Synthesis of CuInS2, CuInSe2, and Cu(InxGa1-x)Se2(CIGS)
Nanocrystal “Inks” for Printable Photovoltaics
Matthew G. Panthani,†Vahid Akhavan,†Brian Goodfellow,†Johanna P. Schmidtke,‡
Lawrence Dunn,§,|Ananth Dodabalapur,§Paul F. Barbara,‡and Brian A. Korgel*,†
Departments of Chemical Engineering, Chemistry & Biochemistry, and Physics and
Microelectronics Research Center, Texas Materials Institute and Center for Nano- and
Molecular Science and Technology, The UniVersity of Texas at Austin, Austin, Texas 78712-1062
Received July 25, 2008; E-mail: firstname.lastname@example.org
Abstract: Chalcopyrite copper indium sulfide (CuInS2) and copper indium gallium selenide (Cu(InxGa1-x)-
Se2; CIGS) nanocrystals ranging from ∼5 to ∼25 nm in diameter were synthesized by arrested precipitation
in solution. The In/Ga ratio in the CIGS nanocrystals could be controlled by varying the In/Ga reactant ratio
in the reaction, and the optical properties of the CuInS2and CIGS nanocrystals correspond to those of the
respective bulk materials. Using methods developed to produce uniform, crack-free micrometer-thick films,
CuInSe2nanocrystals were tested in prototype photovoltaic devices. As a proof-of-concept, the nanocrystal-
based devices exhibited a reproducible photovoltaic response.
I-III-VI2chalcopyrite compounds, particularly copper in-
dium gallium selenide (Cu(InxGa1-x)Se2; CIGS), are effective
light-absorbing materials in thin-film solar cells.1These materials
possess advantageous properties for solar applications: their band
gap energy is at the red edge of the solar spectrum; they are
direct band-gap semiconductors with correspondingly high
optical absorption coefficients;2,3and CIGS materials, in contrast
to other candidate materials for thin-film solar cells such as CdTe
and amorphous silicon (a-Si), are stable under long-term
excitation.4High efficiency CIGS-based devices are typically
fabricated using polycrystalline films,5and single-junction CIGS
solar cells have demonstrated nearly 20% solar energy conver-
sion efficiency,6which is significantly higher than either CdTe
or a-Si based devices.7Furthermore, CIGS devices and manu-
facturing processes may have less environmental impact than
those with thin film materials with large amounts of Cd and
Pb, like CdTe and PbSe based solar cells, although to date the
highest efficiency CIGS photovoltaic (PV) devices have none-
theless required CdS buffer layers.8
One of the hurdles currently impeding widespread com-
mercialization of CIGS-based solar cells is the difficulty in
achieving controlled stoichiometry over large device areas,
leading to high manufacturing costs and poor device yield.9
CIGS layers in state-of-the-art devices are deposited by a
multistage coevaporation process in which alternate copper,
indium, and gallium layers are deposited followed by reacting
with a selenium source, Se, or H2Se gas, in the chamber.10,11
This process is time-consuming and the CIGS stoichiometry is
difficult to controlsintermetallic phases can form and the Se
content can vary significantly in the films.9,12Large material
losses on the deposition chamber walls also increase cost. For
all of these reasons, alternative CIGS layer deposition strategies
One approach with the potential to produce CIGS layers with
controlled stoichiometry without the need for high temperature
annealing is to chemically synthesize CIGS nanocrystals with
controlled stoichiometry and crystal phase and disperse them
in solvents, creating a paint or ink. Such an approachsof
printable CIGS inkssmakes accessible a range of solution-based
processing techniques and may lead to inexpensive fabrication
†Department of Chemical Engineering.
‡Department of Chemistry & Biochemistry.
§Microelectronics Research Center.
|Department of Physics.
(1) Miller, A.; Mackinnon, A.; Weaire, D. Solid State Physics 1981, 36,
(2) Devaney, W. E.; Chen, W. S.; Stewart, J. M.; Mickelsen, R. A. IEEE
Trans. Electron DeVices 1990, 37, 428–433.
(3) Scheer, R.; Walter, T.; Schock, H. W.; Fearheiley, M. L.; Lewerenz,
H. J. Appl. Phys. Lett. 1993, 63, 3294–3296.
(4) Guillemoles, J.-F.; Kronik, L.; Cahen, D.; Rau, U.; Jasenek, A.; Schock,
H.-W. J. Phys. Chem. B 2000, 104, 4849–4862.
(5) Rockett, A.; Birkmire, R. W. J. App. Phys. 1991, 70, R81–R97.
