Energy sources

Energy is indisputably a resource on which humanity has become dependent. Without energy,
our society will not function. Without energy, we cannot find or administer medicine to cure
disease, prepare food, purify water, drive our cars, operate computers, study at night, etc. The
current energy need is roughly 15 TW (15·1012 W) and this number is projected to increase.
Historically fossil fuel (coal, petroleum, and natural gas) have enabled our energy consumption
for the past century, and continues to dominate our energy production. Today roughly 81% of our
energy is supplied by fossil fuel, 2.7% is being supplied by nuclear energy, and the remaining
share is renewable sources, with biomass being the largest source of energy at roughly 12%.
You can see a breakdown of our energy consumption in figure 1.

Figure 1: Current energy mix. The total energy consumption is roughly 15 TW.
Source: Wikipedia.

Fossil fuels are continually being formed via natural processes such as anaerobic decomposition
of buried dead organisms fueled by photosynthesis. They are, however, generally considered
non-renewable resources because fossil fuels take millions of years to form and the known
viable reserves are being depleted faster than new ones are being made. Even if fossil fuels can
cover the energy consumption for many years to come, there are many other reasons to look for
alternatives. CO2 pollution is probably the strongest arguments against fossil fuels, but
combustion of fossil fuels also produces other air pollutants, such as nitrogen oxides, sulfur
dioxide, volatile organic compounds, and heavy metals. Harvesting, processing, and distributing
fossil fuels create their own environmental concerns. Coal mining methods, particularly
mountaintop removal and strip mining, have negative environmental impacts, and offshore oil
drilling pose a hazard to aquatic organisms. Hydraulic fracturing used to extract natural gas
entail its own suit of local environmental concerns.

When we are looking for energy sources available that can replace fossil fuels, it is useful to
consider how much energy each process can deliver. Below you can see estimates of the
available power from each energy process:

 Tide: 0.3 TW
  Earth heat: 2 TW
 Hydro power: 4 TW
 Wind: 75 TW
 Biomass: 6 TW
 Direct radiation: 26,000 TW
 Coal: 900 TWy
 Petroleum: 240 TWy
 Natural gas: 215 TWy
 Uranium: 300 TWy

Notice that numbers for fossil fuel and uranium are in total energy, the remaining numbers are
given as resources available per year (the resources could be given as TWy/y to emphasize
that they are a yearly resource, however they are shown in TW for simplicity). While the specific
numbers may vary from source to source, the magnitude of the numbers are reasonably
accurate.

Available energy
Let us do a calculation of the direct radiation from the sun ourselves. We know the luminosity of
the sun to be 3.828·1026 W, and the distance from the sun is one astronomical unit or
1.496·1011 m. In the last section, we learned how to calculate the solar constant (the radiation
power per square meter or irradiance outside the Earth’s atmosphere), by dividing the luminosity
of the sun with the area of a sphere of radius 1 AU. Now we want the direct radiation over the
entire earth, so we will need to multiply this number with the area of a disc with radius equals
the radius of Earth (6.371·106 m).

PEarth=PSunAsphere⋅AEarth=PSun4⋅π⋅r2SE⋅π⋅r2E=3.828⋅1026[W]4⋅π⋅(1.496⋅1011[m])2⋅π⋅(6.371⋅106[m]

)2=1.75⋅1017W
Notice that the units become W since the two meter terms cancel out. This number represents
the total energy received from the sun, however, we would like to calculate a technical potential.
Therefore, we must consider the amount of energy lost to atmospheric scattering and
absorption, we assume that 51% will be transmitted through the atmosphere. Further we would
like to only count the landmass, so we consider the water to landmass ratio on Earth of 29%. The
total technical potential of direct radiation therefore becomes

PTechnical=1.75⋅1017⋅0.51⋅0.29=2.6⋅1016W=26,000TW.
26,000 TW is a strikingly big number and to provide a visual reference you can look at figure 2,
where the available energy is plotted for all renewable energy sources plus fossil and uranium.
Figure 2. Diagram of energy sources by available resources. Notice that while all renewable
sources are given per year, the fossil and uranium sources are complete available resources.
The direct radiation is only partially shown.

Looking a figure 2, the direct radiation contribution dwarfs all other resources including the total
available reserve of fossil fuel. This should, however, not come as a surprise, as direct radiation
is the primary energy source behind almost all the energy sources listed. Let us take a step
backwards and look at the processes and primary energy sources behind the energy sources on
which we rely. If we consider the primary energy source behind all the energy technologies we
use, they fall under four categories: movement of the planets, earth heat, solar radiation,
and supernova. In figure 2, you can see a flowchart linking the primary energy to its concrete
uses.

Figure 3. All energy sources categorized by their primary energy, process, and concrete use.
From earlier, we know that the original process behind fossil fuel is photosynthesis, a process
fueled by the direct radiation from our sun. The same is true for biomass. Condensation and rain,
the process behind hydropower is also fueled by sunlight, as is atmospheric movements (wind)
and waves. The only primary energy source aside from solar radiation which contributes
significantly to our energy mix is uranium. Together with other heavy elements, uranium was
formed in the supernova signifying the last stellar evolutionary stages of massive stars. The
supply of uranium on Earth is therefore fixed, and uranium must be considered a depletable
resource on Earth.

Looking back at the energy mix, it is remarkable that only about 16% of the world consumption
is based on renewable energy sources. Wikipedia hosts an interesting list of countries by
electricity production from renewable sources. Some countries have an energy mix dominated by
renewable sources even going up to 100% of the electricity production. In all cases these high
numbers arise because of fortunate geographical conditions making hydropower a dominant
solution. The country scoring highest when disregarding hydropower is Denmark with 56.58
percent of its electricity being generated from renewable sources mainly wind, biomass, and
solar. These numbers demonstrate that renewable energy sources can generate a much larger
share of our energy, however, we must also remember that renewable electricity production,
from sources such as wind power and solar power can be criticized for
being variable or intermittent, as weather and day/night cycles dictate. In countries, such as
Denmark, this problem is solved by importing energy from neighboring countries when
production is low. A complete switch to renewable energy will not be easy, and many solutions
need to be implemented. I hope that your takeaway point from this text is that renewable
sources are fully capable of delivering all the energy we can ever need. We need
to deploy and develop renewable technology to optimally use the resources available to us and
at the same time evolve our electrical grid and energy storage options.

Sunlight at Earth
So far, we have discussed the various energy sources available to us without going into depth
with any of them. Since the subject of the course is solar cells, let us now turn our attention to
solar energy. The Sun itself is a massive fusion reactor in which hydrogen atoms is fused into
helium. The energy from this fusion reaction is released into Space in the form of radiation. We
have already encountered the value for the radiation power of the sun (3.828·10 26 W), which we
used to calculate the solar constant (1,367 W/m2). The number we calculated is the value as
measured outside Earth’s atmosphere. In figure 1 you can see the spectrum both inside and
outside the atmosphere. We designate the light as measured outside the atmosphere Air Mass
0 (AM0) since this light has not passed through the atmosphere. The irradiance of this light is the
already mentioned 1,367 W/m2, the solar constant.
Figure 1. Solar radiation spectrum for direct light at both the top of Earth's atmosphere (yellow
area) and at sea level (red area). As light passes through the atmosphere, some is absorbed by
gases with specific absorption bands. Source Wikimedia.

