Multichromophoric energy sensitization of C₆₀ for organic photovoltaics

Organic photovoltaic (OPV) cells have great promise to become a viable alternative to existing solar cell technologies, dominated by silicon-based devices. Nevertheless, OPVs currently produce significantly less photocurrent than their inorganic counterparts because the absorption bandwidth of the active layer is limited. As absorption of the donor is shifted further into the near-infrared (NIR) to enhance photon collection from the solar spectrum,^1–8 it is increasingly necessary to utilize additional absorbing materials to ensure broad spectral coverage.^9–14 This has led to a recent surge in research related to tandem^15–17 and ternary blended bulk-heterojunction^18,19 solar cells where a divide-and-conquer strategy has been employed in which separate regions of the solar spectrum are absorbed by different materials. However, in ternary blended devices, the processes for photocurrent generation can become quite complex as multiple energetic pathways from incident photon to collected electron are possible¹⁹ and in tandem cells current matching requires significant effort.¹⁶ 

To overcome the complexity of ternary blended bulk-heterojunctions and tandem systems and still achieve an increase in photon collection, Trinh et al.¹¹ developed a sensitization strategy to improve the absorption efficiency of a C₆₀-based acceptor layer. In this sensitization approach, energy absorbed by one material, the sensitizer, is transferred to the host, in this case C₆₀, which is responsible for exciton transport, charge separation, and electron conduction. In this scheme, the sensitizers need to be carefully designed to ensure efficient energy transfer and maintain electron conductivity in the blended host–sensitizer layer. In the previous work, a chlorinated zinc dipyrrin compound, ZCl, was designed and established as an excellent sensitizer for the C₆₀ acceptor layer.

In this work, we extend the sensitization approach to include multiple sensitizers absorbing a broader range of the solar spectrum to further extend the absorption of the acceptor layer. We demonstrate the improved performance resulting from the inclusion of multiple sensitizers blended with a single host in donor/sensitized C₆₀ acceptor devices. Two energy sensitizers, ZCl¹¹ and hexachloro boron subphthalocyanine (Cl₆SubPc),²⁰ with intense absorption, are utilized to harvest photons in the visible portion of the solar spectrum and transfer energy to C₆₀. In devices, the simultaneous energy sensitization of C₆₀ with both ZCl and Cl₆SubPc is demonstrated. Through the inclusion of multiple sensitizers, we are able to achieve an increase in photocurrent of 30% from 6.5 mA/cm² to 8.6 mA/cm² for the sensitized device. The power conversion efficiency is increased from 3.8% for the standard device to 4.7% for the sensitized device.

The molecular structures and thin film extinction spectra of the active layer materials are presented in Figure 1. The C₆₀ film shows strong absorption in the UV due to allowed transitions and another feature between λ = 400 nm and 550 nm due to an intermolecular charge transfer transition.²¹ ZCl has an extremely intense absorption between λ = 450 nm and 575 nm with a thin film extinction (α) as high as 4 × 10⁵ cm⁻¹. Cl₆SubPc absorbs between λ = 500 nm and 650 nm with α as large as 2.5 × 10⁵ cm⁻¹. The absorption of the sensitizers is almost an order of magnitude more intense than C₆₀ in the regions where they absorb. The blended C₆₀:ZCl:Cl₆SubPc (2:1:1 by volume) film shows contributions from all three components and has significantly more absorbance than pure C₆₀ between λ = 500 nm and 670 nm. (2,4-bis[4–(N,N-diphenylamino)–2,6-dihydroxyphenyl] squaraine) (DPSQ),^3,9,22,23 a typical NIR absorbing donor, absorbs between λ = 600 nm and 800 nm. The donor absorption in conjunction with that of the codeposited acceptor film should extend the photoresponse of a sensitized device uniformly from the UV to the NIR.