(6) Contreras, Miguel, A.; AbuShama, K. R., J.; Hasoon, F.; Young, D. L.;
Egaas, B.; Noufi, R. Prog. PhotoVoltaics Res. Appl. 2005, 13, 209–
(7) Green, M. A.; Emery, K.; King, D. L.; Hishikawa, Y.; Warta, W.
Prog. PhotoVoltaics Res. Appl. 2007, 15, 35–40.
(8) Eberspacher, C.; Fthenakis, V. M.; Moskowitz, P. D. 25th IEEE
Photovoltaic Specialist Conference, IEEE: Washington, DC, 1996; pp
(9) Powalla, M.; Dimmler, B. Thin Solid Films 2000, 361-362, 540–
(10) Repins, I.; Contreras, M. A.; Egaas, B.; DeHart, C.; Scharf, J.; Perkins,
C. L.; To, B.; Noufi, R. Prog. PhotoVoltaics Res. Appl. 2008, 16, 235–
(11) Contreras, M. A.; Egaas, B.; Ramanathan, K.; Hiltner, J.; Swartzlander,
A.; Hasoon, F.; Noufi, R. Prog. PhotoVoltaics Res. Appl. 1999, 7,
(12) Schock, H.-W.; Noufi, R. Prog. PhotoVoltaics Res. Appl. 2000, 8, 151–
Published on Web 11/17/2008
10.1021/ja805845q CCC: $40.75 2008 American Chemical Society
16770 9 J. AM. CHEM. SOC. 2008, 130, 16770–16777
routes for CIGS light-absorbing layers.13A chemical, solution-
based approach alleviates the need for a high temperature
annealing step under selenium atmosphere and may solve the
CIGS “selenium problem”sthat is, avoiding Se loss and
achieving the correct CIGS stoichiometry in films covering large
substrate areas.14Photovoltaic devices incorporating nanocrys-
talline-based CdSe/CdTe15and CuS16absorber layers have been
reported and demonstrated solar energy conversion efficiencies
as high as 2.9%, although a high temperature anneal at 400 °C
was required. Semiconductor nanocrystals have also been
combined with polymers to produce solution-processed photo-
voltaics, such as hybrid CdSe nanocrystal/poly-3(hexyl-
thiophene) solar cells, which yield reported efficiencies of up
to 1.7%.17Many different semiconductor nanocrystals can be
synthesized by colloidal routes, including groups II-VI,18
III-V,19I-VI.20-22IV-VI,23and IV24-26semiconductors, but
the synthesis of I-III-VI2nanocrystals is much less developed.
Nonetheless, there are literature reports of the synthesis of
ternary chalcopyrite compound nanocrystals, such as CuInS2,
CuInSe2, and other I-II-VI2semiconductor nanocrystals such
as AgInS2.27-32These nanocrystals, however, generally suffer
from relatively low yields, poor crystallinity,33and poor
uniformity in composition and phase.3,34This is not surprising
considering that many of these systems have very complicated
phase diagrams and nanocrystals can further exhibit greater
phase complexity than the corresponding bulk materials.22
Here, we report the high-yield synthesis in solution of phase-
pure nanocrystals of chalcopyrite CuInS2, CuInSe2, and Cu(Inx-
Ga1-x)Se2with controlled In/Ga ratio. The nanocrystals disperse
readily in various nonpolar solvents and can be deposited as
uniform, crack-free micrometer-thick films onto glass and metal
substrates. Prototype photovoltaic devices (PVs) were fabricated
using CuInSe2nanocrystals as the absorber layer and exhibited
a robust and reproducible photovoltaic response. The reported
baseline performance (solar energy power conversion efficien-
cies of ∼0.2%) requires improvement, which might be achieved
by additional processing steps including ligand exchange,
chemical treatment, or high temperature annealing.
Materials. All chemicals were used as received without further
purification. Oleylamine (OLA; >70%) was obtained from Fluka;
copper(II) acetylacetonate (Cu(acac)2; 99.99+%), copper(I) chloride
(CuCl; 99.995+%), indium(III) chloride (InCl3; anyhydrous 99.99%),
indium(III) acetylacetonate (In(acac)3; 99.99+%), elemental sulfur
(99.98%), and o-dichlorobenzene (DCB; 99%) from Aldrich
Chemical Co.; elemental selenium (99.99%) and gallium(III)
chloride (GaCl3; 99.9999%) from Strem Chemicals; and chloroform
(99.99%), ethanol (absolute), and tetrachloroethylene (TCE; spec-
trophotometric grade 99+%) from Fisher Scientific. N2and forming
gas (7% H2, 93%N2) were received from Matheson Tri-Gas.
