Open Photovoltaics Analysis Platform Free for Research

+1 demo: Basic page


Three dimensional solar cell

full artical download

I.                    Optical Geometries used in Organic Solar Cells

A.            Why are Organics not high performance

Solar energy is primarily distributed from 300nm to 1700nm ( Fig. 1 ) [1] . A material with a band gap of 1.1 eV (1100nm) is capable of absorbing 77% of the solar irradiation power at the earth’s surface (for 100% absorptivity). [2] However, most of organic materials have larger band gaps (>1.9 eV), and only a part of the incident solar light is absorbed. Recently, researchers have reported stable, small band gap conjugated polymers [3-4] , such as poly (5,7-bis(4-decanyl-2-thienyl)-thieno(3,4-b)diathiazole-thiophene-2,5) (PDDTT) with a spectral response from 300 nm to 1450 nm ( Fig. 2 ).  These materials and others, [5-7]  are promising, but reported efficiencies are still under 10%.  Certainly it is reasonable that these performances will rise as processing and bulk heterojunction formation dynamics become better understood for the system.  However, there remain some issues in the use of such materials which will remain difficult to answer.  Specifically, by using the donor-acceptor structure to reduce band gap and broaden spectral overlap with solar, the density of states across this energy window is lowered.  This is essentially because the HOMO bands formed by the donor and the LUMO bands formed by the acceptor do not “add” electron states significantly over their wide band gap analogues.  Essentially, this is the same thing as evoking the “oscillator strength sum rule.”  [7-9] Thus, these materials may present a lower optical density and thicker films must be used to effectively absorb the light.  Generally speaking this will require that the overall mobility of the carriers be increased proportionally and in a balanced fashion, because thicker films will lead to space charge buildup and/or high radiative recombination losses. [10] At the same time, for most standard laboratory devices (planar spun caste devices), there are lots of other places where energy is lost: in the, ITO, PEDOT, and even the Al.

Fig. 1 Solar spectrum (AM 1.5). From NREL AM1.5g standard curve.


Fig. 2  Absorbance vs. wavelength of PDDTT compared with PDDTT PCBM blends as well as pure PCBM. From X. Gong, M. Tong, Y. Xia, W. Cai, J. S. Moon, Y. Cao, G. Yu, C. Shieh, B. Nilsson, A. J. Heeger, “High-Detectivity Polymer Photodetectors with Spectral Response from 300 nm to 1450 nm” Science, vol 325, no. 5948, pp. 1665-1666, Sep. 2009


The oscillator strength (optical density) problem leads to a second issue faced by most planar devices.  The mobility of the excitons, and the mobility of the carriers usually require the absorbing layer to be made very thin.  This reduces exciton recombination and allows removal of separated excitons such that the space charge near the contacts is minimized.  However, since the back place of the photoactive absorber is a metal contact (a mirror), the electric field in the absorber (near the contact) is going to zero.  Consequently much of the radiation is reflected away.  As we will discuss below, this is known as the “thin film” effect, and is well known and calculated by several groups. [11-13]

Finally, the low mobility of the charge carriers themselves presents problems in achieving performance.  Even when the charge mobility is well balanced, the relatively low mobility of the carriers leads inevitably to space charge near both electrodes.  This comes from the finite time taken to transfer charge from the organic conductor to the metal/ITO contacts as well as the relatively slow replenishment of charge as current is extracted.  This added “internal resistance” must always limit the total power generated by the device relative to its optimum power (also known as the filling factor).

So, for getting higher efficiencies in organic photovoltaics, it may not be enough to synthesize new materials.   Strategies may also involve ways of mitigating the effects of optical density as well as of space charge and low mobility.  In this work we examine the way in which these issues are being addressed through the use of device geometries.  Specifically, we review several approaches to engineering organic photovoltaics for better optical performance that have been recently explored.  We then focus on a particularly promising approach that has received much attention in the literature lately: the fiber-based solar cell. Finally we will show that such optimum optical engineering can provide organics with performance enhancements that makes them competitive for the larger solar market, not simply niche applications as some have suggested.  Indeed, we anticipate that fiber-based devices will top conventional OPV performance using the new small band gap polymers, providing long-term stable, inexpensive, building integrated PV at more than competitive pricing.

B.              What can be done to help using optical engineering?

1)    Optical spacer

Recently, an optical spacer [14] has been shown to raise efficiency in organic solar cells by allowing the layered structure to avoid destructive interference within the charge separating layer. [5] As mentioned above, the optical interference between the incident (from the ITO side) and back-reflected light, the intensity of the light in the charge separation layer is reduced near the metallic (Al) electrode. This “optical interference effect” is especially important for thin-film structures where layer thicknesses are comparable to the absorption depth and the wavelength of the incident light, as is the case for photovoltaic cells fabricated from semiconducting polymers. In addition, the use of TiOx as the spacer material can also act as a good hole blocking layer, [15] since the top of the valence band of TiOx is at nearly 8.1 eV. Fig. 3 compares the optical field strength distribution between devices with the TiOx optical spacer and a conventional cell.


Fig. 3 Schematic representation of the spatial distribution of the squared optical electric field strength |E|2 inside the devices with a structure of ITO/PEDOT/active layer/Al (left) and ITO/PEDOT/active layer/optical spacer/Al (right). Reprinted with permission from Jin Young Kim, Sun Hee Kim, Hyun-Ho Lee, Kwanghee Lee, Wanli Ma, Xiong Gong, Alan J. Heeger.. New architecture for high-efficiency polymer photovoltaic cells using solution-based titanium oxide as an optical spacer. Adv. Mater. 18, 572–576 (2006);. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.


Dr. Zhao [16] has used MoO3 as optical spacer to focus the optical field on the absorber layer P3HT:PCBM and improve the absorption. In Fig. 4 , it can be seen that the intensity maximum resides close to the center of the active layer, however, the relative optical intensity decreases with the increase of MoO3 thickness. This means that the thickness of MoO3 changes the distribution of optical field and corresponding energy. Thus, it also provides a way to reduce the interference and reflection effects associated with the thin film effect.


