Available online at www.sciencedirect.

com

Energy Procedia 00 (2013) 000000

www.elsevier.com/locate/procedia

SiliconPV: March 25-27, 2013, Hamelin, Germany

A graphical user interface for multivariable analysis of silicon

solar cells using scripted PC1D simulations
Halvard Hauga*, Birger R. Olaisena,b, rnulf Nordsetha, Erik S. Marsteina,c
a
Institute for Energy Technology, Instituttveien 18, 2007 Kjeller, Norway
Norwegian Defence Research Establishment, Instituttveien 20, 2007 Kjeller, Norway
c
Department of Physics, University of Oslo, 0316 Oslo, Norway

Abstract
PC1D is an open source computer program for simulation of crystalline semiconductor devices, and is the most
commonly used simulation software in the solar cell community today. In this work we present a modified version of
PC1D which runs from a command line, thus allowing for scripted simulations using any programming language. In
order to increase the accessibility of the new program, we have developed a free-of-use graphical user interface
which can be run as a standalone executable. We demonstrate the potential of this program by presenting examples of
different simulations. The new features that have been implemented in the new program include multivariable
analysis, visualization of three-dimensional data sets, extraction of relevant parameters not directly available in
PC1D, utilization of optimization algorithms, fitting to experimental data and implementation of additional features
like contact shading, emitter series resistance and passivated rear surfaces with local contacts. Simulations can be
performed in a simple and efficient manner, allowing for fast analysis and optimization of cell performance, as well
as an enhanced fundamental understanding of the physics of solar cell devices.

2013 The Authors. Published by Elsevier Ltd.

Selection and/or peer-review under responsibility of the scientific committee of the SiliconPV 2013
conference
Keywords: Solar cells; simulation; software

1. Introduction
PC1D is one of the most commonly used modelling programs for simulation of crystalline
semiconductor solar cells. The program uses a finite-element numerical method for solving the coupled
nonlinear equations for carrier generation, recombination and transport in the device and can be used both

* Corresponding author. Tel.: +47 994 86 757.

E-mail address: [email protected].

Halvard Haug/ Energy Procedia 00 (2013) 000000

for simulation of device performance and as a tool for new users to understand the fundamentals of solar
cell physics [1]. The main advantages of PC1D include high calculation speeds, an intuitive user interface
and an extensive list of material and physical parameters. By varying the applied bias or the wavelength of
the excitation light source, PC1D can calculate both current-voltage characteristics and spectral quantum
efficiency of the solar cell, but it also has a large number of other options for output data, both in the
spatial domain and in the time domain.
However, PC1D has some limitations in that it is somewhat cumbersome to do parameter variations,
and additional calculations cannot be implemented directly by running it from within another program.
The source code for the program is however freely available, allowing for user-defined modifications that
have the potential to fix these issues. In this work we present a new version of the PC1D called cmd-pc1d
which runs from a command line, thus allowing for scripted simulations using any programming
language. In addition, we have also developed a free-of-use program which we have named PC1D for
Matlab, which have many built-in possibilities for more advanced analysis, using the command line
version of PC1D as the basis for the calculations. This program, which is written in Matlab, is modular
and easy to use, giving the user the possibility of switching between the original PC1D interface and the
new graphical user interface (GUI). The program is available free of charge and can either be run from
Matlab or as a standalone executable.
2. Details of the simulation programs
2.1. The cmd-pc1d program
While PC1D is a great tool for users wanting to manually experiment with solar cell simulations, the
original GUI makes it difficult to use for automated simulation or integration into other systems. The
cmd-pc1d program was developed to allow access to all aspects of PC1D without having to launch the
PC1D GUI. Since the PC1D source code has been made freely available, it has been possible to create a
modified version of the program where all GUI code has been removed, and replaced with a simple
command line interface called cmd-pc1d.exe. The program can be run in two different modes, (i) with a
binary .prm parameter file, in which case the program runs the included simulation, or (ii) with both a
.prm file and a batch file, which causes the program to run in batch mode. The simulation results are
either written to a output file or directly to the command line window, which allows this version of PC1D
to be scripted with common scripting tools like Matlab.

Fig.1. Schematic illustration of a typical workflow using the command-line version of PC1D and the two converter programs
(marked in blue text).