Copper(I) chloride, indium(III) chloride, and gallium(III) chloride
were stored in a nitrogen-filled glovebox to prevent degradation.
CuInS2Nanocrystal Synthesis. A 0.26 g (1 mmol) portion of
Cu(acac)2and 0.41 g (1 mmol) of In(acac)3are added to 7 mL of
DCB in a 25 mL three-neck flask in air. In a separate 25-mL three-
neck flask, 0.064 g (2 mmol) of elemental sulfur is dissolved in 3
mL of DCB in air. Both flasks are then attached to a Schlenk line
and purged of oxygen and water by pulling vacuum at room
temperature for 30 min, followed by N2bubbling at 60 °C for 30
min. Between 0.5 and 2 mL (1.5 to 6 mmol) of OLA are added to
the (Cu, In)-DCB mixture and both flasks are heated to 110 °C
and combined, maintaining a N2 flow. The reaction mixture is
refluxed (∼182 °C) for 1 h under N2flow. The reaction is allowed
to cool to room temperature, and the nanocrystals are separated by
adding excess ethanol. The yield of solution-stable nanocrystals
after purification was ∼90%.
CuInSe2Nanocrystal Synthesis. In a nitrogen-filled glovebox,
1 mmol of CuCl (0.099 g), 1 mmol of InCl3(0.221 g), and 2 mmol
of elemental Se (0.158 g) are combined in a 25-mL three-neck flask
with an attached condenser and stopcock valve. The stopcock valve
is closed before removing the flask from the glovebox, where it is
attached to a Schlenk line and placed on a heating mantle. OLA
(10 mL) stored in air is injected into the flask. The flask is purged
of oxygen and water by pulling vacuum at 60 °C for 1 h, followed
by N2bubbling at 110 °C for 1 h while stirring. The mixture is
then heated to 240 °C, and the reaction proceeds for 4 h under
vigorous stirring. The reaction is cooled to ∼100 °C, where ∼10
mL of chloroform is added to quench the reaction and ∼5 mL of
ethanol is added to precipitate the nanocrystals. After adding the
ethanol, the reaction mixture is immediately removed and placed
in a centrifuge tube. Reactions carried out for less than 4 h yielded
nanocrystals with a larger size distribution and more agglomeration
when dispersed after purification. A significant amount of poorly
capped and large (up to 200 nm diameter) nanocrystals are found
in the crude reaction product, which is separated from the well-
capped nanocrystals. The typical product yield of the well-dispersed
CuInSe2nanocrystals was ∼15%. Arrested precipitation procedures
in which OLA complexes of Cu, In, and Se were formed separately
and then combined at high temperature yielded nanocrystals that
were very unstable when purified and redispersed.
Cu(In,Ga)Se2 Nanocrystal Synthesis. A typical reaction is
carried about by adding 1 mmol of CuCl (0.099 g), 2 mmol of
elemental Se (0.158 g), and 1 mmol total of InCl3(0.00 to 0.221
g) and GaCl3(0.00 to 0.111 g) to a 25-mL three-neck flask with
attached condenser and stopcock valve in a nitrogen-filled glovebox.
The stopcock valve is closed before removing the flask from the
(13) Eberspacher, C.; Pauls, K. L.; Serra, J. P. Mater. Res. Soc. Symp. Proc.
2003, 763, B8.27.1–B8.27.6.
(14) Kumar, A. P.; Reddy, K. V. Thin Solid Films 1997, 304, 365–370.
(15) Gur, I.; Fromer, N. A.; Geier, M. L.; Alivisatos, A. P. Science 2005,
(16) Wu, Y.; Wadia, C.; Ma, W.; Sadtler, B.; Alivisatos, A. P. Nano Lett.
2008, 8, 2551–2555.
(17) Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002, 295,
(18) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993,
(19) Micic, O. I.; Curtis, C. J.; Jones, K. M.; Sprague, J. R.; Nozik, A. J.
J. Phys. Chem. 1994, 98, 4966–4969.
(20) Motte, L.; Billoudet, F.; Pileni, M. P. J. Phys. Chem. 1995, 99, 16425–
(21) Larsen, T. H.; Sigman, M.; Ghezelbash, A.; Doty, R. C.; Korgel, B. A.
J. Am. Chem. Soc. 2003, 125, 5638–5639.
(22) Ghezelbash, A.; Korgel, B. A. Langmuir 2005, 21, 9451–9456.
