This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE JOURNAL OF PHOTOVOLTAICS

Rear Surface Optimization of CZTS Solar Cells

by Use of a Passivation Layer With
Nanosized Point Openings
Bart Vermang, Yi Ren, Olivier Donzel-Gargand, Christopher Frisk, Jonathan Joel, Pedro Salome,
Jerome Borme, Sascha Sadewasser, Charlotte Platzer-Bjorkman, and Marika Edoff

AbstractPreviously, an innovative way to reduce rear interface

recombination in Cu(In,Ga)(S,Se)2 (CIGSSe) solar cells has been
successfully developed. In this work, this concept is established in
Cu2 (Zn,Sn)(S,Se)4 (CZTSSe) cells to demonstrate its potential for
other thin-film technologies. Therefore, ultrathin CZTS cells with
an Al2 O3 rear surface passivation layer having nanosized point
openings are fabricated. The results indicate that introducing such
a passivation layer can have a positive impact on open-circuit voltage (VO C ; +17%rel.), short-circuit current (JS C ; +5%rel.), and
fill factor (FF; +9%rel.), compared with corresponding unpassivated cells. Hence, a promising efficiency improvement of 32%rel.
is obtained for the rear passivated cells.
Index TermsAluminum oxide, Cu(In, Ga)(S, Se)2 , Cu2 (Zn,
Sn)(S, Se)4 , nanosized point contacts, solar cells, surface passivation layer, thin-film.

I. INTRODUCTION
ZTSSe and CIGSSe semiconductor materials exhibit comparable optical and electronic properties and are used in
photovoltaic (PV) thin-film (TF) solar cells with similar structure but fairly different efficiency levels. Indeed, the CZTSSe
kesterite and CIGSSe chalcopyrite structures are known to be
associated with each other, where the advantage of CZTSSe
lays in its composition of nontoxic and earth-abundant materials. Both of these p-type semiconductor materials are used
as absorber layers in TF solar cells, combined with an n-type
buffer layer and front and rear contact layers. At present, top

Manuscript received June 4, 2015; revised September 24, 2015; accepted

October 22, 2015. This work was supported by the Swedish Science Foundation
(VR), the Swedish Energy Agency, and the European Commission via FP7
Marie Curie IEF 2011 under Action No. 300998. The work of B. Vermang was
supported by the Flemish Research Foundation FWO (mandate 12O4215N).
The work of P. Salome was supported by the European Commission through
the FP7 Marie Curie IEF 2012 under Action No. 327367.

B. Vermang is with the Angstr

om Solar Center, University of Uppsala, Uppsala 75121, Sweden, with the Department of Electrical Engineering, University of Leuven, Leuven 3001, Belgium,and also with Thin Film Photovoltaics,
IMEC, Leuven 3001, Belgium (e-mail: bart.vermang@imec.be).
Y. Ren, O. Donzel-Gargand, C. Frisk, J. Joel, C. Platzer-Bjorkman, and

M. Edoff are with the Angstr

om Solar Center, University of Uppsala, Uppsala 75121, Sweden (e-mail: yi.ren@angstrom.uu.se; olivier.donzel-gargand@
angstrom.uu.se; christopher.frisk@angstrom.uu.se; jonathan.joel@angstrom.
uu.se; charlotte.platzer@angstrom.uu.se; marika.edoff@angstrom.uu.se).
P. Salome, J. Borme, and S. Sadewasser are with the Laboratory for Nanostructured Solar Cells, International Iberian Nanotechnology Laboratory, Braga
4715-330, Portugal (e-mail: Pedro.Salome@inl.int; jerome.borme@inl.int;
sascha.sadewasser@inl.int).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/JPHOTOV.2015.2496864