Fig. 4 The simulated optical field distribution (for 520 nm illumination) as a function of the distance from ITO/P3HT:PCBM interface in 19these inverted cells. These cells have the structures of ITO/Ca(1 nm)/P3HT:PCBM(85 nm)/MoO3(x nm) /Ag(100 nm) with x=0, 1, 3, 5, and 15. The ultrathin Ca is neglected in the simulation. Reprinted with permission from D. W. Zhao, P. Liu, X. W. Sun, S. T. Tan, L. Ke, and A. K. K. Kyaw. An inverted organic solar cell with an ultrathin Ca electron-transporting layer and MoO3 hole-transporting layer. Appl. Phys. Lett. 95, 153304 (2009). Copyright 2009, American Institute of Physics


2)    Proprietary cylindrical modules by Solyndra

Because the conventional planar PV panels must be fixed at some angle relative to a moving solar source, they produce optimum power for only a few hours during the day. Of course, a gimbaling mechanism can be added to continuously point the panel at the sun from sun rise to sun set, but this significantly raises install and maintenance costs.  Recently, Solyndra Inc. has introduced panels that employ cylindrical geometries so that the modules can capture sunlight across a 360-degree photovoltaic surface. This self-tracking design allows Solyndra's PV systems to capture more sunlight over the day than traditional planar-surfaced solar panels, which require costly tilted mounting devices to improve the capture of direct light, offer poor collection of diffuse light and fail to collect reflected light from rooftops or other installation surfaces. [17]  ( Fig. 5 )

Fig. 5  The tubular design of Solyndra’s modules optimizes the collection of direct, diffuse, and reflected sunlight [18]


3)    Fiber based solar cell

Another way of stating the above limitations on all photovoltaic devices, including organics, in reaching their theoretical maximum, is that devices are not capable of both; efficient charge transport and optimal optical absorption for a given absorption band. For example, the P3HT:PCBM devices, the ideal thickness for absorption is greater than 250 nm. However, transport of photogenerated excitons and charged polarons in the device is dominated by hopping mechanisms, requiring the active layer thickness to be less than 100 nm for balanced and efficient charge removal even when bulk heterojunctions are employed. Unfortunately, the device cannot be both. [19]

Fig. 6 Schematic of the fiber photovoltaic cell architecture. Reprinted with permission from Jiwen Liu, Manoj A. G. Namboothiry, and David L. Carroll, Optical geometries for fiber-based organic photovoltaics, Appl. Phys. Lett, 90, 133515 (2007) Copyright 2007


Ideally, the radiant energy could bewaveguided” into organic thin film devices such that reflective and transmissive losses were minimized while keeping the electronic properties the same, then the efficiencies could be raised significantly. We have recently reported on an approach to doing exactly this, using fiber-based solar cell geometries. We have examined the geometric optical considerations of coupling light into thin, organic active layers from a waveguiding fiber. ( Fig. 6 ). Utilizing standard multimode optical fibers, we have fabricated thin film devices which we refer to as “claddings” based on the bulk heterojunction blend P3HT: PCBM as an absorbing material and indium tin oxide ITO as the transparent conductor. Surprisingly, this architecture works with active film thicknesses significantly greater than those typical for thin film devices, suggesting that propagating modes within the layer are playing a role in the performance of the device. The experiment shows fiber-based devices are a reasonable approach to creating building blocks for higher performance organic platforms. [20] This has recently been repeated by several groups using dye sensitized cells as the organic absorbing system. [21]

In our work, [22] the front of the fiber was aperatured by fitting the fiber end through a hole in an optical baffle with a diameter equal to the fiber. Thus no light was incident on the fiber’s side and the overall angle of incidence was 90 degree to the face of the fiber. Shown in Fig. 7 is the IV curve from a typical 0.2 mm DIA., 1 mm long fiber device under AM1.5g illumination. The aspect ratio (L/d) in this case is 5. This device uses an ITO inner conductor and was not annealed after fabrication. The resistivity of the internal conductor was measured to be 1.4 /cm (along the fiber). This has clearly resulted in a rather large internal resistance as evidenced by the slope at the intercepts of the IV curve. Further, the thick absorbing layer has contributed to this internal resistance. Never the less, such a device easily achieves a FF of 0.42, and a Voc of 0..456V yielding an external power conversion efficiency of roughly 1.82%. In comparison, a planar device built with the same layer thicknesses and same overall processing parameters, results in devices of around 1.0% efficiency using our PEDOT and our P3HT (not shown). The low Voc is caused by the decreasing  light flux in fiber inner surface.


Fig. 7 the illuminated IV curve from a fiber device with an ITO inner electrode and 300 nm thick P3HT:PCBM absorbing layer.


C.             How do Optically Optimal designs Effects Angular Response?

If we use fibers as a way to confine light in the core, and wrap the PV thin film around the outer surface, the diffuse solar light can be absorbed through a very wide range of incident angles Fig. 8 A recently reported model [23] calculates the angular response of capturing light in an organic fiber based photovoltaic cell. The authors show how the fiber-based cell manages light absorption as a function of several parameters including the incident angle, meridional plane, cross sectional area, path length, and refractive index. They predict the optical angular input to achieve maximum absorption of resonant light and the complexity of how changing refractive indices in a multilayer device can alter the angular dependence when considering the incident input light ( Fig. 8 ).

Fig. 8 (a)Diagram depicting the cylindrical nature of the fiber-ITO interface and the relation between θ1 and θ2 (actual). (b) Contour plot of matrix M2, defined as the product of the overall transmittance (into the P3HT layer), the cross sectional area term, and path length, as a function of the incident angle at the fiber face θ0 and the distance from the meridional plane (x) in which the light rays enter the fiber. Reprinted with permission from Seamus Curran, Jamal Talla, Sampath Dias, and James Dewald, Microconcentrator photovoltaic cell, the m-C cell Modeling the optimum method of capturing light in an organic fiber based photovoltaic cell, JOURNAL OF APPLIED PHYSICS. 104, 064305 (2008). Copyright 2008, American Institute of Physics."


