WO2016081646A1

WO2016081646A1 - Photovoltaic devices having plasmonic nanostructured transparent electrodes

Info

Publication number: WO2016081646A1
Application number: PCT/US2015/061410
Authority: WO
Inventors: Qiuming Yu; Beau J. RICHARDSON
Original assignee: University Of Washington
Priority date: 2014-11-18
Filing date: 2015-11-18
Publication date: 2016-05-26

Abstract

Disclosed herein are improved photovoltaic devices ("solar cells") that benefit from the use of a transparent nanostructure plasmonic electrode. The plasmonic electrode is sized and configured to provide plasmonic resonance at a wavelength matched to a photovoltaic layer. The photovoltaic devices include at least two photovoltaic materials and the plasmonic electrode is patterned accordingly in order to provide plasmonic resonance tailored to each of the photovoltaic materials. The plasmonic devices can be configured in either traditional or inverted configurations, depending on the characteristics of the layers of the device. Accordingly, the plasmonic electrode may be an anode or cathode depending on the device configuration. Methods of fabricating the photovoltaic devices are also provided.

Description

PHOTOVOLTAIC DEVICES HAVING PLASMONIC NANOSTRUCTURED

TRANSPARENT ELECTRODES

CROSS-REFERENCE TO RELATED APPLICATION This application claims the benefit of U.S. Patent Application No. 62/081,428, filed November 18, 2014, the disclosure of which is hereby incorporated by reference in its entirety.

STATEMENT OF GOVERNMENT LICENSE RIGHTS

This invention was made with Government support under CBET-1346859 awarded by the National Science Foundation. The Government has certain rights in the invention.

BACKGROUND

Photovoltaic (PV) technology is a promising clean and sustainable source of energy, which is expected to play a major role in meeting the global renewable energy challenge. Organic photo voltaics (OPVs) is an emerging PV technology with promising properties such as low cost, flexibility, lightweight, transparency and large-area manufacturing compatibility. OPVs based on conjugated polymers as electron donor materials blended with different types of fullerene derivatives as electron acceptor materials have been the leading candidates in OPVs in the past several years and achieved 8-10% power conversion efficiency (PCE) using a single bulk heterojunction (BHJ) device structure. The bottleneck for OPVs to achieve high PCE is primarily due to the intrinsic small exciton diffusion length and low charge-carrier mobility of conjugated polymers that limit the thickness of the active layer, leading to poor solar light absorption and, thereby, low PCE. Effective light trapping schemes offer an attractive approach to ensure full absorption of the incident sunlight, which is critical for achieving high current density and going significantly beyond the 10% PCE barrier, which may open the door for commercialization of OPVs.

Several approaches have been explored in order to increase the absorption of solar radiation, including tandem solar cell architectures have been utilized. In a tandem solar cell, two or more photoactive layers with complementary absorption spectra are stacked vertically in such a way that a large bandgap active layer (front cell) is placed at the side closest to light incidence and a small bandgap active layer (back cell) is on top mediated by an intermediate layer in order to adsorb broad band spectrum. The front and back cells can either be connected in series or in parallel. The tandem cell connected in series has the benefit of a large open circuit voltage (Voc) which equals almost to the sum of the Voc of its two sub-cells but the short circuit current density (Jsc) is limited to the value of the sub-cell with the lowest Jsc- Similarly, the tandem cell connected in parallel has a large Jsc but a limited Voc- To achieve high efficiency, for a serially connected tandem solar cell, balancing sub-cell current is very critical. Recently, a new low-bandgap conjugated polymer was used as the active layer of the back cell. By carefully balancing the current of front and back cells, a PCE of 10.6% was achieved for a serially connected organic tandem solar cell, the current record for PCE in such devices. While the approach of tandem solar cells is promising, balance of sub-cell current, optimization of electron-hole recombination in the intermediate layer, and processing compatibility of each layer pose great challenges to make high efficiency devices.

In order to advance OPV devices beyond the current state of the art, new device configurations are needed, in addition to improved PV materials. While tandem PV devices provide some benefits over traditional, single-absorber devices, their electrical design limitations and fabrication complexity are issues that point towards the need to explore alternative device configurations.

SUMMARY

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This summary is not intended to identify key features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.

In one aspect, a photovoltaic device is provided. In one embodiment, the device includes a photovoltaic layer between a hole-collecting electrode and an electron- collecting electrode;

wherein the photovoltaic layer comprises at least two laterally distinct photovoltaic portions: a first photovoltaic portion having a first absorption peak wavelength and a second photovoltaic portion having a second absorption peak wavelength that is different than the first photovoltaic absorption wavelength;

wherein one of the hole-collecting electrode and the electron-collecting electrode is a transparent plasmonic electrode, wherein the plasmonic electrode is nanostructured to include a plurality of apertures configured to provide wavelength specific surface plasmon resonance, and wherein the transparent plasmonic electrode comprises at least two laterally distinct plasmonic portions that are sized, configured, and aligned with the at least two laterally distinct photovoltaic portions of the photovoltaic layer such that a first plasmonic portion has a first plurality of apertures configured to produce a first surface plasmon resonance at a wavelength matched to the first absorption peak wavelength of the first photovoltaic portion and a second plasmonic portion has a second plurality of apertures configured to produce a second surface plasmon resonance at a wavelength matched to the second absorption peak wavelength of the second photovoltaic portion.

In another aspect, a photovoltaic article is provided that includes a plurality of photovoltaic devices according to the disclosed embodiments. By combining multiple devices in a single article, a photovoltaic "fabric" can be formed.

DESCRIPTION OF THE DRAWINGS

The foregoing aspects and many of the attendant advantages of this invention will become more readily appreciated as the same become better understood by reference to the following detailed description, when taken in conjunction with the accompanying drawings, wherein:

FIGURE 1 (left) Illustration of the inverted configuration of lateral multijunctions with spectrally matched plasmonic pixels for enhancing broad spectrum light harvesting in a wide angle. Three types of plasmonic pixels are arranged in Bayer mosaic layout, (right) Energy diagram of three pairs of polymer: fullerene active layers and materials in inverted solar cell configuration.

FIGURE 2. The performance of inverted BHJ solar cells with a dynamic or a static bake ZnO layer and 8.36% PCE was achieved. AFM topography images shows the morphology of the ZnO layers prepared by dynamic and static bake, respectively.

FIGURES 3 A and 3B. Inverted configuration: FIGURE 3 A: The J_Sc versus the active layer film thickness calculated by TM modeling. FIGURE 3B: The optimal charge generation profile for an inverted solar cell with 90 nm active layer.

FIGURE 4. The performance of conventional and inverted BHJ solar cells with the active layer of PTB7:PC₇iBM. FIGURES 5 A and 5B. Conventional configuration: FIGURE 5 A: Comparison of simulated and experimentally measured reflectance spectra. FIGURE 5B: The exciton generation rate from the TM modeling.

FIGURES 6A and 6B. Comparison of absorbance spectra (FIGURE 6A) and local electric field at the Au/PEDOT:PSS interface (FIGURE 6B) for the BHJ solar cells with and without the plasmonic nanostructure (D =175 nm; A=225 nm).

FIGURES 7A-7D. FIGURE 7A: Illustration of the architecture of a conventional BHJ solar cell with Au or Cu nanohole mesh as anode and P3HT:PCBM blend as the active layer with the incident angle at 0 and 30° with respect to surface normal. FIGURE 7B: FDTD simulated absorbance spectra. FIGURES 7C and 7D: The maximum local electric field at the metal/PEDOT:PSS interface and the middle layer of the active layer, respectively.

FIGURES 8A-8E. SEM images of (FIGURE 8A) silicon master mold; (FIGURE 8B) photoresist nanopillars covered with Au; (FIGURE 8C) Au nanopattem on ITO/glass after lift-off and (FIGURE 8D) Au nanopattem on glass after lift-off. FIGURE 8E: J-V curves of inverted solar cells on ITO/glass and on plasmonic nanopattern/glass measured at the indicated electrode. The Au plasmonic nanopattem area (top rectangle) can be seen in the photo of "Device #10 Back". The scale bars are 1 μιη in FIGURES 8A and 8B, 2 μιη in FIGURE 8C, and 5 μιη in FIGURE 8D.

FIGURE 9. Illustration of the fabrication procedure of a lateral multijunction solar cell with matched plasmonic nanostructures for enhancing light absorption. Layout of one mask for Ag deposition and the number of plasmonic pixels within one electrode for different size pixels and electrodes.

FIGURE 10. Experimental setup for the measurements of reflectance spectroscopy of lateral multijunction solar cells.

FIGURE 11. The basic fabrication process using nanoimprint lithography to make substrates with plasmonic nanohole array electrodes.

FIGURES 12A-12D. SEM images with x20k magnification of (FIGURE 12A) the silicon master mold, (FIGURE 12B) the imprinted ETFE sub mold, (FIGURE 12C) an imprinted resist sample after Cr/Au evaporation and (FIGURE 12D) after solution liftoff.

FIGURES 13A-13C. Tilted SEM images of (FIGURE 13A) an imprinted resist sample after Cr/Au evaporation and solution lift-off (x20k) and (FIGURE 13B) a zoom in view (x40k) of FIGURE 13A. FIGURE 13C: A schematic of the effect seen in FIGURES 13A and 13B where the Au-disks on top of the resist nanopillars are bridged to the Au-nanohole array below. This prevents removal of many of the Au-disks even after solution lift-off to dissolve the resist.

FIGURES 14A-14D. SEM images of Au nanoholes on glass made with different spans of oxygen plasma etching followed by metal deposition, solution lift-off and peel- off Etch time and average diameter are (FIGURES 14A and 14B) 30 s (-172 nm), (FIGURE 14C) 60 s (-159 nm) and (FIGURE 14D) 72 s (-152 nm).

FIGURES 15A and 15B. SEM images of nanopatterns transferred into a composite PDMS/h-PMDS stamp at magnifications of 20kX and 80kX, respectively.

FIGURE 16. The basic fabrication process using solvent-assisted nanomolding (SAN) to make substrates with plasmonic nanohole array electrodes with an MPTMS SAM to replace the Cr "glue" layer.

FIGURES 17A-17D. SEM images of Au-only nanostructures on glass fabricated with SAN showing an area with (FIGURES 17A-17B) open nanoholes and (FIGURES 17C-17D) nanoholes covered by "Au-tops" at magnifications of (FIGURES 17A and 17C) x20k and (FIGURES 17B and 17D) x80k.

FIGURE 18. Measured (solid lines) and FDTD simulated (dashed lines) transmission spectra of bare glass, glass/ITO, glass/ Au unpatterned film, glass/ 2 nm Cr/ 23 nm Au 175-225 nanostructure, and glass/25 nm Au-only 175-225 nanostructure.

FIGURES 19A and 19B. FIGURE 19A: Schematic of an inverted device with a plasmonic Au-nanostructured cathode. FIGURE 19B: Current- voltage curves for inverted devices with ITO (solid lines) and Cr/Au nanostructured (dotted lines) cathodes with sol-gel derived ZnO ETLs of -32 nm thickness and varied active layer thicknesses of -122 nm, -93 nm, and -77 nm.

FIGURE 20. External quantum efficiency (EQE) spectra of inverted devices on ITO (solid line) and Cr/Au nanostructured (dotted line) cathodes with -32 nm thick ZnO and -77 nm thick active layers.

FIGURE 21. Measured reflectance spectra (solid lines) of ITO and Cr/Au-based devices with -32 nm ZnO ETLs and -77 nm active layers shown in Table 4. The dashed lines show FDTD -simulated spectra of ITO and Cr/Au-based devices with similar thicknesses. FIGURES 22A and 22B. SEM images of Cr/Au nanostructures on glass coated with a ZnO sol gel film after baking at 200°C at magnifications of x20k and xl60k, respectively.

FIGURE 23. Current- voltage curves for inverted devices with ITO (solid lines) and Cr/Au nanostructured (dotted lines) cathodes with ZnO ETLs of -24 nm thickness and varied active layer thicknesses of -122 nm, -93 nm and -77 nm.

FIGURES 24A and 24B. FIGURE 24A: Current-voltage curves for inverted devices with thicker 48 nm ZnO ETLs on ITO and Cr/Au nanostructures with varying nanohole diameters of -172 nm, -168 nm, -159 nm and -152 nm. FIGURE 24B: EQE spectra for the ITO device and Cr/Au device with -172 nm nanoholes shown in FIGURE24A.

FIGURE 25. Measured reflectance spectra (solid) of the ITO and Cr/Au-based devices with -48 nm ZnO ETLs shown in Table 6. The dashed lines show the FDTD- simulated spectra of ITO and Cr/Au-based devices with similar thicknesses.