Halvard Haug/ Energy Procedia 00 (2013) 000000

In order to control cmd-pc1d from another program, the other program needs to create a binary .prm
parameter file with specifications and commands that PC1D understands. This can be done with the
original PC1D program, but in order to be flexible, two small tools (called convert_prm_to_ascii.exe and
convert_ascii_to_prm.exe) were made to convert the PC1D .prm parameter files between binary format
and text format. Fig. 1 shows a typical workflow when using cmd-pc1d from an external program.
Starting with a base .prm file, which can be made using the original PC1D program, the file is converted
to text format. Then individual parameters are modified, as required by the problem to be solved, and the
file is converted back again to binary format. Next, cmd-pc1d is run with the modified .prm file, and the
results are either saved in a text file or output directly into the calling program. Depending on the type of
program, the results could be displayed to the user, or a different modification could be done to the
parameter file used to run the next iteration.
2.2. The PC1D for Matlab program
Even though many types of calculations are possible with the cmd-pc1d program, it requires a high
effort for users without knowledge of programming. In order to increase the accessibility of the program
to a wider audience, we have created a graphical user interface using the Mathworks Matlab software [2],
which in addition to setting up simulations using the cmd-pc1d program also implements many of
Matlabs features for data analysis, visualization and optimization. The program can be run from Matlab,
but has also been compiled as a standalone executable which does not require a Matlab license.
An image of the PC1D for Matlab user interface is shown in Fig. 2. A group of menu items and
buttons for different actions like loading and saving .prm files, handling simulation results and changing
plotting options are placed in the top part of the screen. The user can choose between five different tabs
for access to various features of the program, including a complete list of input parameters, single and
batch simulation setup, selection of output data, plotting options for batch results and options for
importing and fitting simulations to experimental data. In the first tab a subset of the parameters used to
define the PC1D prm-file is shown in a scrollable list with descriptive parameters names that can be
defined in a separate text file. The user has the option to change to the original PC1D interface at any
time. The figure window placed on the right part of screen can be used for plotting PC1D output data,
data sets calculated from the output data (e.g. current density vs. applied voltage) or results from batch
simulations (e.g. open circuit voltage as a function of bulk carrier lifetime).
The PC1D for Matlab program has implemented a range of calculations used both to calculate
additional results from PC1D output data and for determining PC1D input parameters based on a list of
secondary parameters. A few simulation examples are given in section 3 below, but the rest of
calculations are not described in detail in this paper. A list of the different calculations and secondary
parameters that are currently included is however shown in Table 1. For a complete and updated
description of the different calculations that are included in the program and the necessary settings for
each of them, please see the documentation file which is provided at the download page.
In addition to this, the PC1D program have capabilities for importing external (experimental) data for
direct comparison with the simulated results. We have also implemented an Nelder-Mead optimization
algorithm available through the fminsearch() function in the Mathworks Optimization Toolbox [3], which
can be used to fit simulated data to an external curve or to optimize the results with respect to e.g. the cell
efficiency. Any set of primary or secondary parameters may be set as fitting variables.

Halvard Haug/ Energy Procedia 00 (2013) 000000

Fig.2. Illustration of the graphical user interface at the time of writing. Several relevant features of the program are indicated.

Table 1. List of different result values that are calculated in the PC1D for Matlab program and the parameters of the secondary
models used to set input values for PC1D simulations. Note that changes and additions may occur in future updates of the program.
Results calculated from PC1D output data
Scalar values

Vector data

Secondary parameters used for

setting PC1D input data

Short circuit current

Current density

Open circuit voltage

Power density

Contact finger spacing

Fill factor

Global AM1.5 solar spectrum

Contact finger width

Efficiency

Reflectance times solar spectrum

Contact finger thickness

Emitter sheet resistance

EQE times solar spectrum

Contact finger length

Emitter junction depth

IQE times solar spectrum

Busbar width

Steady state carrier lifetime

Excess minority carrier density

Busbar thickness

Steady state generation rate

QSS carrier lifetime

Busbar length

Steady state injection level

Transient carrier lifetime

Metal resistivity

Generalized lifetime at n = 1015 cm-3

Generalized carrier lifetime

Area specific contact resistance

Sinton optical constant

Integrated weighted reflectance

Front side metallization grid:

Local rear contacts:

SRV of metallized area
SRV of passivated area
Metallization fraction
Contact pitch
Area specific contact resistance

