Transparentgroup

Figure 1. Cross-section of thin film polycrystalline solar cell. As seen the transparent conducting coating contacts the n-type semiconductor to draw current.

Transparent conducting films (TCFs) for photovoltaic applications have been fabricated from both inorganic and organic materials. Inorganic films typically are comprised of a layer of TCO (transparent conducting oxide), generally in the form of indium tin oxide (ITO), fluorine doped tin oxide (FTO), and doped zinc oxide. Organic films are being developed using carbon nanotube networks and graphene, which can be fabricated to be highly transparent in the infrared, along with polymer networks such as poly(3,4 ethlyenedioxythiophene) and its derivatives. Transparent conducting films act both as a window for light to pass through to the active material beneath (where carrier generation occurs) and as an ohmic contact for carrier transport out of the photovoltaic. Transparent materials possess bandgaps with energies corresponding to wavelengths which are shorter than the visible range (380nm to 750nm). As such, photons with energies below the bandgap are not collected by these materials and thus visible light passes through. However, applications such as photovoltaics may require an even broader bandgap to avoid unwanted absorption of the solar spectra.

Transparent Conducting Oxides

Overview

Transparent conductive oxides (TCO) are doped metal oxides used in optoelectronic devices such as flat panel displays and photovoltaics (including inorganic devices, organic devices, and Graetzel cells). Most of these films are fabricated with polycrystalline or amorphous microstructures. On average, these applications use electrode materials that have greater than 80% transmittance of incident light as well as conductivities higher than 10³ (Ω-cm)^-1 for efficient carrier transport. The transmittance of these films, just as in any transparent material, is limited by light scattering at defects and grain boundaries. In general, TCOs for use as thin-film electrodes in solar cells should have a minimum carrier concentration on the order of 10²⁰ cm^-3 for low resistivity and a bandgap less than 380 nm to avoid absorption of light over most of the solar spectra. Mobility in these films is limited by ionized impurity scattering and is on the order of 40 cm²/V-s. Current transparent conducting oxides used in industry are primarily n-type conductors, meaning their primary conduction is from the flow of electrons. Suitable p-type transparent conducting oxides are still being researched.

To date, the industry standard in TCO is ITO, or tin-doped indium-oxide. This material boasts a low resistivity of ~10^-4 Ω-cm and a transmittance of greater than 80%. However ITO has the drawback of being expensive. Indium, the film’s primary metal, is rare (6000 metric tons worldwide in 2006), and its price fluctuates due to market demand (over $800 US per ingot in 2006). For this reason, doped binary compounds such as aluminum-doped zinc-oxide (AZO) and indium-doped cadmium-oxide have been proposed as alternative materials.1 AZO is comprised of aluminum and zinc, two common and inexpensive materials, while indium-doped cadmium oxide only uses indium in low concentrations.

Binary compounds of metal oxides without any intentional impurity doping have also been developed for use as TCOs. These systems are typically n-type with a carrier concentration on the order of 10²⁰ cm^-3, provided by interstitial metal ions and oxygen vacancies which both act as donors. However, these simple TCOs have not found practical use due to their electrical properties’ high temperature and oxygen partial pressure dependence.

Fabrication

Doped metal oxides for use as transparent conducting layers in photovoltaic devices are typically grown on a glass substrate. This glass substrate, apart from providing a support that the oxide can grow on, has the additional benefit of blocking most infrared wavelengths greater than 2 µm for most silicates, and converting it to heat in the glass layer. This in turn helps maintain a low temperature of the active region of the solar cell, which degrades in performance as it heats up. TCO films can be deposited on a substrate through various deposition methods, including metal organic chemical vapor deposition (MOCVD), metal organic molecular beam deposition (MOMBD), spray pyrolysis, and pulsed laser deposition (PLD), however conventional fabrication techniques typically involve magnetron sputtering (MSP) of the film. The sputtering process is very inefficient, with only 30% of the material actually being deposited on the substrate. In the case of ITO this inefficiency is a significant drawback, however work is still underway to increase the recoverable amount of the unused material. Growth typically is performed in a reducing environment to encourage oxygen vacancy formation within the film, which contribute to the carrier concentration (if n-type).

Theory

Charge carriers in these oxides arise from three fundamental sources: interstitial metal ion impurities, oxygen vacancies, and doping ions. The first two sources always act as electron donors. Indeed some TCOs are fabricated solely using these two intrinsic sources as carrier generators. When an oxygen vacancy is present in the lattice it acts as a doubly-charged electron donor. In ITO for example, each oxygen vacancy causes the neighboring In³⁺ ion 5s orbitals to be stabilized from the 5s conduction band by the missing bonds to the oxygen ion, while two electrons are trapped at the site due to charge neutrality effects. This stabilization of the 5s orbitals causes a formation of a donor level for the oxygen ion, determined to be 0.03 eV below the conduction band. Thus these defects act as shallow donors to the bulk crystal. Common notation for this doping is Kröger-Vink Notation and is written as such:

$O_{O}^{x}\leftrightarrow V_{O}^{\bullet \bullet }+0.5O_{2}(g)+2e'$

Here the “O” in the subscripts indicates that both the initially bonded oxygen and the vacancy that is produced lie on an oxygen lattice site, while the superscripts on the oxygen and vacancy indicate charge. Thus to enhance their electrical properties, ITO films and other transparent conducting oxides are grown in reducing environments, which encourage oxygen vacancy formation.