(23) Murray, C. B.; Sun, S.; Gaschler, W.; Doyle, H.; Betley, T. A.; Kagan,
C. R. IBM J. Res. DeV. 2001, 45, 47.
(24) Heath, J. R.; Shiang, J. J.; Alivisatos, A. P. J. Chem. Phys. 1994,
(25) Holmes, J. D.; Ziegler, K. J.; Doty, R. C.; Pell, L. E.; Johnston, K. P.;
Korgel, B. A. J. Am. Chem. Soc. 2001, 123, 3743–3748.
(26) Lu, X.; Korgel, B. A.; Johnston, K. P. Chem. Mater. 2005, 17, 6479–
(27) Elim, H. I.; Ji, W.; Ng, M.-T.; Vittal, J. J. Appl. Phys. Lett. 2007, 90,
(28) Castro, S. L.; Bailey, S. G.; Raffaelle, R. P.; Banger, K. K.; Hepp,
A. F. J. Phys. Chem. B 2004, 108, 12429–12435.
(29) Banger, K. K.; Jin, M. H.-C.; Harris, J. D.; Fanwick, P. E.; Hepp,
A. F. Inorg. Chem. 2003, 42, 7713–7715.
(30) Castro, S. L.; Bailey, S. G.; Raffaelle, R. P.; Banger, K. K.; Hepp,
A. F. Chem. Mater. 2003, 15, 3142–3147.
(31) Arici, E.; Sariciftci, N. S.; Meissner, D. AdV. Funct. Mater. 2003, 13,
(32) Czekelius, C.; Hilgendorff, M.; Spanhel, L.; Bedja, I.; Lerch, M.;
Mu ¨ller, G.; Bloeck, U.; Su, D.-S.; Giersig, M. AdV. Mater. 1999, 11,
(33) Eberspacher, C.; Pauls, K.; Serra, J. 28th IEEE Photovoltaic Specialist
Conference; IEEE:Anchorage, AK, 2000; pp 517-520.
(34) Choi, S.-H.; Kim, E.-G.; Hyeon, T. J. Am. Chem. Soc. 2006, 128,
J. AM. CHEM. SOC. 9 VOL. 130, NO. 49, 2008
Synthesis of Nanocrystal Inks
glovebox, where it is attached to a Schlenk line and placed on a
heating mantle. OLA (10 mL) is injected into the flask. The flask
is purged of oxygen and water by pulling vacuum at 60 °C for 1 h,
followed by N2bubbling at 110 °C for 1 h while stirring. The
mixture is then heated to 240 °C, and the reaction proceeds for 4 h
under vigorous stirring. The product yield of CuInxGa1-xSe2nano-
crystals with x < 1 ranged from 20%-60% after purification.
Nanocrystal Purification. The nanocrystal products were puri-
fied by precipitation with excess ethanol followed by centrifugation
at 8000 rpm for 10 min. After such a washing step, the supernatant
contains unreacted precursor and byproducts and is discarded. The
nanocrystals are in the precipitate. The nanocrystals are then
redispersed in 10 mL of chloroform and centrifuged at 7000 rpm
for 5 min to remove poorly capped nanocrystals and large
particulates, which settle during centrifugation. The well-capped
nanocrystals remain dispersed in the supernatant. The precipitate
is discarded and a small amount of OLA (0.2 mL) is subsequently
added to the supernatant to ensure complete surface passivation of
the nanocrystals. To remove excess capping ligands and remaining
impurities, the product is again precipitated using ∼5 mL of ethanol
and centrifuged at 8000 rpm for 10 min, then redispersed in
chloroform. This process is done three times to obtain a high-purity
product. The isolated nanocrystals disperse in various nonpolar
organic solvents, including hexane, toluene, decane, chloroform,
Materials Characterization. The nanocrystals were character-
ized using a range of analytical techniques, including transmission
electron microscopy (TEM), energy-dispersive X-ray spectroscopy
(EDS), inductively coupled plasma mass spectrometry (ICP-MS),
scanning electron microscopy (SEM), X-ray diffraction (XRD), and
UV-vis-NIR absorbance spectroscopy.
TEM imaging was performed on nanocrystals drop-cast from
chloroform, hexane, or toluene dispersions on carbon-coated 200
mesh copper or nickel TEM grids (Electron Microscopy Sciences).