When the sunlight passes through the atmosphere the spectrum changes, see figure 1. There
are various reasons for this change

 Reflection of light: Sunlight is reflected in the atmosphere reducing the radiation

reaching the Earth
 Absorption of light: Gases (O2, O3, H2O, CO2, …) with specific absorption bands
absorb a part of the radiation causing gaps in the spectrum (see figure 1).
 Rayleigh scattering: When light falls on particles smaller than the wavelength
Rayleigh scattering occurs. As the effect is strongly wavelength dependent, shorter
wavelengths are scattered strongly causing the blue color of the sky.
 Scattering of aerosols and dust particles: This effect is a Mie scattering event
and concerns particles larger than the wavelength. The number of aerosols and dust
particles depends greatly on location, being greatest in industrial and densely
populated areas.

As the effect of the atmosphere is dependent on the length of the path through the
atmosphere it is necessary to designate different spectra according to the path through the
atmosphere. Therefore, we use the term Air Mass (AM) followed by a number indicating the
distance through the atmosphere. AM0 is the spectrum outside the atmosphere. AM1 is the
spectrum after it has traveled the vertical height of the atmosphere. if the sun is at an angle to the
Earth's surface the effective thickness will be greater. AM1.5 atmosphere thickness, corresponds
to a solar zenith angle of z=48.2° and indicates that the light has travelled 1.5 times the vertical
path through the atmosphere. The specific value of 1.5 has been selected in the 1970s for
standardization purposes, based on an analysis of solar irradiance data in the United States.
Since then, the solar industry has been using AM1.5 for all standardized testing or rating of
terrestrial solar cells or modules.
When we sum the energy according to the AM1.5 spectrum in figure 1 we find that only 835
W/m2 is received. Thus only 61% of the originally available 1367 W/m2 is received at Earth as
direct radiation. However, it is important to note the we are now forgetting a large portion of
radiation, namely the diffuse radiation caused by the scattering of the light in the atmosphere.
To account for this, we use the AM1.5G spectrum, where G stands for global radiation and is a
summation of the direct and diffuse radiation. This is the reason we use 1000 W/m2 as the total
irradiance when we determine the peak power of a solar module.

Area need
Let us move on to the specific topic of the course and see how well solar cells would be able to
cover our energy needs. Let us therefore calculate the area needed to cover the Earths energy
consumption (15 TW) with solar cells at 10% efficiency. Since we know the solar constant to be:
1000 W/m2, we can calculate the area needed as

A=\frac{P_{need}}{\eta \cdot \sigma} = \frac{15 \cdot 10^{12} \left[ W \right] }{ 0.1

\cdot 1000 \left[ \frac{W}{m^2} \right] } = 1.5 \cdot 10^{11} m^2 = 150,000
km^2A=η⋅σPneed=0.1⋅1000[m2W]15⋅1012[W]=1.5⋅1011m2=150,000km2

In this calculation η is the efficiency, while σ is the solar constant. To calculate a realistic area on
Earth needed to cover our energy consumption we need to add some assumptions to consider
the day/night cycle, as well as atmospheric conditions. If we assume eight hours of average
daylight and that 70% of all days have sunshine, we can calculate a realistic area needed to
supply the world with solar power.

A=150,000 \left[km^2\right] \cdot \frac{1}{0.7} \cdot \frac{24}{8} = 640,000

km^2A=150,000[km2]⋅0.71⋅824=640,000km2

This is a big area, in fact it is roughly size of France. But remember this is a realistic number.
We assumed 10% efficiency which is well within reach of current technology. If we place all the
solar panels in the Sahara desert in Africa with its area of 9,200,000 square kilometers, we can
find space for our world energy producing solar park 14 times. Realistically we would need to
distribute the solar panels all over the world to produce power matching demand. We will also
need to find solutions to store energy when the sun is not shining.

Apart from the question of whether we can find space for all the solar cells we would need to
power the world, the next obvious question is: Can we make enough solar cells? We are not
ready to fully answer this question yet, but just to get an idea of the scale, let us look at
something else we produce in large areas: newspapers. Let us imagine that we gathered up all
the issues of the most widely circulated newspapers from every day in 2016. When we lay the
pages next to each other we would find that the issues of the 10 most circulated
newspapers corresponds to the area of solar cells we need (assuming all papers are in
Broadsheet and 100 pages per issue). What we have just demonstrated is that our current
technology and production capacity is not in itself a hindrance for us to produce 640,000 square
kilometers of surface area. We will leave the question of whether todays solar technology can be
scaled to this task for the remainder of the course as we need a specific discussion based on
each solar cell technology.

Additional resources
Before you continue to the exam, you can review these additional resources. Please note that all
of these additional resources are strictly voluntary.

Full lenght interviews

 Economics of solar cells. Watch the full interview where Poul Erik Morthorst talks
about the economics of solar cells.
 Efficiency of solar cells - Light sources. Watch the full interview where Nicholas
Riedel efficiency of solar cells and the importance of using well defined light
sources.

Other great resources

 Renewable energy. Wikipedia article on renewable energy sources.

 Photovoltaics: Fundamentals, Technology and Practice. This book by Konrad
Mertens is a great resource on solar cells. Chapter 1 is especially relevant to this
module.

Virtual instruments

 Incidence angle. This virtual instrument demonstrates the dependence of incidence

angle of a solar panel to the power production. You can use the virtual instrument by
moving the angle slider and observe the shadow and the power produced by the
solar cell.
 Diffuse light. This virtual instrument demonstrates the influence of diffuse light. You
can use the virtual instrument by moving the diffuse/direct radiation slider as well as
the angle slider and observe the shadow and the power produced by the solar cell.
 Complete

Power from a solar cell

To understand and measure how much power is produced from a solar cell, the characteristic
curve of a solar cell is an important concept to understand. The characteristic curve show the
current and voltage (IV) characteristics of a solar cell or module giving a detailed description of
its solar energy conversion ability and efficiency, see figure 1. This characteristics curve is most
often called an IV-curve and is basically a graphical representation of the operation of the solar
cell or module summarizing the relationship between the current and voltage at the existing
conditions of irradiance and temperature. IV curves provide the information required to configure
a solar system so that it can operate as close to its optimal peak power point as possible.

Figure 1. Characteristic IV curve of a solar cell. The short circuit current (I SC), and open circuit
voltage (VOC) is marked along with the maximum power point current (IMPP) and voltage (VMPP).

As we saw in the video, it is easy to measure two of the characteristic values of the IV curve,
namely the open circuit voltage (VOC) and the short circuit current (ISC) with a simple multimeter.
The open circuit voltage of the solar cell is the maximum voltage that the solar cell will supply,
while the short circuit current of a solar cell is the maximum current the solar cell will produce.
The issue with the two states, ISC and VOC, is that the most interesting aspect of a solar cell is not
the current flow with no potential drop, nor the potential drop with no current flow, but the product
of these two; the power. When either the potential drop or the current flow is zero, the power,
being the product of the two, will be zero. Therefore, a more interesting aspect is the maximum
power the solar cell can produce.

Figure 2. Schematic of solar cell with a load. An IV curve is obtained when the load is varied and
the voltage and amperage is recorded.

By increasing the resistive load on a solar cell continuously from zero (short circuit) to a very high
value (equivalent to open circuit) one can determine the maximum power point, see figure 2.
Plotting the power as a function of the voltage results in the blue curve, see figure 3, where the
maximum power point (PMPP) is the peak point of the power curve.
Figure 3. Power curve for a solar cell. The power is the product of the voltage and the current.