While this model captures the essential behavior of very short fibers, it is incomplete in that it ignores multiple internal reflections. In other words, it assumes only one reflection in fiber core, and does not take into account the fiber length, and its corresponding relationships.  Further, this model doesn't use an internal PEDOT layer which is typical for such structures, and it incorrectly calculates the full path length in the absorber for rays which are not along the optical axis (it underestimates the actual path length in the absorber which doesn't matter for short fibers but leads to erroneous results for long fibers). Thus many features are missed in this study, and other are simply predicted incorrectly.  For instance, we should expect that the optimum angle of light collection changes for different fiber aspect ratios, (lengths for a given diameter) since the balance of light reflection at the outer contact and absorption in the photoactive layer will clearly depend on the angle at which the multiple reflections within the device occur.  For a given incident angle of the light, the angle at which these reflections happen internally depends on diameter of the fiber. The number of reflections obviously depends on length.  Different ratios of diameter to length will reasonably produce different optimum angles of illumination. Recently, by using ray tracing and optical path iteration, we have presented a more detailed mathematical model and corresponding experiment for fiber-based organic photovoltaics. [24] Fig. 9 (a) shows a schematic of the device on which the calculations were performed. Besides giving an optimum incident angle, position, and an optimum aspect ratio in terms of other parameters of the fiber photocell, we also show that the predicted relationship between current generation and incident angle correlates well with experimental data for a given fiber length, and the experimental results on current generation versus fiber diameter are well reproduced by our simulation for fiber diameters of the order of the wavelength of incident light.

Fig. 9 Schematic of the fiber solar cell architecture and light illumination, ray diagram of light propagation. Reprinted with permission from Yuan Li , Wei Zhou, Dan Xue, Jiwen Liu, Eric D. Peterson, Wanyi Nie, David Carroll. Origins of Performance in Fiber-Based Organic Photovoltaics. Appl. Phys. Lett. 95, 203503 (2009). Copyright 2009.


Fig. 10  (a) Cross section of light path in fiber. Light enters the front face and is transmitted into each layer with transverse angle θi, tangential angle φi, and optical path i, which could be obtained by Snell’s Law and Geometrical optics (b) Longitudinal section of light path in fiber. The refractive indexes of P3HT:PCBM, PEDOT, ITO and fiber are 1.6, 1.6, 1.9, 1.45 respectively. En is remaining energy on certain refracted point, which comes from the previous refracted point En-1 (previous remaining energy). Reprinted with permission from Yuan Li , Wei Zhou, Dan Xue, Jiwen Liu, Eric D. Peterson, Wanyi Nie, David Carroll. Origins of Performance in Fiber-Based Organic Photovoltaics. Appl. Phys. Lett. 95, 203503 (2009). Copyright 2009.

As Fig. 10 shown, when the incident angle θ0 approaches zero, the solar light will exit from the back end face without reflection on the inner fiber surface, namely no E-field in absorption layer. Another side, if the incident angle θ0 approaches 90 degree, there is no light could enter into fiber, as the Fig. 12 shown. Thus, there exists a optimum incident angle for fiber-based solar cell.

Fig. 11   (a) Illumination of the front of the fiber at  normal incidence.  (b) Illumination of the front of the fiber at incident angle ө, less power is incident on the fiber than with normal incident. Reprinted with permission from Yuan Li , Wei Zhou, Dan Xue, Jiwen Liu, Eric D. Peterson, Wanyi Nie, David Carroll. Origins of Performance in Fiber-Based Organic Photovoltaics. Appl. Phys. Lett. 95, 203503 (2009). Copyright 2009.


Fig. 12 (a) is the calculated normalized power absorbed by the P3HT:PCBM layer as a function of the fiber length. There exists a maximum absorption at incident angle for each length of fiber-cell device. Recall that the solar flux rolls off as the cosine of the angle of incidence from the sun and in these simulations this is accounted for.  However, changes in air mass are not. In the case of short fibers (small aspect ratio), the device presents a fundamentally shorter optical path length in the absorber and consequently, less light is collected overall. Fig. 12 (b) compares experiment of current generation in a 1.4 mm diameter fiber that is between 1 mm to 2 mm (aspect ratio of nearly 1), to the simulations of Fig. 12 (a). For the longer fiber, it can be seen that the added length is beneficial to the absorption of light in Fig. 12 (c), but this effect will approach a constant (maximum absorption), because the fiber has absorbed all light power beyond a certain length. Fig. 13 (b), shows the experimentally determined current densities as the fiber diameter is varied (for normally incident light on long fibers). The current generation increases significantly as the diameter of the fiber is decreased, which corresponds to the absorption predicted by the model, shown in Fig. 13 (a).

Fig. 12  (a) (Simulation) The light absorbed versus incident angle, by different fiber length (1mm~50mm), with a diameter of 1.5mm. (b) (Experiment) Short circuit current versus incident angle, with a length between 1 and 2 mm, and a diameter of 1.4mm. (c)  (Simulation) In terms of (a), the percentage of absorption in fiber with the optimum incident angles for different fiber lengths, with a diameter of 1.4mm, incident angle. θ0=π/4, φ1=π/2. Reprinted with permission from Yuan Li , Wei Zhou, Dan Xue, Jiwen Liu, Eric D. Peterson, Wanyi Nie, David Carroll. Origins of Performance in Fiber-Based Organic Photovoltaics. Appl. Phys. Lett. 95, 203503 (2009). Copyright 2009.


Fig. 13 : (a) (Simulation) Light absorbed versus fiber diameter, with a length of 14 mm (b) (Experiment) Current Density versus fiber diameter, with a length of 14 mm. Reprinted with permission from Yuan Li , Wei Zhou, Dan Xue, Jiwen Liu, Eric D. Peterson, Wanyi Nie, David Carroll. Origins of Performance in Fiber-Based Organic Photovoltaics. Appl. Phys. Lett. 95, 203503 (2009). Copyright 2009.