In order to efficiently funnel excitons from the sensitizers to C₆₀ and to ensure electron transport is also occurring via C₆₀, knowledge of the excited state energies and carrier transport levels is required. The singlet and triplet state energies and oxidation and reduction potentials of ZCl,¹¹ Cl₆SubPc,²⁰ and C₆₀²¹ are given in Figure 2. The arrows in Figure 2(a) outline a schematic of possible pathways for energy transfer from ZCl and Cl₆SubPc to C₆₀. Both ZCl and Cl₆SubPc should function as sensitizers because their singlet and triplet energies are greater than or within k_BT of the values for C₆₀. Additionally, if electron transfer were to occur from the sensitizer to C₆₀, it will result in the formation of a CT state with energy equal to the difference between the oxidation potential of the sensitizer and the reduction potential of C₆₀ less the coulombic stabilization provided by the oxidized and reduced species. Assuming a coulombic stabilization of 0.3 eV,¹¹ the ZCl/C₆₀ and Cl₆SubPc/C₆₀ CTs will have energies of 1.98 eV and 1.55 eV, respectively. These CT states will be higher in energy than the triplet of C₆₀ resulting in recombination to form a triplet on C₆₀ and the net effect of energy transfer from the sensitizer to C₆₀. These considerations will guarantee that any excitons generated on the sensitizers will be transferred to C₆₀ and not trapped on the sensitizer. Finally, the reduction potentials of the sensitizers are more negative than that of C₆₀, ensuring that electrons are conducted out of the device efficiently via C₆₀. The conductivity of C₆₀ in a blended film has been demonstrated previously in C₆₀:Bathocuproine (BCP) blends, where the electron conductivity of the blended film is equivalent to that of neat C₆₀ up to 50% BCP by volume.²³ 

To investigate energy transfer between the sensitizers (ZCl and Cl₆SubPc) and C₆₀, thin film photoluminescence quenching experiments were performed utilizing C₆₀ as a quencher. Measurements were performed on films of the sensitizers blended with either the wide gap material BCP or C₆₀ at volume concentrations of 50% to replicate the compositions used in devices. The ZCl and Cl₆SubPc films blended with BCP are emissive and the spectra can be seen in Figure 2(c). The substitution of C₆₀ for BCP results in substantial luminescence quenching due to energy transfer to C₆₀. The luminescence quenching efficiency (⁠$θ$⁠) for C₆₀ can be calculated from the ratio of the integrated emission spectra through the following equation:

where $PLC60$ is the integrated luminescence spectra for the blend with C₆₀ and $PLBCP$ is the integrated luminescence spectra for the blend with BCP. C₆₀ quenches > 95% of ZCl emission and >80% of Cl₆SubPc, suggesting that energy transfer to C₆₀ occurs efficiently.

As the excitons generated on the sensitizer are singlets, energy transfer will occur via Förster resonant energy transfer (FRET),²⁴ where the rate of energy transfer in three dimensions, $kF$⁠, is given by

where $R$ is the intermolecular separation distance and $R0$ is the Förster radius (the distance at which the rate of energy transfer equals the rate of radiative decay). Inspection of the Förster equation reveals the rate of energy transfer is fastest for a donor-acceptor system separated by a small distance. Based on this fact, the sensitization architecture has been designed to ensure maximum energy transfer efficiency in the solid state. This is achieved by codepositing the materials in blends containing 50% C₆₀ by volume in order to obtain minimum separation between the sensitizer and C₆₀ instead of the lamellar cascade approach which has been applied previously.²⁵ The C₆₀:sensitizer ratio was chosen to maintain the high electron conductivity of C₆₀ in the blend,²³ while simultaneously maximizing the absorption of the sensitizers.