conversion efficiencies for small-area CIGSSe solar cells are

typically above 20% (e.g., NREL, ZSW, HZB, EMPA, Solibro,
Solar Frontier, etc.), whereas CZTSSe solar cells are limited to
efficiencies slightly above 10% (e.g., IBM, Imec, IREC, Solar
Frontier, etc.) [1].
The CIGSSe material quality is already high, which endorsed
the introduction of innovative front and rear surface passivation
concepts in recent years. In the past, CIGSSe solar cell research
has mainly been focused on improving the absorber layer (e.g.,
by Ga grading, Na doping, etc.); as a result, efficiencies above
20% were realized. Such high-efficiency CIGSSe solar cells
essentially have long charge carrier diffusion lengths, which
means that the recombination of these charge carriers atand
thus passivation offront and rear CIGSSe surfaces became a
new topic of attention. On the one hand, the front interface of
CIGSSe absorber layers has been improved by the introduction
of 1) a front surface passivation layer or 2) a postdeposition
treatment with potassium fluoride (KF). HZB used the sprayILGAR (ion layer gas reaction) deposition to fabricate ZnS nanodots embedded in an In2 S3 buffer layer, where the nanodots
appear to reduce recombination at the front absorber interface
[2]. However, EMPA also developed a technique that could
generate a front surface passivation layer with nanosized point
openings, namely the self-assembled alkali-templates approach
where a well-controlled grid of KF islands (< 30 nm) is grown
[3]. Moreover, it has been shown that the K treatment itself already results in enhanced passivation of grain boundaries and
donor-like defects (but enables thinning of the CdS buffer layer,
increased junction depth, and increased bandgap as well) [4]. On
the other hand, recombination at the standard Mo/CIGSSe rear
interface of CIGSSe solar cells has been reduced in a similar
way, i.e., by the implementation of an Al2 O3 rear surface passivation layer with nanosized local point openings for contacting,
as is shown in Fig. 1 [5], [6]. Initial studies indicate that this
passivation layer reduces interface recombination by chemical
(corresponding to a reduction in interface trap density) and field
effect passivation (resulting from a fixed charge density in the
passivation layer that reduces the surface minority or majority
charge carrier concentration) [7], [8].
Further improvements in CZTSSe material quality are
certainly desirable; nonetheless, early front surface passivation
attempts have already been made. CZTSSe solar cell technology
requires significant improvement as the record efficiency is
12.6% only, which is rather low compared with the 21.7% best
CIGSSe efficiency [1]. Since CZTSSe technology is not as

2156-3381 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications standards/publications/rights/index.html for more information.

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
2

IEEE JOURNAL OF PHOTOVOLTAICS

TABLE I
OVERVIEW OF ALL STEPS REQUIRED TO FABRICATE AL2 O3 REAR SURFACE
PASSIVATED CZTS SOLAR CELLS WITH WELL-CONTROLLED GRIDS OF
NANOSIZED LOCAL POINT CONTACTS
Step

Description

Remarks

Start

Low iron soda lime

glass (SLG)
Glass cleaning

1 mm thick

1
2
3

Fig. 1. TEM cross-sectional image of an Al2 O3 rear surface passivated pure

selenide CIGSe solar cell with a well-controlled grid of nanosized local rear
point contacts [5], [6].