Halvard Haug/ Energy Procedia 00 (2013) 000000

3. Simulation examples
In this section we present different examples of simulations that are possible with the command-line
version of PC1D and the accompanying graphical user interface. Note that all these simulations can be
performed in a both fast and understandable manner on a standard personal computer. The estimated time
used for setting up the simulation in the user interface by a trained user and the approximate simulation
time are indicated in the figure captions for each case.
3.1. Optimization of a double layer antireflection coating
In order to illustrate the potential for large scale batch simulations in the PC1D for Matlab program,
we have simulated a planar Si solar cell while changing the layer thicknesses in a double layer
antireflection coating (ARC) placed on a planar Si surface. The two thicknesses were varied between 0
and 150 nm in steps of 2 nm, giving a 76x76 array of result data, with all the characteristics of the I-V
curve saved at each point. A constant refractive index of 2.5 and 1.5 was used for the innermost and
outermost layer, respectively. Fig. 3 shows a colour map of the short circuit current of the simulated cell,
plotted within the PC1D for Matlab program. To illustrate the use of the scripted PC1D simulations in an
optimization process, we tracked the steps of the optimization algorithm described above and plotted it
over the colour map in Fig. 3. The optimization function was used to find the maximum short circuit
current with the two layer thicknesses set as free fitting variables.

Fig. 3. 2D colour map showing the short circuit current density of a Si solar cell as a function of the two layer thicknesses in a
double layer ARC. (5776 points. Setup time: ~30 s. Simulation time: ~10 min). The black lines illustrate the steps of an optimization
algorithm used to find the optimal parameters. (Setup time: ~30 s. Simulation time: ~45 s).

3.2. Simulation of a quasi-steady state photoconductance measurement

Quasi-steady state photoconductance (QSS-PC) is a technique commonly used to measure the effective
minority carrier lifetime of a Si wafer [4]. The measurement process can be simulated using PC1D by
calculating the time dependency of the excess carrier density n and optical generation rate G in a
passivated Si wafer. For the calculations shown in this paper, the spectral distribution and time-varying
intensity of the flash that is mounted in the Sinton WCT-100 measurement setup were used as the

Halvard Haug/ Energy Procedia 00 (2013) 000000

excitation source. The generalized expression for the effective carrier lifetime
as [5]

can then be calculated

( )
( )

( )

(1)

When the time-dependent cumulative excess conductivity and cumulative photogeneration have been
chosen as output values, the program automatically calculates the effective lifetime following both the
quasi-steady state, transient and generalized equations described in Ref. [5]. It therafter lets the user plot
them as a function of the excess carrier density, resulting in the familiar curves of the effective lifetime vs
injection level that are obtained from a QSS-PC measurement. The program will also store the effective
lifetime value corresponding to an injection level of n = 1015 cm-3, which can for instance be plotted as
function of one of the variable input parameters in a batch simulation. Fig. 4 (a) shows the effective
lifetime vs excess minority carrier density with varying bulk doping concentration NA.
Because the effective carrier lifetime is calculated from the photogeneration rate in the bulk of the
sample, it is important to know the sample thickness and the optical properties of the sample surfaces
when performing an QSS-PC measurement. These properties are described in the spreadsheet controlling
the measurement as a so-called optical constant, which is defined as the ratio between the photogenerated
current in the sample and a reference value of 38 mA/cm2. This value depends both on the wafer
thickness and the front side reflectance integrated over the spectral distribution of the excitation source,
and is generally not straight-forward to calculate. The optical constant is another example of a useful
additional output value that is automatically calculated in the PC1D for Matlab program when the correct
output data is chosen. Fig. 4 (b) shows the optical constant for a polished Si wafer as function of the
thickness of a coating layer with constant refractive index of 2.1 for different wafer thicknesses. The
results where obtained by running a single, two-parameter variation from the user interface.

Fig.4.(a) Simulated QSS-PC curves (effective carrier lifetime as a function of excess carrier density) for different acceptor doping
levels. Parameters used for the simulations: Bulk lifetime SRH parameters:
s,
cm/s on both
surfaces, midgap defect levels. The effective lifetime was calculated using the generalized formula described in Ref. [5]. (Setup
time: ~30 s. Simulation time: ~30 s). (b) The optical constant (used in a Sinton QSS-PC measurement [3]) for a polished wafer as a
function of the thickness of the coating layer for different wafer thicknesses. (Setup time: ~30 s. Simulation time: ~1 min).