Fig. 14 is absorption at each point across the fiber face to understand how each ray is contributing to the overall absorption cross-section. For light incident at a specific angle of θ0 and φ1 in the middle of the front face of the fiber, the absorption probability is higher than that on the outer edges. High absorption is shown in green, and lower absorption is shown in other colors. The azimuthal angle φ1 is set to π/2  for the example shown, which causes the transmittance (T) to become small along the “y” edge.

Fig. 14 (Simulation) Absorption for the incident point (x0, y0) with incident angles  θ0=π/4, φ1=π/2  (refer to Fig. 10 (a)). The under surface x0-y0 plane is the end face of fiber. The green part is the area of best absorption, the red is the worst. Where, D is Diameter of end face. Reprinted with permission from Yuan Li , Wei Zhou, Dan Xue, Jiwen Liu, Eric D. Peterson, Wanyi Nie, David Carroll. Origins of Performance in Fiber-Based Organic Photovoltaics. Appl. Phys. Lett. 95, 203503 (2009). Copyright 2009.


II.                    Radiation Confinement in Fiber Devices

A.       There are two kinds of fiber devices, illumination from outside, illumination inside.

Generally, organic PV devices built concentrically on optical fibers, in which the organic active layers were sandwiched between conductive electrodes around the fibers, is a relatively new concept, but it can be implemented in different ways. Specifically, work on this can be divided into two groups: devices illuminated outside and the ones with inside illumination.

For the outside-illuminated fiber cells, devices were thermally evaporated in vacuum on the rotated fibers, with CuPc-C60-Alq3 as the active layers. The light entered into the device through the outer semi-transparent electrode on the fiber. [25]

Fig. 15 Illustration of fibercell device illuminated outside. Reprinted with permission from Brendan O’Connor, Kevin P. Pipe, and Max Shtein, Fiber based organic photovoltaic devices, APPLIED PHYSICS LETTERS 92, 193306 (2008). Copyright 2008


Devices produced relatively stable power for varying angle of incident light, giving the concept an important advantage over planar devices in the same way as the Solyndra devices above.  Similarly, Grätzle cells have been built by Konarka in which the inner conductor is a wire and a transparent outer conductor allows for side-illumination [26-27]

As discussed in detail above, OPV devices can also built as the inside-illuminated fibercells , which are processed by dip-coating [20] (PDC-01 Programmable Dip Coater, and Opticer H10 Dip Coater). Fundamentally, the appearance is the only similarity between the devices. Their function is very different.  The purpose of the fiber-cell geometry is to capture incident light or trap it so that it can be absorbed multiple times by the P3HT:PCBM bulk heterjunction layer coated around the fiber.

Fig. 16  Illustration of fibercell devices illuminated from one end of the fiber. Reprinted with permission from Jiwen Liu, Manoj A. G. Namboothiry, and David L. Carroll, Optical geometries for fiber-based organic photovoltaics, Appl. Phys. Lett, 90, 133515 (2007) Copyright 2007


In this sense the fiber-based cells are far more similar in concept to the very high efficiency solar cell (VHESC) concepts demonstrated by the University of Delaware group. [28] The fiber-based cells attempt to approach the thermodynamic limits allowed by the particular absorber being used.  To this end, the fiber-cavity should be seen more in the light of a thermodynamic black body and the illuminated fiber end is the hole in the traditional black body configuration.  Dependence of fiber-based cell performance on fibers’ diameters has been examined in the literature, showing a increasing efficiency in the smaller diameter fibers ( Fig. 17 (a)). Further, the fibercell’s efficiency dependence on incident angles has been examined between devices that are fully coated with the outer reflective conductor and ones coated only 1/3 by aluminum ( Fig. 17 (b)) thereby comparing what is essentially a fully confining black body to leaking black body.   The results seem clear that light capturing performance of the “optical can” formed by the fiber device is influenced by the ratio of “fiber black body volume” to the size of the exit aperture in the “can.”

Fig. 17  (a) Comparison of the Isc and power conversion efficiency as a function of the fiber diameter for light incident along the axis and perpendicular to the plane of the cleaved surface of the fiber (only 1/3 of the fiber circumference is coated with LiF/Al electrode).(b) Comparison of the Isc for full circumference coated with LiF/Al and 1/3 of circumference coated with LiF/Al on fiber and power conversion efficiency as a function of the angle of incidence with respect to the axis of the optical fiber of fiber diameter of 1.5 mm. Reprinted with permission from Jiwen Liu, Manoj A. G. Namboothiry, and David L. Carroll, Optical geometries for fiber-based organic photovoltaics, Appl. Phys. Lett, 90, 133515 (2007) Copyright 2007


B.       Predicted attributes of fiber-based confinement

Using the absorption spectrum of each material and an Iterative Approach, we can roughly estimate how much energy will be absorbed by continuous transmission and absorption in a fiber-based cell.  This will allow us to know the maximum efficiency for this type of PV.