mature as CIGSSe [9], [10], there are still many difficulties

that need to be addressed so that the electrical performance can
be increased. Typical issues are secondary phase formation,
fluctuating potentials, reproducibility and stability, and also cell
architecture. Consequently, present research is mainly focused
on enhancing the absorber quality [9], [10]. Nevertheless,
interface passivation also deserves attention, as it has been
verified that high-performing CZTSSe solar cells are dominated by interface recombination [11]. Initial efforts to improve
the front absorber/buffer-layer interface have already been
presented: First, also in CZTSSe solar cells, the influence of K
incorporation on its properties has already been studied. This
study showed that K doping can enhance the (1 1 2) preferred
orientation, increase the grain size, and reduce the formation
of ZnS secondary phases [12]. Second, also TiO2 front surface
passivation layers have already been tried in CZTSSe solar cells.
This passivation layer did not have nanosized point openings,
but was thin enough so that the photoexcited electrons could
effectively tunnel through the layer in the device [13].
In this work, for the first time, a novel rear surface passivation
concept is integrated and studied in ultrathin pure sulfide CZTS
solar cells. A prototypical method based on e-beam lithography,
as previously developed for pure selenide CIGSe solar cells
and shown in Fig. 1, is used to generate well-controlled
grids of nanosized point openings in an Al2 O3 passivation
layer. This way, Al2 O3 rear surface passivated CZTS solar
cells with nanosized local rear point contacts are fabricated,
and their performance and cross sections are studied and
compared with unpassivated reference cells. Solar cells with
ultrathin absorber layers are favored ( 500 nm), as these
are ideal characterization devices to investigate charge carrier
recombination at the rear CZTS surface (which becomes well
assessable due to the short distance between the space charge
region and the rear CZTS surface).
II. EXPERIMENTAL DETAILS
The Al2 O3 rear surface passivated CZTS solar cell fabrication
and characterization sequences are summarized in Table I; for
more details concerning the nanostructured Al2 O3 passivation
layer formation, the general cell processing (e.g., the CZTS
absorber layer formation), and a device model of unpassivated
reference CZTS solar cells, see [6], [14], and [15].

Rear contact
deposition
Passivation layer
deposition
Creation of openings
in the passivation
layer
Na precursor
deposition
Cu-Zn-Sn-S
deposition

Annealing

8
9

Absorber etching
Buffer deposition

10

Window deposition

11

Front contact
deposition
Solar cell scribing
Characterization

12
End

Deionized (DI) water and Cole-Palmer Micro-90

detergent in ultrasonic bath at 60 C
DC-sputtering of Mo 0.6 /, 350 nm thick
DC or RF-sputtering of Al2 O3 (30 nm)
E-beam pattern & reactive ion etching BCl3
400 nm opening diameter, 2 m spacing, or
180 nm opening diameter, 1 m spacing. See [6]
Evaporation of NaF 5 or 10 nm thick
Reactive sputtering in mixed Ar:H2 S gas CuS, Zn,
and Sn targets Substrate temperature of 180 C 1.7
Cu/Sn 1.8, Zn/(Cu + Sn) 0.4
Closed graphite box in tube furnace 35 kPa static
Ar atmosphere 80 mg elemental S inclusive 560 C
for 10 min
2 min in 5 wt% KCN solution
Chemical bath deposition (CBD) of CdS 60 C,
50 nm thick
RF-sputtering of (i-)ZnO(:Al) 400 nm thick,
40 /
Evaporation of Ni/Al/Ni 400/3000/400 nm thick
Mechanical by use of a stylus four cells of 0.5 cm2
X-ray fluorescence (XRF) Raman scattering [16]
X-ray diffraction (XRD) [16] Profilometry
Scanning electron microscopy (SEM)
Transmission electron microscopy (TEM)
Energy-dispersive X-ray spectroscopy (EDX)
Standard light current versus voltage (JV)
External quantum efficiency (EQE)

The unpassivated reference cells have the same processing sequence, but without steps 3
and 4. A summary of all characterization techniques typically applied is also given.

III. RESULTS AND DISCUSSION

A. CZTS Solar Cells With Ultrathin Absorber Layer as
Characterization Devices
Thinning down the absorber layer of standard CZTS solar
cells results in reduced VOC and JSC , because of a high rear
interface recombination velocity and a reduced bulk quality (including the presence of secondary phases). Fig. 2 shows the average VOC and JSC of unpassivated reference CZTS solar cells
with ever thinner absorber layers. Indeed, this figure illustrates
a reduction in VOC (24%rel.), JSC (27%rel.), FF (9%rel.),
and power conversion efficiency (PCE; 49%rel.) if the CZTS
absorber layer is reduced from 2000 down to 500 nm. As all cell
processing and X-ray fluorescence results are equivalent, absorber layer thickness (tCZTS ) should be the main variable for
these cells. In that case, the most logical explanation for these
VOC and JSC losses is a high recombination rate at the Mo/CZTS
rear interface and incomplete absorption, respectively, which
both become most obvious in the case of very thin absorber layers (as previously seen for CIGSe solar cells [5]). However, the
overall picture is more complicated for these CZTS solar cells
since experimental results also indicate that the thinnest CZTS
layers are more sensitive to secondary phase segregation (SnS
and ZnS are observed at both the front surface of the CZTS