FIGURES 26A and 26B. FIGURE 26A: Current-voltage curves for inverted devices with -44 nm ZnO NP ETLs on ITO (solid line) and Cr/Au nanostructured (dotted line) electrodes. FIGURE 26B: SEM image of a Cr/Au nanostructure covered with a ZnO NP film that measured -44 nm on a bare glass/ITO substrate.

FIGURE 27. Current-voltage curves for inverted devices with ITO, Cr/Au, and Au-only nanostructured electrodes with -24 nm ZnO ETLs and -77 nm thick active layers. The top inset is an SEM image of a Au-only nanostructure with smaller holes and large areas covered with unlifted Au disks and the bottom inset is that of a Cr/Au nanostructure.

FIGURE 28. Photocurrent ratio versus incident light angle for inverted devices on ITO, and Cr/Au nanostructured electrodes.

FIGURES 29A-29C. FIGURE 29A: Photos of typical inverted devices on flexible PET/ITO substrates. FIGURE 29B: Current- voltage curves of inverted devices on glass/ITO (dotted lines) and on PET/ITO (solid lines) comparing ZnO films from NPs synthesized in-house and purchased from Sigma- Aldrich. FIGURE 29C: Current- voltage curves of devices on PET/ITO with in-house synthesized ZnO NPs but with varied active layer spin speeds for re-optimization.

FIGURES 30A and 30B. FIGURE 30A: Photograph of a bare PET substrate with Cr/Au nanostructures fabricated by NIL. FIGURE 30B: Current-voltage curves of a control device on PET/ITO and an ITO-free device with a Cr/Au nanostructured electrode on PET.

FIGURE 31. Current- voltage curve for an inverted ITO-free device on a conformable Parylene substrate with Cr/Au nanostructured electrodes. The device performance parameters and a photo of the device wrapped around the end of a pen are included in the inset.

DETAILED DESCRIPTION

Disclosed herein are improved photovoltaic devices ("solar cells") that benefit from the use of a transparent nanostructure plasmonic electrode. The plasmonic electrode is sized and configured to provide plasmonic resonance at a wavelength ("color") matched to a photovoltaic layer. The photovoltaic devices include at least two photovoltaic materials and the plasmonic electrode is patterned accordingly in order to provide plasmonic resonance tailored to each of the photovoltaic materials. The plasmonic devices can be configured in either traditional or inverted configurations, depending on the characteristics of the layers of the device. Accordingly, the plasmonic electrode may be an anode or cathode depending on the device configuration. Methods of fabricating the photovoltaic devices are also provided.

While the provided photovoltaic devices can be formed from materials and in configurations known to those of skill in the art, two particular departures from known devices are incorporated in order to provide improved photovoltaic devices.

First, two or more "portions" are included in the devices. Particularly, each portion is sensitized to a specific peak wavelength (and related range of wavelengths). Therefore, the device can absorb two or more peak wavelengths, based on the number of portions. These portions are in the same "layer" of the device but they are laterally distinct. For example, each portion can be deposited separately on the same layer of the device by inkjet printing or other micro-printing methods. Both the photovoltaic layer and the transparent plasmonic electrode are configured particularly to define the portions. The nanostructures of the plasmonic electrode are different between the plurality of portions so as to tailor the plasmonic resonance to match the photovoltaic peak wavelength of absorption of each portion.

Second, the plasmonic electrode is transparent, meaning that the device need not include a traditional, rigid transparent oxide conductor, such as indium-tin oxide (ITO). Accordingly, in one embodiment, the photovoltaic device does not include indium-tin oxide or fluorine tin oxide. ITO and the like are not conformable and therefore structurally limit any devices into which they are incorporated— in particular, ITO-based devices cannot be bendable. By relieving design constraints, the transparent plasmonic electrode allows the disclosed devices to be flexible or formed conformably in non-planar configurations.

Photovoltaic Device Architecture

The disclosed devices can be formed in either traditional (conventional) or inverted configurations. Depending on the type of device and electronic properties of the device layers the plasmonic electrode can act as the anode or cathode of the device.

One distinguishing characteristic of the disclosed devices is the definition of the device into "portions" that are laterally distinct and are defined by a particular combination of photovoltaic material and related plasmonic nanostructure tailored to provide plasmon resonance at a wavelength related to the absorbance characteristics of the photovoltaic material. In particular, in a first portion of a device the PV material absorbs at a first peak wavelength and a portion of the plasmonic electrode disposed above (i.e., in between the source of electromagnetic radiation and the PV layer) the PV material in the first portion is configured to provide plasmonic resonance at the first peak wavelength. Similarly, for a second portion of the device a second PV material absorbs at a second peak wavelength and the plasmonic electrode has a different nanostructured pattern above the second portion that provides plasmonic resonance at the second peak wavelength.

In one embodiment, the first absorption peak wavelength and the second absorption peak wavelength are between 250 nm and 2500 nm. This range includes essentially all significant portions of the solar spectrum, including ultraviolet (UV), visible, and infrared (IR) wavelengths. The devices are only limited by the PV materials available to absorb the impinging wavelengths from the sun.

In one embodiment, first photovoltaic portion and the second photovoltaic portion have a shape independently selected from the group consisting of a triangle, a square, a rectangle, and a polygon. Any shape or combination of shapes can be used as long as the other requirements of the device are met. Preferred shapes are those that can be joined together to form complete coverage of a surface.

In one embodiment, the first photovoltaic portion and the second photovoltaic portion have a largest dimension of 1 mm.

A representative inverted configuration device is illustrated in FIGURE 1. The device has a "pixel" configuration with three different "colors" being absorbed by the PV layer of the device across four different pixels (one color is duplicated twice in this configuration). Light enters the device from the bottom, as pictured, and the transparent plasmonic electrode forms not only the cathode of the device but also enhances light delivered to each pixel via plasmonic resonance. As illustrated in the break-out depiction of the plasmonic "nanomesh" electrode, the nanostructures are holes that are varied in diameter to tune the resonance wavelength to match each of the related PV material. Note that the three PV materials are disposed in a single layer of the device that is laterally varied to include the different PV materials. As used herein, "laterally varied" is used to describe the illustrated configuration wherein a single layer of the device is formed from two or more materials such that the material varies laterally across the device. FIGURE 1 also depicts the electronic energetics of the exemplary device configuration, including BHJ materials incorporated into the three pixel materials of the PV layer. An electron-transport layer and hole-transport layer are also included to facilitate device operation.

While two laterally distinct portions define the most basic configuration of the disclosed devices, any number of portions can be used. The number of portions are limited only by the ability to form the portions laterally. In embodiments when the portions are micron-sized pixels, one limitation is the lateral resolution by which the portions can be defined, both in terms of the PV layer deposition and the fabrication of the nanostructures of the plasmonic electrode. For meso-scale devices, such concerns are not as acute and therefore essentially any number of portions can be used.

In one embodiment, the photovoltaic layer further comprises a third laterally distinct portion that includes a third photovoltaic portion having a third absorption peak wavelength that is different than the first absorption peak wavelength and the second absorption peak wavelength.

In another embodiment, the photovoltaic device further includes a substrate adjacent one of the hole-collecting electrode and the electron-collecting electrode. As used herein, the term "adjacent" refers to a configuration that is either abutting or disposed some distance from one of the electrodes in a direction distal to the PV layer.

In one embodiment, the substrate is deformable without breaking (i.e., plastic or elastic deformation). In one embodiment, the substrate is conformable.

In one embodiment, the substrate is a material selected from the group consisting of glass, polyethylene terephthalate (PET), polyethylene-naphthalate (PEN), fluorinated ethylene propylene (FEP), parylene, cellulose nanocrystal, and other transparent substrates. As used herein, the term "transparent substrate" is defined by 80% transparency at the peak absorbance wavelengths of the device, from 250 nm to 2500 nm.

In one embodiment, the device is hermetically sealed. Hermetic sealing methods and structures are known to those of skill in the art and are useful for packaging and protecting the devices for commercial applications. In one embodiment, hermetic sealing comprises an encapsulating layer disposed covering the device such that oxygen is prevented from contacting the device. Photovoltaic Layers

The core of the disclosed devices is the photovoltaic layer. Essentially any known or future-developed photovoltaic material can be incorporated into the disclosed devices. While the exemplary embodiments disclosed herein are generally described in the context of bulk-heterojunction (BHJ) PV materials, in practice the devices are not limited to BHJ materials. Given the compelling possibility of utilizing inexpensive micro-printing techniques, such as inkjet, to form the different portions of the device (e.g., in a "pixel" design), in certain embodiments the PV layer is formed from solution-processable materials. Accordingly, in one embodiment, the photovoltaic layer is formed from a solution-processable PV material selected from the group consisting of a bulk- heterojunction material, a quantum dot material, a perovskite material, and a combination thereof.

In one embodiment, the photovoltaic layer comprises a bulk-heterojunction material, comprising an electron donor material heterogeneously mixed with an electron accepting material.

In one embodiment, the electron donor material is selected from the group consisting of poly(3-hexylthiophene) (P3HT), poly[2-methoxy-5-(2-ethylhexyloxy)-l,4- phenylenevinylene] (MEH-PPV), Poly[[4,8-bis[(2-ethylhexyl)oxy]benzo[l,2-b:4,5- b']dithiophene-2,6-diyl][3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno[3,4-b]thiophenediyl]] (PTB7), poly[4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)benzo[l ,2-b;4,5-b^*]dithiophene-2,6- diyl-alt-(4-(2-ethylhexyl)-3-fluorothieno[3,4-b]thiophene-)-2-carboxylate-2-6-diyl)] (PTB7-Th), poly(9,9-dioctylfluorene-alt-bithiophene) (F8T2), poly[(9,9-dioctylfluorenyl- 2,7-diyl)-alt-5,5-(40,70-di-2-thienyl-20,10,30-benzothiadiazole)] (APFO-3), poly[bis(4- phenyl)(2,4,6-trimethylphenyl)amine] (PTAA), poly(2,5-bis(3-hexadecylthiophen-2- yl)thieno[3,2-b]thiophene) (PBTTT), polybenzo[l,2-b:4,5- b']dithiophene-4,7-dithien-2- yl-2,l,3-benzothiadiazole (PBnDT-DTBT), poly[N-9"-heptadecanyl-2,7-carbazole-alt- 5,5-(4',7'-di-2-thienyl-2',l ',3'-benzothiadiazole)] (PCDTBT), poly[2,6-(4,4-bis-(2- ethylhexyl)-4H-cyclopenta[2,l-b;3,4-b']dithiophene)-alt-4,7(2,l,3-benzothiadiazole)] (PCPDTBT), fluorinated PCPDTBT (F-PCPDTBT), poly[2,7-(5,5-bis-(3,7-dimethyl octyl)-5H-dithieno[3,2-b:20,30-d]pyran)-alt-4,7-(5,6-difluoro-2, 1 ,3-benzothiadiazole)]

(PDTP-DFBT), poly[(2,5-bis(2-hexyldecyloxy)phenylene)-alt-(5,6-difluoro-4,7- di(thiophen-2-yl)benzo[c]-[ 1 ,2,5]thiadiazole)] (PPDT2FBT), poly[diketopyrrolopyrrole- terthiophene] (PDPP-TTT), poly(di(2-ethylhexyloxy)benzo[l,2-b:4,5-b0]dithiophene- cooctylthieno[3,4-c]pyrrole-4,6-dione) (PBDTTPD), poly{[4,9-dihydro-4,4,9,9-tetra(4- hexylbenzyl)-s-indaceno[l,2-b:5,6-b']-dithiophene-2,7-diyl]-alt-[2,3-bis(3- (octyloxy)phenyl)-2,3-dihydro-quinoxaline-2,2'-diyl] (PIDTDTQx), and other semiconducting materials based on poly-(2,7-carbazoles), carbazole-benzothiadiazole copolymers, diketopyrrolopyrrole (DPP) copolymers, benzodithiophene (BDT) derivatives, and indacenodithiophene (IDT) based copolymers.

In one embodiment, the electron acceptor material is selected from the group consisting of [6,6]-phenyl-C(61)-butyric acid methyl ester (PC60BM), [6,6]-phenyl- C(71)-butyric acid methyl ester (PC70BM), indene-C60 bisadduct (ICBA), and other semiconducting materials based on fullerene derivatives such as bis-adducts, diphenylmethano fullerenes, dimethylphenylmethano fullerene bis-adducts, and endohedral fullerenes, as well as, additional n-type semiconducting materials based on naphthalene diimide-perylene diimide copolymers and inorganic acceptor nanocrystals and quantum dots (e.g. PbS, PbSe, CdS, CdSe, ZnO, and pyrite FeS2).

In one embodiment, the first photovoltaic portion and the second photovoltaic portion each comprise bulk-heterojunction materials.