First, for conventional planar cell with the architecture ITO/PEDOT/P3HT:PCBM(wt 1:1)/Al, when the solar light enters into the thin films, it will be absorbed in every layer as well as reflected by each interface, which will lead to energy loss as the Fig. 18 (a) shown.. As a result, in terms of calculation shown in Fig. 18 (b), 44% of the energy was not absorbed after only one transmission through the conventional planar cell. The loss includes energies reflected in every interface and energies absorbed by AL, PEDOT and ITO. If we were to capture this energy, the photovoltaic efficiency could be enhanced by 1.5X (again this is calculated for the P3HT:PCBM system with absorber thickness 70 nm). [29] In conventional planar cells, the thickness of P3HT:PCBM is around 100nm to 200nm, so thick absorbers will have a higher recombination rate that is harmful to efficiency, as shown in Fig. 19 . [11]  It is necessary to think about how to avoid this natural defect of planar cell. For fiber-based cell, since the incident angle on inner surface of fiber is usually large (near 80 degree), the reflection coefficient on the interface of P3HT:PCBM and Al goes up considerably and this part of reflected energy will be absorbed in continuous reflections in fiber. Using our reported efficiency of P3HT:PCBM, [30] the predicted efficiency would be near 8%. Further, if we adopted a polymer with a wider spectral response, the predicted efficiency will be much higher due to the maximized absorption. We point out above that, many groups are using methods to improve absorption and avoid exciton recombination; for example, tandem architectures, [31] and optical spacers. [32] However, they all also lead to extra absorption by the additional layers such as TiOx and MoO3. In the case of fiber devices, the use of the ITO layer is a limiting factor in the ultimate performance multiplier (1.5 X) because the thick layer supports an addition mode of the cavity that is very lossy.  In the case of small band gap polymers, although their devices with new polymers have high efficiency over 7%, their absorptions in longer wavelength are still not enough high, i.e., a large percentage of the available flux remains uncaptured. However, the fiber-base cell could trap this part of long wavelength light in its cavity, and thereby reach its theoretical maximum. Removing the ITO from the fiber-device architecture in favor of another internal conductor that does not support optical modes of the system, such as a thin layer of carbon nanotubes, can raise the performance to capture this remaining flux. Again, we emphasize this is for normally incident light. When the enhanced angular response is added in (because the illumination angle varies in one day.), such devices will producedouble the power of their planar analogues. Because for conventional planar cells, the reflected light energy rises considerably for large incident angle. However, our fiber-like cell (tube based solar cell [29] ) has no reflection at its end face and allows for black body behavior.  In our cost calculations below, we consider scenarios from the 1.5 X improvement to the 3 X improvement to account for most of the polymer systems under study today.

Fig. 18    (a) The process of solar spectrum being absorbed in solar cell by continuous transmission in the architecture ITO/PEDOT/P3HT:PCBM (70nm, wt 1:1) /Al. (simulated by Iterative Approach and absorption spectrum.  (b) Absorbed energy by P3HT:PCBM (blue line) and remain light energy (red line) in PV, along with continuous transmissions.


Fig. 19  Thickness dependence of the short-circuit current JSC with varying Langevin recombination efficiency reff. A purely optical simulation is also shown. Reprinted with permission from R. Häusermann, E. Knapp, M. Moos, N. A. Reinke, T. Flatz, and B. Ruhstaller. Coupled optoelectronic simulation of organic bulk-heterojunction solar cells: Parameter extraction and sensitivity analysis. J. Appl. Phys. 106, 104507 (2009). Copyright 2009, American Institute of Physics.


III.                  Extending this to real devices

Obviously the fabrication of large area devices in the fiber-device configuration is the greatest technical challenge to the use of such technologies.  One approach to large area devices has recently been demonstrated that utilizes aligned fiber arrays stamped into plastic substrates in a fashion similar to that of a hair brush.  The substrates can then be coated in a variety of ways: sputter deposition of inner contacts, dip coating or spray coating of active materials, and evaporation of outer contacts.  This results in the individual fiber-devices in parallel contact, producing an additive current across the surface as needed.  Shown in Fig. 20 is two examples of such fiber substrates, one using very short fibers (almost lenses with a l/d <1) and one with longer fibers (l/d >>1); where l = fiber length and d = fiber diameter.

Fig. 20 (a) Optical microscope image of plastic substrate with stamped fiber on the surface (b) Photo of aligned plastic fibers on fixed substrate (c) a schematic of the cross-section of the device after coating.  Light enters from the top.


Fabrication of functioning organic devices on such 3-D substrates can be challenging since typically tolerances for film thicknesses are rather tight. However drop casting and spray techniques do allow for simple devices to be made that exhibit reasonable performance.  Indeed the inner conductor places the strictest constraints on performance, not the photoactive absorbers (to which the device seems to be extremely fault tolerant).

A.              Fiber Bundles

Figure 21 Structures of Fibercell boudles in which each single fibercell was connected with each other in parallel by silver paste (ITO-ITO, Al-Al)


Of course, optical fiber processing is a very advanced science and a first thought was to simply draw-coat the fibers using existing technology, then post coating to cut and align fiber segments.  This presents very challenging problems of contacting the inner electrode as well as control over the fiber alignment.  Early devices made from 8-10 fiber-cells were arranged in parallel as a bundle and using silver paste contacts did work, but was so time intensive, that concept was dropped early.  Never the less, they did show that the fiber-device performance was additive, the angular response and high capture efficiency of the individual fibers were reflected in the bundles as well.  Finally, the effects of the packing fraction between fibers lowered performance by approximately ¼. 

B.              Aligned stamped fiber devices (short)

The first truly large area fiber-device cells were built on stamped polycarbonate substrates [33] using P3HT:PCBM as the absorbing layer. Since the plastic substrate could stand no more than 70 oC, highly conductive PEDOT:PSS was used for the inner transparent electrode instead of ITO. All coatings were added by simple drop-casting and finally the aluminum electrode was thermally evaporated on the top directly in a vaccum chamber.  We note that no annealing was used for these devices and the polycarbonate transmitted only 60% of the light in the absorption band of the P3HT. Yet a comparison of the devices to the equivalent planar device shows significant enhancements in performance, even for these relatively poor devices.  Shown in Fig. 22 is a typical IV from the devices described.  On average efficiencies of around 1% can be achieved from this fabrication.  In comparison, efficiencies of around 0.3% are more typical of the non-fiber, planar devices built under the same conditions.  It is important to note that for this illustration, the fiber-devices used were very short (50 microns in diameter and around 30 microns in height) so they were certainly NOT the optimum in optical performance.   Furthermore, the 3-D coatings of the fiber devices were poor compared to those of the planar devices.  However, in comparison, the fiber devices always yield around 1.5 times the performance of the equivalent planar device as we have predicted using the models above.  The active area of the devices tested and compared was of the order of 1 cm2 in x-y plane, which is the overlap part between ITO and aluminum illuminated by solar simulator, so this is clearly reflective of what can be gained from using fiber-substrates to enhance optics.  Clearly, internal resistance and optimum light confinement remain challenges for such substrates as used here. 

