SCAPS User manual

The user manual of SCAPS as it stands now (november 2007) is rather messy!
It consists of the original user manual of the SCAPS 2.1 version offered to the PV research
public in 1998. All mechanisms and algorithms explained there are still valid. Enhancements
and additions of subsequent versions were just listed, and added in front of the original
manual… with a break from 2003 (version 2.4) to early 2007 (version 2.6.1). No consistent
manual for the actual SCAPS version (2.7.3, november 2007) is written yet…
The small ‘Getting Started’ manual, which is distributed with the SCAPS programme, however,
is up to date, and readable without reference to earlier documents. So, it is strongly advised to
start with this 5 pages introductory brochure, and not with this uncompleted historical
archive… which nevertheless contains useful information.

Marc Burgelman
ELIS-University of Gent
10 november 2007

1
SCAPS Version 2.7.3, November 2007
Enhancements of version 2.7.03, compared to version 2.6.1
Most enhancements have to do with the user interface, and should facilitate the interpretation
of the simulations. Version 2.7.03 (november 2007) is the first to be distributed externally
after version 2.6.1 of march 2007: the intermediate versions were internal test versions.
1. During a batch simulation, you see the simulation number within this batch task, and the
total number of simulation asked. You are reminded to keep SHIFT pressed when you
were too ambitious, the computer too slow, or whatever makes you to loose patience.
2. During a C-V simulation, an I-V simulation is automatically done also (this has always
been so). When you do an I-V simulation, followed by a C-V simulation, possibly at other
voltages, the I-V points from both simulations are merged now (in earlier versions, the
points calculated with the I-V action were discarded after the C-V action, though the result
of the I-V action was not immediately erased from the screen).
3. More parameters are now accessible in batch mode: parameters dealing with illumination,
with work point setting (e.g. the frequency f of the C-V simulation, and the bias voltage V
of the C-f or QE simulation), and with tunneling.
4. There are now a few batch parameters of an on/off type: you cannot give these parameters
another value then on or off (the user interface does not allow you to do so). These
parameters are: illumination off/on; without/with spectrum-cutoff filters; tunneling off/on
(intra-band tunneling at an interface; or tunneling to interface states at an interface defect).
After the batch calculation, these parameters are restored to their values in their respective
panels (action panel; interface panel; interface defect panel).
5. As a rule, when you vary a parameter in a batch simulation, all other parameters take their
value from the various input panels. There are however now a few exceptions:
1. When the batch parameter affects the illumination intensity (thus the transmission
T or the density value ND of the neutral density filter), SCAPS will set the
illumination on before changing this parameter. After the T or ND-variation, the
original illumination setting (on/off) is restored. This is so because T and ND have
only effect under illumination.
2. When the settings of the spectral cut-off filters are varied as a batch parameter
(thus on/off, or short λ, or long λ), the illumination is set on, and restored to it’s
original setting afterwards.
3. When one of the effective masses is varied as a batch parameter, the
accompanying tunneling mechanism is turned on, and restored to it’s original
setting afterwards (me* and mh* are only used with tunneling).
6. In the I-V panel, there are now 3 modes to display the J-V curves: (i) J-V ; (ii) the J-V
curves scaled to Jsc, thus J/Jsc vs. V (then all curves pass through the point (0 V, -1). When
a scaled J-V curve cannot be drawn, because Jsc was not determined, a warning is given
and the curve is not shown. And (iii) the J-V curves shifted over Jsc, thus J - Jsc vs. V (then
all curves pass through the point (0 V, 0 mA/cm2). When a shifted J-V curve cannot be
drawn, because Jsc was not determined, a warning is given and the curve is not shown.
The special J-V modes are implemented because they can be informative when studying I-
V curves under a varying illumination (see e.g. J. Marlein amd M. Burgelman, ‘Empirical

2
 J-V modeling of CIGS solar cells’, Proceedings of NUMOS (Gent, march 2007), p. 227-
234.
7. On all panels with graphs, there is at the top, right a button ‘curve info on/off’. This toggle
works on all panels: the information facility is set on or off on all panels simultaneously.
When this mode is on, and you click on (or near) a curve in a graph, a window appears
with information on the graph (what properties are simulated), on the curve (a
measurement (if so, the measurement file and its date are given), or a ‘single shot
simulation’ (if so, the date and def file are given), or a ‘batch simulation’ (if so, the value
and name of all batch parameters is given). Especially this last feature seems very useful
to let you know what you have simulated there. Also, information on the point is given:
the coordinates, with their units, of the cursor point. In the doping profile plot (C-V panel)
you get the voltage as an extra; in the Nyquist plot (C-f panel) you get the frequency as an
extra.
8. There is a manual scale mutton on all graphs. Sometimes it works on several graphs
simultaneously (e.g. the position scale of all 4 plots on the energy band panel). In some
cases, setting auto scale after you have set a manual scaling does not work properly (in
logarithmic graphs; LabWindows/CVI has some problems with these ). On all graphs
where it is appropriate, there is still a fast linear / logarithmic button: there should never
be more then one mouse click between you and your logarithmic thinking ☺ !
9. And there is a mouse controlled zoom and pan function on all graphs! When you press
CRTL and keep the left mouse button pressed, you are zooming in on the graph as long as
you keep the mouse button pressed. With CTRL and the right mouse click, you are
zooming out as long as the mouse button is pressed. With CTRL + SHIFT and keeping the
left mouse button pressed, you can pan the graph. This great LabWindows/CVI feature
has in fact been present for quite a while, but I was not aware of it ☺. Setting the scale to
auto (with the scale button) after a mouse-zoom action might not always work correctly
with logarithmic graphs. Also the zoom function itself does not work completely with
logarithmic graphs .
10. When you display a measurement in the I-V panel, the solar cell parameters Voc , Jsc, FF
and η are given (when determined), in the same way as this is done with simulations.
11. A bug in the calculation of the spectral response, when a neutral density filter with T ≠ 1
or ND ≠ 0 is present, or when spectrum cut-off filters are present, is corrected. The
influence of these filters on the actual illumination is now clearer in the action panel (even
before a calculation is done): you get the power contained in the spectrum of the lamp or
sun, the power after the cut-off filter (if present) and the power after the neutral density
filter. For clarity: the spectral cut-off filters and the neutral density filter only work on the
bias light (dc light), used in the I-V, C-V, avd C-f calculations, and as the bias light source
in the QE calculation. These filters however do not work on the monochromator light used
in the QE calculation. The QE itself is the response of the cell current to monochromatic
light [that is, QE(λ) is derived from: J(all work point conditions, including bias light +
monochromatic light at λ) – J(all work point conditions, including bias light) ]. The
intensity of the monochromatic light can be set in the cell definition panel / the orange
numerical button. It is also accessible as a batch parameter.
12. The panel with the generation plot is now treated as the EB, AC, IV, CV, Cf and QE
panels: it accessible from all these panels. It now also gives the various components of the
thermal recombination (i.e. capture and emission, per defect). And the ft-plot (occupation
probability of the defects, versus position x) is repeated in this Gen/Rec panel (it is also in