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
VERMANG et al.: REAR SURFACE OPTIMIZATION OF CZTS SOLAR CELLS BY USE OF A PASSIVATION LAYER

Fig. 2. Average open-circuit voltage and short-circuit current for unpassivated

reference solar cells as a function of CZTS absorber layer thickness (eight cells
per CZTS thickness). The standard deviation is shown as error bars, and the
dashed line serves as a guide to the eye. Taken from [16].

films and the back contact interface) and nonuniform defect

properties. These secondary phases can be detrimental to the
device performance, since the SnS phase has a lower bandgap
than CZTS phase and could modify the interface properties,
while the ZnS phase can block current flow and introduce dead
areas [16]. Indeed, using the Solar Cell Capacitance Simulator
software, in-depth simulations have been performed and compared with these empirical data, indicating that the lower performance of the ultrathin CZTS solar cells is caused by both a
reduced bulk quality and high rear interface recombination [17].
Introducing a rear surface passivation layer with nanosized
local point contacts in such ultrathin CZTS solar cells has the
potential to reduce the rear surface recombination velocity and
the impact of secondary phase segregation significantly. As previously shown in ultrathin CIGSe solar cells [5][8], an adequate rear surface passivation layer leads to a low rear surface
recombination velocity ( 100 cm/s), resulting in a substantial
VOC and JSC increase. Additionally, such a passivation layer
with nanosized point contacts reduces the Mo/CZTS contacting
area to about 2% of the total rear area (see Table I), which also
means that the detrimental impact of SnS and ZnS secondary
phases may be reduced. Hence, these solar cells with an ultrathin absorber layer ( 500 nm) are excellent study devices to
investigate rear surface passivation layers in CZTS solar cells.
B. Mo(/Al2 O3 )/CZTS Rear Interface Study of Rear Passivated
CZTS Solar Cells
In this work, a NaF precursor layer is required to fabricate Al2 O3 rear passivated CZTS solar cells, andupon
completionthe final passivation layer seems to be quite rough
but present. An attempt to make rear passivated CZTS solar
cells with point contacts but without the use of NaF precursor
resulted in detachment of the absorber layer during the KCN
etching step. Even more, another attempt to make rear passivated CZTS solar cells without point contacts but with the use
of NaF precursor also resulted in similar peeling of the absorber
layer during KCN etching. Hence, a NaF precursor and local
point contacts (which thus seem to act as anchoring points
for the CZTS absorber layer) are required to fabricate rear pas-

Fig. 3. TEM cross-sectional image of an Al2 O3 rear surface passivated CZTS

solar cell with a well-controlled grid of nanosized local rear point contacts.

Fig. 4. TEM cross-sectional picture of an Al2 O3 rear surface passivated

CZTS solar cell with a well-controlled grid of nanosized local rear point contacts, including an EDX map for the elements Zn, Al, and Mo at the rear
Mo/(Al2 O3 /)CZTS interface.