The performance of the multichromophoric sensitized device was optimized and compared to both a singly sensitized device and a control with pure C₆₀. For the purpose of probing solely the performance of the acceptor layer, the wide energy gap donor N,N′-di-[(1-naphthyl)-N,N′-diphenyl]-1,1′-biphenyl)–4,4′-diamine (NPD) was used. The control device had the structure ITO/MoO₃(10 nm)/NPD(11 nm)/C₆₀(50 nm)/BCP(10 nm)/Al, the singly sensitized device has the structure ITO/MoO₃(10 nm)/NPD(11 nm)/C₆₀(15 nm)/C₆₀:ZCl(1:1 50 nm)/BCP(10 nm)/Al, and the multichromophoric devices had the structure ITO/MoO₃(10 nm)/NPD(11 nm)/C₆₀(15 nm)/C₆₀:ZCl:Cl₆SubPc(2:1:1 x nm)/BCP(10 nm)/Al, where x was varied from 30 to 90 nm in 20 nm increments. The neat layer of C₆₀ is placed between the sensitized layer and the donor/acceptor (D/A) interface in order to ensure that all excitons that reach the D/A interface are transported through C₆₀ and to guarantee charge transfer occurs between NPD/C₆₀ eliminating any changes in the kinetics or thermodynamics of charge separation and recombination due to the presence of the sensitizers.²⁶ The device performance was probed as a function of the thickness of the blended C₆₀:ZCl:Cl₆SubPc layer and the results are summarized in Figure 3(a). Compared to the control device with 50 nm C₆₀, the devices containing the C₆₀:ZCl:Cl₆SubPc layer all exhibited an increase in photocurrent without substantial change in V_OC or FF. The increase reaches its maximum between a thickness of 50 nm and 70 nm for the C₆₀:ZCl:Cl₆SubPc layer where the photocurrent plateaus at 4.2 ± 0.1 mA/cm², 1.2 mA/cm² greater than the control with pure C₆₀. Figure 3(b) summarizes the performance of the optimized devices containing C₆₀, C₆₀:ZCl, and C₆₀:ZCl:Cl₆SubPc and Figure 3(c) compares the external quantum efficiencies (EQEs). The sensitized devices both show increased J_SC compared to the C₆₀ reference with the C₆₀:ZCl:Cl₆SubPc device outperforming the C₆₀:ZCl device. From the EQE, it is apparent that the increase in photocurrent is due to contribution from the sensitizers. For the C₆₀:ZCl device, contribution from ZCl is evident between λ = 500 nm and 600 nm. In the C₆₀:ZCl:Cl₆SubPc device, the ZCl enhancement is evident between λ = 500 nm and 575 nm, while the Cl₆SubPc signal can be seen between λ = 575 nm and 650 nm. These data clearly show that the sensitizers contribute to photocurrent production without negatively impacting other device characteristics and that the inclusion of multiple sensitizers allows the absorption of the acceptor layer to be shifted further to longer wavelengths. However, while the application of the sensitizers with a transparent donor shows the feasibility of multichromophoric sensitization, the utility of sensitization can be shown in devices with an intensely absorbing donor.

The continuing development of NIR absorbing donors has led to a gap in the absorption of devices containing fullerenes. This causes an EQE droop which is typified in DPSQ/C₆₀ devices^3,9,22,23 where the C₆₀ photoresponse ends at λ = 550 nm and the DPSQ photoresponse is strongest between λ = 700 nm and 800 nm. Devices were fabricated to further illustrate the impact of sensitization on device performance with a NIR donor (see device structures in Figure 4(a)). The current–voltage (J–V) and EQE are shown in Figures 4(b) and 4(c) and performance parameters are summarized in Table I. The V_oc of the devices remains unchanged at 0.92 ± 0.01 V for both the reference (DPSQ) and sensitized (DPSQ(s)) devices. The FF of the reference and sensitized devices are 0.60 ± 0.02 and 0.56 ± 0.02, respectively. Sensitization leads to a marked increase in photocurrent from Jsc = 6.5 ± 0.2 mA/cm² to 8.6 ± 0.2 mA/cm² for DPSQ and DPSQ(s), respectively. The EQE reveals that the increase in response is due to both ZCl and Cl₆SubPc sensitization. Compared to the previous work with a single sensitizer, the inclusion of multiple sensitizers results in a larger increase in photocurrent, 2.1 mA/cm² for the multichromophoric device compared to 1.2 mA/cm² for sensitization with a single chromophore.¹¹ The sensitizers completely fill the droop in absorption, resulting in a dramatically enhanced η_p from 3.6 ± 0.2% to 4.4 ± 0.3%. The final device has achieved broadband spectral coverage with EQE in excess of 20% from λ = 350 nm to 800 nm.