3
 the EB panel). Especially with this ft-plot you will appreciate the curve information
facility!

Marc Burgelman, UGent - ELIS

August - October 2007

4
13. SCAPS Version 2.4, September 2003

Enhancements of version 2.4, compared to version 2.3

External series resistance:

It is now possible to specify an external (discrete) series resistance Rs in the action panel. The
option series resistance: no is equivalent to setting series resistance: yes with Rs = 0
Ω.cm2. When a series resistance is present, the start and end voltages in the action panel are to
be understood as the voltage Vj over the solar cell without series resistance. The applied
voltage is then calculated as Va = Vj + J.Rs.A. All graphs display properties as a function of
the external, applied voltage Va; also, the voltage in the output lists is the applied voltage Va.
In the capacitance C and conductance G calculations, the series is taken into account by
correcting the values of C and G = 1/R calculated with the voltage Vj according to:
⎛ G 1 + ω 2C 2 RR// ⎞
⎜ ⋅ ⎟
⎛ G ⎞ ⎜ 1 + GRs 1 + ω 2C 2 R//2 ⎟ RRs Rs
⎜C ⎟ ← ⎜ ⎟
where R// = =
R + Rs 1 + GRs
⎝ ⎠ C 1
⎜⎜ ⋅ 2 2 2⎟
⎟
⎝ 1 + G Rs 1 + ω C R// ⎠
2 2

No series resistance is taken into account in the calculation of the quantum efficiency QE (λ).

Solar spectra
In previous versions, the data in the spectrum files AM1_5G.spe and AM1_5D.spe summed up
to a power density of 962.6 W/m2 and 767.2 W/m2 respectively. In the efficiency calculation
however, an incident power of 1000 W/m2 was assumed, resulting in a negligible error for the
air mass 1.5 global spectrum, but a significant error in the air mass 1.5 direct spectrum. This
is now corrected. The total light power contained in the spectrum, and the light power
incident on the cell are now listed in the output files (both are equal when there is no external
neutral density filter (ND = 0), when the option spectrum cut off is disabled, and when there
are no filter characteristics defined at the illuminated contact). SCAPS now comes with
several additional spectra: an air mass 1.5 global and direct spectrum scaled to exactly 1000
W/m2 (“one sun”) light power, several monochromatic spectra with one sun light power, and
several “white” spectra (constant photon flux at all wavelengths, or constant incident power at
all λ; wavelength range 400 nm – 800 nm or 300 nm – 1300 nm). All these files have been
better documented, which should it make easier to the user to make her/his own spectrum
files in a valid SCAPS format.

Generation profile g(x) from an external file

There is now a possibility to input the generation profile g(x; spectrum) from a user specified
file, instead of letting it calculate simply by g(x, λ) = α(λ) Φ0(λ) exp(-α(λ).x), and integrating
over all wavelengths of the spectrum; Φ0(λ) is here the incoming photon flux. The advised
way to create a g(x) file is as follows: set up a problem in SCAPS; perform a calculation with
the internally calculated g(x). Save the results in the EB-panel (energy bands panel), or,
alternatively, press show in the action panel and copy/paste the desired data in a user file (e.g.
Origin or Excel). The you have the x-coordinates of the nodes used by SCAPS. Make a
generation file with the g(x) data from your own model, calculation or measurement… For an
example of a valid format, see the examples coming with SCAPS (file extension .gen). Click

5
Generation of eh pairs determined from file in the action panel, and select your file. When
the g(x) data in your file do not use the x values used by SCAPS, the program will have to
interpolate between your x values, and this might cause inaccuracy. To judge the
interpolation, you can visualize the g(x) data in the Solar Cell Definition panel, by clicking
Show generation profile. The g(x) profile you will see then is, however, only updated after a
calculation has been performed with the calculated or user specified g(x) . Also, the actual
g(x) data are saved or shown in the Energy Bands panel. To further assist the user with the
optical management, the α(λ) data used are also shown (as a list), when you click show in the
Cell Parameters panel (that is, where you define the properties of one individual layer of the
solar cell structure). Also this α(λ) shown is only actualized after having performed a
calculation with it.

Varia
The use of “common units” is now consistent in all graphs and output lists: the cm replaces
now the m, and the mA/cm2 replaces the A/m2. It is a pity for the advocates of the SI system,
but a commodity for most users.
There have been small esthetical changes to the panels and output. We regret that the utilities
offered by the National Instruments software (LabWindows/CVI) do not seem to work in our
hands when it comes to making the look of the program independent of the screen settings of
the user computer: what is perfectly aligned on the computer of one of us, is not displayed
perfectly on the other computer, as it should.