Dopant ionization within the oxide occurs in the same way as in other semiconductor crystals. Shallow donors near the conduction band (n-type) allow electrons to be thermally excited into the conduction band, while acceptors near the valence band (p-type) allow electrons to jump from the valence band to the acceptor level, populating the valence band with holes. It is important to note that carrier scattering in these oxides arises primarily from ionized impurity scattering. Charged impurity ions and point defects have scattering cross-sections that are much greater than their neutral counterparts. Increasing the scattering decreases the mean-free path of the carriers in the oxide, which leads to poor device performance and a high resistivity. These materials can be modeled reasonably well by the free electron gas theory assuming a parabolic conduction band and doping levels above the Mott Criterion. This criterion states that an insulator such as an oxide can experience a composition-induced transition to a metallic state given a minimum doping concentration n_c, determined by:

$n_{c}^{1/3}a_{H}^{*}=0.26\pm 0.05$

where a_H* is the mean ground state Bohr radius. For ITO this value requires a minimum doping concentration of roughly 10¹⁹ cm^-3. Above this level, the typically-electrically insulating material becomes metallic and is capable of allowing carrier flow.

Transparent Conductive Polymers

Overview

Figure 2. Polymer photovoltaic cell using transparent conducting polymers.

Conductive polymers were reported as early as the 1900s as derivatives of polyaniline. Research continued on such polymers in the 1960s and 70s and continued into the turn of the century. Most conductive polymers are derivatives of polyacetylene, polyaniline, polypyrrole or polythiophenes. These polymers have conjugated double bonds which allow for conduction. By manipulating the band structure, polythiophenes have been modified to achieve a HUMO-LUMO separation that is transparent to visible light. In a polymer, the band structure is determined by the molecular orbitals. The effective bandgap is the separation between the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO). This gap can be approximated with understanding of the Huckel method.

Applications

Transparent conductive polymers can be used as electrodes on light emitting diodes and photovoltaic devices. They have conductivity below that of transparent conducting oxides and have low absorption of the visible spectrum allowing them to act as a transparent conductor on these devices. The transparent conductive polymers can be made into flexible films making them desirable despite their lower conductivity. This makes them useful in the development of flexible electronics where traditional transparent conductors will fail. However, because transparent conductive polymers do absorb some of the visible spectrum and significant amounts of the mid to near IR, they lower the efficiency of photovoltaic devices.

Poly(3,4-ethylenedioxythiophene) (PEDOT)

Poly(3,4-ethylenedioxythiophene) (PEDOT) has conductivity of up to ~ 1000 S/cm. Thin oxidized PEDOT films have ~10% or less absorption in the visible spectrum and excellent stability. However, PEDOT is insoluble in water making processing more difficult and costly.

The bandgap of PEDOT can be varied from 1.4 - 2.5 eV by varying the degree of π-overlap along the backbone. This can be done by adding substituents along the chain, which result in steric interactions preventing π-overlap. Substituents can also be electron-accepting or donating which will modify the electronic character and thus modify the bandgap. This allows for the formation of a wide bandgap conductor which is transparent to the visible spectrum.

PEDOT is prepared by mixing EDT monomer with an oxidizing agent such as FeCl3. The oxidizing agent acts as an initiator for polymerization. Research has shown that increasing the ratio of [FeCl3]/[monomer] decreases the solubility of the PEDOT. This is thought to be a result of increased crosslinking in the polymer making it more difficult to dissolve in a solvent.

Poly(3,4-ethylenedioxythiophene) PEDOT: Poly(styrene sulfonate) PSS

Doping PEDOT with poly(styrene sulfonate) can improve the properties over the unmodified PEDOT. This PEDOT:PSS compound has become the industry leader in transparent conductive polymers. PEDOT:PSS is water soluble making processing easier. PEDOT:PSS has a conductivity ranging from 1 to 10 S/cm while still transmitting visible light well. Treatment in air at 100oC for over 1000 hours will result in a minimal change in conductivity.

PEDOT:PSS is prepared by polymerizing EDT monomer in an aqueous solution of PSS using Na₂S₂O₈ as the oxidizing agent. This aqueous solution is then spin coated and dried to make a film.