A high quality monocrystalline silicon solar cell may produce 0.55 V open-circuit (V OC) , at 45 °C
cell temperature (and a slightly higher voltage at lower temeratures), however, the maximum
power is typically produced with 75% to 80% of the open-circuit voltage and 90% of the short-
circuit current. If a solar cell vendors only rates their solar cell "power" as V OC · ISC, without giving
load curves, the actual performance can be seriously distorted.

The value linking the product of VOC and ISC to Pmpp (maximum power point) is the fill factor (FF)

FF = \frac{P_{MPP}}{V_{OC} \cdot I_{ISC}}FF=VOC⋅IISCPMPP

The fill factor is used as a quality parameter for solar cells and has a value around 80% for a
normal silicon PV cell. The primary parameter extracted from the IV curve is the power
conversion efficiency (PCE), which describes the general efficiency of the solar cell; that is the
ratio of generated electricity to incoming light energy. It can also be expressed in terms of the
open circuit voltage, the short circuit current, and the fill factor

η=PoutPin=VOC⋅ISC⋅FFPin
This formula builds upon last week, in that we now focus on the power output of the solar cell. As
you may remember we can express the input power in terms of the area of the solar cell and the
solar constant as Pin=σ·A.

Important terms

 Open Circuit Voltage (VOC) The open circuit voltage of the solar cell is the maximum
voltage that the solar cell will supply; that is the voltage when an infinite load is
applied.
 Short Circuit Current (ISC) The short circuit current of a solar cell is the maximum
current of the solar cell under conditions of a zero resistance load; a free flow or
zero volt potential drop across the cell.
 Maximum power output (PMPP) The maximum power point is the maximum product of
voltage and current along the IV curve.
  Fill factor (FF) The fill factor is the ratio between the maximum power produced by
the solar cell and the product of VOC and ISC.
 Power conversion efficiency (η or PCE) The power conversion efficiency is the ratio
of output power to input power.

ned in the previous section, we need to apply a resistive load to the solar cell, before we can
measure its power production. In this section, we will substitute this resistive load with
an electronic load, by using a source measure unit.

A source measure unit (also called a SMU or source-meter) outputs a voltage and measures

the current that flows. A bench-top power supply would be able to fulfill this role; however, most
source measure units are programmable and allows the user to sweep the voltage over a
defined range making the IV curve measurement automatic. It is also possible to supply a
current and measure the voltage, however this is not typically done for solar cells, since the
voltage is typically predictable for a given device.

Figure 1. An illuminated solar cell causes a current to flow into the load (left). The source meter
acts as a load and sinks the current (right).

A specific advantage of using a source measure unit over a resistive load is that the source
measure units can drive the solar cell with negative voltage or voltages higher than the open
circuit voltage of the solar cell. This results in an extended IV curve, see figure 2.

Figure 2. Full IV curve. The maximum power point is shown with a red dot.

In the following module, we will focus on how insights into the solar cell can be obtained by
modeling the solar cell and using the full IV curve.
 Taking the sun inside
Measurements of efficiency of solar cells requires a stable light source that closely matches the
conditions of sunlight. It is important that not only the intensity, but also the spectrum is matched
to a standard. The most direct option is to use the sun itself, but cloud covers and variations in
atmospheric conditions complicates test and require correction to compare measurements over
time. The spectrum also changes throughout the day further limiting the time for testing.

In place of sunlight, an artificial light source that simulates the sun can be substituted. We call
these light sources solar simulators (or artificial suns) and define them as devices that provide
illumination approximating natural sunlight. The artificial light sources are classified according
to three criteria:

1. Spectral match
2. Spatial uniformity
3. Temporal stability

Classification Spectral Match Irradiance Spatial Non-Uniformity Temporal Instability

Class A ±25% 2% 2%

Class B ±40% 5% 5%

Class C ±60-100% 10% 10%

A common classification (EN 60904-9) rates each in one of three classes: A, B, or C (see table
1). A solar simulator meeting class A specifications in all three dimensions is referred to as
a Class A solar simulator, or sometimes a Class AAA.

Table 1. Solar simulator classification. The spectral match is evaluated in six defined spectral
regions between 400 and 1100 nm.

It is difficult to make a light source that exactly matches the AM1.5 spectrum and with the
necessary illumination intensity. The list below reviews commonly used light sources for solar
simulators

 Xenon arc lamps offer high intensities and an unfiltered spectrum that matches
reasonably well to sunlight, see figure 1. However, many undesirable sharp atomic
transitional peaks, making the spectrum less desirable for some spectrally sensitive
applications, characterize the Xe spectrum. See the spectrum in figure 1.
 Metal halide arc lamps were primarily developed for use in film and television
lighting with a high temporal stability and daylight color match are required.
  Quartz tungsten halogen lamps offer spectra that very closely match black body
radiation, although typically with a lower color temperature than the sun. This means
the UV part of the spectrum is lacking, see figure 1.
 Light-emitting diodes (LEDs) are used in research laboratories to construct solar
simulators. They offer a promise for energy-efficient spectrally tailored artificial
sunlight using multiple LEDs. Depending on the number of diodes any spectrum can
be approximated with an LED based solar simulator.

As figure 1 shows there is often a considerable amount of variation between the spectrum of the
lamp and the required AM 1.5G spectrum. There are two approaches for correcting for the
differences between the AM1.5 spectrum and the actual spectrum from a solar simulator. The
first approach is to use a calibration cell that has the same spectral response as the cell under
test. The light intensity is adjusted so the measured ISC of the calibration cell matches the ISC as
measured when the cell was calibrated at an external laboratory. However, slight changes in cell
processing cause changes in spectral response and the need for a new calibration standard.
Further for some solar cell technologies calibration cells may not be possible to obtain and
maintain for novel device materials and architectures.

Figure 1. Spectrum for a halogen lamp (green), xenon lamp (blue), and AM1.5G (red).

The second approach to compensate for the differences in spectrum is to measure the spectral
response of the device under test. Then this response is used to correct for the known
difference between the spectrum of the light source and the standard spectrum. This approach
is more time consuming and to some extent error prone. However, this method will work for any
solar cell technology including novel devices.

Modelling a solar cell

To understand the electronic behavior of a solar cell, it can be useful to create an electrically
equivalent model. In figure 1 left, you can see a schematic representation of a solar cell, and to
the right is a simplified electrical equivalent. We call this the simplified equivalent circuit.

Figure 1. Schematic representation of a solar cell (left) and a simplified electrical equivalent
circuit with one diode (right).

The electrical behavior of the equivalent circuit can be expressed mathematically by

considering Kirchhoff’s circuit law

I = I_{Ph} – I_DI=IPh–ID

IPh is the photo current and ID is the current running through the diode. By substituting in the
formula for the Shockley diode equation, we get the following

I = I_{Ph} –I_{S}\left(e^{\frac {V_{D}}{nV_{T}}}-1\right)I=IPh–IS(enVTVD−1)

IS is the reverse bias saturation current, VD is the voltage across the diode, VT is the thermal
voltage (kT/q, Boltzmann constant times temperature divided by electron charge), and n is the
diode ideality factor. The ideality factor n typically varies from 1 to 2 (though can in some cases
be higher), depending on the fabrication process and semiconductor material. The ideality factor
accounts for imperfect junctions as observed in real transistors and mainly accounts for carrier
recombination. When n is 1, the equation is called the Shockley ideal diode equation.

