Nanostructures for photovoltaics

https://doi.org/10.1016/j.mser.2008.06.002Get rights and content

Abstract

The use of various nanostructures in new solar cell designs and modes of enhancing conventional solar cells are described. The cell designs and enhancements are categorized by the type of nanostructure utilized. These include: (a) bulk nanostructured materials [3D]; (b) quantum wells [2D]; (c) nanowires [1D]; and (d) quantum dots/nanoparticles [0D]. The methods of fabricating such structures are first described, followed by examples from the literature of how they have been utilized in a photovoltaic application. Scientific challenges associated with nanostructured photovoltaic devices are also discussed, followed by the prospects for use in real applications.

Introduction

In recent years there has been a significant, resurgent interest in renewable energy sources. This has been partially motivated by the increase in oil prices worldwide as a result of geopolitical and economic factors, and the general concern associated with global warming that is exacerbated by the emission of greenhouse gases during the production of primary power by conventional means [1], [2]. While many technologies are being considered to supplement oil as a primary energy source, renewable energy sources are seen as the key to long-term weaning of industrialized economies from strict reliance on oil, coal, and natural gas. These include wind, fuel cells, solar cells, geothermal, biofuels, etc. Solar energy conversion is perhaps the most appealing of all these solutions, since the energy source is readily available. Indeed, in 1 h the sun radiates upon the earth as much energy as is used in 1 year by humanity [1]. Solar technologies also easily capture the understanding and imagination of the general public, as they are intuitively attractive owing to the abundant nature of the source.

The concept of converting light to electricity was first introduced with the discovery of the photovoltaic (PV) effect by Edmond Becquerel in 1839 [3]. However, the first commercially viable demonstration of a solar cell did not occur until over 100 years later, with the invention of the crystalline Si-based cell first revealed to the world by researchers at Bell Labs in 1954 [4]. The first Si cell had an efficiency of ∼4% and since then researchers have been attempting to demonstrate, and companies to commercialize, solar cells with high conversion efficiency and low manufacturing cost. In the ensuing 50 years, Si cells have been demonstrated with an efficiency of nearly 25% [5], very close to the theoretical limit for a single junction under one sun illumination of ∼31% [6], [7]. Such high efficiency bulk Si solar cells are, unfortunately, prohibitively expensive for mass production. Indeed, most of today's commercial Si solar cells packaged in modules are ∼14–17% efficient. There is a realization in the PV technical and customer/user communities that increasing cell efficiency while decreasing cost will be critical if PV technology is to be widely utilized for primary or secondary energy needs.

Silicon is not the ideal semiconducting material for solar energy conversion. The indirect bandgap of Si makes optical absorption inefficient due to the requirement of phonon emission/scattering with photons in order to conserve crystal momentum. As a result, the thickness of Si required to absorb 90% of sunlight (of all photons of energy above the bandgap) is ∼125 μm, whereas the thickness requirement for a direct bandgap material such as GaAs is ∼0.9 μm [8]. The key reason why Si is the leading material used in the PV industry today is the fact that it is the second most abundant element in the earth's crust, making it a relatively inexpensive semiconductor (incidentally, it is most likely the most scientifically and technologically studied material in the history of science). Furthermore, while silicon was still an emerging semiconductor material when the first silicon solar cell was demonstrated, today there is a strong technological base for silicon due to the success of the electronics industry that can be directly applied to the mass manufacturing of Si-based solar cell technologies. Even though it is so abundant, it does require a minimum purity level in order to be useful for solar applications [9], though not to the level of purity required for the electronics industry. Until recently, silicon used in the PV industry today was recycled from the high performance, computing electronics industry, which lead to a shortage of suitable Si PV material. This trend continues, though there are now significant investments occurring in Si wafer production factories for the solar industry [10], [11].

The above discussion highlights three key questions facing PV: (1) How can the efficiency of solar cells be increased to competitive levels with other energy sources? (2) How can the cost of solar cells be decreased to a level suitable first for both secondary and primary power generation? (3) How can both of these goals be achieved in a single solar cell device/module technology and related manufacturing process? Fig. 1 shows the well-known plot, as defined by Green [12], of cell efficiency vs. cost per unit area for the key general PV technologies, mainly bulk Si, thin films, and the so-called Generation III concepts, the technologies of which are still being defined. Reference to this figure will be made in the subsequent discussion. The above questions lead to yet another question that is the central theme of this paper: can nanotechnology be used to address the above three questions, and if so, how?

Owing to the fact that the ∼$11 billion PV industry (in 2006, and estimated at ∼$17B in 2008) has been growing at a rate of 30–40% in the last few years and is expected to continue to do so in the next decade [13], it is of great interest for the PV research and industrial communities to address these questions. A brief summary of leading approaches to the first two questions will be described, followed by a general description of how proposed approaches employing nanoscale structures are capable of answering the third question.