Transparent Plasmonic Electrode

The transparent plasmonic electrode forms either the anode or cathode of the device and is structured and configured to enhance absorption of a particular wavelength at the PV layer below the plasmonic electrode.

In one embodiment, the transparent plasmonic electrode is formed from a metal selected from the group consisting of gold, silver, copper, aluminum, platinum, palladium, and combinations thereof. The material of the plasmonic electrode depends partially on whether the plasmonic electrode is used in the device as the anode or cathode. Those of skill in the art will appreciate the design considerations related to device energetics and design. In the disclosed embodiments, in the inverted device configuration illustrated in FIGURE 1, the transparent plasmonic electrode is a cathode formed from gold and the relatively lower work function silver is used as the anode.

In one embodiment, the transparent plasmonic electrode is defined by a transmission of at least 40% at the first absorption peak wavelength. In another embodiment, the transparent plasmonic electrode is defined by a transmission of at least 40% at both the first absorption peak wavelength and the second absorption peak wavelength. In yet another embodiment, in which there are more than two portions, the transparent plasmonic electrode has at least 40% transmission at each of the absorption peak wavelengths.

The plasmonic nanostructures in each of the portions can be tailored to a particular wavelength of plasmonic resonance in accordance with known techniques. In one embodiment, the first plurality of apertures and the second plurality of apertures have a shape independently selected from the group consisting of a circle, a distorted circle, a square, a rectangle, a triangle, an x-shape, and a polygon. Any shape can be used as long as it produces the necessary plasmonic resonance.

In one embodiment, the first plurality of apertures and the second plurality of apertures have the same shape in the two laterally distinct portions. As illustrated in FIGURE 1 , both pluralities of apertures are circular in one embodiment.

In one embodiment, the first plurality of apertures and the second plurality of apertures have a smallest dimension of from 5 nm to 5 microns. As used herein, the smallest dimension is defined as the shortest point between opposing sides of an aperture. In a circle the smallest dimension is a diameter. In a triangle it is the length of the shortest leg of the triangle. In a square it is the width. In a rectangle it is the shortest of the length and width.

In one embodiment, the first plurality of apertures and the second plurality of apertures are arranged in a pattern independently selected from the group consisting of square, rectangular, hexagonal, and triangular. Any arrangement can be used in order to produce the desired effect. The exemplary device of FIGURE 1 has a square pattern.

In one embodiment, the first plurality of apertures and the second plurality of apertures have a center-to-center pitch dimension of from 50 nm to 5 microns.

In one embodiment, the transparent plasmonic electrode has a thickness of 5 nm to 1 micron.

In one embodiment, the hole-collecting electrode is a high-work-function electrode and the electron-collecting electrode is a low-work-function electrode.

Inverted Device Configuration

In one embodiment, the photovoltaic device is an inverted photovoltaic device and the transparent plasmonic electrode is the electron-collecting electrode. An exemplary inverted device configuration is illustrated in FIGURE 1 and other FIGURES and Examples. In one embodiment, the hole-collecting electrode is a material selected from the group consisting of Al, Ag, Cu, Au, Pd, and Pt.

In one embodiment, the photovoltaic device further includes an electron transport layer between the transparent plasmonic electrode and the photovoltaic layer.

In one embodiment, the photovoltaic device further includes a hole transport layer between the hole-collecting electrode and the photovoltaic layer.

Convention Device Configuration

As noted above, the plasmonic electrode can be incorporated into conventional devices as well, particularly as a substitute for ITO or other rigid transparent oxide conductor. Accordingly, in one embodiment, the photovoltaic device is a conventional photovoltaic device and the transparent plasmonic electrode is the hole-collecting electrode. Conventional device architecture is illustrated in various FIGURES, notably FIGURE 4 which illustrates a representative conventional device that includes ITO as illustrated but it will be appreciated that substituting a plasmonic electrode for ITO in such a structure would be an embodiment of the disclosed devices.

In one embodiment, the electron-collecting electrode is a material selected from the group consisting of Al, Ag, Au, and Cu.

In one embodiment, the photovoltaic device further includes a hole transport layer between the transparent plasmonic electrode and the photovoltaic layer.

In one embodiment, the photovoltaic device further includes an electron transport layer between the electron-collecting electrode and the photovoltaic layer.

Photovoltaic Device Operation

The disclosed devices are operated by exposure to an electromagnetic light source that includes at least the first and second absorption peak wavelengths. Typically the lights source is the sun but operation is not limited as such. Artificial light sources can also be used.

The power generated by the devices can be delivered to an electrical load in order to power an electrical circuit or store charge in a battery, capacitor, or other charge- storage device. The use of photovoltaic devices is well known to those of skill in the art and the disclosed devices can be used in any manner known PV devices can be used.

Articles Incorporating Photovoltaic Devices

In another aspect, a photovoltaic article is provided that includes a plurality of photovoltaic devices according to the disclosed embodiments. By combining multiple devices in a single article, a photovoltaic "fabric" can be formed. The devices can be connected in parallel or series on the article. Connections between devices can be made by any structures known to those of skill in the art, such as conductive traces. A single battery or multiple batteries can be used to store electricity generated by the article. In one embodiment the battery or batteries are on mounted on the article.

In one embodiment the article includes a substrate of the type disclosed elsewhere herein. Flexible and conformable substrates are particularly desirable. In one embodiment, the article is bendable. In one embodiment, the article is conformable.

In one embodiment, the article is perforated so as to allow gaseous transport across the article. Perforation of the article allows gas transport across the article so as to allow it to "breath." Breathability is desirable in many applications, including clothing and gear for outdoor activities (e.g. jackets, pants, and shoes) where physical exertion increases body heat and moisture underneath the article. Breathable articles allow body moisture and vapors to be released preventing overheating or discomfort.

The following examples are included for the purpose of illustrating, not limiting, the disclosed embodiments.

EXAMPLES

Example 1 : Exemplary Photovoltaic Devices Having Plasmonic Nanostructured Transparent Electrodes

There is growing interest in developing plasmonic BHJ solar cells involving either metallic nanoparticles (NPs) or nanostructures. For the metallic NP approach, it mainly utilizes the strong local electric field induced by localized surface plasmon resonance (LSPR) to enhance the light absorption in the active layer without increasing the active layer thickness and to influence charge transport. By mixing Au NPs in the anode buffer layer poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) of conventional BHJ solar cells, both Jsc and fill factor (FF) were enhanced significantly. A better enhancement in Jsc was obtained by mixing both Au and Ag NPs in the PEDOT:PSS layer due to the dual resonances caused by two different NPs, which broaden the wavelength range of light absorption enhancement. Although colloid NPs are compatible with solution processing, it is difficult to control the NP distribution and therefore, the plasmonic properties. Since plasmonic properties strongly depend on the metallic material, shape, dimension, separation, and spatial arrangement of plasmonic elements as well as the surrounding dielectric media, plasmonic nanostructures must be rationally designed and precisely fabricated in order to achieve the most successful applications. Advances in nanofabrication technologies have allowed the fast and low-cost fabrication of large area sub- 100 nm scale features, creating research potential on integration of plasmonic nanostructures in BHJ solar cells. In addition, the fabrication of plasmonic nanostructures on soft substrates could also be integrated in the large area roll-to-roll process. Plasmonic nanostructures offer many advantages such as acting as a transparent electrode, polarization-insensitivity, and omnidirectional light scattering, which enables wide angle light absorption. Thin gold films with sub-wavelength nanoholes have been made on glass substrates to serve as transparent electrodes and in the meantime to utilize their plasmonic effects. In a conventional BHJ solar cell containing

ITO/PEDOT:PSS/P3HT:PCBM/TiO_x/Al, external quantum efficiency (EQE) was significantly increased from 27% to 69% by using a plasmonic gold mesh electrode to replace ITO, resulting in an increased Jsc and FF and thus, about 52% increase of PCE. In addition, the high broad band light absorption can be maintained in a wide light incident angle (+/- 65°) and polarization independence. With the support of EAGER program, we have explored integration of a wide angle light concentrator enabled by plasmonic nanostructures into organic BHJ solar cells to enhance the solar energy conversion efficiency even at large oblique incident angles.

In this Example, we disclose a new PV device architecture that includes laterally arranged active layers of different bandgaps and spectrally matched plasmonic nanostructures for each junction. This lateral multijunction organic photovoltaic cell with spectrally matched plasmonic pixels is expected to achieve > 10% PCE by harvesting broad spectrum light in a wide angle.

As illustrated in FIGURE 1 , a pixelated plasmonic cathode is fabricated on a glass substrate for inverted configuration. The plasmonic cathode contains thousands of three types of plasmonic pixels arranged in Bayer mosaic layout. Each pixel is specifically designed to enhance a desired band of light that matches the bandgap of the active layer on top of it. The polymer :fullerene active layers with low, medium, and large bandgaps are selected to absorb the broad spectrum of solar radiation. Since each active layer shares the same cathode and anode, they act as in parallel solar cells. It is expected that this device will have a large Jsc by adding the current from each tiny active layer while the Voc is limited by the lowest one among different active layers. In this work, three pairs are selected. The energy diagram of three possible starting pairs along with other materials in the inverted structure is shown in FIGURE 1. Table 1 shows the polymer bandgap, light absorption range, and the inverted single junction solar cells made by them. We estimate the required Jsc in order to achieve 10-15% PCE with a fixed FF of 65%o and Voc of 0.62 and 0.8 V, respectively. Table 2 shows that even for the lowest Voc of 0.62 V among the three active layers, the required Jsc is 29.78 mA/cm² for 12% PCE. This Jsc is achievable based on the Jsc of the devices made by each of the active layers listed in Table 1.

Table 1. Three active layers and the inverted device performance.

Table 2. Estimated Jsc for fixed FF and Voc in order to achieve different desired

This device architecture has many advantages. First, forming lateral multijunctions not only enables the absorption of broad spectrum light but also eliminates the fabrication complexity in vertical tandem solar cell structures. The lateral multijunction organic solar cells are much easier to fabricate compared to inorganic ones since different active layer solutions can be quickly and accurately printed on desired places in large area via inkjet printing technology. Second, the spectrally and spatially separated plasmonic nanostructures (plasmonic pixels) can be designed to specifically enhance the desired spectrum of light to match the active layer bandgap. A plasmonic pixel as small as 3 x 3 μιη² can still retain its unique plasmon resonance (i.e., color). There is no interference between pixels, which provides sharp color display. Third, this device architecture simplifies the fabrication procedure and allows fast, low cost, and large area production. Lastly, this device architecture provides great flexibility in designing high efficiency solar cells.

The objective of this work is to develop lateral multijunction organic solar cells with spectrally matched plasmonic pixels to enhance broad spectrum light harvesting to achieve high energy conversion efficiency of > 10% PCE.

Inverted BHJ solar cells with up to ~8.36% PCE and TM optical modeling We first made inverted BHJ solar cells on ITO-coated glass substrates using ZnO as the electron selective layer, PTB7:PC7iBM as the active layer, M0O3 as the hole selective layer, and silver as the top hole collecting electrode. As shown in FIGURE 2, we varied the thickness and morphology of the ZnO layer as well as the thickness of the active layer to optimize the solar cell performance. The ZnO layer was deposited using a sol gel method by spin coating a precursor solution and subsequently baking the films. The film thickness was controlled by varying the precursor concentration and spin rate while the morphology was controlled by varying the baking condition. A static bake was performed by placing the substrates on a hotplate at 200 °C immediately after spin coating. Alternately, a dynamic bake was performed by letting the coated substrates sit at room temperature for ~10 minutes before placing them on a hotplate and then ramping the temperature from ambient to 200°C over ~1 minute. All ZnO films were baked for 1 hour. The active layer solution consisted of a 1 : 1.5 wt. ratio of PTB7:PC7iBM in chlorobenzene with 3 vol% diiodooctane added. The AFM images in FIGURE 2 show that the static bake condition produced a smoother film consisting of spheroidal ZnO nanoparticles while the dynamic bake condition produced a nanoridge structure. A better PCE of 8.36% was obtained for the solar cell with the static bake ZnO layer compared to the one with the dynamic bake ZnO layer. A smoother and thicker ZnO layer could better prevent current leakage, resulting in a larger fill factor. In this optimized inverted solar cell, the ZnO layer is -48 nm and the active layer is 85 nm. This PCE is higher than the similar inverted solar cells reported.