Fig. 22 IV curve of the stamped fiber device described in the text, illuminated by an AM1.5g standard. The inset is a photographic image of the device. 50 microns in diameter and around 30 microns in height


C.             Aligned stamped fiber devices (long)

To improve light confinement, longer fiber must be utilized.  The stamping process above, as well as numerous photolithographic processes can be used to give large area aligned arrays of fibers from 0.2 mm to 2 mm in diameter, with lengths of 0.25 to 5 mms and separations of no more than 100 microns.  We have investigated substrates like these fabricated from poly methyl methacrylate (PMMA), and a number of photolithographic resins, all with excellent transmission properties.  Because these devices are much longer, internal resistance is very high when ITO is not utilized.   Never the less, we have begun optimization of such substrates for applications that require both a flexible and rigid solar cell.  The upper support of the fiber array is made very thin to achieve flexible device configurations.  Thicker upper supports allows for ease in coatings, however makes the device more rigid.  Again, using spray coating of P3HT:PCBM, and dip coated PEDOT as the internal conductor, devices with longer fibers have been fabricated.  IV curves from flexible substrates are shown in Fig. 23 in which annealed and unannealed devices fabricated with chlorobenzene and toluene are compared. 

Fig. 23 Comparison of fibercell IV curve (a) in dark (b) under illumination for flexible substrates using different solvents


We note that for these device, the longer fibers do not allow for efficient use of the lower conductivity PEDOT inner electrode and consequently space charge begins to build as evident in Fig. 23 (b) as the “hump” in the IV near zero current.  Further, as can be seen, device performance is strongly dependent on processing of the absorber in this case.  This is not true in the shorter fiber case and probably derives because the longer fibers make the device particularly sensitive to growing internal resistance.


IV.                  Summary: cost effectiveness and the emergence of solar energy

Over the past 50 years, the PV community was growing so fast that the annual production capabilities were well beyond 1 GWp per year and the cumulative installed PV capacity will grow to a 10 GWp limit before 2010. Growth rates are still in the 30-40% regime. [34]

Fig. 24 Transforming the global energy mix: The exemplary path until 2050/2100. (Source:WBGU, World in Transition Towards Sustainable Energy Systems, at http://www.wbgu.de/)

In the next 50 years, it is anticipated by many that solar power will take a leadership role in global energy production, as the Fig. 24 shows. (taken from the scientific advisory committee for the German government for transforming of sustainable energy systems) Depending which study one reads, energy demand is expected to at least double by 2050, and quadruple by 2100. This means a need in an additional production capacity of approximately 500EJ till 2050. Error! Reference source not found.  In terms of Fig. 24 , we would estimate that solar energy will supply 140 TW by 2050.  This simply means that the market is very large for such technologies and is likely to embrace several different technology platforms.  Thus, while organics may not replace silicon anytime soon, we argue that given the size of the market, the coming need for energy production, and the likely bottlenecks in production of any given technology, organics will not need to replace any technology to be truly profitable.  However they will need to provide equivalent power and service life to be viable.

A.              Estimated Efficiency

Above we calculate performance enhancements for a general photoactive material used in the fiber-device configuration.  Typically we expect a x1.5 to x3 improvement in efficiency and a x3 to x6 overall power generation rate (kWhrs/day) over planar solar cells with a simple architecture.   However it is important to note that polymer synthesis is marching forward quickly, with planar efficiencies already reaching > 7%. These advances then translate into much greater performance for the fiber-devices.  This is because the optical architecture of fiber-devices as well as for some of the other devices discussed here, addresses fundamental problems of organics.  However, do advances in polymer synthesis translate into a larger or smaller $Wp (cost per peak watt of power generated)?  Given that the substrates are far more complex that planar, this is more difficult to answer.

B.              Cost model for Fiber Devices

There are two kinds of cost for photovoltaics business: costs for the module and the costs for the installation [35] . The second cost point is approximately the same for any of the organic technologies discussed here.  The first part includes materials, production, overhead, area related and energy related costs. [36] In terms of the estimation from Gilles Dennler and Christoph J. Brabec, [35] for conventional OPV, it has a direct cost potential between 25 and 100 m-2, whereas the 30-60 m-2 regime appears as a very reasonable, fast to realize, cost scenario at rather low volumes. Indirect costs of <10 m-2, but more likely even <5 m-2 are reasonable at large volumes. For fiber solar cell, the substrate (PET) would be replaced by fiber matrix substrate as Fig. 20 shown. Consequently, the cost for this part is increased as is the costs of coating due to the higher effective surface area.  It is most simple to estimate how the added “internal” area of the fiber-geometry effects the overall cost by calculating this cost in terms of the known costs associated with planar devices.


To calculate the cost of a fiber-cell we take the cost of a planar device and wrap it around a fiber.  For fibers of radius r, and length l;

Active area covered by one fiber = p r2.

Cost of one FiberCell = 2 p r l C

Where the cost of planar device = C in $/m2


1 m2 of FiberCell = (number of fibers) ¾ p r2 (where ¾ is the packing fraction)


The packing fraction of the fibers is the ratio of the area occupied by a fiber face surface to the overall area the fibers are packed into. 