sivated CZTS solar cells. Cross sections of such cells are shown
in Figs. 3 and 4, displaying transmission electron microscopy
(TEM) and energy-dispersive X-ray spectroscopy (EDX) measurements. In CIGSSe solar cells, Na doping has been shown
to enhance film morphology, grain growth, conductivity, VOC ,
and FF, but the detailed processes have yet to be completely
understood. Preliminary studies of Na doping of CZTSSe solar cells have already been performed and show improvements
in film morphology, grain orientation, grain boundaries, shunt
resistance, FF, JSC , and VOC [18]. This work does not aim to
study the fundamental impact of NaF precursors on CZTS solar cell performance; therefore, equivalent NaF precursors have
been used in unpassivated reference solar cells. The TEM pictures shown in Figs. 3 and 4 noticeably display the Al2 O3 rear
surface passivation layer and its point openings, wherein MoS2
formation occurred. However, these pictures also show a few
voids and it remains ambiguous to distinguish between these
voids and the Al2 O3 . Therefore, in Fig. 4, an EDX map of this
Mo(/Al2 O3 )/CZTS rear interface region is shown, indicating
that the Al2 O3 passivation layer is fairly complete but quite
rough compared with equivalent passivation layers in CIGSe
solar cells (as, e.g., shown in Fig. 1). Indeed, it seems that by
some means, chemical interaction occurs at the rear interface
during sulfurization of the CZTS precursor, which could be expected as Mo/Al2 O3 is likewise used as sulfurization catalyst
[19], [20]. However, also this study is outside the scope of this
work, where the Al2 O3 layer shown in Figs. 3 and 4 is satisfactory to study its effect as rear surface passivation layer in

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
4

IEEE JOURNAL OF PHOTOVOLTAICS

TABLE II
AVERAGE VALUES AND STANDARD DEVIATION OF CELL CHARACTERIZATION
RESULTS (AM1.5 G) FOR 0.5-CM2 AL2 O3 REAR SURFACE PASSIVATED
ULTRATHIN (400 NM) CZTS SOLAR CELLS AND CORRESPONDING
UNPASSIVATED ULTRATHIN AND THICK (2000 NM) REFERENCE CELLS
[CZTS; pass.; pitch; NaF]

# cells

VO C
(mV)

JS C
(mA/cm2 )

FF (%)

PCE (%)

[Thin; N.A.; N.A.; 5 nm]

[Thin; Al2 O3 ; 2 m; 5 nm]

[Thin; Al2 O3 ; 1 m; 5 nm]

[Thin; Al2 O3 ; 2 m; 10 nm]

[Thick; N.A.; N.A.; 0 nm]

461
53
540
09
540
30
522
05
584
21

14.7
0.6
15.4
0.2
13.6
0.2
15.2
0.1
18.6
0.5

45.6
4.4
49.9
0.7
43.8
3.4
42.9
3.0
50.8
4.1

3.1
0.3
4.1
0.1
3.2
0.2
3.4
0.3
5.5
0.6

The rear passivation layers have a well-controlled grid of nanosized local point openings
with a pitch of 1000 or 2000 nm.

taken from [16].

CZTS solar cells with ultrathin (and, thus, lower bulk quality
[16]) absorber layers.
C. Analysis of the Rear Passivated CZTS Solar Cell Results
Compared With Corresponding Unpassivated Reference Cells
Ultrathin CZTS solar cells with a nanostructured Al2 O3 rear
surface passivation layer have been fabricated and show reduced rear surface recombination and enhanced optical confinement, as compared with corresponding unpassivated reference cells. Table II gives an overview of measured ultrathin
CZTS solar cells (tCZTS = 400 nm) with and without an Al2 O3
rear surface passivation layer having nanosized point openings,
but also thick (tCZTS = 2000 nm) unpassivated reference cells
(taken from [16]). Representative JV and EQE curves for these
cells are presented in Fig. 5(a) and (b), respectively. Note that
the ultrathin unpassivated reference cell measurements with
tCZTS = 400 nm show slight improvements in FF, JSC , and
VOC compared with the results obtained in Fig. 2, as expected
due to the use of 5 nm of NaF precursor [18]. First, this discussion will concentrate on the Al2 O3 rear passivated cells with a
pitch (distance between the point openings) of 2 m, and where
5 nm of NaF is used. As shown in Table II, VOC (+ 17%rel.), JSC
(+ 5%rel.), FF (+ 9%rel.), and PCE (+ 32%rel.) are positively
impacted in these rear passivated cells, as compared with corresponding unpassivated reference cells. The conceivable reasons
are 1) a reduction in rear surface recombination velocity due
to the passivation effect of the Al2 O3 layer (chemical and field
effect passivation [5][7]) and 2) a reduced impact of the SnS
and ZnS secondary phases due to the reduced rear contacting
area (merely 2% of the total rear surface is contacting area). Additional interpretations can be made from Fig. 5: 1) The series
resistance (Rs ) has been estimated from the one-diode model,
indicating an increase from 1.1 to 14.2 cm2 for the rear passivated cells, compared with their corresponding reference cells,
as expected from the shape of the JV curves. Even more, the J
V curve of this rear passivated cell also shows a slight roll-over
effect, which could be caused by a lack of Na or a barrier effect [5]. Indeed, the used Al2 O3 passivation layer largely blocks
Na diffusion from the substrate ([5], which could result in Na

