptor interface; cycloaddition; organic photovoltaics; atom probe tomography Organic electronic technologies, encompassing devices such as organic photovoltaics (OPVs) and organic light-emitting diodes, hold the promise of mass-manufacturability with significantly lower embedded energy costs and higher application-specific tunability compared to their conventional inorganic counterparts. Understanding interfaces in OPV systems is crucial for improving device performance and developing viable OPV material systems. Here, we show that the acene/fullerene donor-acceptor (D-A) interface chemically reacts, which is of particular interest because the acene/fullerene bilayer is a model system for testing our understanding of bilayer OPVs and acenes are used extensively for singlet fission. 1 A fullerene molecule can make its closest contact with the edge, face, or end of an acene molecule, or any position between these extremes. Models of acene/C60 contact configurations predict that the intermolecular coupling, which determines charge transfer (CT) state recombination and dissociation rates, is significantly different for each configuration; 2 furthermore, face-on interfaces are likely morphologically unstable with a high probability of the acene and fullerene mixing. 3 Meanwhile, experimental work has struggled to directly confirm the effects of interface orientation on device parameters. 4 At the same time, efforts to understand how steric hindrance affects device parameters such as open circuit voltage (VOC ) in tetracene (Tc)/C60 and rubrene/C60 based devices remain an area of study. 5 Part of the reason for the difficulty in experimentally reproducing the model results is that the high electron affinity of the fullerene cage makes it reactive with electron-rich acene molecules through Diels-Alder cycloaddition. 6–9 This reaction is commonly used for func2

ACS Paragon Plus Environment

Page 3 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

tionalization of C60 with polycyclic aromatic hydrocarbons. 9–12 Of particular relevance to this work is the well understood [4+2] cycloaddition of C60 to short linear acenes (i.e. anthracene, tetracene and pentacene) in solution. 8,9,12–19 The reaction of Tc and C60 produces a Diels-Alder cycloadduct (DAc), which has been shown to occur in the solid state during vibration milling, 20 and is thermally stable up to 300 ◦C. 8 This reaction is most likely to proceed when the π-system of the acene is tangent to the fullerene, obscuring the expected effects of interface orientation. 21 The presence of the DAc at acene-fullerene interfaces was predicted by J. E. Anthony, 22 but due to the difficulty of detection, only recently has experimental evidence for a Diels-Alder cycloaddition reaction between pentacene and C60 bilayers been shown. 23 Here, using atom probe tomography (APT) based mass spectrometry and Fourier transform infrared (FTIR) spectroscopy, we unequivocally show that Tc/C60 interfaces undergo a Diels-Alder cycloaddition reaction creating DAc species, even in the typical end-on interface configuration most commonly found in devices. 23,24 Furthermore, we show that this reaction lowers the number of sites at the interface available for CT exciton recombination, thereby reducing the losses in open circuit voltage (VOC ) in our bilayer cells by as much as 140 mV relative to typical organic photovoltaic systems. 5 Controlling the D-A interface to minimize CT exciton recombination while maintaining efficient charge separation can serve as an important device engineering tool, with the potential for generalization to other organic material systems. Because the DAc is difficult to detect using conventional techniques, we employ APT in the local electrode atom probe configuration. APT is well-suited for investigations of the interface due to its high mass-resolving power and atomic-level spatial resolution. 25,26 Application of APT to organic semiconductor materials has so far been limited to the poly(3hexylthiophene-2,5-diyl) (P3HT):C60 system, in which the polymer, P3HT, was found to fragment at numerous locations, preventing further analysis. 27 However, in the Tc/C60 system, both constituents evaporate as whole molecules, obviating this fragmentation problem