To design inverted devices with optimal performance, we performed transfer matrix (TM) optical modeling. We measured the optical properties of ZnO films and active layers prepared under different conditions via ellipsometry and use them in TM optical modeling. FIGURE 3A shows the TM modeling results of Jsc versus the active layer film thickness. Two Jsc maxima were predicted for the active layer film thickness around 70-90 nm and 200-250 nm. We fabricated the solar cells with the active layers in these ranges. For the thinner active layer, the experimental Jsc (~16 niA/cm²) is close to the value predicted by TM modeling and the experimental FF averaged -68% producing a maximum PCE of 8.36%. For the thicker active layer, however, the experimental Jsc (~16 niA/cm²) is much lower than the value predicted by TM modeling and the experimental FF averaged 51% caused by increased charge recombination, which the modeling does not take into account. In this system, the optimal active layer thickness is -70-90 nm according to the TM model which is validated by our experimental results. The charge generation profile calculated from TM modeling is shown in FIGURE 3B.

FDTD simulations and TM modeling for conventional BHJ solar cells

We also made conventional structure solar cells with PTB7:PC₇iBM as the active layer and obtained 8.07% PCE, which is similar to those reported in literature. For comparison, the J-V curve and the device structures for both conventional and inverted solar cells are shown in FIGURE 4. We made thin film of each layer involved in the BHJ solar cells, measured their optical properties via ellipsometry, and input them in the FDTD simulations and TM modeling. FIGURE 5A shows that the reflectance spectra of a conventional device with a 200 nm PTB7:PC₇iBM active layer obtained from the FDTD simulation and TM modeling are very similar to that measured experimentally. FIGURE 5B shows the exciton generation rate in the active layer of the conventional device calculated by TM modeling. The current density (Jsc) calculated from the TM modeling is 15.60 niA/cm² with the assumption of 100% internal quantum efficiency, which is very close to the value of 14.94 mA/cm² measured from the BHJ solar cell. We further conducted the FDTD simulations on two BHJ solar cells: one having the plasmonic nanohole array as anode while another having ITO as anode. Simulation results show the enhanced absorption in 570-800 nm (FIGURE 6A), which matches the energy band gaps of the donor (PTB7) and acceptor (PC₇₁BM) in the active layer, and a strong local electric field at the gold/PEDOT:PSS interface (FIGURE 6A) for the solar cell with the plasmonic nanohole array.

Another set of FDTD simulations showed that for the active layer containing P3HT:PCBM (FIGURE 7A), light absorption (FIGURE 7B) and local electric field at the metal/PEDOT:PSS interface (FIGURE 7C) and in the middle of the active layer (FIGURE 7D) were greatly enhanced when a gold or copper nanohole array was used as the transparent anode. Moreover, the enhancement was retained as the incident angle was changed from 0 to 30° respective to surface normal.

Inverted devices with plasmonic nanoholes as cathodes

The gold nanohole array with the diameter of 175 nm and the period of 225 nm as an anode enhances the light absorption in the range of 600-800 nm as shown in FIGURE 6 A from FDTD simulations. We have fabricated the gold nanoholes on glass or ITO/glass substrates through a multistep procedure utilizing thermal nanoimprint lithography (NIL) at the Washington Nanofabrication Facility (WNF) on campus. Briefly, the nanopillar patterns with an area of 5 x 3.5 mm² were formed on a silicon master mold using electron beam lithography (EBL) followed by reactive ion etch (RIE) (FIGURE 8A). ETFE daughter molds with nanoholes were made via NIL after cleaning and silanizing the Si master mold. Next, Nanonex NXR-1025 thermal imprint resist was spun onto the glass substrate and baked. Then, a photoresist nanopillar array on the substrate was made via NIL using the ETFE daughter molds. The resist residuals between pillars were removed via plasma etching. A 30 nm thin layer of gold was deposited on the substrate with 2 nm chromium as the adhesion layer. The SEM image (FIGURE 8B) of the substrate covered with a gold thin film shows that the nanopatterns are successfully made on the glass substrates. Finally, the plasmonic nanohole array was obtained by lift-off of the resist nanopillars using a 1 : 1 :5 volume ratio solution of ΝΗ₄ΟΗ:Η₂θ₂:Η₂0 under sonication. The SEM images show that photoresist nanopillars were not fully lifted off under the conditions we used. Nevertheless, we used these plasmonic substrates and made our first batch of devices. We obtained 2.51% PCE using the Au nanopatterned film on glass as the cathode in the inverted organic solar cell. As shown in FIGURE 8E, the Voc and FF of the plasmonic solar cell are similar to those of the control device using the ITO as anode but the Jsc is smaller. This could mainly be due to the un-lifted nanopillars blocking the light transmission and therefore, reducing light absorption in the active layer. This is a very promising result for the first try. To the best of our knowledge, this is the first time that gold plasmonic nanostructure was made on the glass substrate (not ITO/glass) and used as cathode to make a working inverted BHJ solar cell.

Table 3 summarizes the characterization of the devices tested in FIGURE 8E.

Table 3. Characterization of devices in FIGURE 8E.

Design of plasmonic nanostructures and optimization of device architecture Selection of polymer: fullerene pairs. In order to design the plasmonic nanostructures for spectrally matching the bandgaps of different active layers used in the lateral multijunction OPVs, the appropriate polymer: fullerene pairs are selected. In this work, we start with three pairs of polymer: fullerene with the polymers having large, medium and low bandgaps. In addition, these active layers should have the similar Voc since they act as thousands of micron-scale solar cells connected in parallel. The overall device Voc is limited by the lowest one. We start with three pairs listed in Table 1. All materials are commercially available. Large Voc and Jsc were obtained from the individual OPVs made by these materials as active layers. The estimation in Table 2 shows that PCE over 10% could be achieved by using these three pairs for the proposed device architecture.

Design plasmonic nanostructures with enhanced light absorption and local electric field in the spectral range matching each of three active layers. For both FDTD simulation and TM modeling, the system is considered as one active layer. The interaction between adjacent pixels or active layers will not be considered. The light absorption ranges for each of the three active layers are listed in Table 1. We conducted FDTD simulations to design the plasmonic nanostructures with the following design criteria: (1) light absorbance has to be above 0.8 in the wavelength range of corresponding to each active layer; (2) local electric fields in the active layer has to be enhanced at least two-fold; (3) high absorbance and electric field have to be maintained for incident angle varied from 0 to 60° with respect to surface normal.

The unit cell for FDTD simulations has the same configuration as the inverted or conventional solar cells to be made experimentally. The Bloch periodic boundary conditions are applied to the x- and j-direction to describe an infinite square array of nanoholes and uniaxial perfectly matched layers along the z-direction, with the z boundaries being placed 100 nm above the top cathode layer and 600 nm below the glass substrate. A light source is placed 400 nm below the glass substrate to mimic the sun light shining from the glass side. One monitor is placed 100 nm below the light source to collect the reflectance spectra. Two monitors are placed at the bottom and the top of the active layer in order to obtain the absorption spectra. Another three monitors are used to collect the electric field profiles at the metal/ZnO interface, the middle of the active layer and the central cross section of the nanohole, respectively. A continuous wavelength (cw) plane wave light source with the wavelength from 400 to 1000 nm that covers the photovoltaic-relevant spectrum range is used for illumination. The wavelength- dependent refractive indices and dielectric constants of each layer is measured using ellipsometry and used in the simulations.

The entire architecture of the conventional and inverted devices is optimized using TM optical modeling by varying the thickness of each layer in the device to ensure the maximum light absorption in the active layer. For TM optical modeling, the configurations and the thickness of each layer that mimic the real devices are used. The measured wavelength-dependent complex indices of refraction is input to the simulations. The total electric field distribution inside the device is mapped out and the exciton generation rate is calculated.

Typical values of the parameters to be used in FDTD simulations and TM modeling are discussed in the following. The gold film thickness is fixed at 30 nm. The diameter of nanoholes is 150-500 nm. The nanoholes are arranged in a square or triangular grid with the periodicity between 200-600 nm. The thickness of the active layers will adopt those in typical devices, which are between 80 and 200 nm. Since the two metal layers (i.e., the plasmonic nanostructure and the top electrode) form a Farby- Perot cavity, the thickness of each layer inside the device affects the standing waves formed in the cavity, and thus the plasmonic properties, light absorption and ultimately, device performance. The thickness of the hole and electron transfer layers is varied between 10 and 50 nm. Finally, the incident angle is varied from 0 to 80° with respect to surface normal to investigate the incident angle dependence of absorbance spectra and electric field distributions using FDTD simulations.

Fabrication of lateral multij unction solar cells with spectrally and spatially matched plasmonic nanostructures

Array of plasmonic pixels. In the plasmonic color filter work, plasmonic pixels as small as 3 x 3 μιη² were made in blue, green and red colors and were arranged in Bayer mosaic layout to cover an area of 2016 x 1792 μιη². In this Example, the optimized plasmonic nanostructures for each of three active layers is fabricated in pixel sizes of 30 x 30 μιη² and 120 x 120 μιη², respectively, and arranged in the similar way as Bayer mosaic layout to cover the area of 4200 x 2640 μιη². This area is a little larger than the rectangle electrode shown in FIGURE 9. The number of pixels for the small and large size pixels within a large rectangle or a small circle electrode is listed in the table in FIGURE 9. The selection of the plasmonic pixel size takes into account the drop size of inkjet printing, which is discussed next. Different size plasmonic pixels allow the investigation of the dependence of device performance as well as optical and optoelectronic properties on the plasmonic pixel size. As illustrated in FIGURE 9, two "green" color pixels are used in Bayer mosaic layout. This "green" color is equivalent to the active layer with the pair of PTB7:PCyiBM. In this way, the ratio of three active layers in a device is 1 :2: 1 for P3HT:IC₆₀BA, PTB7:PC₇iBM, and pDPP5T-2:PC₆₀BM. The coverage of 50% by the PTB7:PC₇iBM active layer could provide better device performance. We use the same fabrication method described above to make the plasmonic nanostructures on glass substrates. The quality of the nanostructures is examined using AFM and SEM.

Optimization of the inkjet printing conditions. Spin-coating is a fast and simple method to produce homogeneous thin films, which has been widely used in laboratories for making organic solar cells. However, spin-coating is unable to be used in an automated fabrication and cannot pattern a substrate selectively. Inkjet printing is, in contrast to spin-coating, able to pattern in an automated and variable manner. In addition, inkjet printing also has the advantages of being non-contact, mask-less, consuming little amounts of materials, and generating a low amount of waste. Inkjet printing has been used to print high efficiency organic solar cells as well as for combinatorial screening best combinations and conditions for making organic solar cells. No doubt, inkjet printing requires different processing conditions to yield optimal layer properties.

In this Example, we use inkjet printing to print different active layers on the places having spectrally matched plasmonic pixels underneath the ZnO layer as shown in FIGURE 9. Prior to printing on the substrates with plasmonic nanostructures, we investigate the printing conditions using ITO coated glass substrates. The ZnO layer is made via the sol gel method under the condition of static bake with the precursor concentration to generate the desired film thickness as discussed above. The MoOx and Ag layers are thermally evaporated as we did before.

The vapor pressure, boiling point, and surface tension of the solvents are critical parameters for inkjet printing technology. Previous studies showed that the best droplet formation was achieved by using 68% of ortho-dichlorobenzene (oDCB) and 32% 1,3,5- trimethylbenzene (mesitylene) as a solvent to dissolve P3HT:PC60BM and a dense, smooth film was obtained. Compared to oDCB, mesitylene has higher vapor pressure and lower boiling point and surface tension, which can accelerate the evaporation of solvent. Since oDCB is the solvent used for making P3HT:IC₆₀BA, PTB7:PC₇iBM, and pDPP5T-2:PC₆oBM active layers using spin-coating, we also add mesitylene to oDCB as solvent for making the blends of active layers for inkjet printing. Briefly, after making the ZnO layer, the samples are placed 1 mm below the inkjet print head for the coating of the photoactive layer by a commercial inkjet printing tool (Fujifilm Dimatix, Inc.). The piezoelectric-driven inkjet head is filled with solution, a 10 pL drop is delivered at once with a droplet spacing 30-35 μιη and speed 6-9 m/s. The motorized xyz stage will control the position accurately. A fiducial camera is used for the substrate alignment and a drop watcher camera is used to monitor the drop shape. We also vary the weight ratio and concentration of each blend to further control the thickness and morphology of the active layer. The substrate is placed on a plate with temperature control. The substrate temperature from 25 to 60°C is tested. The film thickness is measured by optical interferometer and the morphology is investigated using AFM. Finally, the device performance is evaluated by taking the J-V curves illuminated by a solar simulator calibrated to the AM 1.5 Global spectrum with the light intensity of 100 mW/cm².