Fig. 5. Representative (a) currentvoltage and (b) external quantum efficiency curves for Al2 O3 rear surface passivated ultrathin (tC Z T S = 400 nm)
CZTS solar cells and corresponding unpassivated ultrathin and thick (tC Z T S =
2000 nm, taken from [16]) reference cells. The rear passivation layers have a
well-controlled grid of nanosized local point openings with a pitch of 1 or 2 m.
All EQE spectra are smoothed using a 50-point SavitzkyGolay smoothing filter.

deficient solar cells), and the distance between its point openings
could be too large (and as a result, it acts as a barrier layer). 2)
Compared with corresponding reference cells, the rise in JSC in
the rear passivated CZTS cells is mainly realized at higher wavelengths (from about 700 to 900 nm), while this wavelength range
was much broader for rear passivated CIGSe solar cells (from
about 550 to 1050 nm; see, e.g., [6]). This difference could be
related to the formation of secondary phases at the front and rear
of ultrathin CZTS absorbers, an issue not present in the CIGSe
solar cells, or to the shorter diffusion length of CZTS absorbers.
Indeed, the passivation layer with nanosized point openings is
effective in reducing the impact of ZnS and SnS secondary
phase segregation but primarily at the rear CZTS interface, and
the electron diffusion length is anticipated to be between 250
and 500 nm only (from the simulations performed in [15][17]).
Note thatpossiblythese EQE curves may even provide an
opportunity to distinguish between bulk and rear interface improvement of the rear passivation layer in ultrathin CZTS solar
cells. Second, the other rear passivated solar cells will be discussed. As mentioned before, the Al2 O3 rear passivated cells
with a pitch of 2 m and 5 nm of NaF show high Rs and might
be slightly Na deficient or have too large of a pitch. Therefore,
rear passivated cells with smaller pitch (1 m) and thicker NaF
(10 nm) are also prepared. Despite a significant increase in VOC

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
VERMANG et al.: REAR SURFACE OPTIMIZATION OF CZTS SOLAR CELLS BY USE OF A PASSIVATION LAYER

