The sun radiates light at a characteristic temperature of approximately
6,000°K, resulting in a broad solar spectrum spanning ultraviolet
to mid-infrared. This very high-grade source of heat results in
the possibility of extremely high power conversion efficiencies,
ultimately reaching a thermodynamic limit of 93%  – the so-called
Landsberg limit.1 However, the breadth of emission photon energies
makes it difficult to efficiently convert sunlight into electricity. For this
reason, conventional solar cells generally demonstrate much lower
efficiencies. This can be traced to the property of the material used to
absorb sunlight – the semiconductor absorber.
Semiconductor Absorbers
All semiconductor materials absorb sunlight up to a threshold energy,
below which the solar cell becomes transparent. This threshold for
absorption is a fundamental property of all semiconductor materials
and is commonly known as the band-gap energy. This can vary
from the ultraviolet (GaN 3.4eV) to the mid-infrared (InAs 0.36eV). As
optical absorption can only take place for photon energies above the
band-gap energy, the choice of semiconductor (and hence band-gap)
controls how much of the solar spectrum can be absorbed. Its optical
absorption is also closely related to the optimum voltage at which the
solar cell will operate.2 In general, if a photon has an energy in excess
of the band-gap energy, the excess will be dissipated as heat. A low
band-gap semiconductor capable of absorbing much of the solar
spectrum and giving a large electrical current will therefore dissipate a
great deal of energy as heat, resulting in a low voltage. Conversely, a
solar cell with a high band-gap will preserve much of the energy from
the photons absorbed, resulting in a high voltage. However, it will be
transparent to much of the incident radiation, resulting in a low
current. Since electrical power is the product of current and voltage,
there is an optimum band-gap for the semiconductor solar cell that
maximises this product, resulting in greater efficiency.
The Efficiency Limit
The efficiency limit for solar cells was first defined by Shockley and
Queisser.3 Applying the principle of detailed balance, they proposed
that any solar cell capable of absorbing sunlight would also be capable
of emitting light, resulting in a fundamental limit for any solar cell
technology. Making the reasonable assumption that one photon would
result in one photogenerated electron, they were able to place limits
on the ultimate efficiency of a semiconductor solar cell, finding a
limit of 31% for a single semiconductor absorber under standard,
unconcentrated solar illumination.
Figure 1 shows the Shockley–Queisser limit, together with a number of
research laboratory solar cells4 with efficiencies of up to 28%. Typically in
industrial, large-area, mass production, the efficiencies of these solar cells
drop to between 10 and 20%. It is evident that while the conventional
single semiconductor solar cell can still be improved – especially in terms
of industrial manufacturing – these cells are nearing their theoretical limit.
Considering that the Landsberg limit is so high, it is likely that there
are alternative and fundamentally more efficient ways of converting
sunlight into electricity. This article illustrates some of these
fundamentally more efficient routes to photovoltaic solar energy
conversion, often dubbed ‘third-generation photovoltaics’.5
Routes to Improve Solar Cell Efficiency
Concentration of Sunlight
One effect that can lead to an increase in these conventional solar cell
efficiencies is the concentration of sunlight, whereby lenses or mirrors
are used to focus light onto small, but highly efficient, solar cells.6 In
general, doubling the incidence upon a solar cell results in an
approximate doubling in the current delivered, but also to a
modest increase in voltage. The latter stems from a fundamental
thermodynamic loss regarding the absorption and emission of light
from the solar cell.7,8 By fully concentrating sunlight 46,200 times, this
loss can be eliminated, increasing the efficiency of an otherwise ideal
Shockley–Queisser solar cell from 31 to 41%. In practice, optical
collectors are generally built to concentrate sunlight a few hundred
times and, exceptionally, several thousand times.9
Multijunction
The problem of absorbing the spectrally-broad incident sunlight yet
maintaining high voltage at the solar cell arises because a single
semiconductor material only efficiently converts photons at its
band-gap energy; excess energy is lost as heat. This situation can be
remedied by engineering several semiconductor absorbers with
© T O U C H B R I E F I N G S 2 0 1 1
Daniel J Farrell received his MPhys from the University
of Hull in 2004 and his PhD from Imperial College
London in 2009. He currently holds the position of
Research Associate in the Department of Physics at
Imperial College London. Daniel's main research
interest lies in harnessing optical and electronic
nanoscale processes to fundamentally improve the
efficiency of photovoltaic solar energy conversion, and
in particular the hot-carrier solar cell. He is the author
of pvtrace: an advanced optical Monte-Carlo
simulation program that simulates photovoltaic optics
and luminescent materials.
E: daniel.farrell@imperial.ac.uk
Nicholas J Ekins-Daukes currently holds a Senior
Lectureship at Imperial College London, where he
works on photovoltaic devices for highly efficient solar
energy conversion. Previously, he lectured at the
School of Physics at the University of Sydney in
Australia, before that he held a JSPS fellowship at the
Toyota Technological Institute, Japan, working on
advanced photovoltaic devices.
a report by
Daniel J Farrel l 1 and Nicholas J Ekins-Daukes2
1. Research Associate, Department of Physics; 2. Senior Lecturer, Department of Physics and the Grantham Institute for Climate Change, Imperial College London
Routes to High-efficiency Photovoltaic Conversion of Solar Energy
84
So
la
r