If we draw an IV curve using this formula we can see, we get the same characteristic solar cell
curve that we saw in the last module, see figure 2. Let us now look at the short circuit current and
the open circuit voltage as expressed by the formula above.
Figure 2. Characteristic IV curve of a solar cell. The short circuit current (I SC), and open circuit
voltage (VOC) is marked.

Short circuit current

As we know from the last module, a short circuited solar cells (the contacts are directly
connected) delivers the short circuit current (ISC) and the voltage is 0. With the simplified
equivalent circuit equation, this results in

ISC=I(V=0)=IPh–IS(e0–1)=IPh
Thus, we know that the short circuit current (ISC) is equal to the photocurrent (IPh). We can in fact
see this result directly from looking at the equivalent circuit (figure 1), since an external short
circuit also short circuits the internal diode (meaning ID=0).

The photo current is directly proportional to the number of photons hitting the solar cell (learn
more in week 3) and thus the short circuit current is proportional to the irradiance. In figure 3, you
can see the characteristic solar cell curve as a function of increased incident irradiance.

Figure 3. Characteristic curve (IV curve) with increasing incident radiation.

Open circuit voltage

The other extreme case aside from the short circuit current is the case when the current
becomes zero. We call this case the open circuit voltage, see figure 2. To determine the V OC, we
need to solve the equivalent circuit equation for the case when the current is 0.

VOC=V(I=0)=n⋅VT⋅ln(IPhIS+1)=n⋅VT⋅ln(ISCIS+1)
Remember that the photocurrent equals the short circuit current (I Ph=ISC). The +1 term inside the
logarithm can be ignored for anything above extremely small currents, so an approximate version
of the equation can be written as
VOC≈V(I=0)=n⋅VT⋅ln(ISCIS)
Thus, we can see that the irradiance dependency of VOC to incident radiation is much lower for
VOC as compared to ISC. In fact, the open circuit voltage changes with the natural logarithm of the
irradiance (remember that the photocurrent is proportional to the irradiance). You can also see
this in figure 3, as the VOC is only slightly affected by decreasing irradiance.

Parasitic resistances
In the previous section, we saw how we could model a solar cell with a simple equivalent circuit.
In this section, we will take the model a step further by considering parasitic resistances in the
solar cell. Resistive effects in solar cells reduce the efficiency of the solar cell by dissipating
power in the resistances. In figure 1, you can see an equivalent circuit including series
(RS) and shunt (RSH) resistances.

Figure 1. Parasitic series (RS) and shunt (RSH) resistances in a solar cell circuit.

The equivalent circuit diagram consist of a current source arising from the absorbed light, a diode
from the directional properties of the solar cell stack, a series resistor Rs, for the resistance in the
open diode state, and a shunt resistance RSH, for the resistance in the closed diode state. The
total current through the circuit can be described as

I = I_{Ph} – I_D – I_{SH}I=IPh–ID–ISH

I is the output current of the solar cell, IPh is the photo-generated current, ID is the diode current
and ISH is the shunt current. Expanding the equation with the the diode current equation and the
shunt current with the resistive loss terms results in

I = I_{Ph} - I_{S} \left( e^{\frac{V + I \cdot R_{S}}{nV_T}} - 1 \right) - \frac{V + I

\cdot R_{S}}{R_{SH}}I=IPh−IS(enVTV+I⋅RS−1)−RSHV+I⋅RS

The equation is a function of the internal parameters and the voltage drop between the two
contact points. I appears on the left as well as the right hand side of the equation, so we will
solve it numerically. In figure 2 and 3, you can see the influence of series resistance and shunt
resistance on the IV curve. With an increasing RS, the curve flattens and the fill factor decreases
considerably. For decreasing values of RSH, the situation is similar; however, here the open circuit
voltage is affected as the rising shunt current causes the diode voltage to drop. As you can see
the key, impact of parasitic resistance is to reduce the fill factor, which also explains why the fill
factor is a good quality control value for solar cells.

Figure 2. Influence of series resistance (RS) on the solar cell characteristic curve.

Figure 3. Influence of shunt resistance (RSH) on the solar cell characteristic curve.

Determining the parasitic resistances

It is possible to approximately determine the shunt (RSH) and series (RS) resistance from the IV
curve of a solar cell. If you look at figure 2, you can see that the slope of the IV curve varies
drastically at the VOC point when the series resistance is changed. The same can be said for the
point of zero voltage in figure 3 as the shunt resistance is changed. At the point where the
voltage is 0, the current over the diode can be ignored since we are effectively at short circuit
conditions. We can write the equation for the current as

I = I_{Ph} - \frac{V + I \cdot R_{S}}{R_{SH}}I=IPh−RSHV+I⋅RS

The slope of the curve is found as the derivation

\frac{dI}{dV} = 0 - \frac{1}{R_{SH}} - \frac{R_S}{R_{SH}} \frac{dI}{dV} dVdI=0−RSH1
−RSHRSdVdI

We can resolve this equation as

\frac{dI}{dV} = -\frac{1}{R_S + R_{SH}}dVdI=−RS+RSH1

If we assume the series resistance to be much smaller than the shunt resistance (R S « RSH),
which would be correct for any reasonable solar cell, we can finally get

R_{SH} = - \left. \frac{dV}{dI} \right|_{V=0}RSH=−dIdV∣∣∣V=0

This shows that the shunt resistance RSH can be determined directly from the slope of the tangent
in the short circuit point, see figure 4. A similar evaluation can be made for the open circuit case
showing the series resistance can be determined by the slope of the tangent in the open circuit
point, see figure 4.

Figure 4. RS and RSH determination from the solar cell characteristic curve. The gradient in the
short circuit and open circuit point can be used to determine the shunt and series resistance
respectively.

Series and parallel connections

When we use solar cells we typically connect more than one solar cell together to increase the
power produced. When we make these connections it is important to understand how the current
and voltage adds together depending on the connection. There are two ways to connect solar
cells, parallel and series connections.
Parallel connection

In a parallel connection all cells are forced to have the same voltage, while the currents are
added up.

V=V_1=V_2=…=V_nV=V1=V2=…=Vn and I=I_1+I_2+…+I_nI=I1+I2+…+In

Figure 1 shows a module of solar cells connected in parallel. On the left you can see the
characteristic curves of modules consisting of three, two, and one cell.

Figure 1. Parallel connection of solar cells. For each cell the voltage will be the same, while the
currents is added from each cell.

Series connection

When cells are connected in series the current for each cell will be the same, while the voltages
are summed.

I=I_1=I_2=…=I_nI=I1=I2=…=In and V=V_1+V_2+…+V_nV=V1+V2+…+Vn

In figure two, you can see a module with serially connected solar cells and a characteristic curve
for three, two, and one connected cells.

Figure 2. Series connection of solar cell. For each cell the current is the same, while the voltage
is added from each cell.

We typically see series connections of cells in modules since this ensures higher voltages. If we
try to imagine a solar cell module consisting of 100 cells each having a voltage of 0.6 V and a
current of 4 A, we can connect all cells in either parallel or serial connection. In the case of a
parallel connection the module voltage would be 0.6 V and the current 400 A. In the case of a
serial connection the voltage would be 60 V and 4 A. In either case the power is the same,
however, the transport losses would be significantly different. If we consider Ohm’s law and the
power law, we see that the power lost to a wire scales with the current squared. Therefore we
would need extremely thick cables if we were to transport high current at low voltages.
Conversely high voltages and low currents give much lower transmission losses high voltages
and low currents give much lower transmission losses and thinner cables are required.