TEM images were acquired on either a Phillips 208 TEM with 80
kV accelerating voltage or a JEOL 2010F TEM operating at 200
keV. The JEOL 2010F TEM is equipped with an Oxford INCA
EDS detector, which was used to collect EDS data. The nanocrystal
composition was also measured using inductively coupled mass
spectrometry (ICPMS) for nanocrystal films and powders using a
GBC Optimass 8000 ICP-TOF-MS. The ICPMS samples were
prepared by digesting a dried nanocrystal powder or thin film in
SEM images were acquired using either a LEO 1530 or Zeiss
Supra 40 VP SEM operated between 1 and 10 keV. For SEM
imaging, the nanocrystals were deposited by evaporation from a
solvent on a Si or Mo-coated glass substrate.
XRD data was acquired using a Bruker-Nonius D8 Advance
θ-2θ Powder Diffractometer equipped with a Bruker Sol-X Si(Li)
solid-state detector and a rotating stage. Cu KR (λ) 1.54 Å)
radiation was used. For XRD, the nanocrystals were evaporated
from concentrated dispersions onto quartz (0001) substrates as ∼0.5
mm thick films. Diffraction data was collected by scanning for 4
to 12 h with an angle increment of 0.01° or 0.02° at a scan rate of
6°/min and a rotation speed of 15 rpm.
UV-vis-NIR absorbance spectra were obtained with a Varian
Cary 500 UV-vis-NIR spectrophotometer using hexane-dispersed
nanocrystals in a quartz cuvette. Film thicknesses were found using
a Veeco Dektak 6 M stylus profiler.
Nanocrystal Film Deposition and Photovoltaic Device Fabri-
cation. Thick films (∼1 µm) of nanocrystals were deposited onto
12 × 25 mm glass or Mo-coated glass substrates by dropping 150
µL of TCE dispersions with nanocrystal concentrations of 5 mg/
mL. The film was fully dried by placing the substrate in a vacuum
chamber at room temperature for 12 h.
Photovoltaic test structures were fabricated with a conventional
sandwich-type Mo/CuInSe2/CdS/ZnO/indium tin oxide (ITO) con-
figuration. The molybdenum (Mo) backcontact was first deposited
on soda lime glass (Delta Technologies, 25 × 25 × 1.1 mm3
polished float glass) by radio frequency (rf) sputtering from a pure
Mo target (99.999%, Lesker) in ultrapure Ar (99.999%, Praxair)
at 5 mTorr. Radio frequency sputtering was used instead of DC
sputtering because it has been reported to provide a film with a
better combination of substrate adhesion and good conductivity.35
The CuInSe2nanocrystal layer was deposited by dropcasting it from
TCE dispersions and then placing the film under vacuum overnight
at room temperature to dry. A CdS buffer layer was deposited from
solution using a procedure and parameters outlined by McCandless
and Shafarman.36Stock aqueous solutions of 0.015 M cadmium
sulfate (Aldrich, 99.999%), 1.5 M thiourea (Fluka, 99%), and 14.28
M ammonium hydroxide (Fisher Scientific, Certified ACS) were
made and used in preparation of working solutions by mixing 1.25
mL of the CdSO4solution, 2.2 mL of the CS(NH2)2solution, and
2.8 mL of the NH4OH solution. Substrates were placed on a hot
plate for 10 min that had been preheated to 90 °C, after which
∼0.5 mL of the working solution was deposited on each substrate.
The substrates were immediately covered to reduce the loss of
ammonia from the solution. After 2 min the substrates were
removed from the hot plate, rinsed with DI water, and laid flat to
dry. The i-ZnO/ITO top contact was deposited by rf sputtering from
pure targets of each material: ZnO (99.9%, Lesker) was deposited
using 0.5% O2in Ar (99.95%, Praxair) and ITO (99.99% In2O3:
SnO290:10, Lesker) was deposited in Ar. The final active region
of the device was 8 mm2(a 4 mm × 2 mm rectangle).
The electrical properties of the PV devices were characterized
using a Karl Suss Probe station and an Agilent 4156C Parameter
Analyzer. Detailed studies of power conversion efficiencies were
done using a Keithley 2400 General Purpose Sourcemeter and a
Xenon Lamp Solar Simulator (Newport) equipped with an AM1.5
filter. Incident photon conversion efficiency (IPCE) spectra were
gathered using a lock-in amplifier (Stanford Research Systems,
(35) Scofield, J. H.; Duda, A.; Albin, D.; Ballard, B. L.; Predecki, P. K.
Thin Solid Films 1995, 260, 26–31.