3

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and permitting subsequent analysis. In APT, a large voltage is applied to a sample prepared with a sub-micrometer radius of curvature, generating a large local electric field. The sample, held at cryogenic temperatures, is excited with a laser pulse of adequate intensity to induce field evaporation of a single molecule on the surface. The ionized molecule is then accelerated by the electric field towards a two-dimensional position- and time-sensitive detector. This process is repeated until the desired thickness of sample has been evaporated and analyzed. The mass-to-charge ratio of each ion is extracted through the time of flight and the initial positions of the ions can be reconstructed in three dimensions. 25,26,28 APT testing and reconstruction parameters are provided in the Methods section. Samples for APT analysis were prepared using vacuum thermal evaporation (VTE) to deposit Tc and C60 directly onto tungsten probe tips with ∼500 nm radii and ∼30◦ shank angles. Compared to typical surface feature sizes and device thicknesses of ∼100 nm, this tip radius is sufficiently large that significant differences in interface morphology between these samples and those on planar substrates are not expected. APT mass spectra are shown in Figure 1 for a bilayer sample of a 40 nm layer of C60 followed by a 60 nm layer of Tc and a co-deposited sample of Tc:C60 in a 1:1 volume blend with a total layer thickness of 100 nm. The bilayer sample for Figures 1 and 2 was deposited in the reverse order typical of devices, as collecting APT data for a high evaporation-field material (C60 ) on top of a low evaporation-field material (Tc) is more difficult than the reverse. The APT mass-spectrum (see Figure 1) contains peaks corresponding to the masses of singly and doubly oxidized Tc, C60 , the covalently bonded DAc, and singly and doubly oxidized species of each; an extended table of peak identifications is provided in the Supporting Information (Table S1). The peak at 948 Da, the sum of the masses of C60 and Tc, indicates that the DAc species is present in our samples. Bilayer samples have a sub-monolayer of this species (∼0.36 % of molecules in the entire film, as detected in APT), making the absorption signal difficult to confidently discern using established methods (e.g. FTIR). APT was also

4

ACS Paragon Plus Environment

Page 4 of 19

Page 5 of 19

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

performed on samples with C60 deposited on Tc, with similar concentrations of the DAc observed; however, the effects of the substantially higher evaporation field of the C60 on top of the Tc prevents faithful tomographic reconstruction. To increase the interfacial area and number of DAc molecules present, the Tc and C60 were co-deposited (1:1 by volume), increasing the DAc concentration to ∼11 % of the total ion count. FTIR measurements on the co-deposited film confirm the presence of vibrational modes previously attributed to the Diels-Alder cycloadduct (see Supporting Information, Figure S1), eliminating the possibility that van der Waals bonded pairs of C60 and Tc are co-evaporating during APT analysis. 8 Detection of the submonolayer of the DAc in the bilayer sample demonstrates the impressive sensitivity of APT to observe chemical changes and impurities in molecular organic systems. Spatial information is also provided by APT: the concentration profile (Figure 2a) shows the DAc is confined to the interface, while the distribution of the DAc in the plane of the interface (Figure 2b) shows no obvious structure. FTIR measurements on bilayer Tc/C60 samples, both as-deposited and annealed at 75, 100, and 125 ◦C, are shown in Figure 3. The increased magnitude of the strength of the vibrational mode at 700 cm−1 (see Figure 3, inset) indicates the amount of DAc increases on annealing. Furthermore, samples co-deposited (1:1 volume ratio of Tc:C60 ) at substrate temperatures of −75 ◦C show an increased DAc concentration, which indicates that the adduct forms at room temperature or below, making avoidance of this reaction difficult in VTE (see Supporting Information, Figure S2). Because the D-A heterojunction interface determines much of the performance of OPVs and the DAc is known to have a lowest unoccupied molecular orbital (LUMO) energy ∼0.16 eV shallower than C60 (see Figure 4a, where the highest occupied molecular orbital (HOMO) value is estimated), 20 the presence of the DAc at the interface is expected to affect the device properties. Figure 4b,c shows a device stack and current-voltage (J-V ) data respectively for bilayer devices (glass / ITO / 20 nm MoOx / 100 nm Tc / 60 nm C60 / 8 nm BPhen / 100 nm Ag ) deposited at a substrate temperature of 25 ◦C that were tested as-

6

ACS Paragon Plus Environment

Page 6 of 19

Page 7 of 19

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment

Nano Letters

(a)

(b)