compared with the corresponding unpassivated reference cells,

no decrease in Rs or JV roll-over is obtained, compared with
the rear passivated cells with 2-m pitch and 5-nm Na. Contrarily, Fig. 5(a) shows an increase in Rs and even more JV
roll-over both for smaller pitch and thicker NaF, resulting in a
JSC and FF reduction. Hence, two final remarks can be made:
1) As the charge carrier diffusion lengths are anticipated to be
short (laterally, it might even be lower due to recombination at
grain boundaries), a pitch of 1 m is probably still too large.
Thus, the slight roll-over seen in the JV curves of the rear passivated cells could be an indication of a barrier layer between the
rear contact and absorber, which would make sense in the case
of a too short diffusion length combined with a point contact
spacing that is too wide. Note that such short diffusion lengths
also designate that characterization devices with a CZTS thickness of 400 nm are still on the thick side to study rear interface
recombination. 2) Using a thicker NaF layer did not diminish
the JV roll-over, but then again, it remains difficult to draw
Na-related conclusions as the influence of NaF and its thickness
on the fabrication of CZTS solar cells is not fully understood
and requires more investigation.
IV. CONCLUSION AND OUTLOOK
An Al2 O3 rear surface passivation layer is established and
studied in CZTSSe solar cells, where it provides a positive impact on cell performance. Rear passivated pure sulfide CZTS
TF solar cells with nanosized point openings have been fabricated, which showed a clear increase in VOC , JSC , FF, and,
hence, efficiency, as compared with corresponding reference
cells. These cells have ultrathin (tCZTS = 400 nm) absorber
layers and, therefore, are exemplary characterization devices
to investigate rear interface recombination. The reasons for the
improvement in cell performance are a reduction in 1) charge
carrier recombination at the rear absorber surface and 2) the
impact of secondary phase segregation. Theoretically, this approach could also be interesting for CZTS solar cells with
thicker absorber layers if these would exhibit long diffusion
lengths, whichunfortunatelyis not yet the case.
This is a very promising result, but better understanding is
essential. Therefore, new experiments (fabrication of solar cells,
but also simplified test devices for optoelectrical characterization [7], [8]) with alternative passivation layers and reduced
point contact spacing are scheduled.
Note that layer deposition techniques and methods to generate nanosized features appropriate for upscaling do exist, in case
such passivation layers with nanosized point openings would
become tempting for industrial application. For the deposition of
passivation layers, one could think of spatial ALD [21], [22], or
lower quality deposition methods such as sputtering or plasmaenhanced chemical vapor deposition [23], [24], while hole colloidal lithography and nanoimprint lithography in combination
with either dry or wet chemical etching have industrial potential
for nanostructuring [25].