Fabrication of lateral multij unction solar cells with spectrally and spatially matched plasmonic nanostructures. The optimal inkjet printing conditions is applied to print different active layers on the glass substrates with spectrally and specially arranged plasmonic pixels covered by the sol gel made ZnO layer. The fabrication procedure is illustrated in FIGURE 9. For the plasmonic pixel size of 30 x 30 μιη², one drop per pixel is printed. For the plasmonic pixel size of 120 x 120 μιη², 4 x 4 drops with a spacing of 30 μι ίβ printed. With a single piezoelectric-driven inkjet head, we print the active layer of PTB7:PCviBM first, then pDPP5T-2:PC₆₀BM, and finally P3HT:IC₆₀BA. After printing, a thin layer of MoOx (~10 nm) is thermally evaporated followed by evaporation of 100 nm Ag as the anode. The number of device and the size of device on each glass substrate is defined by the mask used in the Ag deposition. One mask has 4 rectangular openings with the area of 2.5 x 4 mm² as shown in FIGURE 9 and another has 16 circular openings with 2 mm diameter. The number of plasmonic pixels within each device is also listed in the table in FIGURE 9.

Evaluation of the device performance and investigation of the optical and opto-electronic properties

Device performance. The performance of lateral multijuntion solar cells with spectrally matched plasmonic pixels is evaluated by measuring the current density versus voltage (J-V) curves under dark and illuminated by a solar simulator calibrated to the AM 1.5 Global spectrum with the light intensity of 100 mW/cm² in a N₂-filled glove box. Illumination with normal and varied incident angles is tested by mounting the device on a rotation stage. The device performance in terms of Jsc, Voc, FF, and PCE are determined and correlated to the device fabrication conditions, active layer morphology and structure, and plasmonic properties.

Optical and opto-electronic properties. The reflectance spectra of lateral multijunction solar cells is measured using a home-built apparatus shown in FIGURE 10. The experimental measured reflectance spectra is compared to those obtained from FDTD simulations. The absorbance spectra is estimated using A = 1 - R, given that there is no transmission through the top metal electrode. In order to understand how and why plasmonic nanostructures can enhance the solar energy conversion efficiency in lateral multijunction solar cells, comprehensive optical and opto-electrical experiments is conducted on the devices. The EQE, UV-Vis-NIR, and PL spectra is measured in order to understand whether the enhanced light absorption and electric field help with the exciton generation. Time-resolved PL spectra is measured and the spectra is fitted to estimate the lifetime of photogenerated excitons. The maximum exciton generation rate Gmca) and exciton dissociation probabilities P(E, T) is estimated from the plot of photocurrent density versus effective voltage and the equation to correlate their relationship. The charge carrier density (n) and the effective charge carrier lifetime (¾) is estimated from impedance spectroscopy. The devices performance and the optical and optoelectronic properties is correlated to the plasmonic pixel size in order to elucidate the pixel size effect, which could be related to the light scattering and re-adsorption by neighboring active layers.

Example 2: Fabrication Methods and Testing of Exemplary Photovoltaic Devices Having Plasmonic Nanostructured Transparent Electrodes

Fabrication of Plasmonic Nanostructured Electrodes

FDTD simulation results provide the initial design for the nanostructured electrodes. A silicon master mold with 175 nm diameter and 225 nm pitch nanopillars was made using electron beam lithography (EBL) and reactive ion etching (RIE) methods. Using this master mold, we performed a multi-step procedure involving nanoimprint lithography (NIL) to fabricate the actual electrodes for devices. FIGURE 1 1 shows the primary fabrication steps. The basic steps include NIL of an ETFE sub-mold from the silicon master, NIL of resist on glass from the ETFE, etching, metal deposition, and lift-off to leave the final pattern.

FIGURE 1 1 illustrates the basic fabrication process using nanoimprint lithography to make substrates with plasmonic nanohole array electrodes. The fabrication procedure shown in FIGURE 1 1 may appear straightforward; however, several processing challenges were encountered before quality nanopatterns could be achieved. Significant effort was put forth on procedural tuning and troubleshooting steps to solve these challenges. For instance, the concentration and spin coating speed of the resist were varied to find an optimal film thickness that had enough material present to mold sufficiently to the silicon master while also keeping the film thickness around the pillars thin enough to eliminate overly long dry etching times that damaged the pillars. The dry- etching time was also varied in conjunction with these parameters to find the optimal balance. Initially, only Au was deposited on the structures but the solution lift-off steps were harsh enough to remove the resist and the Au requiring the addition of a Cr "glue" layer. FIGURES 12A-12C show SEM pictures of the nanopillars on the silicon master mold, the imprinted nanoholes on the ETFE sub-mold, and the imprinted nanopillars in resist on glass after Cr (2 nm) /Au (28 nm) deposition, respectively. The pattern from the silicon master mold transfers nicely to the ETFE and on to the resist sample. However, as shown in FIGURES 12A-12C, the nanopillars and Au disks on top of them appear to be intact and relatively unaffected by the solution lift-off process. FIGURES 12A-12C illustrate SEM images with x20k magnification of (FIGURE 12 A) the silicon master mold, (FIGURE 12B) the imprinted ETFE sub mold, (FIGURE 12C) an imprinted resist sample after Cr/Au evaporation and (FIGURE 12D) after solution lift-off. The lift-off procedure is designed to dissolve the leftover resist essentially lifting off the nanopillars along with the top Au disks leaving behind only the Au nanohole pattern. Note that some areas of the film had exposed nanoholes without the "Au-tops" but much of the film looked as shown in the figure.

FIGURES 13A-13C illustrate tilted SEM images of (FIGURE 13 A) an imprinted resist sample after Cr/Au evaporation and solution lift-off (x20k) and (FIGURE 13B) a zoom in view (x40k) of FIGURE 13A. FIGURE 13C: A schematic of the effect seen in FIGURES 13A and 13B where the Au-disks on top of the resist nanopillars are bridged to the Au-nanohole array below. This prevents removal of many of the Au-disks even after solution lift-off to dissolve the resist.

Upon closer inspection, it was observed that the resist does appear to have dissolved away in the lift-off process but the "Au-tops" were left in place. FIGURES 13A-13B show tilted views of the nanopattern to reveal apparent bridging elements connecting the "Au-tops" to the Au-nanohole pattern underneath. For clarity, this bridging effect is depicted in the schematic in FIGURE 13C. Intuitively, this structure would seem to reduce light transmission through the nanostructure as the "Au- tops" block and scatter light coming through the holes underneath. Further troubleshooting was required to fabricate the open Au-nanohole structure.

Additional steps were taken to fabricate the desired nanohole structure and included altering the Au evaporation procedure and developing a peel-off procedure. To reduce the effect of shadowing during evaporation, a different evaporation machine with more precise sample alignment was used and the Au layer thickness was reduced from 28 to 23 nm with a Cr thickness of 2 nm. This helped reduce the bridging but did not solve the problem alone. Next, a peel-off procedure was added after the solution lift-off step. This peel-off procedure had to be optimized itself and involved covering the nanopattems in an optical adhesive, UV-curing the adhesive, and carefully peeling off the cured material. This peel-off essentially removed the more loosely held "Au-tops" while leaving the Au nanohole structure in place. The combined processing steps and peel-off procedure greatly improved the final nanopattern quality, as shown in FIGURES 14A- 14D. FIGURES 14A-14D illustrate SEM images of Au nanoholes on glass made with different spans of oxygen plasma etching followed by metal deposition, solution lift-off and peel-off. Etch time and average diameter are (FIGURES 14A and 14B) 30 s (-172 nm), (FIGURE 14C) 60 s (-159 nm) and (FIGURE 14D) 72 s (-152 nm). Also, varying the dry etching time for resist nanopillars allowed for nanoholes of different sizes to be made. FIGURES 14A-14B show the nanopattern with a 30 s etch time and average hole diameter of -172 nm. FIGURES 14C-14D show the nanopatterns with etch times of 60 and 72 sec giving slightly smaller nanoholes with average diameters of -159 and -152 nm, respectively. This is a valuable tool to experimentally examine the effects of hole size with constant pitch.

As mentioned previously, simulation results suggest that device performance would benefit from eliminating the Cr "glue" layer. Thus, we deployed a self-assembled monolayer (SAM) of 3-mercaptopropyltrimethoxysilane (MPTMS) as a molecular adhesive to bond Au to our substrates without the need for Cr. In order to form the SAMs, the substrates were first cleaned and treated with oxygen plasma to generate surface hydroxyl groups (-OH) and then placed under vacuum in a desiccator with a small volume of MPTMS for several hours. This procedure allows co-condensation of the methoxy (-OCH₃) groups of MPTMS with the surface hydroxyl groups of the substrate. The NIL process described previously works well to make quality nanopatterns using Cr as the "glue" layer. However, the use of MPTMS SAMs was combined with the NIL procedures under various fabrication pathways but did not produce patterns of sufficient quality.

As an alternative to NIL, we explored a method known as microcontact printing

^CP) which may also be better suited to increase nanostructured electrode area with our current silicon mold. This method allows a composite polydimethylsiloxane (PDMS)/hard-polydimethylsiloxane (h-PDMS) stamp to be made from the silicon master mold. FIGURES 15A and 15B show SEM images of nanopatterns transferred to a composite PDMS/h-PDMS stamp. Note that these stamps are quite robust as this one had been used greater than 20 times when these images were captured. The composite PDMS stamp is coated with alkanethiol molecules and then stamped onto the gold-coated substrate. The thiol molecules form a SAM on the Au surface that mirrors the nanopattern from the stamp. The SAM protects the Au layer so that when the substrate is etched only the unprotected areas are removed. This method allows for the nanopattern to be easily stamped onto a substrate of any size at the desired locations. Adaptation of this method to our procedures and materials system required testing and optimization of numerous conditions. For example, the stamp time needed to be tuned so that enough alkanethiols transfer to the Au to protect the metal while not oversaturating the surface causing loss of nanopattern resolution. Additionally, the etching time required tuning to find the condition to fully develop the nanopattern without damaging it. The thiol and etching solution concentrations also required testing and careful control. Nanopattern transfer was achieved with μCP, however, the nanohole patterns lacked fine resolution and it was determined that this method is not practical or reliable for use with nano- features as small as ours. One possible reason for this is that the alkanethiol molecules inside the nanoholes on the PDMS stamp migrate down to the substrate due to the small nanohole depth (less than -140 nm). Other methods related to μCP were also explored, including nanotransfer printing (nTP) and solvent-assisted nanomolding (SAN).

FIGURES 15A and 15B illustrate SEM images of nanopatterns transferred into a composite PDMS/h-PMDS stamp at magnifications of 20kX and 80kX, respectively.

FIGURE 16 illustrates the basic fabrication process using solvent-assisted nanomolding (SAN) to make substrates with plasmonic nanohole array electrodes with an MPTMS SAM to replace the Cr "glue" layer. The SAN method is similar to μCP in that it uses a composite PMDS/h-PDMS stamp to transfer the nanopattern from the silicon master mold to the desired substrate. However, SAN does not use alkanthiols on the stamp as the transfer medium. Instead, the substrate is silanized with MPTMS and then spin coated with poly(methyl methacrylate) (PMMA), which is used as a resist. The stamp is coated with a small drop of acetone and placed on the heated PMMA-coated substrate allowing the PMMA to soften and mold around the nano-features of the stamp. The stamp is left in place for several minutes with heating to evaporate the acetone and cure the PMMA. Then the stamp is removed leaving behind PMMA nanopillars. The residual PMMA at the foot of the pillars is removed by etching in oxygen plasma followed by deposition of Au and lift-off of the nanopillars in acetone to leave behind the final nanohole array. FIGURE 16 summarizes the fabrication steps using SAN. Similar to the μCP method, much work was required to tune the procedure for our nanopatterns including finding the optimal PMMA thickness, etching time and power. FIGURES 17A-17D illustrate SEM images of Au-only nanostructures on glass fabricated with SAN showing an area with (FIGURES 17A-17B) open nanoholes and (FIGURES 17C-17D) nanoholes covered by "Au-tops" at magnifications of (FIGURES 17A and 17C) x20k and (FIGURES 17B and 17D) x80k. High quality, Au- only nanostructures were achieved with the SAN method as seen in FIGURES 17A-17D. However, only small areas of open nanoholes are seen after the solution lift-off process in acetone as shown in FIGURES 17A and 17B. Most of the nanopatterned area is still covered by the "Au-tops" as seen in FIGURES 17C and 17D, similar to what is observed with the NIL method. This problem may be solved by a peel-off procedure as used in the NIL method. However, the peel-off method used with NIL was performed with these glass/MPTMS/Au nanostructures and the entire nanopatterns were removed. This is likely due to weaker bonding of the MPTMS SAMs to the substrates compared to Cr which may stem from less dense MPTMS coverage compared to the 2 nm thick evaporated Cr layers used with NIL. A new peel-off procedure was developed by exploring alternatives to the optical epoxy normally used. It is also clear from FIGURES 17A-17D that the nanoholes fabricated from SAN are smaller compared to those from NIL (see FIGURE 14B). Further tuning of the SAN procedure is required to produce nanoholes of the target diameter of -175 nm and reduce bridging of these "Au-tops" to the underlying nanohole array.