As noted above, the limiting efficiency of a single junction is ∼31% [6]. Interestingly, early papers published shortly after Bell Lab's demonstration of the first Si cell estimated the efficiency to be 21.6% [14] until Shockley and Queisser established the first detailed balance theory showing ∼31% [6], which has since been modified by Kerr et al. [7] using additional calculations of Si solar cells incorporating Auger recombination to be ∼29%.

If one performs a more generalized calculation of solar energy conversion efficiency, it is possible to show that the thermodynamic limit is ∼85%, though as will be shown below we have not reached that level and will almost certainly never do so. The single bandgap thermodynamic efficiency assuming blackbody radiation is ∼44%, the so-called ultimate efficiency. The detailed balance limit efficiency of 31% derived by Shockley and Queisser further takes into account the shape and refractive index of the solar cell, a realistic solar spectrum, concentration, and radiative recombination.

There are a number of well-described concepts for increasing the efficiency of solar cells above the single junction limit (∼31%) so as to begin to approach the thermodynamic limit. These fall into the general category of Generation III solar cell concepts as defined by Green (see Fig. 1). Bulk Si solar cells are considered to be Generation I technologies, whereas Generation II cells are based on thin film technologies that allow for the use of thinner absorber materials deposited on lower cost substrates and hence reduced cell cost, often at the expense of efficiency (see below). Generation III cells are based on several new band structure and energy conversion concepts (to be described below) that have the potential to achieve limiting efficiencies greater than the single junction limit. It is expected that such technologies, which are still emerging, will also be able to achieve cost levels similar to or better than Generation II cell technologies.

The main reason why single junction solar cells are limited to ∼31% is that they do not absorb the significant fraction (∼20%) of the photons in the solar spectrum that are below the bandgap in energy. These photons are simply lost. On the other hand, high-energy photons are lost due to thermalization of high-energy charge carriers in the conduction band duo to phonon scattering. This can be shown graphically by plotting the solar flux absorbed by a semiconductor as a function of bandgap energy and overlaying a plot of the work performed by a device that includes radiative recombination, as shown by Henry [15]. Such an analysis also yields the ∼31% efficiency expected for the optimum bandgap of ∼1.35–1.45 eV at one sun concentration. If we add to this other device-level loss mechanisms such as reflection, non-radiative recombination, it is apparent that the efficiency quickly deteriorates from the ∼31% entitlement value.

The multi-junction cell (MJC) allows absorption of a wider range of wavelengths in the solar spectrum by combining solar cells of varying bandgap in a series (tandem) stack. A generalized theory based on 2–4 and beyond (infinite) number of bandgaps shows that the theoretical efficiency for four junctions is ∼71% under maximum concentration [16]. This is the only high efficiency solar cell concept that has been shown experimentally to work, with a 3-junction device recently performing at ∼40.7% under 240 sun illumination [17], the current world record. The typical structure for such cells is a multi-layer epitaxial thin film stack grown on a Ge substrate with tunnel junctions in between to match the currents between each bandgap cells (Fig. 2) [18]. At present these cells are expensive (>$7 W−1) and used in space applications, as well as terrestrial concentrator systems in solar power stations in which small area is necessary.

Another leading concept for high efficiency is the intermediate band (IB) solar cell [19]. Again, the idea is to absorb more of the solar spectrum, though in this case by introducing states within the bandgap of a semiconductor material such that low-energy photons can be absorbed in a two-photon process that promotes charge carriers to the conduction band (Fig. 3). Such a band structure would have a limiting efficiency of ∼60% and thus has great promise. However, to date very few bulk materials have been shown to possess such intermediate states. One of these is the ZnCdTe system demonstrated by Yu et al. using bulk samples using a combined high-energy oxygen implantation and laser recrystallization process [20]. There are still challenges for such an approach: (1) a way of p–n doping to separate charge carriers must be developed; (2) the problem of potentially significant non-radiative recombination in such materials must be addressed; (3) assuming the above are solved, a manufacturable process would need to be developed. Other materials have been studied theoretically [21], but not demonstrated experimentally. The use of quantum dot assemblies to form an effective intermediate band structure (mini-bands) is being pursued by Luque and co-workers [22], [23], who have shown promising initial results. Technical challenges for this concept include absorption, charge transport, and manufacturability issues that must be addressed by further research (see below). At present, IB solar cell concepts are in the early research stage, though are of great interest for future PV devices.