Poly(4,4-dioctylcyclopentadithiophene)

Poly(4,4-dioctylcyclopentadithiophene) can be doped with both iodine and 2,3-dichloro-5,6-dicyano-1,4-benzoquinone(DDQ) to form a transparent conductor. The doped polymer has low absorption of the visible spectrum with an absorption band centered around 1050 nm. When doped with iodine a conductivity of 0.35 S/cm can be achieved. However, the iodine has a tendency to diffuse out in air, making the iodine-doped poly(4,4-dioctylcyclopentadithiophene) unstable.

A second doping is done with and 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ). This polymer exhibits a conductivity of 1.1 S/cm, however; the DDQ doped compound does have a tendency to lose its doping in air and thus have slight decreases in conductivity. DDQ doped polymer has better stability than the iodine doped polymer but is still below that of PEDOT. Due to its poor properties relative to PEDOT and PEDOT:PSS, Poly(4,4-dioctylcyclopentadithiophene) must improve before it is feasible for solar cell applications.

Poly(4,4-dioctylcyclopentadithiophene) is solution polymerized by combining monomer with iron (III) chloride. Once the polymerization is complete the doping is done by exposing the polymer to iodine vapor or DDQ solution.

Carbon Nanotubes

Advantages

Transparent conductors are fragile and tend to break down due to fatigue. The most commonly used TCF is Indium-Tin-Oxide (ITO) because of its good electrical properties and ease of fabrication. However, these thin films are usually fragile and such problems as lattice mismatch and stress-strain constraints lead to restrictions in possible uses for TCFs. ITO has been shown to degrade with time when subject to mechanical stresses. Recent increases in cost are also forcing many to look to carbon nanotube films as a potential alternative.

Carbon nanotubes (CNTs) were discovered in 1991 by Ijima of NEC and since then have attracted attention because of their materials properties, including a high elastic modulus (~1 - 2 TPa), a high tensile strength (~13 - 53 GPa), and a high conductivity (metallic tubes can theoretically carry an electrical current density of 4x10⁹ Å/cm², which is ~1000 times higher than for other metals such as Cu). CNT thin films have been used as transparent electrodes in TCFs because of these good electronic properties. However, it is difficult to obtain uniform dispersions of the CNT thin films, and the films are hard to characterize. AFM measurements are not representative of the sample as a whole. TEM and SEM measurements, on the other hand, require pre-treatment with gold or carbon sputtering, which may introduce impurities or cause defects in the original pattern of the composite.

Preparation of CNT thin films

The preparation of CNT thin films for TCFs is composed of three steps: the CNT growth process, putting the CNTs in solution, and, finally, creation of the CNT thin film. Nanotubes can be grown using Laser ablation, electric-arc discharge, or different forms of chemical vapor deposition (such as PeCVD). However, nanotubes are grown en-masse, with nanotubes of different chiralities stuck together due to van der Waal’s attraction. Density gradient ultracentrifugation (DGU) has recently been used to get rid of this problem by Hersham et. al of Northwestern University. Using DGU, transparent conductors were constructed using only metallic tubes. Because DGU allows for separation by density, tubes with similar optical properties (due to similar diameters) were selected and used to make CNT conductive films of different colors.

Figure 3. CNTs of various diameters seperated within a centrifuge tube. Each distinct diameter results in a different color.

In order to separate the grown tubes, the CNTs are mixed with surfactant and water and sonicated until satisfactory separation occurs. This solution is then sprayed onto the desired substrate in order to create a CNT thin film. The film is then rinsed in water in order to get rid of excess surfactant.

Comparing CNTs to TCOs

CNTs can also be used in addition to transparent conducting oxides (TCOs) in thin-film photovoltaic devices. Two TCOs which are often used are ZnO/Al and In2O3/Sn (ITO). Devices made with these TCOs attained energy-conversion efficiencies of 19.5% in CuIn_1-xGa_xSe₂-based (CIGS) solar cells and 16.5% in CdTe-based solar cells. These photovoltaic devices had much higher efficiencies compared to the devices made with CNT thin films: Britz et. al. report an efficiency of 8%, with an open circuit voltage (V_oc) of 0.676 V, a short circuit flux (J_sc) of 23.9 mA/cm², and a fill factor of 45.48%. However, CNT thin films show many advantages over other transparent electrodes in the IR range9. CNT thin films were reported to have a transmittance of over 90% in this range (400 nm - 22 µm). This paves the way for new applications, indicating that CNT thin films can be used as heat dissipaters in solar cells because of this high transmittance.

As stated previously, nanotube chirality is important in helping determine its potential aid to these devices. Before mass production can occur, more research is needed in exploring the significance of tube diameter and chirality for transparent conducting films in photovoltaic applications. It is expected that the conductivity of the SWNT thin films will increase with an increase in CNT length and purity. As stated previously, the CNT films are made using randomly-oriented bundles of CNTs. Ordering these tubes should also increase conductivity, as it will minimize scattering losses and improve contact between the nanotubes.