The transmission spectra of various substrates were measured and calculated using FDTD simulations as shown in FIGURE 18. FIGURE 18 graphically illustrates measured (solid lines) and FDTD simulated (dashed lines) transmission spectra of bare glass, glass/ITO, glass/ Au unpatterned film, glass/ 2 nm Cr/ 23 nm Au 175-225 nanostructure and glass/25 nm Au-only 175-225 nanostructure.

The measured transmission of bare glass substrates shows broad-spectrum transmission of greater than 90% and the measured and simulated spectra for glass/ITO substrates show reasonable agreement with greater than 80-90% transmission. Measured and simulated spectra of unpatterned Au films on glass show good agreement and transmission less than 40%. Once a 175-225 nanostructure is put into an Au-only film on glass, the simulated spectra shows increased transmission to 50-60%) from 400-580 nm with a pronounced dip due to plasmonic absorption at -610 nm and greatly increased transmission of 70-90% from -630-900 nm. However, once a thin Cr "glue" layer is put in place, the simulated transmission of a 175-225 nanostructured substrate dampens out to -40% across the spectrum. The measured transmission of a glass/Cr/Au 175-225 nanostructured substrate shows similar spectra with broad transmission of 40-60%. The Cr layer appears to lower the plasmonic absorption and character of the nanostructure which will dampen the high intensity electric fields of SPPs that act to couple light into the device. This explains the trends seen in FIGURE 18 that devices with Au-only nanopatterns can have enhanced active layer absorption compared to those with Cr/Au nanostructures. However, even with lower light transmission of these nanostructures compared to ITO, light trapping inside the device by the resonant cavity effect can still yield active layer absorption that is comparable to ITO-based devices. Additionally, these spectra are for bare substrates before they are made into full devices. The properties of plasmonic nanostructures are highly sensitive to the surrounding materials and it has been observed that light reflection and absorption by a bare plasmonic nanostructure reduced 2 to 6 fold when the structures were formed into a complete device. Unlike ITO where the substrate transmission is directly correlated to active layer absorption, the plasmonic properties and resonant cavity effects of these nanostructured electrodes in completed-OPV devices have significant impacts on device performance that are not made evident by bare substrate transmission alone.

Plasmonic Nanostructured Electrodes Deployed in OPVs on Glass Substrates High quality plasmonic Au-nanostructures with Cr "glue" layers were fabricated using NIL and tuned lift-off/peel-off procedures. These Cr/Au nanostructures, as seen in FIGURES 14A-14B, were deployed on bare glass substrates as cathodes in ITO-free inverted devices with the structure shown in FIGURE 19A. ITO-free OPV devices with a power conversion efficiency (PCE) as high as 5.70% were achieved. FIGURE 19B shows the current-voltage curves of devices on ITO versus those on Cr/Au nanostructures with ~32 nm thick ZnO ETLs and varied active layer thicknesses. These ZnO ETLs were formed using the sol gel method with a 0.3 M precursor solution spun at 2000 rpm and baked under static conditions. For each active layer thickness, the ITO and Cr/Au nanostructure-based devices had approximately equal Voc's and FF's, shown in Table 4, indicating that the electrical characteristics of these two types of electrodes are very similar in these OPVs. The primary difference in PCE between these two cathodes was due to a lower J_Sc for devices with Cr/Au nanostructures. For a thicker active layer of -122 nm, both types of devices show lower FF's due to increased recombination as compared to the devices with thinner active layers. The Jsc of the nanostructured device is 61% of the ITO-based devices (i.e. 9.82 vs. 15.98 mA/cm²). As the active layer thickness decreases closer to the optimum at -93 nm, the Jsc's of the devices increase to 11.80 mA/cm² and 17.43 mA/cm² for Cr/Au and ITO, respectively. The nanostructured device Jsc increased to 68% of that of the ITO control at the optimal active layer thickness. When the active layer thickness is decreased further down to -77 nm, the Jsc of the ITO device drops to 15.90 mA/cm² while that of the Cr/Au device remained approximately the same at 11.39 mA/cm², which is now 72% of the ITO device. As seen in our previous work, the Jsc of these ITO-based devices is very sensitive to active layer thickness with decreased photocurrent observed as the thickness varies above or below the optimum. However, due to the resonant cavity effect created by the nanostructured electrodes, the Jsc's of these ITO-free devices is retained as the active layer becomes optically thin. These performance trends agree with those in the FDTD calculated active layer absorption spectra.

FIGURE 19A illustrates schematics of an inverted device with a plasmonic Au-nanostructured cathode. FIGURE 19B presents current- voltage curves for inverted devices with ITO (solid lines) and Cr/Au nanostructured (dotted lines) cathodes with sol- gel derived ZnO ETLs of -32 nm thickness and varied active layer thicknesses of -122 nm, -93 nm, and -77 nm. Results are summarized in Table 4.

Table 4. The performance parameters corresponding to the J-V curves in FIGURE 19B for inverted devices on ITO and Cr/Au nanostructured cathodes with -32 nm thick sol-gel derived ZnO ETLs and varied active layer thicknesses. Active layer film thickness measurements are from profilometry of films on bare glass/ITO.

ITO -32 122 +/- 6 0.75 62.2 15.98 7.46

ITO -32 93 +/- 16 0.76 64.5 17.43 8.55

ITO -32 77 +/- 7 0.76 65.9 15.90 7.98

Cr(2nm)/Au(23nm) -32 122 +/- 6 0.74 61.2 9.82 4.45

Cr(2nm)/Au(23nm) -32 93 +/- 16 0.75 64.5 11.80 5.70

Cr(2nm)/Au(23nm) -32 77 +/- 7 0.75 65.2 11.39 5.59 Despite these effects, the Cr/Au-based devices show lower Jsc's and PCE's compared to the ITO controls. The external quantum efficiency (EQE) of a Cr/Au versus ITO device with -32 nm ZnO ETLs and -77 nm active layers is shown in FIGURE 20. The EQE of the devices with nanostructured electrodes is over 20% lower than the ITO control at wavelengths of -460-680 nm, in agreement with the simulation results discussed previously. To maximize Jsc and EQE, it is important to ensure that the ZnO thickness is optimal for the resonant cavity and to eliminate the need for the Cr "glue" layer, which can interfere with the plasmonic properties of the Au film and reduce active layer absorption as mentioned before.

FIGURE 20 illustrates external quantum efficiency (EQE) spectra of inverted devices on ITO (solid line) and Cr/Au nanostructured (dotted line) cathodes with -32 nm thick ZnO and -77 nm thick active layers.

The reflectance spectra of the ITO and Cr/Au-based devices with -32 nm ZnO ETLs and -77 nm thick active layers were measured and compared to spectra of FDTD-simulated devices with similar thicknesses in FIGURE 21. FIGURE 21 illustrates measured reflectance spectra of the ITO and Cr/Au-based devices with -32 nm ZnO ETLs and -77 nm active layers shown in Table 4. The dashed lines show the FDTD-simulated spectra of ITO and Cr/Au-based devices with similar thicknesses. The measured reflectance (solid lines) of ITO and the nanostructured based devices is very close between 550-700 nm averaging -12%. Despite similar reflectance, the Cr/Au nanostructured device showed a Jsc that is ~28%> lower than the ITO device. This shows that the difference in active layer absorption between these devices is not due to higher reflectance but likely due to parasitic Au-layer absorption that is not offset by light trapping effects of the resonant cavity. The FDTD-simulated spectra (dashed lines) have similar trends to the measured data but deviate in some wavelength ranges likely due to error in film thickness and refractive index as well as imperfect nanohole patterns in the experimental devices.

In addition, any damage to the nanostructures could adversely affect light transmission and trapping in these OPVs. The ZnO sol gel films made here use a baking temperature of 200°C which can lead to damage of nanometer scale features in thin Au films. However, when baked with the sol gel layer on top, these nanostructures appear intact and unchanged as seen in the SEM images in FIGURES 22 A and 22B. FIGURES 22A and 22B illustrate SEM images of Cr/Au nanostructures on glass coated with a ZnO sol gel film after baking at 200°C at magnifications of x20k and xl60k, respectively.

FDTD calculations suggest that minimizing ZnO thickness is ideal for maximizing active layer absorption. The ZnO ETLs in the previously discussed devices were ~32 nm thick and so the thickness was reduced to determine if performance could be improved. FIGURE 23. Current- voltage curves for inverted devices with ITO (solid lines) and Cr/Au nanostructured (dotted lines) cathodes with ZnO ETLs of ~24 nm thickness and varied active layer thicknesses of -122 nm (dark blue), ~93 nm (blue) and -77 nm (light blue). FIGURE 23 shows the current- voltage curves and Table 5 shows the performance parameters for devices with ~24 nm thick ZnO ETLs. Similar to the devices with ~32 nm thick ZnO, the Voc's and FF's are approximately constant between the ITO and Cr/Au nanostructured devices at each active layer thickness. However, in this case the Jsc's remain fairly constant for the nanostructured devices as active layer thickness is decreased and are significantly lower than the previous devices with ~32 nm ZnO with values ranging from 8.39-8.86 mA/cm². One possible reason for this difference in behavior is that the Cr/Au nanostructures have a total thickness of 25 nm, i.e. the nanoholes are ~25 nm deep. The ZnO thickness of ~24 nm is based on measurements taken from bare glass/ITO substrates. The ZnO sol gel solution appears to fill the nanoholes, as seen in FIGURES 22A and 22B, which means that the ZnO thickness on top of the nanostructure itself may be non-uniform and much thinner than 20 nm. In addition to this, small variations in nanopattern quality may also lead to these unpredicted trends. Table 5. Performance parameters corresponding to the J-V curves in FIGURE 23 for inverted devices on ITO and Cr/Au nanostructured cathodes with ~24 nm thick sol-gel derived ZnO ETLs and varied active layer thicknesses.

ITO -24 122 +/- 6 0.74 58.9 15.41 6.75 ITO -24 93 +/- 16 0.75 63.9 16.52 7.89 ITO -24 77 +/- 7 0.76 64.4 16.25 7.91

Cr(2nm)/Au(23nm) -24 122 +/- 6 0.74 60.9 8.86 3.97

Cr(2nm)/Au(23nm) -24 93 +/- 16 0.74 63.9 8.39 3.95

Cr(2nm)/Au(23nm) -24 77 +/- 7 0.74 63.2 8.73 4.07

FIGURE 24A presents current-voltage curves for inverted devices with thicker 48 nm ZnO ETLs on ITO and Cr/Au nanostructures with varying nanohole diameters of -172 nm, -168 nm, -159 nm and -152 nm. FIGURE 24B presents EQE spectra for the ITO device and Cr/Au device with -172 nm nanoholes shown in FIGURE 24 A. Nanopattems with varied nanohole diameters were fabricated using NIL by varying the resist etching time after imprinting yielding sizes of -172 nm, -168 nm, -159 nm and -152 nm with etching times of 30 s, 48 s, 60 s and 72 s, respectively. These Cr/Au nanostructures were used in inverted devices with thicker ZnO ETLs of -48 nm. The current- voltage curves and performance parameters of these devices are shown in FIGURE 24A and Table 6, respectively. The nanostructured device etched for 30 s (i.e. the same condition for all previous Cr/Au devices mentioned) shows a Voc and FF that are approximately the same compared to the ITO device but the Jsc is 51% lower (see Table 6). This device's Jsc of 7.52 mA/cm² is lower than those of the devices with thinner ZnO ETLs in agreement with FDTD simulations. The EQE spectra in FIGURE 24B clearly show this lower absorption across a broad wavelength range. Also, decreasing the nanohole diameter generally decreased the Jsc while it increased the FF. This is in agreement with the calculations mentioned previously and with what we expect intuitively. Decreasing the nanohole size blocks more light from entering the device while the increased metal electrode area from thicker "walls" between nanoholes decreases the series resistance of the electrode. The primary impact of a decreased series resistance is to increase the device FF as we observe. Therefore, electrode sheet resistance and optical transmission, along with the various layer thicknesses, need to be balanced for peak performance.

Table 6. The performance parameters corresponding to the J-V curves in

FIGURE 24A for inverted devices with sol-gel derived thicker 48 nm ZnO ETLs on ITO and Cr/Au nanostructures with varying nanohole diameters by changing the resist etching time.