−1.8 eV

Tc

−3.4 eV

−3.5 eV

DAc

C60

−6.0 eV

−6.2 eV

Ag (100nm) BPhen (8nm) C60 (60nm) Tc (100nm) MoO3 (20nm) ITO Glass

−5.1 eV

(c)

Current Density (mA cm2)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 19

Voltage (V) 0.2 0.4 0.6

0 0.5

0.8

bilayer interlayer ● ● As dep. 75°C 100°C 125°C (N/A)

0.0

−0.5 −1.0 ●















−1.5 −2.0

















−2.5

Figure 4: (a) Schematic of the materials stack used in devices. (b) Energy diagram showing the highest occupied molecular (HOMO) and lowest unoccupied molecular orbital (LUMO) of the Tc, DAc, and C60 . 20,29–31 DAc LUMO value is estimated. (c) Representative current density-voltage (J-V ) curves for Tc/C60 bilayer devices under 1 sun illumination (AM1.5), tested as-deposited and after annealing at 75, 100, and 125 ◦C for 30 min (solid lines). Equivalent devices with a 10 nm Tc:C60 interlayer are also shown in dashed lines with symbols and colors that correspond to annealing conditions for the bilayer devices.

8

ACS Paragon Plus Environment

Page 9 of 19

(a)

Energy (eV)

1.0

2.0

3.0

4.0 30 20

−1

10

10

EQE (%)

EQE (arb. units)

100

exp

As dep. 75°C 100°C 125°C

ECT fit

10−2

10−3 1.2

1.3

1.4 1.5 1.6 Energy (eV)

(b) 1.3

1.8

exp

ECT

1.1 1.0

VOC ●

● ●

320



280



240

160

120

40

0.8

200

● ●

As deposited 100°C



80

0.9 0.7

1.7

● ● ● ● ● ● ●

1.2

0

qVOC (V), ECT (eV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Temperature (K)

Figure 5: (a) Semilog external quantum efficiency (EQE) curves, normalized to the DAc absorption peak at 1.72 eV, for representative devices as-deposited and annealed at 75, 100, and 125 ◦C; the dashed line shows a representative fit to the charge-transfer (CT) state according to Equation 2. (inset) Linear, non-normalized, EQE (%). (b) Measured qVOC and exp as functions of temperature with linear extrapolations to 0 K. ECT

9

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 19

work on Tc/C60 . 4,5,32–34 We do, however, observe the commonly reported results of VOC ≈ 0.55 V for samples exposed to air during J-V testing; 4,33 furthermore, FTIR and J-V results suggest that the samples previously exposed to air do not undergo additional DAc formation on thermal annealing. Short circuit current (JSC ) also increases relative to an as-deposited device for the 75 and 100 ◦C annealed bilayer devices, as previously observed by Shao et al.; this increase in JSC was explained by a morphology change leading to higher hole mobility in the Tc layer. 33 To understand the changes in VOC on annealing, we combine the formalisms of Giebink et al. and Burke et al., where VOC for an ideal two-component OPV is given by: 35,36

VOC

  exp ∗ ∗ ECT nkB T kCT NHOMO NLUMO ≈ − ln q q ζmax JX /a0

exp exp = ECT − is the apparent CT exciton energy, ECT Here, ECT

2 σCT , 2kB T

(1)

including both the

temperature-independent CT exciton energy, ECT , and a term containing the CT disorder width, σCT . 36 q is the electron charge, n is the diode ideality factor, kB is the Boltzmann constant, T is temperature, kCT is the CT exciton recombination rate, ζmax is the maximum possible CT exciton concentration, JX is the exciton current reaching the interface, a0 defines the spatial extent of the CT state, and N ∗HOMO and N ∗LUMO are the number of donor HOMO and acceptor LUMO states or their respective defect states at the interface. The origin of the change in VOC can be elucidated through measurements of the CT state energy using external quantum efficiency (EQE) measurements. 37–40 The CT state EQE peak Table 1: Values of open circuit voltage (VOC ), short circuit current (JSC ), fill factor (F F ), power conversion efficiency (P CE), ideality (n), and number of samples (num.) for bilayer Tc/C60 and interlayer devices both as deposited and subjected to post-deposition annealing at 75, 100, and 125 ◦C for 30 min. Errors reported are one standard deviation for samples grown in a single evaporation run. Anneal VOC (mV) JSC (mA/cm2 ) FF P CE (%) n num.