Shadow effects
In the last section we saw how solar cells are connected electrically. It is important to consider
what happens when one or more solar cells in such connections are in shadow.

In figure 1 you can see how a shaded cell in a module affects the overall module performance.
To draw this plot it is assumed that the shaded cell is 90% shaded. From earlier we know that
the open circuit voltage of a shaded cell only changes slightly, while the short circuit
current will drop with around 90%. The overall IV curve for a module decays with roughly the
amount of current loss from the one shaded cell. Thus the overall power loss of the module
corresponds to the shaded area. This means that a parallel connection is nicely behaved in
partial shading losing only the shaded area.

Figure 1. A module consisting of three cells connected in parallel is shown. When all cells are
illuminated the blue IV curve is obtained. When one module is in shadow the module IV curve
corresponds to the green curve. You can see the IV curve of a single unshaded cell (dotted
orange) and a single shaded cell (dotted grey).

While the parallel connection is relatively good-natured when it comes to partial shading, the
same cannot be said for a series connection, as can be seen from figure 2. With one shaded cell
in the series connection the two other cells must push their current through the shaded cell. This
causes a negative voltage over the shaded cell. The current is limited by the shaded cell and
therefore the overall current of the module is limited. As can be seen from figure two, this effect
reduces the power output from the module with a single shaded cell significantly.
Figure 2. A module consisting of three cells connected in series is shown. When all cells are
illuminated the blue IV curve is obtained. When one module is in shadow the module IV curve
corresponds to the green curve. You can see the IV curve of a single unshaded cell (dotted
orange) and a single shaded cell (dotted grey).

A solution to the problem with series connection is to employ a bypass diode over the shaded
cell, see figure 3. Bypass diodes allow current to pass shaded cells and thereby reduce the
voltage losses through the module. When a module becomes shaded its bypass diode begins to
conduct current through itself. With the bypass diode the current is no longer limited by the
shaded cell, and the impact on the overall IV curve is greatly reduced, see figure 4.

Figure 3. Bypass diode mounted over shaded cell. The bypass diode drastically reduces the
losses caused by the partial shading.

Bypass diodes also serve another purpose, namely the prevention of hotspots. In modules with
one shaded cell and no bypass diode, there will be a massive heating of the shaded cell caused
by the other cells. The usage of bypass diodes alleviates this problem.

In real solar modules bypass diodes are not employed for each cell. The reason is that the
diodes would have to be incorporated into the encapsulation if the solution should be practical.
Instead bypass diodes are employed for strings of cells (e.g. 12, 18, or 24 cells). It is therefore
important that modules are not mounted in places were partial shading is expected during the
day.
Figure 4. A module consisting of three cells connected in series is shown. When all cells are
illuminated the blue IV curve is obtained. When one module is in shadow the module IV curve
corresponds to the green curve without any bypass diode and to the red curve with a bypass
diode. You can see the IV curve of a single unshaded cell (dotted orange) and a single shaded
cell (dotted grey).

How do solar cells work?

In order to understand how solar cells work it is important to understand the structure and
properties of semiconductors. We will go through this topic in this module, however, before we
get so far let us look at a simplified step-by-step explanation of how a solar cell works. Figure
1 shows the four steps graphically.

1. When sunlight shines on the solar cell, photons (light particles) bombard the upper
surface.
2. The photons carry their energy down through the cell.
3. Upon absorption the photon gives its energy to an electron creating an electron /
hole pair.
4. The electron moves across the barrier into the upper n-type layer and escapes out
into the circuit.
5. Flowing around the circuit, the electron charges the battery.
Figure 1. Simplified explanation of the operation of a solar cell.

The explanation above is highly simplified and it leaves a list of questions

 How much energy does each photon carry?

 Why and how far are the photons moving through the materials?
 What happens at the absorption event and what is a hole?
 Why and how does the electron move to the electrode?

Semiconductors
As its name implies, a semiconductor is a material that conducts current, but only partly. The
conductivity of a semiconductor is somewhere between that of an insulator, which has no
conductivity, and a conductor, which has full conductivity. Most semiconductors are crystals and
in this section, we will focus on silicon as an example of a semiconductor.

To understand how silicon acts as a semiconductor, we turn to how electrons are organized in an
atom. The electrons in the atom are organized in layers called shells. The outermost shell is
called the valence shell. When the atoms bonds with neighboring atoms, the electrons in
valence shell are the ones that form covalent bonds. Silicon has four electrons in its valence
shell making it possible for each silicon atom to bind with the valence electrons from four other
silicon atoms forming a lattice structure. In figure 1, a two-dimensional description of the lattice
structure is shown.
Figure 1. Each green circle represents a silicon atom, and the red circles between the atoms
represent the shared electrons. Each of the four valence electrons in each silicon atom is shared
with one neighboring silicon atom. Thus, each silicon atom is bonded with four other silicon
atoms.

The electrons in the covalent bonds formed between each atom in the crystal structure are held
in place by these bonds and hence they are localized to the region surrounding each atom.
These bonded electrons cannot move or change energy, and thus are not considered free. They
cannot participate in current flow, absorption, or other physical processes of interest in solar
cells. However, the situation where all electrons are stuck in this way is only true at a
temperature of absolute zero. At elevated temperatures, electrons can gain enough energy to
escape from their bonds. When this happens, the electrons are free to move about the crystal
lattice and participate in conduction. At room temperature, a semiconductor has enough free
electrons to allow it to conduct current. At or close to absolute zero a semiconductor behaves
like an insulator.

Bandgap

For an individual atom, there are defined discrete energy levels for the electrons to occupy.
Electrons in atoms (and molecules) can change energy levels by emitting or absorbing a photon,
with energy exactly equal to the energy difference between the two levels.

When we bring two atoms together a mutual coupling of the atoms occurs. Thus, the energy
conditions change and each energy level is divided into two individual levels. In the case of
three coupled atoms we get three levels and so on, see figure 2. When we look at our silicon
crystal lattice from before we practically get an infinite number of atoms coupled together, and
thus a continuum of energy levels. We call these energy bands, where each band is an energy
state permitted for an electron.
Figure 2. Origin of energy bands in a semiconductor crystal. As the number of coupled atoms
increases the number of energy levels follows, and when the number of atoms is infinite
continues energy bands are formed.

The existence of the energy bands explains the conduction behavior of the

semiconductor (remember the semiconductor is an insulator at low temperatures, and a
conductor at high temperatures). The reason being that the highest occupied energy band is
filled with electrons, while the lowest unoccupied energy band has no electrons, see figure 3. We
call these bands the valence band (highest occupied band) and the conduction band (lowest
unoccupied band) respectively. If the electrons remain in the valence band they are bound and
the semiconductor behaves as an insulator, which is the case at zero-degree kelvin (-273.15 °C).
When the temperature is increased, individual electrons can loosen from their bonds and
become available in the crystal as free electrons, see figure 3. In the band model this
corresponds to the case where the electrons transitions from the valence band to the conduction
band. These free electrons increase the conductivity of the crystal, making it a conductor at
elevated temperatures. The width of the gap between the valence band and the conduction band
is called the bandgap and in the case of silicon its value is Eg = 1.12 eV.