Stefaan Degrave and Marc Burgelman, UGent-ELIS, September - December 2003

6
SCAPS Version 2.3, December 2001
Enhancements of version 2.3, compared to version 2.2

• New possibility to introduce optical layers at both back and frontcontact. These layers
only act as optical filters for the incoming light. At both contacts, the filters specified in
the contact panel can be used as transmission or reflection spectra. The most probable use
of this new feature is the introduction of a transmission filter at the window side of the
solar cell and the introduction of a reflection filter at the opposite side. In this case, one
reflection of the incoming light is taken into account at the back contact. No interference
effects are taken into account. When simulating quantum efficiency with a transmission
filter at the window side of the cell, the external QE is calculated.
• When the ‘Set problem’ button is used on the ‘Action panel’, one can either load an
existing file, or create a new definition file or edit the previously loaded definition file. By
choosing one of these possibilities, the ‘Cell panel’ is edited. On this panel, modifications
to all parameters specifying the cell can be done. These modifications are only stored
when the ‘OK’ button is used. When the ‘cancel’ button is used, these modifications are
lost. In both cases, the definition file is stored into the working memory of the program.
No calculations are done by clicking on the ‘OK’ or ‘cancel’ button, is contrast to earlier
versions of the programme. In addition one has to use the ‘Calculate’ button on the
‘Action panel’ any time one wants to perform simulations.
• On all simulation panels, a comment window is added that contains the name of the
definition file, the date it was last saved and the date and time the simulation was done.

Stefaan Degrave, RUG-ELIS, December 2001

7
SCAPS Version 2.2, September 2001
Enhancements of version 2.2, compared to version 2.1

Small esthetical changes to the panels and output.

Major add-ons:
• The optical absorption of the semiconductor layers now can be taken from a user file.
Examples of such user files are distributed with the program: CdS.abs, CdTe.abs, and
ZnO.abs. The authors welcome absorption files from other relevant thin film materials.
• in SCAPS, quantum efficiency was calculated by calculating two times the current
density: once in the “working point” (i.e. at a given bias voltage and bias illumination, set
by the working point), and once in the working point conditions, with a number of
monochromatic photons included. Both values are subtracted and presented as quantum
efficiency. In previous SCAPS versions, the number of monochromatic photons was fixed
at 1015 /cm2.s . In SCAPS 2.2, This number of photons is made available in the user
interface (under “edit problem”, orange button “numerical”). In certain conditions, the
subtraction method suffers from inaccuracy (subtracting two large, nearly equal numbers)
(e.g. at large forward voltage, especially when interface states are present). Increasing the
number of incident photons can alleviate this problem.
• Quantum efficiency can now be plotted as an absolute value |QE| or as the actual value
QE; at a fixed scale between 0% and 100% or at an automatic scale. This is by magenta
buttons at the right of the plot. This can be interesting when the actual value of the
spectral response is negative (occurring e.g. beyond cross-over), or when QE is very low
to see the details, or when the apparent QE exceeds unity (which can occur at forward bias
voltage).

Stefaan Degrave, RUG-ELIS, September 2001

8
 SCAPS Version 2.1, june '99

Enhancements of version 2.1, compared to version 2.0

Version 2.0 was distributed in september 1998, after the World PV Conference in Wien.
Version 2.1, distributed from january 1999 on, is not to be considered as a finished version: it
is under continuous evolution. The user manual of SCAPS 2.0 is still valid. Enhancements of
version 2.1 up to now are shortly listed here (not exhaustive):

• Convergence is improved: the termination criteria for the electrostatic potential and the two fermi-
levels are made available to the user (edit problem → numerical), and are set to 10-3 kT/q or kT by
default (version 2.0: 10-8 kT/q or kT, not under control of the user: this was much too severe).
• When divergence occurs, the points calculated so far are not lost, but are shown.
• The status of the program is indicated in the lower left corner: “OK”, “calculating”, “paused”.
• During calculation, a status line is shown at the bottom of the screen, giving information of what is
being calculated, and how far calculation has proceeded (e.g. “calculating I-V from 0 V to 0.6 V;
V = 0.45 Volt”)
• The possibility of illumination from both the p (left) and the n side (right) is built in now (both on
the action panel and in the problem definition editing panel). The illuminated side is saved in the
.def file. When a calculation is done, illumination information is given in the lower right corner:
“dark”, “←light” or “light→” (i.e. illumination from left or from right).
• The interpretation of the graphs is facilitated by the addition of the meaning of a graph in words in
the same color as the graph (e.g. under the recombination graph, the words “SHR
semiconductors”, “front contact”, “back contact” and “interfaces” in the colors of the
corresponding graphs).
• When an I-V curve is calculated, the cell parameters of the curve which was last calculated are
shown (i.e. Voc , Jsc , FF, η). If extrapolation was needed for the calculation, a warning LED is
lighting under the value. If extrapolation seems unrealistic to the program, the value is not shown
(e.g. trying to find Voc when the calculation is only done from 0 V to 0.1 V: extrapolation would
result in an unrealistically high value).
• In the problem definition panel, a comment window is added. The user can freely write comment
in it, and this comment is saved (“e.g. problem file used for the simulation of fig. 3 in My Big
Article; february 99; A.B.”): this helps the user or his colleagues to find their way in old .def files
(at least, this is hoped for).
• In the C-f panel, when the cursor in the Nyquist plot (i.e. the Im(Z) versus Re(Z) plot) is clicked,
also the frequency value of the point indicated by the cursor is shown.
• Many graphs can be shown on a logarithmic or a linear scale, by simply clicking the log/lin switch
nearby the axis under consideration (magenta colored switches)
• The user can choose to see the G(ω) plot or the G(ω)/ω plot in the C-f panel, by clicking the
magenta colored switch.
• In all panels, a “Show” button is added under the “Save” button. The Show function is identical
with the Save function, but the data are shown on screen, instead of saved in a text file. Data
shown on screen, or parts of it, can be copied (ctrl-C) and pasted (ctrl-V) into other Windows
based programs, e.g. an Excel data sheet.
• The term “calculate new” is replaced by “set problem”, and “calculate next” by “calculate” (action
panel)

Marc Burgelman, january 1999

9
SCAPS-1D User Manual

Version 2.0, August '98

1. GENERAL INFORMATION:

Copyright:

SCAPS-1D is a development of the Department of Electronics and Information Systems (ELIS)

of the University of Gent, Belgium. For any further information, contact:
Prof. Marc Burgelman
Department of Electronics and Information Systems (ELIS)
University of Gent
Pietersnieuwstraat 41
B-9000 Gent, Belgium
Phone: +32 9 264 3381
Fax: +32 9 264 3594
e-mail: burgelman@elis.rug.ac.be

SCAPS-1D is written as an application program of the “LabWindows/CVI” software

(Copyright National Instruments Corporation. All rights reserved).
You are totally free to use SCAPS. However, you must not further distribute or sell it. Please
direct all requests for further distribution or installation to the address above.