as-deposited 765.6 ± 0.7 2.09 ± 0.03 65.3 ± 0.6 1.04 ± 0.01 1.67 ± 0.04 7

Bilayer (25 ◦C) 75 ◦C 100 ◦C 787.0 ± 8 805.6 ± 0.8 2.29 ± 0.10 2.38 ± 0.10 64.0 ± 3.0 65.0 ± 0.6 1.17 ± 0.09 1.24 ± 0.06 1.65 ± 0.10 1.48 ± 0.16 15 6

125 ◦C 770.0 ± 40.0 1.63 ± 0.10 52.0 ± 5.0 0.65 ± 0.11 1.32 ± 0.47 10

10

Bilayer (−30 ◦C) as-deposited 768.0 ± 0.8 1.31 ±0.06 56.8 ± 0.4 0.57 ± 0.03 1.68 ± 0.13 3

ACS Paragon Plus Environment

Interlayer (−30 ◦C) as-deposited 75 ◦C 100 ◦C 801.8 ± 1.3 840.1 859.7 ± 0.6 1.21 ± 0.03 1.25 1.19 ± 0.03 51.3 ± 0.6 48.1 47.1 ± 0.3 0.50 ± 0.02 0.51 0.48 ± 0.01 1.51 ± 0.05 1.29 1.30 ± 0.08 6 1 7

Page 11 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

is described by: 2,5,36,37,40,41 f EQE ∝ √ exp E 4πλexp kB T



exp + λexp − E)2 −(ECT 4λexp kB T

where λexp is the apparent reorganization energy, λexp = λ0 +



2 σCT , 2kB T

(2)

including both a

temperature-independent reorganization energy term, λ0 , and an energetic disorder term. The electronic coupling term, f, is proportional to the number of available interface sites with a given CT state energy, and the square of the electronic coupling matrix element beexp , f, and λexp are determined by fitting tween the ground state and the excited CT state. ECT

Equation 2 to the EQE response, as shown in Figure 5a. 5,37 We observe that the apparent energy of the CT state peak does not change with annealing temperature (±0.008 eV across 31 samples measured at room temperature), suggesting that the second term in Equation 1 is responsible for the change in VOC . The measurements of EQE and VOC at temperatures from 200 to 320 K shown in Figure 5b confirm that the second term in Equation 1 is responsible for the observed increase in VOC ; the linear fits to the qVOC and ECT data coincide at T = 0 K (i.e. the 95 % confidence intervals of the value at T = 0 K overlap), as predicted by Equation 1. Fitting  exp h i  ∗ ∗ ECT kPPr NHOMO NLUMO nkB T the data in Figure 5b to the quantity − V ln results in ≈ OC q q ζmax JX /a0 −1.88 ± 0.06 and −1.56 ± 0.04 mV K−1 for the as deposited and 100 ◦C annealed bilayer deexp /q is unchanged for these devices, the difference in these vices, respectively; because ECT

fits is directly related to the difference in VOC . The mean room temperature ideality factors for these devices (n =1.67±0.04 and 1.48±0.16, from 7 and 6 devices, respectively) explains 70 ± 7 % of the change in VOC . This improvement in ideality on annealing should occur because face-on Tc/C60 interfaces have the smallest CT-exciton energy 42 and are the most reactive; 21 therefore, formation of the DAc should remove the deepest trap sites from the density of states, reducing the effective trap temperature, or altering the reorganization enσ2

), and improving the ideality, and thus the VOC . 35,43–45 ergy (i.e. decreasing λexp = λ0 + 2kCT BT

11

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 19

Indeed, we do observe that the CT state narrows from λexp = 134 ± 20 meV for the as deposited devices to 100 ± 12, 99 ± 7, and 95 ± 8 meV for the devices annealed at 75, 100, and exp 125 ◦C, respectively. Furthermore, we extract σCT from the temperature dependence of ECT