A third, and perhaps longest-term, approach for high efficiency solar cells is the concept of carrier multiplication. This implies that for a single absorbed photon, more than one electron–hole pair is generated, mainly by an avalanche-type process employing a high local field. Nozik first proposed this theoretically [24]; recent photoluminescence experiments on PbSe and PbS nanocrystals have shown that it is indeed possible to multiply the number of excitons to 7 for a single absorbed photon [25]. The practical aspects of this, including charge separation [26], etc. require additional basic research.

Finally, there are various efforts in up [27] or down [28] energy conversion of photons into a suitable energy regime for a single bandgap solar cell. This is primarily done by use of another layer integrated above or behind the cell. The layer typically consists of micron or nano-sized phosphor particles that absorb part of the spectrum and convert it to a more suitable energy for the solar cell in the composite system. To date, efforts in down-conversion have been successful using bulk materials and with limited success when using nanostructures, with attempts using Si nanocrystals [29] and quantum dots [30] being the main approaches. Recent work in up-conversion [31] has shown a small effect that is promising.

While research continues in developing such ultra-high efficiency Generation III solar cells, there are also significant efforts to develop Generation II-type thin film, single junction cells in order to decrease cost.

Thin films solar cells are widely recognized as a key solution to reducing the manufacturing cost of PV cells in the near to medium term. In 2007, thin film solar cells comprised ∼5–8% of the PV market, and their share is growing. Thin film solar cells are able to be produced at low cost by removing the bulk active substrate and using an additive deposition process on top of a low cost substrate such as glass, metal foil, plastic, etc. However, by using a thin film the efficiency is generally reduced compared to the bulk for two main reasons: (1) the thinner cell cannot absorb as much of the solar spectrum without more elaborate light trapping schemes or use of exotic compositions; (2) thin film cells are typically polycrystalline or amorphous and therefore contain a significantly higher density of non-radiative recombination centers at grain boundaries or in the bulk of the active semiconductor grains. Typical thin film cells and thin film cell-based modules, with a few exceptions (see below) [32], [33], [34], have an efficiency of less than 10%. [35], [36]. A brief review of the leading technologies is now given.

The leading technology is the CdTe system that is typically combined with CdS to form a heterojunction [37]. Such cells have been shown in the lab to produce a 16.5% power conversion efficiency [32], though manufacturing efficiency is expected to be on the order of 10–11%. A suitable recycling program is also typically required with CdTe technologies.

Another leading thin film material is Cu(In,Ga)Se2 (CIGS). This materials system, which sometimes includes substitution of sulfur for selenium and has now been studied for over 20 years [38], has been shown to produce a lab efficiency of ∼19.9% [39] and module efficiencies on the order of 8–12%. CIGS is able to absorb >90% of the solar spectrum above its bandgap with a thickness of only 2–3 μm [40], and reasonable efficiencies are also achievable with a 1-μm thick film [41]. The high cost and low abundance of In is a concern with CIGS in the long-term, as is the relatively high degradation observed under long-term exposure to the atmosphere [42] that requires high quality encapsulation to minimize.

Si-based thin film cells have been explored in various formats, including polycrystalline and amorphous materials. The efficiency of such cells has been ∼8% or less, making them suitable for specific PV applications, though hydrogenated amorphous silicon (a-Si:H) thin films cells make up a large fraction of the thin film PV market share. Efforts to grow either large grain films or well-oriented (crystallographically) films have been made, with limited success. Recently, tandem structures based on amorphous Si-based materials have yielded maximum efficiencies of ∼15.5% in laboratory cells, with stabilized values of ∼13% [43]. It is noted that a-Si:H cells are subject to the so-called Staebler–Wronski effect, in which the power conversion efficiency of the cells degrades after initial exposure (∼1000 h) to sunlight until a stable lower efficiency value is obtained [44]. This is related to the creation of additional point defects upon excitation of light that act as non-radiative recombination centers in the a-Si:H thin film material. A summary of best results achieved to date for thin film cells [36] and modules is given in Table 1.

Section snippets

Nanostructure-based concepts

Many types of nanostructures have been applied to solar cells. The nanostructures to be discussed will be classified into four types: (a) nanocomposites [3D], (b) quantum wells [2D], (c) nanowire and nanotubes [(quasi) 1D], and (d) nanoparticles and quantum dots [(quasi) 0D]. These structures have been employed in various functions and for various performance/energy conversion enhancement strategies. What follows is a brief review of the historical development for the nanostructure class under

General technical challenges

The above discussion on charge transport in QDs highlights the general problem that the absorption, charge separation/conversion, and transport mechanisms in nanostructured solar cells are not well understood and require further addressing. This was made evident by the work of Mora-Sero et al. [168], who showed preliminary data that the electrical characteristics of nanostructured solar cells may be fundamentally different from conventional bulk or thin film p–n junction solar cells. The