Purpose of the program:

SCAPS-1D (a Solar Cell Capacitance Simulator in one Dimension) has been developed to
realistically simulate the electrical characteristics (dc and ac) of thin film heterojunction solar
cells. It has been tested extensively for thin film CdTe and Cu(In,Ga)Se2 solar cells.

The program SCAPS-1D finds dc solutions for structures consisting of an arbitrary number of
semiconductor layers, with arbitrary doping profiles (as a function of position), with arbitrary
energetic distributions of deep donor and/or acceptor levels (single level as well as uniform,
Gauβ or tail distribution) in the semiconductor bulk and at the heterojunction interfaces (surface
states), for arbitrary light spectra (e.g. the superposition of monochromatic light and AM1.5G
bias light). Admittance spectra (within an arbitrary frequency band) can be calculated for every
dc working point (bias voltage, illumination and temperature), provided that the dc algorithm
converges.

System requirements:

SCAPS-1D runs on a personal computer (PC) under Microsoft Windows. The minimum system
requirements are: PC at least 25 MHz 80386 (33 MHz 80486 recommended minimum);
Operating systems Windows 3.1, Windows 3.11 or Windows 95 or higher; VGA resolution
video adapter or higher; Math coprocessor; 4 MB of memory; free hard disk space equal to 2
MB, plus the size of the installed files needed (about 2.5 MB). A mouse is essential for
convenient navigation through the program. The department ELIS of the University of Gent

10
provides to university users an installation diskette to install SCAPS-1D.

Installation:
The “installation diskette” contains four files:
info.doc this text in Word 6.0 format
info.txt this text in text format (ASCII) (on request)
setup.exe the installation program
scaps21.001 a compressed file
scaps21.002 a compressed file
Normally, these four (five) files are e-mailed. You must place them in one directory. When
tough difficulties arise, the files can be sent on a diskette. Double-click on setup.exe. The
run-time engine of LabWindows/CVI (© National Instruments) will be installed. Also, a
directory c:\scaps (or another directory, as you specify) will be created, containing the
following files:
scaps21.exe the program file
Gui_v21.uir the graphical user interface file
am1_5d.spe a text file with the AM1.5 direct spectrum
am1_5g.spe a text file with the AM1.5 global spectrum
example cigs.def a text file with the problem definition for an example CIGS cell
example cdte.def a text file with the problem definition for an example CdTe cell
Also, an icon for SCAPS will be created. You can start SCAPS by double clicking this icon, or
by double clicking scaps21.exe in the Explorer (File Manager).
The disk space occupied by all created files is less than 4 MB.

References:

SCAPS-1D was introduced in the open literature, in reference

[1] M. Burgelman, P. Nollet and S. Degrave, “Modelling polycrystalline semiconductor solar

cells”, Thin Solid Films, 361-362, 527-532, 2000
[2] A. Niemegeers and M. Burgelman, “Numerical modelling of ac-characteristics of CdTe and
CIS solar cells”, Proc. 25nd IEEE Photovoltaic Specialists Conference (Washington D.C.,
april 1996), pp. 901-904, IEEE, New-York, 1996

The algorithms used in SCAPS were further explained in:

[3] A. Niemegeers, S. Gillis and M. Burgelman, “A user program for realistic simulation of
polycrystalline heterojunction solar cells: SCAPS-1D”, Proceedings of the 2nd World
Conference on Photovoltaic Energy Conversion (Wien, Österreich, july 1998), pp. 672-675,
JRC, European Commission, 1998
[4] M. Burgelman, P. Nollet and S. Degrave, “Modelling polycrystalline semiconductor solar
cells”, Thin Solid Films, 361-362, 527-532 (2000). Also presented (invited) at the European
Materials Research Society, E-MRS, Strasbourg, France, 1-4 june 1999

SCAPS-simulation results were published in:

[5] [3] A. Niemegeers and M. Burgelman, “Effects of the Au/CdTe back contact on IV-

11
 and CV-characteristics of Au/CdTe/CdS/TCO solar cells”, Journal of Applied Physics, march
15, 1997 issue 6, pp. 2881-2886.
[6] A. Niemegeers, Marc Burgelman, R. Herberholz, U. Rau, D. Hariskos and H.-W. Schock,
“A model for the electronic transport in Cu(In,Ga)Se2”, accepted for publication in Progress
in Photovoltaics ( june 1998).
[7] Niemegeers, S. Gillis and M. Burgelman, “Interpretation of capacitance spectra in the
special case of novel thin film CdTe/CdS and CIGS/CdS solar cell device structures”,
presented at the 2nd World Conference on Solar Energy Conversion, Wien, july 1998.
[8] M. Burgelman, P. Nollet and S. Degrave, “Electronic behaviour of thin film CdTe solar
cells”, Applied Physics, A69, 149-153, 1999

When you publish results obtained with SCAPS, please refer to one of the above articles.

2. INTRODUCTION TO SCAPS-1D

The program consists of several panels (or windows, or pages, … in other jargon). The main
panel is the “action panel”. It allows to set the problem, to set the action list of required
calculations, to execute the calculations, to navigate to other panels, to save and plot results,
and to quit the program. When starting SCAPS, one enters the action panel.
→ The QUIT button on the action panel is the only correct way to leave the program.
Other panels are the “EB, IV, AC, CV, Cf and QE panels”. They are used to display results of
energy band calculations, and of IV, AC, CV, Cf and QE calculations if these were carried out.
Finally, there are minor panels which only are in use when the user is asked to input data, to
select files or to confirm choices. Navigation between the panels is by clicking the
corresponding buttons with the mouse.