(36) McCandless, B. E.; Shafarman, W. N. Chemical Surface Deposition
of Ultra-thin Semiconductors. U.S. Patent 6,537,845, March 25, 2003.
Table 1. Measured CIGS Nanocrystal Composition
composition measured by EDSa
composition measured by ICPMSb
aEDS measurements have an error of ca. (2 atom %.bICPMS measurements have an error of (0.1 atom % for Cu, (0.2 atom % for In, (0.1
atom % for Ga, (0.5 atom % for Se.
16772J. AM. CHEM. SOC. 9 VOL. 130, NO. 49, 2008
Panthani et al.
model SR830), a monochromator (Newport Cornerstone 260 1/4M),
and a Si photodiode calibrated by the manufacturer (Hamamatsu).
Results and Discussion
CuInS2Nanocrystals. CuInS2nanocrystals were synthesized
using a variation of the procedure developed by Ghezelbash
and Korgel22for CuS nanocrystals, by adding In(acac)3as an
In source to the reaction:
Elemental sulfur dissolves in dichlorobenzene and could be used
directly as the sulfur source. Figure 1 shows TEM images of
CuInS2nanocrystals synthesized using the reaction scheme in
eq 1. The nanocrystal size could be roughly controlled by
varying the OLA/metal ratio. Figure 1 shows CuInS2nano-
crystals with two different average diameters obtained by
varying the OLA/metal ratio in the reaction. The average
nanocrystal diameter was increased from 6 to 12 nm as the OLA/
metal ratio by decreasing the ratio from 6:1 to 3:1. The
nanocrystal shape was not perfectly spherical, which contributed
to the relatively broad size distributions of the nanocrystals. High
resolution TEM (Figure 2) showed the crystallinity of the
nanocrystals, with lattice spacings corresponding to tetragonal
CuInS2. XRD (Figure 3) confirmed that the nanocrystals are
chalcopyrite (tetragonal) CuInS2and that no other phases are
produced in the reaction. EDS from fields of nanocrystals gave
an average Cu/In/S composition of 0.29:0.25:0.46, which is near
the target 0.25:0.25:0.5 ratio, considering the error of the EDS
detector (approximately (2 atom %) and that Cu is slightly
overrepresented in the EDS spectra because of signal from the
Cu sample holder. There was no compositional variation from
particle to particle within the error of the EDS detector. The
band gap energy determined from absorbance spectra (Figure
4) of optically clear (i.e., nonscattering) dispersions of nano-
crystals was found to be 1.29 eV (960 nm), which is within the
range of the CuInS2band gap energy (which has been reported
to lie between 1.2 and 1.5 eV) reported in literature.3,37
CuInSe2 Nanocrystals. CuInSe2 nanocrystals could not be
synthesized using an approach similar to CuInS2because unlike
S, Se does not dissolve in dichlorobenzene. After exploring a
variety of different reaction approaches, one effective route was
(37) Berger, L. I. In CRC Handbook of Chemistry and Physics, 79 ed.;
CRC Press, Boca Raton, FL, 1999; pp 12-84.
Figure 1. TEM images of CuInS2nanocrystals synthesized with varying
OLA/(Cu+In) mole ratios: (a,b) 6:1, 8 nm diameter; (c,d) 3:1, 12 nm
Figure 2. HRTEM images of a CuInS2 Nanocrystals (a,b) and their
respective fast Fourier transforms (FFTs) (c,d). The d-spacings correspond
to chalcopyrite (tetragonal) CuInS2.
Figure 3. XRD and (inset) elemental composition measured by EDS of 8
nm diameter CuInS2nanocrystals. The peak labels correspond to those of
chalcopyrite (tetragonal) CuInS2(JCPDS No. 085-1575).
Figure 4. Room temperature absorbance spectrum of 8 nm diameter CuInS2
nanocrystals dispersed in hexane.
J. AM. CHEM. SOC. 9 VOL. 130, NO. 49, 2008
Synthesis of Nanocrystal Inks
a direct combination of Cu and In salts and solid Se in a flask
with oleylamine followed by heating to 240 °C for 4 h:
Figure 5 shows TEM images of a typical CuInSe2nanocrystal
preparation. The nanocrystals are approximately 15 nm in
diameter. Both high-resolution TEM (Figure 6) and XRD
(Figure 7) confirmed that the nanocrystals are crystalline with
tetragonal chalcopyrite CuInSe2 structure. The d-spacings
observed in TEM and the FFTs of the TEM images are also
consistent with tetragonal CuInSe2. No other crystal phases were
observed in the XRD patterns of the product. Compositional
analysis by ICPMS showed that the average composition of the
nanocrystals in the sample has a molar Cu/In/Se ratio of 1:1:2
and the composition of individual particles measured by EDS
was 1:1:2 with a variation from particle to particle less than
the experimental error of ca. (2 atom %.