Figure 3. Conduction and valance bands for silicon. At a temperature of absolute zero no
electrons are free (left), while at elevated temperatures electrons, can gain enough energy to
escape from their bonds (right) becoming free electrons capable of conducting a current.

Insulators, semiconductors, and conductors

To round out the introduction to semiconductors, we will look at how the energy band applies to
the two other classes of materials, namely insulators and conductors. In the case of insulators,
the bandgap is very large (typically greater than 3 eV), meaning that almost no free electrons will
be available even at high temperatures. In the case of metals (a conductor), we see a special
case, namely that the valence band and the conduction band overlap. In this case numerous
electrons are available as free electrons and therefore metals possess high conductivity even at
very low temperatures.
Figure 4. Energy bands for insulator, semiconductor, and conductor (metal) materials.

We are seeing the spectral response of the solar cell, i.e. number of electrons
output compared to the number of photons incident on the device.

Light interaction

A photon can be absorbed by a semiconductor when the energy of the photon is equal to or
greater than the bandgap, see figure 1. Since the energy levels form a continuum, any photon
with an energy greater than the bandgap will be absorbed.

Figure 1. When a photon is absorbed by the semiconductor, an electron is raised to the

conduction band.

The photon energy is exclusively a function of the wavelength of the photon. Other factors, such
as the intensity of the radiation, do not affect photon energy. In other words, two photons of light
with the same color (and, therefore, same wavelength) will have the same photon energy, even if
one was emitted from a candle and the other from the Sun. The energy and momentum of a
photon depend only on its frequency (ν) or inversely, its wavelength (λ) according to the
following formula

E=h(w)=h*v=hc\lambda E=ℏω=hν=hc/λ

ω = 2πν is the angular frequency, h is the Planck constant, and ħ = h/2π is the reduced Planck
constant. The higher the photon's frequency, the higher its energy. Equally, the longer the
photon's wavelength, the lower its energy. We know both Planck’s constant (h = 6.626·10 -34 J·s)
and the speed of light (c = 2.998·10 8 m/s)

hc=6.626⋅10−34[J⋅s]⋅2.998⋅108[m/s]=1.9865⋅10−25J⋅m
Since we commonly report wavelength in nanometers (1 nm = 10 -9 m) and the bandgap and
photon energy in electron volts (1 eV = 1.602·10-19 J) we get this expression which relates the
energy and wavelength of a photon

E=1240/λ

where the energy is expressed in electron volts and the wavelength in nanometers.

Since the illumination power is fixed, the number of photons becomes inversely proportional to
the wavelength as the energy of each photon is given by

E=hc/λ

This explains the fact that a lower current is observed when the wavelength is

decreased (energy of each photon is increased and therefore there are fewer photons). Excess
photon energy does not contribute to the photo-current.

At long wavelengths the energy of each photon is insufficient to overcome the band-gap and

the solar cell in incapable of absorbing the photons. This explains the cutoff in current above

\lambda=\frac{1240}{E_g} = \frac{1240}{1.12} = 1107 nmλ=Eg1240=1.121240=1107nm

At 1107 nm the photon energy has equal energy to the band-gap of silicon, and wavelengths
below 1107 nm will be absorbed, while wavelengths above will not.

Figure 1. Spectral response of idealized solar cell. Notice the cutoff at the wavelength equivalent
to the band-gap.
Conductivity and doping
When an electron is released from its bond, figure 1 left, a gap is left in the crystal structure,
which we call a hole. This process is called electron-hole pair generation. The reverse
process, where a free electron falls back into a hole is called electron-hole pair recombination,
and is depicted in figure 1 right. We use the term “hole” to describe the lack of an electron at a
position where one could exist in an atom or atomic lattice because we can model the hole as a
positively charged particle in the semiconductor.

Figure 1. Generation (left) and recombination (right) of electron-hole pairs.

If an electric voltage is applied across a semiconductor an electric field is established. This

means that the negatively charged electrons will be accelerated in the direction of the positive
pole of the voltage source, and the holes are accelerated in the direction of the negative pole.
In this way the semiconductor acts as a conductor of current.

We use the notion of a hole as a way to conceptualize the interactions of the electrons within a
nearly full system, which is missing just a few electrons. The movement of a single hole
is conceptually simpler to work with, than monitoring every electron in the valance band. To
more simply explain the concept of hole mobility, you can imagine a row of people seated in
an auditorium, where there are no spare chairs. Someone in the middle of the row wants to
leave, and climbs over the back of the seat into an empty row, and walks out. The empty row is
the conduction band in this analogy, while the person leaving is a free electron. Now
someone else comes by and wants to sit down in the crowded row (the valence band). To create
space, a person in the crowded row moves into the empty seat the first person left behind and
the empty seat moves one spot closer to the edge. As every person follows suit and moves into
the empty seat, the empty seat effectively moves towards the edge of the row. Once the empty
seat reaches the edge, the new person can sit down.

Like we described the movement of an empty seat, hole mobility is actually the movement of
many separate electrons in the valance band.
Doping

As a starting point semiconductors (intrinsic semiconductors) are poor electrical conductors,

since the electrical conductivity relies on thermal electron excitation. Further the number of
electrons in the conduction band is equal to the number of holes in the valence band. However,
we can change this situation by introducing foreign atoms to the semiconductor crystal; this
process is referred to as doping. The addition of a small percentage of foreign atoms in the
regular crystal lattice produces dramatic changes to their electrical properties and we can
therefore produce n-type and p-type semiconductors, where there is an abundance of free
electrons or holes.

If for instance we introduce a phosphorus atom into a silicon crystal lattice, see figure 2, we
create a free electron, since phosphorus has one additional valence electron as compared to
silicon. This fifth valence electron finds no open bonds, and is instead only weakly connected to
the crystal lattice, meaning it is available as a free electron at room temperature. We can
illustrate this in the band diagram as the doping atom generates an additional energy level just
below the conduction band. This means that only very low energy (far less than room
temperature) is required to lift the electron into the conduction band. We call the foreign atom
(phosphorus in this case) a donor atom, as the donor atom “gives” a free electron to the crystal
lattice. The resulting semiconductor material is said to be n-doped, since the concentration of
free electron rises drastically and there is therefore a majority of negative charge carriers.

Figure 2. If a phosphorus atoms is introduced into a silicon crystal a free electron becomes
available, since phosphorus has five valence electrons compared to silicon's four. This is called
n-doping. Because of the doping a new energy level is available in the band diagram just below
the conduction band edge.

It is likewise possible to do p-doping of a semiconductor by introducing foreign atoms with less

valence electrons (e.g. boron) than the silicon semiconductor, see figure 3. In the p-doped
material hole conductance becomes possible. The boron atom is in this case referred to as
an acceptor atom. In any case, doping densities are very low, e.g. only 1 in 100,000 silicon
atoms are replaced. Yet the effect of doping can increase the conductivity of the material by
orders of magnitude.
Figure 3. If a boron atoms is introduced into a silicon crystal one electron is missing as the boron
atom only offer three valence electrons. This is called p-doping. Because of the doping a new
energy level is available in the band diagram just above the valence band edge.