We now first discuss how to set up a problem. Then we'll discuss how calculations are
performed, and how the results are displayed.

3. MAKING, MODIFYING AND USING A “PROBLEM DEFINITION

FILE ”

When the program is started the first screen is the action panel. The complete problem set-up
is contained in a “problem definition file”; this is a text file with extension .def.
By pressing the button “calculate new” in the action panel, one can load an existing
definition file (“from file”), or create a new definition file (“create new”). When there is
already a definition file in the working memory, one can modify it (“edit”). On starting up,
“create new” and “edit” are equivalent. We will now discuss in detail how to set up a
problem.

3.1 Building a new definition file, or editing an existing one

In the following text, left means the side of the back contact, right the side of the front
contact. The light is always incident on the front contact, thus from the right.

After clicking the “create new” or the “edit” options, the “solar cell definition panel” is
active. It allows the user to create a structure and to define all its characteristics. He can

12
define structures consisting of up to 9 layers. The first layer is the back contact, the last one is
the front contact. The user can specify the properties of the intermediate semiconductor layers
(max. 7). For each layer, up to three different SRH recombination centers (discrete or
distributed in energy) can be defined, and for each interface up to three different interface
recombination centers (discrete or distributed in energy).

A. BACKCONTACT

If one presses the “back contact” button, the properties of the back contact can be set: a
velocity parameter for electron and holes and information about the metal work function. If
the “OK” button is pressed the data are accepted, and the solar cell definition panel
reappears. The meaning of the parameters is as follows:

1) For the majority carriers, the model is that of thermionic emission. The meaning of the
velocity parameter is then “thermal velocity” vth. The majority carrier current jc,maj at
back contact is then given by:
⎛ qΦ B ⎞ ⎛ ⎛ qV ⎞ ⎞
jc, maj = qv th N C exp⎜ − ⎟ ⎜ exp⎜ c ⎟ − 1⎟ or jc, maj = qv th NV "
⎝ kT ⎠ ⎝ ⎝ kT ⎠ ⎠

depending on the type of the majority carriers (electrons or holes respectively); Vc is the
voltage step over the contact. The barrier height is calculated from the metal work
function Φm, the electron affinity χ and the bandgal Eg :

ΦB = Φm − χ ( n − type; electron barrier height )

Φ B = χ + Eg − Φ m ( p − type; hole barrier height )

2) For the minority carriers, the model is the classical description of minority carrier surface
recombination. The meaning of the velocity parameter is then “minority carrier
recombination speed”. The minority carrier current jc,min at back contact is thus given by

jc,min = q S c ( c − c0 )

where c denotes electron or hole concentration, and c0 denotes the equilibrium

concentration of the electrons or holes at the back contact.
3) The metal work function φm (for majority carriers) can be input by the user. However, the
user can also choose the option “flat bands”. In this case, SCAPS calculates for every
temperature the metal work function Φm in such a way that flatband conditions prevail:
⎛ NC ⎞
Φ m = χ + ln⎜ ⎟ ( n − type)
⎝ Nd − Na ⎠
⎛ NC ⎞
Φ m = χ + E g − ln⎜ ⎟ ( p − type)
⎝ Na − Nd ⎠
Here, Nd stands for the shallow donor concentration and Na for the shallow acceptor
concentration, specified in the layer definition. Though this option does not correspond to
a physical situation, it allows the user to conveniently define a perfect ohmic contact.
4) All parameters relevant to the back contact are assumed to be independent of temperature.

13
B. SEMICONDUCTOR LAYERS

If the “add layer” button is pressed in the solar cell definition panel, the characteristics of the
layer can be defined in a new screen. The user can now enter the name and the properties of
the layer.
When the right mouse button is clicked, with the cursor positioned at one of the
semiconductor layers, a menu appears which allows the user to duplicate and to remove an
already existing layer. When the option “duplicate” is chosen, the new layer is inserted below
the already existing layer. Also an new interface is created between the newly created layer
and the preceding one. When the option “remove” is chosen, the selected layer and the
interface preceding it are removed from the structure.

Remarks:

1) The densities of states in conduction band and valence band, NC and NV, depend on
temperature as
15
.
⎛T⎞
N C (T ) = N C 0 ⎜ ⎟
⎝ T0 ⎠
15
.
⎛T⎞
NV (T ) = NV 0 ⎜ ⎟
⎝ T0 ⎠
where the reference temperature T0 equals 300 K. The reference values of NC0 and NV0 at
room temperature are the values to be specified. The thermal velocity vth(T) depends on
temperature as
0.5
⎛T⎞
vth (T ) = vth0 ⎜ ⎟
⎝ T0 ⎠
Again, the value at room temperature vth0 is the specified value.

2) The diffusion coefficient is set to kT0 times the mobility. Both diffusion coefficient and
mobility are independent on temperature.

3) The optical absorption constant α(λ) of a layer with bandgap Eg is given by

⎛ B⎞
α (λ ) = ⎜ A + ⎟ hν − E g
⎝ hν ⎠

where hν is the photon energy, and A en B constants to be defined.

If the user presses the “defect 1” button, he can define the properties of the Shockley-Read-
Hall (SRH) recombination in the bulk of the semiconductor layer. One can remove the defect
by clicking with the right mouse button on the defect te be deleted and choose the “remove”
button. When defect 1 is specified, the user may specify a “defect 2”, and afterwards a “defect

14
3”.

General remark on defects and recombination:

In a diode, current is converted from hole current at the p-contact to electron current at the n-
contact. This means that somewhere in the diode recombination must take place, even in the
most ideal device. So the user must specify recombination somewhere, at least at one place
(in a layer or at an interface). If he does not do so, a convergence failure will result in non-
equilibrium conditions (non zero voltage, and/or illumination).