(σCT = 64 ± 12 and 67 ± 11 meV for the as-deposited and 100 ◦C annealed devices shown in Figure 5b), respectively, suggesting that the it is the reorganization energy, λ, that decreases on annealing. Vandewal, et al. show, using an alternate form of the equation for VOC , that  exp  ECT exp − V − λexp ), so that a reduction in the reorganization energy is indeed ∝ ln (ECT OC q expected to increase VOC . 37

The remainder of the difference in



exp ECT q

− VOC



can be explained by a reduction in the

∗ ∗ number of sites that can support CT excitons (e.g. NHOMO and NLUMO ), and therefore

the net CT recombination rate, in the annealed device. This reduction in the number of sites available for CT excitons is directly proportional to the reduction in CT state EQE shown in Figure 5a. Together with the FTIR data shown in Figure 3, this indicates that the Diels-Alder cycloaddition reaction proceeds along the interface, reducing the number of sites available that can support CT excitons. This reduction in the EQE of CT state absorption without change in energy is consistent with the results of Vandewal et al., in which the increased VOC was ascribed to reduced D-A interfacial area in bulk heterojunction OPVs. 46 Here, we calculate the percentage of the EQE at 1.35 eV (CT state absorption, EQECT ) relative to 1.72 eV (a known DAc absorption edge, 8 EQEDAc ) as a quantitative metric of the number of sites available to support CT excitons. For the as-deposited bilayer devices, this percentage is 45.7 ± 11.6 %, and decreases progressively with annealing temperature: 19.7 ± 4.2, 15.8 ± 2.0, and 12.2 ± 1.3 % for annealing temperatures of 75, 100, and 125 ◦C, respectively. This percentage decreases further in the interlayer devices: 2.5, 2.2, 1.1, and 0.4 % for as-deposited, 75, 100, and 125 ◦C annealed devices respectively (only one EQE data point was taken for each interlayer device), indicating even lower concentrations of available CT sites.

12

ACS Paragon Plus Environment

Page 13 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

The devices annealed at 125 ◦C and those containing a co-deposited Tc:C60 “interlayer” with higher concentrations of the DAc do show further enhanced VOC , but at the expense of JSC , FF, and power conversion efficiency. In this work, we use APT, FTIR, temperature-dependent EQE, and illuminated J-V data to show that the Tc/C60 interface is chemically unstable and undergoes an Diels-Alder cycloaddition reaction, altering the density of states that can support CT excitons and their reorganization energy, thereby increasing VOC towards the limit of ECT /q. Manipulating the density of CT states at the D-A heterojunction interface may serve as a valuable tool which can be generalized to other organic material systems. Here, our results show VOC values for bilayer cells up to 140 mV higher than the qVOC ≈ ECT −600 ± 70 meV trend described by Graham et al., suggesting that control of CT exciton recombination is a viable strategy to exceed this empirical upper limit. 5 We also show that APT is valuable for the study of small-molecule semiconducting materials, providing three-dimensional chemical information at the nanometer scale using molecular mass to distinguish chemical species. Using APT, we directly detect an interfacial solid-state Diels-Alder cycloaddition product at a threshold below a confident detection threshold for FTIR, and reconstruct the spatial distribution of this product in an OPV device to confirm it is confined to the D-A interface.

Methods APT and FTIR Sample Preparation. Thin films of bilayer and blended fullerene C60 (MER Corp., sublimed grade, purified again with a vacuum gradient sublimation) and tetracene (Sigma Aldrich, sublimed grade, used as received) were prepared by vacuum thermal evaporation on silicon wafer substrates (FTIR) or 500 nm Micromanipulator brand tungsten tips (APT) to a final thickness of 100 nm. The film thickness and ratio of Tc:C60 (1:1 by volume) was controlled by the deposition rate, monitored by a quartz crystal oscillator. OPV Device Fabrication and Testing. OPV devices (3 mm2 area) were deposited via

13

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

thermal evaporation on indium tin oxide (ITO)-coated glass substrates (Bay View Optics,