Like the CuInS2nanocrystals, the CuInSe2nanocrystals are
not spherical and exhibit significant faceting. The faceting has
thus far been difficult to control, but this might be addressed
by the optimization of several factors, including the capping-
ligand chemistry and the way reactants are added to the reaction.
The relatively broad size distribution of the nanocrystals (ranging
from as small as 5 nm to as large as 25 nm) is largely the result
of this irregularity in particle shape.
Cu(Inx,Ga1-x)Se2 (CIGS) Nanocrystals. CIGS nanocrystals
were synthesized following the approach developed for CuInSe2
nanocrystals, but with the addition of GaCl3 to the reaction
mixture in the desired In/Ga mole ratio:
The In/Ga ratio could be tuned across the entire stoichiometric
range with x from 0 to 1 using this approach. Figure 10 shows
TEM images of CuInxGa1-xSe2with x ranging from 0.79 to 0.
Figures 8 and 9 show XRD data of CIGS nanocrystals
synthesized with Ga/In ratios varying from 0 to 1. All of the
Figure 5. TEM images of CuInSe2nanocrystals with an average diameter
of 15 nm.
Figure 6. (a,b) HRTEM images of CuInSe2nanocrystals and (c,d) their
FFTs. The observed d-spacings and the indexed FFTs are consistent with
chalcopyrite (tetragonal) CuInSe2.
Figure 7. XRD pattern of chalcopyrite CuInSe2nanocrystals (JCPDS No.
00-040-1487). The scattering intensity is plotted on a logarithmic scale to
elucidate the (211) peak. Dashed boxes indicate reflections that are unique
to chalcopyrite (CuInSe2).
Figure 8. XRD patterns of CIGS nanocrystals synthesized with varying
In:Ga ratios: (a) CuInSe2(b) CuIn0.79Ga0.21Se2(by EDS) (c) CuIn0.51Ga0.49Se2
(by EDS) (d) CuGaSe2nanocrystals. The diffraction patterns correspond
to those of the tetragonal chalcopyrite phases of the respective compounds.
The indexing of the peaks noted in (a) correspond to the expected peaks
positions of the chalcopyrite compounds.
16774 J. AM. CHEM. SOC. 9 VOL. 130, NO. 49, 2008
Panthani et al.
patterns are consistent with chalcopyrite (tetragonal) crystal
structure and exhibit the expected amount of peak broadening
due to their nanoscale crystal domain size. The diffraction peaks
shift to higher 2θ with increasing Ga content, due to the
decreased lattice spacing with smaller Ga atoms substituting
for larger In atoms. The In/Ga ratio of the nanocrystals
determined by ICPMS and EDS were consistent with the In/
Ga mole ratio in the reaction mixture (see Supporting Informa-
tion for sample EDS data). Additionally, EDS measurements
on different nanocrystals on the substrate did not show any
noticeable variation in Cu/In/Ga ratio from particle to particle
in the sample. Table 1 summarizes the synthesis results. The
band gap energies of the Cu(InxGa1-x)Se2nanocrystals deter-
mined from room temperature absorbance spectra (Figure 11)
(CuInSe2, 0.95 eV; CuIn0.56Ga0.44Se2, 1.14 eV; CuGaSe2, 1.51
eV) of nanocrystal dispersions were also consistent with energies
of the corresponding bulk compounds: 0.95, 1.23, and 1.6 eV.40
The only noticeable difference in the nanocrystals with varying
In/Ga ratio was that nanocrystals with higher Ga content were
more difficult to stabilize in solution without aggregation.
Particularly the CuGaSe2nanocrystals were not easily dispersible
after isolation from the reaction mixture. More effective capping
approaches to Ga-rich nanocrystals are desirable.
Deposition and Photovoltaic Response of CuInSe2 Nano-
crystal Layers. As a proof-of-concept, films of OLA-coated 15
nm CuInSe2nanocrystals were tested as the absorber layer in
PV devices fabricated with a conventional layered Mo/CuInSe2/
CdS/ZnO/ITO configuration. An illustration of the device
configuration is provided in the Supporting Information. Typical
PV devices require relatively thick absorber layers (>1 µm),
and therefore a strategy was first developed to deposit uniform,
crack-free nanocrystal films.