Remarks:

1) The “neutral” defect is an idealization of a defect which contributes to the Shockley-

Read-Hall recombination but does not contribute to the space charge. In the case of a
“neutral” defect, only the product of σ and Nt affects the dc and ac solutions (through the
carrier lifetimes; τn e.g. is given by 1/(σn.Nt.vth) ): in this case, the defect centers
contribute to recombination but not to the space charge. This type of defect can be chosen
if one wants to specify electron and hole lifetimes without specifying a defect density
which affects the space charge.
2) The lifetime is only displayed when there is one single defect level, uniform in space.
3) All parameters in the ‘defect’ menu are assumed to be independent on temperature.

Name Description Value Default Units

Type type of the defect donor neutral

acceptor
neutral
Sigma_n capture cross section numeric 1.0E-15 cm2
for electrons
Sigma_p capture cross section numeric 1.0E-15 cm2
for holes
Energydistribution type of the defect single single
distribution uniform
cb tail
vb tail
gauss
Et energy of the trap numeric 0.5 eV
level with respect to
the valence band
Ekar Characteristic energy numeric 0 eV

Profile spatial variation of constant constant

defect density linear
exponential
Lkar characteristic length numeric 0.1 μm

Nleft max. defect numeric 1.0E+14 /cm3

concentration to the or
left of the layer /cm3.eV

15
 Nright max. defect numeric 1.0E+14 /cm3
concentration to the or
right of the layer /cm3.eV

Depending on the nature of the energy distribution, some parameters have different meanings:

Energy single Uniform Valenceband Conductionband Gauss

Distribution Tail tail
Et energy level mean energy of bottom of the tail top of the tail mean energy of
the distribution (see energy axis) (see energy axis) the distribution
Ekar no meaning total width of the characteristic characteristic width of the
distribution energy (see energy (see distribution (see
energy axis) energy axis) energy axis)
Nleft, density in constant density density of the density of the density of the
Nright /cm3 of the distribution distribution at distribution at distribution at
in /cm3eV E=Et in /cm3eV E=Et in /cm3eV E=Et in /cm3eV

For every distribution of defects an energy axis is calculated with 41 equidistant energies:

Distribution type Energy ranges N(E) =

From To
Uniform Et – Ekar/2 Et + Ekar/2 N

vb tail Et Et + 7.Ekar N.exp( - (E-Et)/Ekar )

cb tail Et Et - 7.Ekar N.exp( - (Et-E)/Ekar )

Gauss Et - 3.Ekar Et + 3.Ekar N.exp( - ((E-Et)/Ekar )2 )

The concentration N in the table above is calculated from Nleft, Nright and Lkar:

Spatial profile N(x) = Remark

16
 Constant Nleft Nright and Lkar are ignored

Linear Nleft + (Nright-Nleft).(x-x0)/d Lkar is ignored

Exponential A.exp((x-x0)/Lkar) + B A and B are calculated such that

N(x0) = Nleft and
N(x0+d) = Nright

Here x0 is the spatial coordinate of the left side of the layer, and x0+d the one of the right
side. If, in the case of an exponential profile, no value for Lkar is specified, a defect
concentration equal to 0 will be assumed.
When all parameters are set, one can add the layer with all its properties to the already
existing structure by pressing the button “accept”.

C. INTERFACE BETWEEN DIFFERENT LAYERS

Between two layers one can define the characteristics of the interface by clicking on the
rectangle right from two layers. Again 3 possible defects can be set by clicking on the
“defect1,2,3” button.

When the structure is completely defined, the definition file can be saved by clicking on the
“save” button. When the “OK” button is clicked the energy band panel is showed with
different graphics.

Name Description Value Default Units

Type type of the defect donor neutral

acceptor
neutral
Sigma_n capture cross section numeric 1.0E-15 cm2
for electrons of the
conduction band
Sigma_p capture cross section numeric 1.0E-15 cm2
for holes of the valence
band
Energydistribution type of the defect single single
distribution uniform
cb tail
vb tail
gauss
Et energy of the trap level numeric 0.5 eV
with respect to the left
valence band
Ekar characteristic energy numeric 0 eV
(remark)
N max. defect numeric 0 /cm2
concentration or
/cm2.eV

17
Remarks:

1) The energy parameters have the same meaning as in B.

2) The interface recombination velocities Sn = Niσnvth,n and Sp = Niσpvth,p are only shown
when there is a single interface state energy level. When the thermal velocities of the two
layers forming the interface have different values, there are different values for the
interface recombination speeds of carriers coming from the left or from the right.

3) If not stated explicitly otherwise, all parameters of the semiconductor layers and the
interfaces are assumed to be independent on temperature.

4) The model for interface transport is thermionic emission. The thermal velocity of the
interface transport equals the smallest thermal velocity of the two neighboring layers.

5) Reflection of photons at interfaces is not taken into account.

D. FRONT CONTACT

The front contact is specified similarly to the back contact: see A. Backcontact.

E. NUMERICAL PARAMETERS

Some numerical parameters are also made available in the user interface by the “numeric”
button in the solar cell definition panel: the maximum number of iterations, and clamping
values for the electrostatic potential and for both quasi Fermi potentials. During numerical
integration, these potentials are not allowed to vary from step to step by more than the
corresponding clamping value. Setting of these parameters influences the convergence
behavior in extreme cases, where convergence is difficult to obtain.

3.2 Loading and changing an existing definition file

For loading an already existing definition file one clicks on the “from file” button in the
calculate new menu. In this screen the user can browse and select his definition file. If he
presses the “OK” button he gets immediately the energy band panel with the equilibrium
graphs. To go back to the action panel one clicks on the “action” button.
If one wants to create a new definition file based on an old one, one has to click on the
“create new” button in the calculate new menu as for a new definition file. In this screen the
“load” button allows one to load an existing definition file an afterwards one can change it
like a new definition file.

3.3 Making changes in an already loaded definition file

If the user clicks on the “edit” button in the calculate new menu he can make changes in the

18
properties of his structure in the same way as he has defined the properties by making the
definition file.