ITO n/a n/a 0.74 69.8 15.24 7.89

Cr(2nm)/Au(23nm) 30 -172 0.72 68.3 7.52 3.70

Cr(2nm)/Au(23nm) 48 -168 0.72 71.1 6.79 3.47

Cr(2nm)/Au(23nm) 60 -159 0.71 69.9 6.44 3.21

Cr(2nm)/Au(23nm) 72 -152 0.72 72.2 6.57 3.43

The reflectance spectra were measured for the ITO and Cr/Au-based devices with -48 nm thicknesses shown above and compared to the FDTD-simulated spectra in FIGURE 25. FIGURE 25 illustrates measured reflectance spectra of the ITO and Cr/Au- based devices with -48 nm ZnO ETLs shown in Table 6. The dashed lines show the FDTD-simulated spectra of ITO and Cr/Au-based devices with similar thicknesses. Similar to the devices with the thinner ZnO thicknesses shown previously, these devices show measured reflectance spectra that are close to -10% from -550-680 nm. The Cr/Au device shows lower reflectance than ITO from 450-550 nm and from 700 nm to the PTB7 E_g. Yet the Cr/Au device's Jsc is -51% lower than the ITO device indicating that device reflectance is not the primary source of lost active layer absorption between the two cases. The FDTD-simulated reflectance (dashed lines) follows the same general trends as the experimental devices with differences likely due to error in film thickness and refractive index as well as imperfect nanohole patterns in the experimental devices.

ZnO films fabricated with the sol gel method typically require baking at 200°C which is not compatible with many flexible substrates. Therefore, we fabricated devices with ZnO NPs synthesized in our lab that do not require a baking step and are formed via a layer-by-layer spin coating method. More detailed discussion of these ZnO NP films will follow in the next section when used with flexible substrates. FIGURE 26A presents current-voltage curves for inverted devices with -44 nm ZnO NP ETLs on ITO (solid line) and Cr/Au nanostructured (dotted line) electrodes. FIGURE 26B presents SEM image of a Cr/Au nanostructure covered with a ZnO NP film that measured -44 nm on a bare glass/ITO substrate. The current-voltage curves and performance parameters of inverted devices with -44 nm thick ZnO NP ETLs and active layer thickness of -103 nm are shown in FIGURE 26 A and Table 7. These devices performed similarly to devices with ZnO sol gel ETLs with a similar thickness of -48 nm. The Voc and FF are close to the ITO control devices while the Jsc is 46% of ITO. Based on simulation, the ZnO film should be thinner to increase the active layer absorption in these devices as discussed previously. FIGURE 26B shows an SEM image of a ZnO NP film coating a Cr/Au nanopatterned electrode. This film was fabricated by spin coating three sequential layers of the filtered stock ZnO NP solution at 3000 rpm to build up the film that measures ~44 nm on bare glass/ITO substrates. The SEM image shows how the NP coverage is non-uniform with areas of thinner and thicker coverage. Regardless, this condition appears to adequately coat the nanostructure as seen by the normal Voc and FF. However, devices with thinner ZnO NP films (e.g. only one layer spun), were fabricated and measured at ~32 nm on bare glass/ITO substrates (not shown). The control ITO devices at this condition showed typical performance with a PCE as high as 7.06%, however, the devices on Cr/Au nanostructures were shorted and showed no Voc- This is possibly due to a combination of inadequate NP film coverage over the nanostructured electrode with nanostructure defects that penetrate the overlying layers to make contact with the top anode. Devices with thicker ZnO NP ETLs of ~66 nm were also made (not shown) and both the ITO and Cr/Au nanostructure-based devices showed diminished Offs of 58%o and 48%, respectively, and lower Jsc's leading to low PCEs of only 5.26% and 2.09%), respectively. Recall that ITO-based devices with ZnO sol gel films showed no sensitivity to thicknesses in this range. The ZnO NP films are more porous with less densely packed crystalline character compared to the sol gel films leading to higher series resistance for these thick films.

Table 7. The performance parameters corresponding to the J-V curves in FIGURE 26 A for inverted devices with ZnO NP ETLs on ITO and Cr/Au electrodes.

ITO 44 +/- 8 103 +/- 15 0.74 71.7 14.31 7.58

Cr(2nm)/Au(23nm) 44 +/- 8 103 +/- 15 0.72 68.0 6.60 3.25

In addition to controlling the ZnO layer thickness, simulation results suggest that eliminating the Cr "glue" layer in the nanostructured electrodes can further enhance active layer absorption in these devices. A new fabrication technique, i.e. SAN, was used to make Au-only nanostructures as seen in FIGURES 17A-17D. As discussed, these Au- only nanostructures have small areas of open nanoholes but are largely composed of nanoholes still covered by "Au-tops" and new procedures are being developed to resolve this issue. Still, these preliminary Au-only nanostructured electrodes were deployed in inverted devices and the current-voltage curves and performance parameters are shown in FIGURE 27 and Table 8, respectively. FIGURE 27 illustrates current- voltage curves for inverted devices with ITO, Cr/Au, and Au-only nanostructured electrodes with -24 nm ZnO ETLs and -77 nm thick active layers. The top inset is an SEM image of a Au-only nanostructure with smaller holes and large areas covered with unlifted Au disks and the bottom inset is that of a Cr/Au nanostructure. These devices show similar Voc to the ITO and Cr/Au-based devices and a higher FF likely due to smaller nanohole diameters and the presence of more Au area. The Jsc of the Au-only device is 6.11 mA/cm², which is only 38% of that of the ITO device. Still, even with diminished transmission into the device due to the non-ideal nanostructure, the PCE reached 3.14% showing promise for making high efficiency devices with optimized Au-only nanostructures.

Table 8. The performance parameters corresponding to the J-V curves in FIGURE 27 for inverted devices on ITO, Cr/Au and Au-only nanostructured electrodes.

ITO -24 77 +/- 7 0.76 64.4 16.25 7.91

Cr(2nm)/Au(23nm) -24 77 +/- 7 0.74 63.2 8.73 4.07

Au (25nm) -24 77 +/- 7 0.73 69.8 6.11 3.14

It has been observed that plasmonic nanostructures can enhance light coupling into a device over a broad range of light incident angles and wavelengths. This is an obvious advantage for PV devices to help improve PCE by harvesting more diffuse light and possibly eliminating the need for expensive solar-tracking systems. We tested the performance of inverted devices with -48 nm ZnO ETLs on ITO and Cr/Au nanostructured electrodes with the incident light angle varied from 0° to 60°. The photocurrent ratio (i.e. ratio of Jsc at an angle to the Jsc at 0° or normal to device surface) at various incident angles for these devices is shown in FIGURE 28. FIGURE 28 illustrates photocurrent ratio versus incident light angle for inverted devices on ITO and Cr/Au nanostructured electrodes. The photocurrent ratios for both types of devices drops below -90% past a 30° incident angle to 47% and 45% for ITO and Cr/Au, respectively, at 60°. Both devices retain Voc and FF over the various incident angles but show very similar decreases in Jsc as shown in Table 9. The lack of any observed absorption enhancement at different angles of incidence for the Cr/Au nanostructures is possibly due, in part, to the presence of the 2 nm thick Cr layer. Simulation has shown that this layer dampens the plasmonic absorption, which creates the strong fields associated with SPPs that couple to light at various angles.

Table 9. The performance parameters corresponding to devices tested under varied light incident angle.

Angle of Voc FF Jsc PCE Jsc

Substrate

Incidence (V) (%) (niA/cm²) (%) Ratio

ITO 0° 0.73 0.69 12.84 6.45 100%

15° 0.73 0.69 12.25 6.18 95%

30° 0.73 0.68 11.55 5.73 90%

45° 0.72 0.69 9.50 4.73 74%

60° 0.71 0.69 6.00 2.92 47%

Cr/Au 0° 0.70 0.69 6.52 3.14 100%

15° 0.70 0.68 6.46 3.09 99%

30° 0.70 0.68 5.95 2.83 91%

45° 0.69 0.68 4.87 2.31 75%

60° 0.68 0.66 2.91 1.29 45%

Plasmonic Nanostructured Electrodes Deployed in OPVs on Flexible Substrates

Compatibility with flexible substrates is a big advantage of OPVs as high throughput R2R processes require flexibility and devices made on these materials bring the possibility of unique and widespread usage cases. As mentioned, replacing ITO as a transparent electrode is an important requirement for these devices to be produced on a large scale and plasmonic electrodes are a prime candidate for replacement as they are mechanically robust which is essential for flexible solar cells.

Before making flexible devices with plasmonic electrodes, we fabricated PET/ITO based devices to optimize our inverted structure with this new substrate. We typically use the sol gel method to deposit high quality ZnO electron transport layers. However, the sol gel method uses a high baking temperature of 200°C, which will warp and damage the PET/ITO substrates. Alternate methods for making ZnO films include using pre-synthesized ZnO nanoparticles (NPs), which require lower baking temperatures.

We compared devices with glass and PET substrates and also with different types of ZnO NPs. ZnO NPs are commercially available and can be an easy and relatively cheap option for making ZnO films. We deposited ZnO NPs from Sigma Aldrich and baked the films at 140°C, which worked reasonably well with the PET substrates. These NPs have a silane-based ligand and are dispersed in water requiring a baking step for sufficient drying. We also synthesized ZnO NPs in-house, which have a much smaller - OH group ligand and are dispersed in a mixture of n-butanol, methanol and chloroform according to procedures in the literature. The mixture of solvents for these NPs dries very quickly while spin coating and a baking step is not required which is advantageous. These ZnO NP films are very stable and can be formed by spin coating in a layer-by-layer fashion while the purchased NPs can be rinsed off with water. Also, it is expected that these in-house synthesized, bake-free NPs will form higher quality films with a smoother surface and lower sheet resistance brought about by less insulating ligands.

We compared devices using both types of ZnO films on glass and PET to find an optimal device design for use with plasmonic electrodes. FIGURES 29A-29C illustrate: FIGURE 29A: Photos of typical inverted devices on flexible PET/ITO substrates; FIGURE 29B: Current- voltage curves of inverted devices on glass/ITO (dotted lines) and on PET/ITO (solid lines) comparing ZnO films from NPs synthesized in-house and purchased from Sigma-Aldrich; FIGURE 29C: Current- voltage curves of devices on PET/ITO with in-house synthesized ZnO NPs but with varied active layer spin speeds of 800 rpm, 1000 rpm, 1200 rpm and 1400 rpm for re-optimization. FIGURES 29A and 29B show some photos of typical devices on PET and the J-V curves for devices using these different NP films on glass (dotted lines) versus PET (solid lines), respectively. FIGURE 29B includes data obtained from devices using the ZnO NPs purchased from Sigma- Aldrich and used with both glass and PET substrates. It is interesting to see that the average PCE of the PET device at 6.00% was comparable and even slightly higher than that of the glass device at 5.87% (see Table 10). The FF was slightly lower for the PET device as expected because the PET/ITO sheet resistance is lower than that of glass/ITO. The slight increase in PCE stemmed from an increase in Jsc for the PET device, which may be due to an optical change, however, the standard deviation in the Jsc for these devices was relatively high making it comparable to the glass device.

When comparing the glass devices with the purchased ZnO NPs versus the synthesized NPs, there is a clear increase in both FF and Jsc, which boosts the average PCE from 5.87% to 6.72%, respectively, with a PCE as high as 7.03%>, when using the synthesized NPs. This agrees with our expectation that the synthesized NPs provide a higher quality ZnO film.

However, when comparing these different ZnO NPs on flexible PET/ITO substrates, the FF and Jsc dropped lowering the average PCE from 6.00% to 5.36%> with the synthesized NPs. This trend is opposite to what was seen with the glass devices leading us to check the active layer optimization with this new ZnO film. Using the purchased ZnO NPs, it was previously determined that a spin speed around 1400 was optimal. Indeed, the PET devices with synthesized ZnO NPs showed improved FF and Jsc with a PCE as high as 6.91% when the active layer was made thicker by lowering the spin speed from 1400 rpm to 1200 rpm (see FIGURE 29C and Table 10). Similar to earlier work, the change in ZnO film conditions and morphology requires re-optimization of the active layer. It is also noteworthy that these flexible PET devices showed performance comparable to the glass devices and approached 7% PCE.

Table 10. Performance parameters for inverted devices with varied substrate, ZnO NP film, and active layer spin speed corresponding to the J-V curves shown in FIGURES 29B and 29C.