(number of fibers) = 1 (m2)/(¾ p r2)

Cost of 1 m2 of Fiber Cells = 2prlC (number of fibers)


Cost of FiberCell = 8prlC/3pr2 ($/ m2)

= (Geometric) C l/r


So the cost advantage of the fiber-device using organic absorbers is cancelled out by the high materials use unless the fiber-device can produce significant power, or if the geometric factor (roughly equivalent to the thickness of the most expensive layers) can be made small.  As we discussed above the power advantage can be significant.  The fiber-geometry can produce a factor of 1.5 to 3 greater efficiency for a given absorber depending on the losses in the conductor.  For a given efficiency, the fiber-based device can also produce twice the overall power generation: kWhr/day.  To see if the power increase is enough to offset the added cost, Fig. 25 (a) plots the aspect ratio against that of cost ratio of fiber-cells to planar cells.  Notice that when the length of the device rises, more light is captured and the device approaches maximum performance for a given absorber.   This is plotted for a geometry factor of 1.  However, it is not necessary to use absorbers that are 100 nm thick for fiber devices, since multiple passes through the absorber will allow light to be absorbed for nearly any thickness of polymer.  Therefore we can consider the cases where the absorber is much thinner corresponding to a geometry factor G of < 1.  Plotted in Fig. 25 (b) is a comparison of the ratio of cost fiber device to planar device, for different values of G.  Notice that the price is dropped as G is reduced, making it possible to achieve a device with 3 times the efficiency and 6 times the power generation but at a similar cost to the planar device if one uses absorbing layers that are 1/10th the thickness of the planar device.  For the fiber devices, this is reasonable.

Fig. 25 (a) . The aspect ratio against that of cost ratio of fiber-cells to planar cells. (b) Comparison of the ratio of cost fiber device to planar device


So what this seems to suggest is that there are routes to the creation of much higher performance organic devices through the use of optical engineering.  These approaches can increase device function two ways, through light confinement allowing for better solar collection and absorption in general.  Secondly, they may modify angular response of the device allowing for better power generation daily.  More surprisingly, a reasonable argument can be put forward that such approaches can be made cost competitive with their planar counterparts.  This surprising statement comes simply from the fact that due to light confinement, materials requirements are changed.

A recent PIRA report showed thin film solar now acquiring 12% of total market share even though they do not provide the same performance as the competing technology: Si.  However, as these new polymers are paired with novel optical geometries that allow for easy building integrated power generation that is efficient and inexpensive, it is likely that the uptake of organic technologies will dramatically accelerate.  


The authors gratefully acknowledge funding from AFOSR grant number: FA9550-04-1-0161, and DOE grant number: DE-FG02-07ER46428. 



[1]           "ASTM Standard G173, 2003e1, " Standard Tables for Reference Solar Spectral Irradiances: Direct Normal and Hemispherical on 37° Tilted Surface" ASTM International, West Conshohocken, PA, 2003, DOI: 10.1520/G0173-03E01, www.astm.org.."

[2]           "Nunzi, J. M. C. R. Physique 2002, 3, 523.."

[3]           X. Gong, M. Tong, Y. Xia, W. Cai, J. S. Moon, Y. Cao, G. Yu, C.-L. Shieh, B. Nilsson, and A. J. Heeger, "High-Detectivity Polymer Photodetectors with Spectral Response from 300 nm to 1450 nm," Science, vol. 325, pp. 1665-1667, September 25, 2009 2009.

[4]           R. C. Coffin, J. Peet, J. Rogers, and G. C. Bazan, "Streamlined microwave-assisted preparation of narrow-bandgap conjugated polymers for high-performance bulk heterojunction solar cells," Nat Chem, vol. 1, pp. 657-661, 2009.

[5]           S. H. Park, A. Roy, S. Beaupre, S. Cho, N. Coates, J. S. Moon, D. Moses, M. Leclerc, K. Lee, and A. J. Heeger, "Bulk heterojunction solar cells with internal quantum efficiency approaching 100%," Nature Photonics, vol. 3, pp. 297-U5, May 2009.

[6]           H. Y. Chen, J. H. Hou, S. Q. Zhang, Y. Y. Liang, G. W. Yang, Y. Yang, L. P. Yu, Y. Wu, and G. Li, "Polymer solar cells with enhanced open-circuit voltage and efficiency," Nature Photonics, vol. 3, pp. 649-653, Nov 2009.

[7]           L. J. Huo, J. H. Hou, H. Y. Chen, S. Q. Zhang, Y. Jiang, T. L. Chen, and Y. Yang, "Bandgap and Molecular Level Control of the Low-Bandgap Polymers Based on 3,6-Dithiophen-2-yl-2,5-dihydropyrrolo[3,4-c]pyrrole-1,4-dione toward Highly Efficient Polymer Solar Cells," Macromolecules, vol. 42, pp. 6564-6571, Sep 8 2009.

[8]           J. Berkowitz, "Sum rules and the oscillator strength distribution in helium," Journal of Physics B-Atomic Molecular and Optical Physics, vol. 30, pp. 881-892, Feb 28 1997.

[9]           M. Kamohara, Y. Izumi, M. Tanaka, K. Okamoto, M. Tanaka, F. Kaneko, Y. Kodama, T. Koketsu, and K. Nakagawa, "Optical oscillator strength distribution of amino acids from 3 to 250 eV and examination of the Thomas-Reiche-Kuhn sum rule," Radiation Physics and Chemistry, vol. 77, pp. 1153-1155, Oct-Dec 2008.

[10]         F. Zhang, A. Vollmer, J. Zhang, Z. Xu, J. P. Rabe, and N. Koch, "Energy level alignment and morphology of interfaces between molecular and polymeric organic semiconductors," Organic Electronics, vol. 8, pp. 606-614, Oct 2007.

[11]         R. Hausermann, E. Knapp, M. Moos, N. A. Reinke, T. Flatz, and B. Ruhstaller, "Coupled optoelectronic simulation of organic bulk-heterojunction solar cells: Parameter extraction and sensitivity analysis," Journal of Applied Physics, vol. 106, p. 104507, 2009.

[12]         L. A. A. Pettersson, L. S. Roman, and O. Inganas, "Modeling photocurrent action spectra of photovoltaic devices based on organic thin films," Journal of Applied Physics, vol. 86, pp. 487-496, 1999.

[13]         M. G. Harrison, J. Gruner, and G. C. W. Spencer, "Analysis of the photocurrent action spectra of MEH-PPV polymer photodiodes," Physical Review B, vol. 55, pp. 7831-7849, Mar 15 1997.

[14]         J. Y. Kim, S. H. Kim, H. H. Lee, K. Lee, W. L. Ma, X. Gong, and A. J. Heeger, "New architecture for high-efficiency polymer photovoltaic cells using solution-based titanium oxide as an optical spacer," Advanced Materials, vol. 18, pp. 572-+, Mar 3 2006.