The p-n junction

At this point we have spent considerable time discussing semiconductors, how they interact with
light, and their conduction behavior. We are still missing one crucial element before we can
understand how a solar cell operates. If we take two pieces of silicon, one n-doped and one p-
doped, and combine them, we create what is called a p-n junction. Figure 1 depicts the situation
where the two parts are being combined. There is a surplus of free electrons on the n-side,
which will diffuse to the p-doped region where they will recombine with the holes.
Reversely holes diffuse from the p-doped side into the n-doped side and recombine with
electrons. There are thus almost no free electrons and holes in the region near the junction and
because of the fixed charges in the junction region an electric field is established. This field
leads to the electrons being pushed to the left and the holes to the right, and we are left with
a space charge region that exists at the p-n junction.

Figure 1. The p-n junction is formed when a piece of n-doped semiconductor is put in contact
with a piece of p-doped semiconductor. Electrons will flow from the n-side to the p-side and there
occupy the holes. On the n-side fixed positive charges remain, while on the p-side fixed negative
charges are generated.

When we combine our knowledge of semiconductors, the p-n junction, and the interaction of
photons in semiconductors, we are ready to understand how a solar cell operates. In figure 2,
you can see a lighted p-n junction. We know that if the incoming photons has more energy than
the bandgap, it is absorbed and promotes an electron from the valance band to the conduction
band creating a free electron-hole pair. This electron-hole pair is separated by the prevailing
electric field: electrons to the n-side and holes to the p-side. There the electrons and holes
are majority carriers with little chance of recombination. When the electrons and holes are
drawn of at the contacts, we have generated power.

Figure 2. An illuminated p-n junction. The free electron and hole generated by light absorption
are separated by the electric field.

Recombination and diffusion length

When we consider the operation of a solar cell it is important to understand some of

its limitations. When electron-hole pairs are formed, there is a chance that they will recombine.
The mechanism of importance for this is typically imperfection recombination. When the
theoretically ideal crystal is impure due to foreign atoms, crystal structure errors or similar, the
recombination probability becomes extremely large. In order to compare materials we can
calculate a carrier lifetime. The carrier lifetime defines how long a generated electron-hole pair
exists on average until it recombines. In the case of silicon carrier lifetimes range
from microseconds to milliseconds, depending on the quality of the silicon and the doping
concentration. We can also express the carrier lifetime in terms of diffusion length, which
describes the distance a generated electron travels in the semiconductor before it recombines.
Typically diffusion lengths for silicon range from 50 to 500 micrometers.

When we design solar cells, this diffusion length becomes very important since it in many ways
dictate the constitution of the specific solar cell technology. Typically the layer thickness in a
solar cell should be on the order of the diffusion length, but this consideration must also be
weighed against the absorption characteristics of the specific materials.

In an n-doped semiconductor the concentration of free electron is drastically increased and there
is therefore a majority of negative charge carriers. A n-type silicon semiconductor is made by
introducing for example Phosphorus, that has one additional valence electron as compared to
silicon.

A p-doped semiconductor is created by introducing foreign atoms with less valence electrons
(e.g. boron) than the silicon semiconductor. In the p-doped material hole conductance becomes
possible.

With both types of doped silicon semiconductor it is possible to form a pn-junction. With the pn-
junction a static electric field is formed allowing the charges to be seperated and current
extracted from the solar cell.

Additional resources
Before you continue to the exam, you can review these additional resources. Please note that all
of these additional resources are strictly voluntary.

Other great resources

 How Does a Transistor Work? Great and intuitive educational video where Derek
from Veritasium explains in simple terms how a transistor works. The part about
semiconductors and doping from 1 minute to 3 minute and 20 seconds is especially
relevant to this course.
 How Solar Panels Work. Brilliantly explained and quit colorful video from Doc
Schuster explaining how solar cells work. You might want to skip the
introduction and go directly to the explanation. The video is a continuation from the
video: How Does a Diode Work which is also recommendable.
 pveducation.org. Check out this great online resource. The sections on the pn-
junction is especially relevant to this module.
 Photovoltaics: Fundamentals, Technology and Practice. This book by Konrad
Mertens is a great resource on solar cells. Chapter 3 is especially relevant to this
module.

Virtual instruments

Color response. Using this virtual instrument you can see how the current behaves when you
change the illumination wavelength (color). Notice that the illumination power is constant and
you only change the wavelength. At both low and high wavelengths (infrared and ultraviolet), the
light falls outside the visible spectrum and is shown as dark.
Theoretical solar cell efficiency
When we create solar cells, the efficiency is the parameter that typically gets the most attention.
The higher the efficiency, the better we use the available solar resource, and the smaller an area
we need. Ideally, a solar cell would use every single photon available and use all of the energy
from each photon ensuring that we have a 100% efficient solar cell. However, there are limits to
the efficiency we can achieve.

Losses in efficiency include optical losses and electrical losses such

as reflection, transmission, contact resistance, recombination sites, and so on. While we can
improve on all of these mechanisms, there are limits to a solar cells efficiency that we cannot
overcome, and from these we can determine a theoretical solar cell efficiency.

Spectral efficiency

A fundamental limit to the efficiency of a solar cell has to do with the bandgap of the
semiconductor used. Photons with energy less than the bandgap does not get absorbed and
photons with energy exceeding the bandgap lose their excess energy through thermalization,
see figure 1.

Figure 1. If a photon has less energy than the bandgap (EG) it will not be absorbed (left). Photons
with higher energy than the bandgap lose their excess energy in a thermalization process.

We can define the bandgap wavelength, λG, in terms of the bandgap

λG=h⋅c/ EG

This wavelength corresponds to the wavelength of light that is just absorbed. The portion of the
solar spectrum above λG cannot be used for electrical energy, because the photons have less
energy than the bandgap. The portion of the light is thus lost as transmission, see figure 1.

For photons with energy exceeding the bandgap (λ < λG), the energy is sufficient for absorption.
The surplus energy of the photons are, however, given up as the electron relaxes down to the
bottom of the conduction band. We call this process thermalization loss. This means that any
surplus energy cannot be used for electrical energy. In figure 2, you can see a visualization of the
spectral usage for a silicon solar cell with a bandgap of 1.11 eV.

Figure 2. Spectral losses in a silicon solar cell device with a bandgap of 1.11 eV. All photons with
an energy lower than the bandgap are lost to transmission, while higher energy photons lose
their excess energy to thermalization.

Theoretical efficiency

When we talk about determining the efficiency limit of a solar cell there are two things we have
excluded when we calculated the spectral efficiency. Firstly, in a real solar cell it is not possible to
use the full voltage, VMax = EG/q. Secondly, the fill factor cannot be 100% and therefore the
maximum power point current and voltage will be lower than ISC and VOC. Both these limitations
come from the fact that a real solar cell has a p-n junction. When we include these two limitations
it is possible to calculate a theoretical efficiency limit for any given bandgap energy, see figure 3.
It is important to note the we assume that all photons are absorbed and contribute to the photo
current. For any real solar cell there will be some reflection losses, recombination losses, etc.
However, it is remarkable that the record laboratory efficiency for silicon solar cells today
is 26.6%, when we know the theoretical limit is 29.4%.
Figure 3. Theoretical efficiency and its dependence on the bandgap energy. The wiggles in the
curve are a result of the IR absorption bands in the atmosphere.

With a monochromatic light source, we would create a solar cell with a semiconductor material
with a bandgap equal to the wavelength equal of the illumination source. In this way every single
photon would get absorbed, and none of their energy would be lost to thermalization.