4. SPECTRUMFILES

C:\SCAPS\AM1_5G.spe : AM1.5 global spectrum

C:\SCAPS\AM1_5D.spe : AM1.5 direct spectrum

Other spectrum files can be user defined. The spectrum file is an ASCI-file. The first line
contains the number of wavelengths that appear in the spectrum file. This line is immediately
followed by two columns. The first column specifies the wavelength λ (nm). Suppose the
spectral density (in Watt/m2.nm) is given by P: the incident power in the wavelength interval
[λ-dλ/2 , λ+dλ/2] is then given by Pdλ (in Watt/m2). The second column in the spectrum file
contains this incident power Pdλ (in Watt/m2) in a wavelength interval dλ around λ. As
such, the total incident power is given by the sum of the values in the second column. The
wavelength interval dλ considered around a wavelength λi is: dλ = (λi+1 - λi-1)/2. For the first
wavelength λ1 it is dλ = λ2 - λ1, and for the last wavelength λN it is dλ = λN - λN-1. The
maximum number of wavelengths appearing in the spectrum file equals 100. The number of
incident photons, per m2 and per s, of wavelength between λ-dλ/2 and λ+dλ/2 is then set
equal to 5.035 1015 × λ × Pdλ , λ in nm, Pdλ in Watt/m2).The numerical factor equals 10-9/
hc where Planck's constant h and the light velocity c are expressed in SI units.
For convenience, the spectrum file may be preceded by comment lines, which begin with the
character “>” or “/”. These lines are not read by SCAPS. They can be used to identify the
source or the purpose of a given spectrum file.

5. CARRYING OUT NUMERICAL SIMULATIONS

The problem is set in the problem definition file. The requested calculations are set in the
“action list” in the action panel (see below). After creating, editing or loading a problem
definition file, one can go back to the action panel. When “cancel” is pressed, no calculation
is performed, but the edited cell structure stays in the working memory. When “OK” is
pressed, the calculation of the defined problem is immediately carried out, and the “energy
band panel” is displayed. One can return to the action panel by pressing the “action” button,
to edit the action list of required calculations, to modify the problem definition, or to quit the
program.

The action panel allows to specify a “working point” defined by the temperature (K), a DC-
bias voltage (V), the illumination conditions (dark, full spectrum, neutral density and/or band
pass filters) and an ac-frequency.

The option “spectrum cut-off” simulates the use of a long pass filter and a long stop filter in
the dc-illumination condition. This option is only effective when “spectrum cut-off” is set to
yes, and when illumination is on (i.e. “dark” = no). The spectrum cut-off affects all calculated
curves (though many cell types are not very sensitive to it). However the quantum efficiency
does not fall to zero in the regions where the spectrum is cut-off. The spectral cut-off affects
the bias illumination condition (dc-bias light source), whereas the quantum efficiency is

19
calculated as the response to a small ac light source, superimposed to the bias light
(simulating the measuring set up with chopped light from a monochromator and bias light).

The following characteristics can be calculated:

1) Current vs voltage: from a specified voltage V1 to V2 (V) in n steps, for the illumination
conditions and temperature specified in the working point (wp).

2) Capacitance vs voltage: from a specified voltage V1 to V2 (V) in n steps, for the

illumination conditions, temperature and ac-frequency, specified in the working point (wp).

3) Capacitance vs. frequency: from a specified frequency f1 to f2 (kHz) in n steps (equidistant

on a logarithmic scale) for the illumination conditions, dc-bias voltage and temperature
specified in the working point.

4) Quantum efficiency vs. wavelength: from a specified wavelength λ1 to λ2 (nm) in n steps

for the (bias) illumination conditions, dc-bias voltage and temperature of the working point.

Every new calculation starts with the calculation of the equilibrium band diagram (V=0,
dark); if the working point specifies illumination, next the short circuit band diagram is
calculated. Every next calculation starts from the previously calculated solution. The
calculation order is: first the working point, then the IV, CV, Cf and QE curves, if requested.
When the “pause” button in the action panel is enabled, the program halts after every
calculation, and resumes when the “continue” button is pressed in the action panel or in the
energy bands panel. When “pause” is disabled, all requested calculations are performed one
after the other.

Convergence problems may occur when the voltage of the working point differs too much
from zero (= the voltage of the previously equilibrium and possibly Jsc calculation). To
overcome this problem, SCAPS allows to specify a number of steps to be taken to evolve
from zero voltage to the working point voltage (this number is specified in the field next to
the dc-bias field in the working point specification). Problems may also occur if the starting
voltage of the IV or CV calculation differs too much from the working point voltage. In that
case one can set the dc-bias voltage of the working point equal to the starting point of the IV
or CV calculation, so that the IV or CV calculations can start with the correct solution.

Once a calculation is finished, it is possible to view the corresponding IV, CV, Cf and QE
panels. The energy band panel displays the most recently calculated band diagram. If a Cf
calculation is made, SCAPS automatically displays the “AC panel” (see below).

By pressing the “calculate next” button in the action panel, a new calculation with the same
definition file, but with a different action list can be made. If the working point is the same as
for the previous calculation, the equilibrium, short circuit and dc bias conditions will not be
recalculated.

6. THE “ACTION PANEL”

The following items are available on the action panel:

20
• The definition of the working point (see above)
• The definition of the action list (see above)
• Setting the calculation mode: “pause” between calculations or not (see above)
• Setting the problem definition: button “calculate new” (see above)
• Carrying out calculations according to a new action list: button “calculate next” (see above)
• “continue” to calculate next item from action list, when “pause” is enabled
• “stop” to halt a calculation being carried out
• “clear previous” to clear previously calculated graphs
• “graphs” to activate the energy bands panel, from which in turn one can navigate to all other
graphical panels (IV, CV, Cf and QE)
• “save all” to save the results of all calculations yet preformed in a text file; the data in the
text file are separated by tabs, and can be cut and pasted in a spread sheet or graphical
program
• “quit” to quit from SCAPS-1D. This is the only correct way to leave SCAPS !