Summary and outlook

The major approaches to applying nanostructures to photovoltaics have been discussed, following a review of the status of standard solar cells based on silicon and thin films, as well as the major high-efficiency concepts that have been proposed in the literature. It was shown that 3D nanocomposite cells have already shown promise in a commercial context with dye-sensitized solar cells having shown a champion cell efficiency of ∼11% and modules have been produced with an efficiency of ∼6%. 3D

Acknowledgements

The author would like to thank D. Merfeld, E. Butterfield, T. Feist, G. Trant, and M.L. Blohm for their support of this work. Thanks also to B.A. Korevaar, O. Sulima, J. Rand, R. Corderman, J. Balch, J. Fronheiser, and R. Rohling for technical collaboration and useful discussions.

References (168)

  • W.G.J.H.M. van Sark et al.

    Energy Policy

    (2007)
  • A. Marti et al.

    Sol. Energy Mater. Sol. Cell

    (1996)
  • A.J. Nozik

    Physica E

    (2002)
  • V. Svrcek et al.

    Thin Solid Films

    (2004)
  • A. Shalav et al.

    Sol. Energy Mater. Sol. Cell

    (2007)
  • X. Wu et al.
  • M.V. Yakushev et al.

    Thin Solid Films

    (2004)
  • M. Powalla et al.

    Thin Solid Films

    (2001)
  • S. Zhang et al.

    Surf. Coat. Technol.

    (2003)
  • K.H. Chung et al.

    Mater. Sci. Eng.

    (2003)
  • J. Gang et al.

    Scr. Mater.

    (2003)
  • J.-B. Han et al.

    Sol. Energy Mater. Sol. Cell

    (2005)
  • S. Guha et al.

    J. Non-cryst. Solids

    (2006)
  • G.L. Araujo et al.

    Sol. Energy Mater. Sol. Cell

    (1994)
  • J.C. Rimada et al.

    Microelectron. J.

    (2007)
  • D.C. Johnson et al.

    Sol. Energy Mater. Sol. Cell

    (2005)
  • D.C. Johnson et al.
  • J.R. Heath et al.

    Chem. Phys. Lett.

    (1993)
  • N. Lewis (California Institute of Technology):...
  • M.I. Hoffert et al.

    Nature

    (1998)
  • E. Becquerel, La lumi_ere: ses causes et ses e_ets, tome second, Paris (1867), p....
  • D.M. Chapin, C.S., Fuller, G.S. Pearson, A new silicon p–n junction photocell for converting solar radiation into...
  • J. Zhao et al.

    Appl. Phys. Lett.

    (1998)
  • W. Shockley et al.

    J. Appl. Phys.

    (1961)
  • M.J. Kerr et al.
  • J. Singh

    Electronics and Optoelectronic Properties of Nanostructures

    (2003)
  • Y.S. Tsuo, P. Menna, T.H. Wang, T.F. Ciszek, New opportunities in crystalline silicon R&D, NREL Report #CP-590-25612...
  • J. Carey, What's raining on solar's parade, Business Week, February 6,...
  • M.A. Green

    Third Generation Photovoltaics: Advanced Solar Energy Conversion

    (2003)
  • W. Hoffman

    Sol. Energy Mater. Sol. Cell

    (2006)
  • M.B. Prince

    J. Appl. Phys.

    (1955)
  • C.H. Henry

    J. Appl. Phys.

    (1980)
  • R.R. King et al.

    Appl. Phys. Lett.

    (2007)
  • H. Yoon et al.

    Prog. Photovolt: Res. Appl.

    (2005)
  • A. Luque et al.

    Phys. Rev. Lett.

    (1997)
  • K.M. Yu et al.

    Phys. Rev. Lett.

    (2003)
  • P. Wahnón et al.
  • A. Martı et al.

    Phys. Rev. Lett.

    (2006)
  • A. Luque et al.
  • A.J. Nozik

    Annu. Rev. Phys. Chem.

    (2001)
  • R.D. Schaller et al.

    Nano Lett.

    (2006)
  • T. Trupke et al.

    J. Appl. Phys.

    (2002)
  • T. Trupke et al.

    J. Appl. Phys.

    (2002)
  • W.G.J.H.M. van Sark et al.

    Sol. Energy Mater. Sol. Cell

    (2005)
  • B. Yan et al.
  • M.A. Contreras et al.

    Progr. Photovolt.: Res. Appl.

    (1999)
  • A. Shah et al.

    Science

    (1999)
  • B. von Roedern et al.
  • B.M. Basol
  • U. Rau et al.

    Appl. Phys. A: Mater. Sci. Process

    (1999)
  • Cited by (0)

    View full text