Halvard Haug/ Energy Procedia 00 (2013) 000000

3.3. Si solar cell with passivated rear surface and local contacts.
In the last section we showed examples of how additional calculations based on PC1D output data
have been implemented into the PC1D for Matlab program. Another possibility is to include calculations
in the beginning of the process, determining parameters which are then used as input for the PC1D
calculations. One example of a case where this approach is relevant is when calculating the rear side
surface recombination velocity (SRV) of a passivated Si solar cell with local point contacts. This is
normally achieved with Fischers equation, giving the rear side SRV as [6]

(2)

where D is the minority carrier diffusion constant, W is the wafer thickness, p is the pitch between the
contacts, f is the metallization fraction and Smet and Spass are the SRVs of the metallized and passivated
regions, respectively. These and other parameters used for such secondary models are shown in a separate
list in the GUI (see Fig. 2), and can be edited and used for parameter variations and fitting in the same
way as the normal PC1D parameters. The GUI also allows for setting a value for the specific contact
resistance, which together with the metallization fraction and the cell area is used to determine the total
internal series resistance of the base contact. Fig. 5 shows the simulated cell efficiency of a Si solar cell as
a function of the rear side metallization fraction for different values of the SRV of the metallized area. For
the lowest metallization fractions the cell efficiency drops quickly because of series resistance in the
contacts. The efficiency is also reduced at higher metallization fractions for large values of Smet caused by
the increase in SRV at the rear surface, resulting in a peak in efficiency at approximately 3 % for this
particular case.

Fig. 5. Simulated cell efficiency of a PERC-type Si solar cell as a function of the rear side metallization fraction for different values
of the SRV of the metallized area Smet. Values for the SRV of the passivated area
in units of cm/s, the contact pitch and the
area-specific contact resistance
are indicated. (Setup time: ~30 s. Simulation time: ~1 min).

Halvard Haug/ Energy Procedia 00 (2013) 000000

4. Program availability and documentation

The files for the graphical user interface, the command line version of PC1D and the associated
conversion programs can be downloaded free of charge from the following URL:
www.ife.no/en/ife/departments/solar_energy/downloads.
The cmd-pc1d program runs on 32-bit and 64-bit windows, but versions for the Mac and Linux
operating systems may be added in the future. The standalone version of PC1D for Matlab program is
currently only compiled for 64-bit Windows, and requires the installation of a library of Matlab files
called Matlab Compiler Runtime, which can be downloaded and installed free of charge (see download
page for details).
For more information regarding the various options and calculations implemented in the PC1D for
Matlab program, readers are referred to the documentation file shown on the download page. New
updates will be documented in a separate changelog which can be found at the same location. Please cite
this conference paper in addition to the original PC1D publications if you plan to use the software in
published material.
5. Summary
In this paper we have presented a modification of PC1D to create a version called cmd-pc1d which can
be run from a command line. We also present a new graphical user interface for the program, which
includes many new features for multivariable analysis, advanced plotting, optimization, fitting to
experimental data and implementing additional models and parameters not available in the original PC1D
program. The new GUI can be run as a standalone executable without the need for any paid license files,
and can be downloaded free of charge.

Acknowledgements
This work has been funded by the Research Council of Norway through the project "Thin and highly
efficient silicon-based solar cells incorporating nanostructures," NFR Project No. 181884/S1.

References
[1]
[2]
[3]
[4]
[5]
[6]

D. A. Clugston and P. A. Basore, PC1D version 5: 32-bit solar cell modeling on personal computers. Conference Record
of the 26th IEEE Photovoltaic Specialists Conference, 1997. pp. 207210.
The Mathworks, Inc, MATLAB 2012b, Natick, Massachusetts, United States.
The Mathworks, Inc, Optimization Toolbox Users Guide, 2009. [Online]. http://www.mathworks.se/help/toolbox/optim/
R. A. Sinton, A. Cuevas, and M. Stuckings, Quasi-Steady-State PhotoConductance, A new method for solar cell material
and device characterization, Conference Record of the 25th IEEE Photovoltaic Specialists Conference, 1996, p. 457.
H. Nagel, C. Berge, and A. G. Aberle. Generalized analysis of quasi-steady-state and quasi-transient measurements of
carrier lifetimes in semiconductors. Journal of Applied Physics, vol. 86, no. 11, pp. 62186221, 1999.
B. Fischer. Loss analysis of crystalline silicon solar cells using photoconductance and quantum efficiency measurements.
Dissertation. Universitat Konstanz, Konstanz, 2003.