[15]         S. Cho, J. Yuen, J. Y. Kim, K. Lee, A. J. Heeger, and S. Lee, "Multilayer bipolar field-effect transistors," Applied Physics Letters, vol. 92, pp. -, Feb 11 2008.

[16]         D. W. Zhao, P. Liu, X. W. Sun, S. T. Tan, L. Ke, and A. K. K. Kyaw, "An inverted organic solar cell with an ultrathin Ca electron-transporting layer and MoO3 hole-transporting layer," Applied Physics Letters, vol. 95, pp. -, Oct 12 2009.

[17]         "Solyndra emerges from stealth mode with cylindrical approach to CIGS thin-film photovoltaics, http://www.solarserver.de/solarmagazin/anlage_0409_e.html."

[18]         "Proprietary cylindrical modules by Solyndra, http://www.solyndra.com/Products/More-Electricity."

[19]         J. W. Liu, M. A. G. Namboothiry, and D. L. Carroll, "Fiber-based architectures for organic photovoltaics," Applied Physics Letters, vol. 90, pp. -, Feb 5 2007.

[20]         J. W. Liu, M. A. G. Namboothiry, and D. L. Carroll, "Optical geometries for fiber-based organic photovoltaics," Applied Physics Letters, vol. 90, pp. -, Mar 26 2007.

[21]         B. Weintraub, Y. G. Wei, and Z. L. Wang, "Optical Fiber/Nanowire Hybrid Structures for Efficient Three-Dimensional Dye-Sensitized Solar Cells," Angewandte Chemie-International Edition, vol. 48, pp. 8981-8985, 2009.

[22]         W. Zhou, W. Nie, Y. Li, J. Liu, and D. L. Carroll, "Fabrication considerations in fiber-based organic photovoltaics," 2009, p. 741613.

[23]         S. Curran, J. Talla, S. Dias, and J. Dewald, "Microconcentrator photovoltaic cell (the m-C cell): Modeling the optimum method of capturing light in an organic fiber based photovoltaic cell," Journal of Applied Physics, vol. 104, pp. -, Sep 15 2008.

[24]         Y. Li, W. Zhou, D. Xue, J. W. Liu, E. D. Peterson, W. Y. Nie, and D. L. Carroll, "Origins of performance in fiber-based organic photovoltaics," Applied Physics Letters, vol. 95, pp. -, Nov 16 2009.

[25]         B. O'Connor, K. P. Pipe, and M. Shtein, "Fiber based organic photovoltaic devices," Applied Physics Letters, vol. 92, pp. -, May 12 2008.

[26]         M. R. Lee, R. D. Eckert, K. Forberich, G. Dennler, C. J. Brabec, and R. A. Gaudiana, "Solar Power Wires Based on Organic Photovoltaic Materials," Science, vol. 324, pp. 232-235, Apr 10 2009.

[27]         "Solar fiber – another step towards photovoltaic textiles. http://www.talk2myshirt.com/blog/archives/2289."

[28]         A. Barnett, D. Kirkpatrick, C. Honsberg, D. Moore, M. Wanlass, K. Emery, R. Schwartz, D. Carlson, S. Bowden, D. Aiken, A. Gray, S. Kurtz, L. Kazmerski, M. Steiner, J. Gray, T. Davenport, R. Buelow, L. Takacs, N. Shatz, J. Bortz, O. Jani, K. Goossen, F. Kiamilev, A. Doolittle, I. Ferguson, B. Unger, G. Schmidt, E. Christensen, and D. Salzman, "SHORT COMMUNICATION: ACCELERATED PUBLICATION Very high efficiency solar cell modules," Progress in Photovoltaics: Research and Applications, vol. 17, pp. 75-83, 2009.

[29]         Y. Li, E. D. Peterson, H. Huang, M. Wang, D. Xue, W. Y. Nie, W. Zhou, and D. L. Carroll, "Tube-Based Geometries for Organic Photovoltaics," Applied Physics Letters, accepted for publication., 2010.

[30]         M. Reyes-Reyes, K. Kim, and D. L. Carroll, "High-efficiency photovoltaic devices based on annealed poly(3-hexylthiophene) and 1-(3-methoxycarbonyl)-propyl-1-phenyl-(6,6)C-61 blends," Applied Physics Letters, vol. 87, pp. -, Aug 22 2005.

[31]         J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T. Q. Nguyen, M. Dante, and A. J. Heeger, "Efficient tandem polymer solar cells fabricated by all-solution processing," Science, vol. 317, pp. 222-225, Jul 13 2007.

[32]         J. K. Lee, N. E. Coates, S. Cho, N. S. Cho, D. Moses, G. C. Bazan, K. Lee, and A. J. Heeger, "Efficacy of TiO[sub x] optical spacer in bulk-heterojunction solar cells processed with 1,8-octanedithiol," Applied Physics Letters, vol. 92, p. 243308, 2008.

[33]         "Substrates were prepared by Diamond Films in Auckland NZ using a proprietary stamping process."

[34]         C. Brabec, V. Dyakonov, and U. Scherf, "Organic Photovoltaics: Materials, Device Physics, and Manufacturing Technologies," WILEY-VEH, vol. ISBN 978-3-527-31675-5, p. 554, 2008.

[35]         C. Brabec, V. Dyakonov, and U. Scherf, "Organic Photovoltaics: Materials, Device Physics, and Manufacturing Technologies," WILEY-VEH, vol. ISBN 978-3-527-31675-5, p. 557, 2008.

[36]         K. Zweibel, "Thin Film Solar Cells (eds J. Poortmans and V. Arkhipov)," Wiley-VCH Verlag GmbH, Weinheim, Germany, 2006.


Supported by


opvap Matlab SimElectronics Wolfram Demonstrations Project  
OPVAP Gallery Matlab SimElectronics Wolfram Demonstrations Project Chinese Bulletin Board System