[2] Y. Fu et al., ZnS nanodot film as defect passivation layer for

Cu(In,Ga)(S,Se)2 thin-film solar cells deposited by spray-ILGAR (ionlayer gas reaction), Adv. Energy Mater., vol. 1, no. 4, pp. 561564, 2011.
[3] P. Reinhard et al., Alkali-templated surface nanopatterning of chalcogenide thin films: A novel approach toward solar cells with enhanced
efficiency, Nano Lett., vol. 15, no. 5, pp. 33343340, 2015.
[4] P. Reinhard et al., Cu(In,Ga)Se2 thin-film solar cells and modulesA
boost in efficiency due to potassium, IEEE J. Photovoltaics, vol. 5, no.
2, pp. 656663, Mar. 2015.
[5] B. Vermang, V. Fjallstrom, X. Gao, and M. Edoff, Improved rear surface
passivation of Cu(In,Ga)Se2 solar cells: A combination of an Al2 O3 rear
surface passivation layer and nanosized local rear point contacts, IEEE
J. Photovoltaics, vol. 4, no. 1, pp. 486492, Jan. 2014.
[6] B. Vermang et al., Introduction of Si PERC rear contacting design to
boost efficiency of Cu(In,Ga)Se2 solar cells, IEEE J. Photovoltaics, vol.
4, no. 6, pp. 16441649, Nov. 2014.
[7] R. Kotipalli et al., Investigating the electronic properties of
Al2 O3 /Cu(In,Ga)Se2 interface, AIP Adv., vol. 1, no. 5, pp. 107101
1107101-6, 2015.
[8] J. Joel, B. Vermang, J. Larsen, O. Donzel-Gargand, and M. Edoff, On
the assessment of CIGS surface passivation by photoluminescence, Phys.
Status Solidi RRL, vol. 9, no. 5, pp. 288292, 2015.
[9] I. L. Repins et al., Kesterite successes, ongoing work, and challenges:
A perspective from vacuum deposition, IEEE J. Photovoltaics, vol. 3,
no. 1, pp. 439445, Jan. 2013.
[10] A. Redinger et al., Route toward high-efficiency single-phase
Cu2 ZnSn(S,Se)4 thin-film solar cells: Model experiments and literature
review, IEEE J. Photovoltaics, vol. 1, no. 2, pp. 200206, Oct. 2011.
[11] D. A. R. Barkhouse, O. Gunawan, T. Gokmen, T. K. Todorov, and
D. B. Mitzi, Device characteristics of a 10.1% hydrazine-processed
Cu2 ZnSn(Se,S)4 solar cell, Prog. Photovoltaics: Res. Appl., vol. 20,
no. 1, pp. 611, 2012.
[12] Z. Tong et al., Effects of potassium doping on solution processed kesterite
Cu2 ZnSnS4 thin film solar cells, Appl. Phys. Lett., vol. 105, pp. 223903
1223903-4, 2014.
[13] W. Wu et al., Characterization of CZTSSe photovoltaic device with
an atomic layer-deposited passivation layer, Appl. Phys. Lett., vol. 105,
pp. 0421081042108-4, 2014.
[14] Y. Ren, J. J. Scragg, T. Ericson, T. Kubart, and C. Platzer-Bjorkman, Reactively sputtered films in the Cux SZnSSnSy system: From metastability
to equilibrium, Thin Solid Films, vol. 582, pp. 208214, 2015.
[15] C. Frisk et al., Combining strong interface recombination with bandgap
narrowing and short diffusion length in CZTS device modeling, Sol.
Energy Mater. Sol. Cells, vol. 144, pp. 364370, 2015.
[16] Y. Ren et al., Influence of the Cu2 ZnSnS4 absorber thickness on thin film
solar cells, Phys. Status Solidi A, 2015, DOI: 10.1002/pssa.201532311.
[17] C. Frisk, Y. Ren, S. Li, and C. Platzer-Bjorkman, CZTS solar cell device simulations with varying absorber thickness, in Proc. 42nd IEEE
Photovoltaic Spec. Conf., 2015.
[18] A. D. Collord, H. Xin, and H. W. Hillhouse, Combinatorial exploration
of the effects of intrinsic and extrinsic defects in Cu2 ZnSn(S,Se)4 , IEEE
J. Photovoltaics, vol. 5, no. 1, pp. 288298, Jan. 2015.
[19] M. Tang et al., Presulfidation and activation mechanism of Mo/Al2 O3
catalyst sulfided by ammonium thiosulfate, Korean J. Chem. Eng., vol.
31, no. 8, pp. 13681376, 2014.
[20] T.-M. Chen, C.-M. Wang, I. Wang, and T.-C. Tsai, Promoter effect of
vanadia on Co/Mo/Al2 O3 catalyst for deep hydro-desulfurization via the
hydrogenation reaction pathway, J. Catalysis, vol. 272, pp. 2836, 2010.
[21] (2015, Sep. 7). The InPassion ALD Tool. [Online]. Available: http://www.
solaytec.com/
[22] (2015, Sep. 7). The Levitrack ALD System. [Online]. Available: http://
www.levitech.nl/
[23] B. Vermang et al., Employing Si solar cell technology to increase efficiency of ultra-thin Cu(In,Ga)Se2 solar cells, Prog. Photovoltaics: Res.
Appl., vol. 22, no. 10, pp. 10231029, 2014.
[24] G. Dingemans and W. M. M. Kessels, Review: status and prospects of
Al2 O3 -based surface passivation schemes for silicon solar cells, J. Vac.
Sci. Technol. A, vol. 30, no. 4, pp. 0408021040802-27, 2012.
[25] C. Trompoukis et al., Photonic nanostructures for advanced light trapping
in thin crystalline silicon solar cells, Phys. Status Solidi (a), vol. 212,
no. 1, pp. 140155, 2015.

REFERENCES
[1] M. A. Green, K. Emery, Y. Hishikawa, W. Warta, and E. D. Dunlop,
Solar cell efficiency tables (version 45), Prog. Photovoltaics: Res. Appl.,
vol. 23, no. 1, pp. 19, 2015.

Authors photographs and biographies not available at the time of publication.