Tandem and multi-junction solar cells

We have determined that a single junction solar cell makes poor use of the solar spectrum.
There are two reasons for this. First, photons with energies less than the bandgap do not
contribute to the photo current at all. Second, each photon with higher energy contributes one
electron to the current, all the energy exceeding the bandgap is lost. A method to overcome this
limitation is to split the spectrum and use a solar cell that is optimized to each section of
the spectrum.

The simplest way to achieve this is to stack two solar cells on top of each other. A structure of
this kind is a so-called tandem solar cell where two absorbers are stacked on top of each other.
The illumination should first strike the absorber with the larger bandgap it will absorb the light
with a high output voltage. This material will furthermore be transparent for low energy
light which passes to the second absorber with a lower bandgap. In figure 1, you can see
how a stacked tandem solar cell can optimize the utilization of the solar spectrum.
Figure 1. Spectral usage for a stacked tandem solar cell. The top cell has an absorber layer with
a bandgap of 1.63 eV, while the bottom absorber layer has a bandgap of 0.90 eV.

The fabrication of tandem devices can be quite demanding because the extraction of the
currents is non-trivial. Two designs exist, the monolithically integrated tandem and
the mechanically stacked tandem, see figure 2. In the case of monolithically integrated tandem
solar cells, the two cells are connected in series and therefore the current of both cells must
be matched (see the section on series connections from week 2). The absorber thicknesses
must be adopted to yield equal currents from both cells for a given spectral distribution. Spectral
changes during the day or throughout the year may give rise to spectral mismatch and
therefore the cell with less current will act as a load on the second cell and lower the overall
efficiency.

Figure 2. Schematic design of a stacked (right) and a monolithic (left) tandem device. The
illumination occurs from the top into a high bandgap absorber for high-energy photons (blue
light). Light with lower energy passes through to the lower bandgap absorber (red). The stacked
devices (right) consist of two isolated cells while the monolithic approach (left) requires a
transparent tunneling junction.
In stacked solar cells, the top and the bottom cells are contacted separately and we refer to
these as four terminal devices. Due to the separate connections of top and bottom cells,
current matching for the two cells are not required, making the combination of bandgaps more
flexible. The main problem of stacked cells is that they require transparent contacts so the light
transmitted from the top cell can pass to the bottom cell.

High efficiency tandem cells are fabricated from crystalline materials where the required flexibility
in band gap energies is best met by the III-V semiconductors. It is also possible to stack even
more junctions thereby making multi-junction solar cells. You may remember from week 1, that
multi junction GaAs solar cells are used in space applications where their high cost can be
justified.

In a monolithic tandem device the individual solar cells are connected with a tunnel junction,
meaning that we effectively make a series connection between the two cells. Therefore, the
current is restricted in that the device is limited to the cell with the lowest current production.

The top cell needs to be transparent to the light that the bottom cell must absorb. A high bandgap
material will transmit light with energies below the bandgap, leaving it for a lower cell.

Reflection and absorption

When we talk about optimizing the efficiency of solar cells, one extremely important aspect is
how much light gets absorbed by the solar cell. We will talk about two aspects affecting the
amount of light absorbed in the solar cell. One part of this is the reflection losses when the light
hits the surface of the solar cell. The other is the absorption by the material itself.
Reflection losses

When light hits a solar cell, it has to pass from one material (air) to another (silicon). Each of
these two materials have a different refractive index (n) meaning that light travels at different
speeds in either material. This change in material properties results in reflection from the
surface of the second material and the larger the difference in refractive index from one material
to another, the larger the reflection loss. We can write the strength of the reflection given by the
reflection factor R

R= ER /E0

Where E0 is the incident irradiance and ER is the reflected irradiance. For incoming light
perpendicular to the surface, the reflection factor is calculated as

R=[(n1−n2)/(n1+n2)]2

In the case of silicon the index of refraction in the visible spectrum is around 3.9 while air has a
refractive index of 1, so we get

R=[(1-3.9)/(1+3.9)]2

This means that we stand to lose approximately one third of the light. If we consider the case of
oblique incidence of radiation the reflection losses only increases. It is therefore clear that we
need to reduce the reflective loss if we want to make an efficient solar cell, and for this reason,
we apply anti-reflective coatings. The simplest way to create an anti-reflective coating is to
introduce a material in-between air and silicon with an in-between index of refraction. There are
also other strategies including interference anti-reflective coatings, textured surfaces, etc. We will
not go much deeper with the topic of anti-reflective coatings in this course, but it is important to
notice that solar cells using semiconductors with large indexes of refraction require anti-reflective
coatings in order to be efficient.

Figure 1. Anti-reflective coatings improve the light transmittance by reducing the amount of
reflection on the surface.
Absorption coefficient

When we have previously talked about a photon being absorbed if it has enough energy cause
an electron to transition from the valance band to the conduction band, we made a significant
simplification. In our depiction of the valance band and conduction band, we forgot the crystal
momentum. For some semiconductors, our simplification has no consequences if they are direct
bandgap materials, see figure 2.

Some materials have an in-direct bandgap and in those cases, an electron cannot shift from the
highest energy state in the valence band to the lowest energy state in the conduction band
without a change in momentum. A crystal vibration causes this change in momentum and the
important consequence is that absorption is far less likely. A photon with energy close to the
band gap can penetrate much deeper into the semiconductor before being absorbed in an in-
direct band gap material as compared to a direct band gap one. This fact is very important for
solar cells, as silicon is an indirect-gap material and therefore does not absorb light very well.
Silicon solar cells are typically hundreds of microns thick and if they were much thinner, light
(particularly infrared) would simply pass through.

Figure 2. Depiction of a direct bandgap (left) and an in-direct bandgap (right).

Additional resources
Before you continue to the exam, you can review these additional resources. Please note that all
of these additional resources are strictly voluntary.

Full length interviews

 Antireflective coatings, texturing, and nanostructuring . Watch the full interview where
Rasmus Schmidt Davidsen talks about avoiding reflective losses from solar cells.
The interview is focused on silicon solar cells.

 Making the best possible silicon solar cell. In this interview Rasmus Schmidt
Davidsen talks about making the best possible solar cell and the consequences of
using a in-direct bandgap material. It is advisable that you wait with watching this
interview until you have completed the silicon solar cell module in the next week.
Other great resources

 Tabulated values of the Shockley–Queisser limit for single junction solar cells .
Article describing in detail the Shockley–Queisser limit for single junction solar cells.
The article is not open access.

Virtual instruments

 Spectral efficiency. This virtual instrument lets you change the bandgap of a
semiconductor material while plotting the spectral utilization. When the bandgap is
changed the spectral efficiency is calculated along with the theoretical efficiency
limit.
 Tandem solar cells. This virtual instrument lets you change the bandgap of both
semiconductor materials in a tandem solar cell. When the bandgap is changed the
spectral efficiency is calculated along with the theoretical efficiency limit and the
spectral utilization is plotted. Please notice that the calculated efficiency is valid for
an unconstrained tandem device (a stacked device).
 Reflection losses. This virtual instrument demonstrates the effect of index of
refraction on the reflection loss on a solar cell. You can change the index of
refraction and see how the power output is affected.
 Angle of incidence and reflection loss. This virtual instrument demonstrates the
effect of the angle of incidence in combination with the index of refraction on the
power production of a solar cell. You can change the angle of incidence and index of
refraction to see how the power output is affected.