In summary, we have demonstrated that Cl₆SubPc can be utilized as a sensitizer in conjunction with ZCl in blends with C₆₀. Photoluminescence quenching experiments illustrate that C₆₀ quenches excited states on the sensitizers. The sensitizers function by absorbing photons and transferring energy to C₆₀ where C₆₀ then serves to transport excitons and electrons. In OPVs, it was shown that multiple sensitizers can be employed within a single acceptor layer to compliment absorption and enhance photocurrent without deleterious effects to V_OC or FF resulting in increased efficiency. The extension of the acceptor layer response out to λ = 670 nm fully fills the EQE minima exhibited by C₆₀/DPSQ devices and results in a significantly increased η_p of 4.7% for the champion device. These data demonstrate that the sensitization scheme is tolerant to the introduction of additional sensitizers allowing for facile tuning of the acceptor layer absorption. Although in this work the sensitization scheme is enacted in the acceptor layer, a similar strategy could be utilized for donors.

Devices were grown on glass substrates with 150 nm indium tin oxide patterned in 2 mm stripes. Prior to deposition, the substrates were cleaned in a surfactant and a series of solvents as described previously,¹¹ and then exposed to ozone atmosphere for 10 min immediately before loading into the high vacuum chamber (base pressure < 10⁻⁶Torr). MoO₃ was thermally evaporated at 0.02 nm/s. DPSQ was spin-coated from 1.5 mg/ml solutions in chloroform. N,N′-di-[(1-naphthyl)–N,N′-diphenyl]-1,1′-biphenyl)–4,4′-diamine (NPD) was thermally evaporated at 0.1 nm/s. C₆₀ was thermally evaporated at 0.1 nm/s. The sensitized devices contained a blended layer of C₆₀:ZCl deposited at C₆₀ (0.05 nm/s):ZCl (0.05 nm/s) or C₆₀:ZCl:Cl₆SubPc deposited at C₆₀ (0.05 nm/s):ZCl (0.025 nm/s):Cl₆SubPc (0.025 nm/s). All devices were capped with a buffer layer consisting of 10 nm BCP deposited at 0.1 nm/s. Finally, a 100 nm thick Al cathode was deposited at 0.2 nm/s through a shadow mask with a 2 mm slit defining a device area of 0.04 cm². Current density vs. voltage (J–V) characteristics were measured in the dark and under simulated AM 1.5 G solar illumination from a filtered 300 W Xe lamp. J-V measurements were carried out with an illumination area larger than that defined by the cathode. Routine spectral mismatch corrections were performed using a silicon photodiode calibrated at National Renewable Energy Laboratory.²⁷ Chopped and filtered monochromatic light (250 Hz, 10 nm fwhm) from a Cornerstone 260 1/4M double grating monochromator (Newport 74125) was used in conjunction with an EG&G 7220 lock-in amplifier to perform all EQE and spectral mismatch correction measurements. EQE measurements were carried out with an illumination area smaller than that defined by the cathode.

Extinction coefficients for the thin films were calculated from optical constants measured by variable angle spectroscopic ellipsometry. Steady-state emission measurements on thin films were performed using a Photon Technology International QuantaMaster Model C-60SE spectrofluorimeter.

The authors would like to acknowledge the following agencies for funding of this work: the Department of Energy, Office of Basic Energy Sciences as part of Energy Frontier Research Center program, the Center for Energy Nanoscience (DE-SC0001013, Bartynski and Trinh), and NanoFlex Power Corporation (Kirlikovali and Thompson).