7. THE “ENERGY BAND PANEL”

The following buttons are available:

• “continue”: to resume a calculation halted by the pause option
• “save”: to save the results of the energy bands calculation in a text file
• “plot”: to plot the contents of the panel to the Windows default printer. Remark: this is a
low quality “print screen”, only for viewing and discussion purposes. For presentation
purposes, it is better to save the calculation in a text file, and to use this file to make
presentation graphs with one's preferred graphical program.
• “action”: to return to the action panel
and…
the buttons “IV→”, “QE→”, “AC→”, “CV→” and “Cf→”, to navigate to the graphical
panels indicated.

Also, the results of the last energy band calculation are shown in four graphs: this is thus in the
illumination and voltage conditions of the last calculation performed (thus the working point or
the last point of the IV, QE,… curves).

There is a cursor utility available in the EB-panel, and in the other graphical panels as well. If you
click with the mouse on a point of the graph of your choice you can see the x and y value of this
point in the two little windows in the middle of the screen. The units are the same as the user
units of the axes of the relevant graph.

The first graph (upper left) shows the (dc) band diagram: black for conduction and valence bands,
blue for the electron fermi level, red for hole fermil evel or only red if the fermi levels coincide.
The defect levels, if present, have also their color: green for a neutral defect, blue for an acceptor
type defect and red for a donor type defect. (in fact, the characteristic energy Ekar of the defects
is shown). It is easy to discern between the defect levels and the fermi levels, since the defect
levels exactly follow the band edges, whereas the fermi levels do not.

The second graph (upper right) shows the carrier densities as a function of distance: blue for
electrons, red for holes, and black for the total charge.

21
The third graph (lower left) shows the current density as a function of distance: blue for electrons,
red for holes, and green for the total current

The fourth graph (lower right) shows the occupation probability of the defects as function of
distance: green for neutral defects, blue for acceptor type, red for donor type defects.

8. THE “IV- PANEL”

The buttons “continue”, “save”, “plot” and “action”, and the navigation buttons “EB→”,
“QE→”, “AC→”, “CV→” and “Cf→” have the same meaning as in the energy bands panel.

The first graph (left) shows the current density as function of the applied voltage. The first
calculation (that is, after “calculate new”, the problem set up and “OK”), The IV-curve (when
calculated) is plotted in red. Every next calculation (thus after a “calculate next” or after
another “calculate new”), a new IV-curve is added (when calculated), in a different color. The
color order is: red, blue, green, cyan, magenta, yellow, … (14 colors). When “clear previous”
is pressed on the action panel, all IV curves are erased, and the color is starting again from
red.

The second graph (right) shows the recombination current as function of the applied voltage:
recombination of minority carriers at the back contact (blue), recombination of minority
carriers at the front contact (red), recombination at the interface(s) (black), and SRH
recombination in the bulk and depletion layers of all semiconductor layers (green). The total
recombination is the sum of these for terms, and it equals the total current. As this sum often
coincides on the plot with the curve of the dominant recombination mechanism, it is not
shown. Plots of previous performed calculations are erased from screen, to avoid too chaotic
graphs.

9. THE “AC-PANEL”

The buttons “continue”, “save”, “plot” and “action”, and the navigation buttons “EB→”,
“QE→”, “IV→”, “CV→” and “Cf→” have the same meaning as in the energy band panel.

In this panel, the small signal variations (ac-variations) of some properties is shown: the
currents, the negative of the electrostatic potential ψ, the electron fermi level Fn and the hole
fermi level Fp. Remark that these are complex ac-quantities, not the dc-values of these
properties. Both the amplitude and the phase of these quantities are shown.

The first graph (upper left) shows the amplitude of the ac-current density as function of the
distance: blue for electrons, red for holes, and green for total ac carrier current (the total ac-
current is this current plus the displacement current, not shown).

The second graph (upper right) shows the amplitude of -ψ, Fn, Fp as a function of distance:
blue for Fn, red for Fp and green for -ψ.

The third graph (lower left) shows the phase of the current densities as a function of distance
(same colors as the first graph).

22
The fourth graph (lower right) shows the phase of -ψ, Fn, Fp as a function of distance: (same
colors as the second graph).

10. THE “CV-PANEL”

The buttons “continue”, “save”, “plot” and “action”, and the navigation buttons “EB→”,
“QE→”, “IV→”, “AC→” and “Cf→” have the same meaning as in the energy band panel.

The first graph (upper left) shows the capacitance as a function of the applied voltage.

The second graph (upper right) shows the conductance as a function of the applied voltage.

The third graph (lower left) shows the Mott-Schottky plot (1/C2 versus dc-voltage V)

The fourth graph (lower right) shows the doping profile as a function of εsε0/C.

The color convention for these graphs is the same as for the IV-curves. Thus, an IV-curve and
a CV-curve plotted in the same color were calculated with the same problem set up and the
same working point.

11. THE “Cf-PANEL”

The buttons “continue”, “save”, “plot” and “action”, and the navigation buttons “EB→”,
“QE→”, “IV→”, “AC→” and “CV→” have the same meaning as in the energy band panel.

The complex impedance Z(ω) or (admittance Y(ω)) is presented as the equivalent parallel
circuit of a capacitance C and a conductance G.

The first graph (upper left) shows the capacitance as a function of frequency: C(f).

The second graph (upper right) shows the conductance as a function of frequency: G(f).

The third graph (lower left) shows the Nyquist plot: the imaginary part of Z(f) versus the real
part. Each point in this plot corresponds to one frequency.

The color code is the same as for the IV and CV-curves.

12. THE “QE-PANEL”

The buttons “continue”, “save”, “plot” and “action”, and the navigation buttons “EB→”
and “IV→” have the same meaning as in the energy band panel.

The only graph shows the quantum efficiency as a function of the wavelength.
The color code is the same as for the IV , Cf and CV-curves.

23
The authors welcome remarks, identification of bugs or errors, comments....
Enjoy your work with SCAPS-1D !

Marc Burgelman, Alex Niemegeers, Sofie Gillis

Gent, 17th of august 1998

24