Glass/ITO NP (SA) 1400 0.73 +/- 0.00 62.9 +/- 1.1 12.86 +/- 0.19 5.87 +/- 0.19 6.16

NP

Glass/ITO 1400 0.74 +/- 0.00 67.9 +/- 0.6 13.42 +/- 0.27 6.72 +/- 0.16 7.03

(Synth)

PET/ITO NP (SA) 1400 0.72 +/- 0.00 60.6 +/- 1.0 13.71 +/- 0.79 6.00 +/- 0.30 6.25

NP

PET/ITO 800 0.72 +/- 0.01 56.5 +/- 4.7 13.43 +/- 0.58 5.45 +/- 0.70 6.01

(Synth)

NP

PET/ITO 1000 0.64 +/- 0.07 44.6 +/- 10.8 12.15 +/- 0.54 3.55 +/- 1.17 4.89

(Synth)

NP

PET/ITO 1200 0.73 +/- 0.00 63.2 +/- 1.7 13.98 +/- 0.59 6.45 +/- 0.27 6.91

(Synth)

NP

PET/ITO 1400 0.73 +/- 0.01 56.9 +/- 6.4 13.01 +/- 0.42 5.36 +/- 0.50 6.00

(Synth)

The optimized inverted structure using the bake-free synthesized ZnO NPs and the fiexible PET substrates was then deployed to make the first ITO-free flexible devices with Cr/Au nanostructured electrodes. The nanostructured electrodes on bare PET were made with the same procedures used for the rigid glass substrates. FIGURES 30A and 30B show a photograph of the nanopatterns on PET and the current- voltage curves for the device with the nanostructured electrode versus a control on PET/ITO. The PET/Cr/Au nanostructured device shows a PCE of 1.82%, which is much lower, compared to the control device at 5.93%. Note that even the control device in this experiment did not show peak performance, which was due to non-optimal active layer thickness. The primary cause for the lower PCE was a ~67%> decrease in Jsc- This is most certainly due to lower light transmission because of the nanopattern quality. Similarly to the rigid devices, these nanopatterns on PET have "Au-tops" remaining after the solution lift-off step that block the nanoholes underneath. For the glass substrates, this was remedied by using the peel-off procedure described previously. However, this peel-off process did not work with these PET substrates and a new procedure is being developed. Therefore, the Cr/Au nanostructured electrode shown in FIGURE 3 OA and the device using this electrode in FIGURE 30B did not have the "Au-tops" removed.

FIGURES 30A and 30B. FIGURE 30A: Photograph of a bare PET substrate with Cr/Au nanostructures fabricated by NIL. FIGURE 30B: Current-voltage curves of a control device on PET/ITO and an ITO-free device with a Cr/Au nanostructured electrode on PET. To the best of the inventors' knowledge, this is the first report of flexible and ITO-free devices using plasmonic nanostructured electrodes. Table 11 summarizes these results. Table 11. Performance parameters for a control device and a device with a Cr/Au nanostructured electrode on bare PET corresponding to the J-V curves in FIGURE 3 OB.

Voc FF Jsc PCE

Substrate

(V) (%) (mA/cm²) (%)

PET/ITO 0.73 63.6 12.79 5.93

PET/Cr/Au 0.70 62.0 4.22 1.82

OPVs offer compatibility with ultrathin and conformable substrates which can lead to a wide range of potential products and uses for solar energy harvesting devices, such as solar powered clothing and gear for ubiquitous energy generation in remote areas or non-obtrusive, low-power applications for hidden sensors and the "internet of things". As mentioned, ITO is not suitable for such substrates while the nanostructured electrodes discussed in this work have great potential as an alternative.

FIGURE 31 illustrates current- voltage curve for an inverted ITO-free device on a conformable Parylene substrate with Cr/Au nanostructured electrodes. The device performance parameters and a photo of the device wrapped around the end of a pen are included in the inset. FIGURE 31 shows the current- voltage curve of an ITO-free inverted device using Cr/Au nanostructured electrodes on a thin (-100 μιη) parylene substrate. This substrate is highly flexible and conformable in that it can be wrapped around an object like plastic wrap. The inset to FIGURE 31 shows the substrate wrapped over the end of a pen. The device shows a relatively low Voc, FF and Jsc yielding a PCE of ~1%. However, this performance is in spite of several conditions that have yet to be optimized. The Cr/Au nanopatterns used in these devices have the "Au-tops" covering the nanoholes because the peel-off procedure used for glass substrates is not compatible. These devices also used a new amorphous ZnO sol gel condition with a baking step at 100°C. This new sol gel procedure has not been optimized yet and the ZnO film quality at this condition requires further study. This device also has a thicker ZnO film, which should be made thinner as discussed before. Therefore, these conformable devices have several conditions requiring optimization and show promise for increased efficiencies in new conformable OPV devices.

Conclusions

In conclusion, we designed and fabricated plasmonic nanostructures and deployed them as electrodes to replace ITO in OPV devices. The Au-nanohole arrays were designed using FDTD simulations to maximize both active layer light absorption and plasmon-induced electric field enhancements while minimizing device reflectance and showed nanopatterns consisting of 175 nm nanoholes with a 225 nm pitch were optimal. We found that active layer absorption is significantly impacted by the ZnO thickness, which affects the optical field distribution inside the resonant cavity formed between the plasmonic nanostructured electrode and the top Ag electrode. High quality Cr/Au nanostructured electrodes were fabricated by nanoimprint lithography and deployed in ITO-free inverted devices on glass. Devices with thinner ~32 nm ZnO ETLs showed a PCE as high as 5.70% and higher Jsc than devices on thicker ZnO, in agreement with simulation. In addition, as the active layer was made optically thin, ITO-based devices showed a diminished Jsc while the resonant cavity effect from plasmonic nanostructured electrodes retained the Jsc- Simulation also indicates that the Cr "glue" layer dampens the plasmonic properties of the nanostructures limiting active layer absorption. A new method, i.e. solvent-assisted nanomolding, was used to fabricate Au-only nanostructures eliminating the Cr "glue" layer and is still under development. Initial devices fabricated with these Au-only nanostructures showed PCEs as high as 3.14%, even with nanoholes mostly covered by "Au-tops" and non-optimal layer thicknesses, showing promise for improved ITO-free devices. Nanostructured electrodes were used in ITO-free inverted OPVs on flexible substrates, which required different ZnO films with lower temperature treatments. Bake-free, in-house synthesized ZnO NPs made higher quality films than purchased NPs and produced devices on glass/ITO with a PCE as high as 7.03% and on PET/ITO with a PCE as high as 6.91%, which is noteworthy for a flexible device. Preliminary ITO-free, flexible devices on PET showed a PCE of 1.82%. ITO-free devices were also fabricated on highly flexible, ultrathin and conformable Parylene substrates yielding an initial PCE of over 1%. To our knowledge, this is the first time nanopatterned plasmonic electrodes have been applied to highly flexible ITO-free OPVs. While illustrative embodiments have been illustrated and described, it will be appreciated that various changes can be made therein without departing from the spirit and scope of the invention.

Claims

CLAIMS The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. A photovoltaic device, comprising a photovoltaic layer between a hole- collecting electrode and an electron-collecting electrode;

2. The photovoltaic device of Claim 1, wherein the photovoltaic layer is selected from the group consisting of a bulk-heterojunction material, a quantum dot material, a perovskite material, and a combination thereof.

3. The photovoltaic device of Claim 1, wherein the photovoltaic layer comprises a bulk-heterojunction material, comprising an electron donor material heterogeneously mixed with an electron accepting material.

4. The photovoltaic device of Claim 3, wherein the electron donor material is selected from the group consisting of P3HT, MEH-PPV, PTB7, PTB7-Th, F8T2, APFO- 3, PTAA, PBTTT, PBnDT-DTBT, PCDTBT, PCPDTBT, F-PCPDTBT, PDTP-DFBT, PPDT2FBT, PDPP-TT-T, PBDTTPD, PIDTDTQx, and other semiconducting materials based on poly-(2,7-carbazoles), carbazole-benzothiadiazole copolymers, diketopyrrolopyrrole (DPP) copolymers, benzodithiophene (BDT) derivatives, and indacenodithiophene (IDT) based copolymers.

5. The photovoltaic device of Claim 3, wherein the electron acceptor material is selected from the group consisting of PC₆oBM, PC₇₀BM, ICBA, and other semiconducting materials based on fullerene derivatives such as bis-adducts, diphenylmethano fullerenes, dimethylphenylmethano fullerene bis-adducts, and endohedral fullerenes, as well as, additional n-type semiconducting materials based on naphthalene diimide-perylene diimide copolymers and inorganic acceptor nanocrystals and quantum dots (e.g. PbS, PbSe, CdS, CdSe, ZnO, and pyrite FeS₂).

6. The photovoltaic device of Claim 3-5, wherein the first photovoltaic portion and the second photovoltaic portion each comprise bulk-heterojunction materials.

7. The photovoltaic device of Claim 1, wherein the transparent plasmonic electrode is formed from a metal selected from the group consisting of gold, silver, copper, aluminum, platinum, palladium, and combinations thereof.

8. The photovoltaic device of Claim 1, wherein the transparent plasmonic electrode is defined by a transmission of at least 40% at the first absorption peak wavelength.

9. The photovoltaic device of Claim 1, wherein the first plurality of apertures and the second plurality of apertures have a shape independently selected from the group consisting of a circle, a distorted circle, a square, a rectangle, a triangle, an x-shape, and a polygon.

10. The photovoltaic device of Claim 1, wherein the first plurality of apertures and the second plurality of apertures have the same shape in the two laterally distinct portions.

11. The photovoltaic device of Claim 1 , wherein the first plurality of apertures and the second plurality of apertures have a smallest dimension of from 5 nm to 5 microns.

12. The photovoltaic device of Claim 1, wherein the first plurality of apertures and the second plurality of apertures are arranged in a pattern independent selected from the group consisting of square, rectangular, hexagonal, and triangular.

13. The photovoltaic device of Claim 1, wherein the first plurality of apertures and the second plurality of apertures have a center-to-center pitch dimension of from 50 nm to 5 microns.

14. The photovoltaic device of Claim 1, wherein the transparent plasmonic electrode has a thickness of 5 nm to 1 micron.

15. The photovoltaic device of Claim 1, wherein the first photovoltaic portion and the second photovoltaic portion have a shape independently selected from the group consisting of a triangle, a square, a rectangle, and a polygon.

16. The photovoltaic device of Claim 1, wherein the first photovoltaic portion and the second photovoltaic portion have a largest dimension of 1 mm.

17. The photovoltaic device of Claim 1, wherein the first absorption peak wavelength and the second absorption peak wavelength are between 250 nm and 2500 nm.

18. The photovoltaic device of Claim 1, wherein the photovoltaic layer further comprises a third laterally distinct portion that includes a third photovoltaic portion having a third absorption peak wavelength that is different than the first absorption peak wavelength and the second absorption peak wavelength.

19. The photovoltaic device of Claim 1, wherein the hole-collecting electrode is a high-work-function electrode and the electron-collecting electrode is a low-work- function electrode.

20. The photovoltaic device of Claim 1, wherein the photovoltaic device is a conventional photovoltaic device and the transparent plasmonic electrode is the hole- collecting electrode.

21. The photovoltaic device of Claim 20, wherein the electron-collecting electrode is a material selected from the group consisting of Al, Ag, Au, and Cu.

22. The photovoltaic device of Claim 20, further comprising a hole transport layer between the transparent plasmonic electrode and the photovoltaic layer.

23. The photovoltaic device of Claim 20, further comprising an electron transport layer between the electron-collecting electrode and the photovoltaic layer.

24. The photovoltaic device of Claim 1, wherein the photovoltaic device is an inverted photovoltaic device and the transparent plasmonic electrode is the electron- collecting electrode.

25. The photovoltaic device of Claim 24, wherein the hole-collecting electrode is a material selected from the group consisting of Al, Ag, Cu, Au, Pd, and Pt.

26. The photovoltaic device of Claim 24, further comprising an electron transport layer between the transparent plasmonic electrode and the photovoltaic layer.

27. The photovoltaic device of Claim 24, further comprising a hole transport layer between the hole-collecting electrode and the photovoltaic layer.

28. The photovoltaic device of Claim 1, wherein the photovoltaic device does not include indium-tin oxide or fluorine tin oxide.

29. The photovoltaic device of Claim 1, further comprising a substrate adjacent one of the hole-collecting electrode and the electron-collecting electrode.

30. The photovoltaic device of Claim 29, wherein the substrate is rigid or deformable.

31. The photovoltaic device of Claim 29, wherein the substrate is a material selected from the group consisting of glass, PET, PEN, FEP, parylene, cellulose nanocrystal, and other transparent substrates.

32. The photovoltaic device of Claim 1, wherein the device is hermetically sealed.

33. A photovoltaic article comprising a plurality of photovoltaic devices according to Claim 1.

34. The photovoltaic article of Claim 33, wherein the article is bendable.

35. The photovoltaic article of Claim 33, wherein the article is conformable.

36. The photovoltaic article of Claim 33, wherein the article is perforated so as to allow gaseous transport across the article.