Dip-coating worked well to deposit uniform, crack-free
nanocrystal films, but only up to a maximum thickness of ∼300
nm. (See Supporting Information for an example of the thickness
profile of a dip-coated film.) The maximum film thickness of
∼300 nm appeared to be related to the thickness of the fluid
layer that formed on the vertically dipped substrate as it was
pulled from the solvent. Multiple dipping steps could not
improve the film thickness either as it appeared that previously
deposited nanocrystals would redisperse as a new layer of
particles was deposited. This fluid layer thickness depends on
the substrate wettability and the dispersion viscosity and perhaps
could be further improved with more study.
Uniform nanocrystal films in the appropriate thickness range
could be formed by drop-casting from dispersions in high boiling
point organic solvents. (See Supporting Information for SEM
images of uniform, nearly crack-free films deposited from
concentrated tetrachloroethylene (TCE) dispersions, along with
an example of a nonuniform, heavily cracked CuInSe2nano-
crystal film deposited by drop-casting from chloroform, which
has a high volatility.) Nanocrystal films as thick as 3 µm could
be deposited by drop-casting from concentrated TCE dispersions
and the film thickness could be controlled by varying the
nanocrystal concentration in the dispersions as shown in Figure
12. Figure 13 shows pictures of a TCE dispersion of CuInSe2
nanocrystals and the deposition process used to make multiple
films of CuInSe2nanocrystals on 12 mm × 25 mm soda-lime
glass or Mo-coated glass substrates.
Figure 9. Magnification of the (112) XRD peaks from Figure 8 of the
CIGS nanocrystals: (a) CuInSe2 (b) CuIn0.79Ga0.21Se2 (by EDS) (c)
CuIn0.51Ga0.49Se2 (by EDS) (d) CuGaSe2 nanocrystals. The reference
positions are for CuInSe2(JCPDS#00-040-1487), CuIn07Ga0.3Se2(JCP-
DS#00-035-1102), CuIn0.5Ga0.5Se2(JCPDS#00-040-1488), and CuGaSe2
Figure 10. TEM images of CuInxGa1-xSe2nanocrystals with (a) x ) 0.79,
(b) 0.56, (c) 0.21, and (d) 0.
Figure 11. Room temperature absorbance spectra of Cu(InxGa1-x)Se2
nanocrystals dispersed in hexane. The curves correspond to In/Ga stoichi-
ometries of (a) x ) 0, (b) x ) 0.56, and (c) x ) 1. An extrapolation of the
spectra to identify the band edge is shown in the inset. The small feature
at ∼1400 nm is related to the absorbance of hexane.
J. AM. CHEM. SOC. 9 VOL. 130, NO. 49, 2008
Synthesis of Nanocrystal Inks
The current-voltage characteristics and the incident photon
conversion efficiency (IPCE) of a typical PV device made with
CuInSe2nanocrystals are shown in Figure 14. The measured
power conversion efficiencies (η) of 32 devices ranged between
0.01 and 0.24%. The IPCE matches approximately the absor-
bance spectra of the CuInSe2 nanocrystals (Figure 14b),
confirming that the device response results from the nanocrys-
tals. The relatively high IPCE of ∼22% for wavelengths between
400 and 500 nm tails off at higher wavelengths. The long-
wavelength IPCE cutoff at ∼1050 nm corresponds approxi-
16776 J. AM. CHEM. SOC. 9 VOL. 130, NO. 49, 2008
mately to the optical gap of the CuInSe2 nanocrystals as it
should, and the sharp drop in IPCE at wavelengths <400 nm is
the result of ZnO light absorptionsthe ZnO layer is essentially
serving as a photon cutoff filter in the device.
The photovoltaic response of devices made from CuInSe2
nanocrystal layers was reproducible and demonstrated that these
nanocrystals have potential as light absorbing materials in PVs.
However, the PV efficiencies in these particular devices are
relatively low and require significant improvement for practical
applications. Device efficiencies might be improved increasing
the CuInSe2film thickness to absorb more photons, and the
device structures themselves are relatively complicated with
many factors that can decrease efficiency. The open circuit
voltages (Voc) of the CuInSe2nanocrystal devices were actually
quite reasonable, typically near 300 mV, which is getting close
to the high-efficiency vapor-deposited CuInSe2devices (typical
Vocvalues are ∼400 mV).38The short circuit current densities
(Jsc) and fill factors (FF), however, were quite low, with typical
Jscvalues of ∼3 mA cm-2(compared to Jscof ∼35 mA cm-2
for the highest efficiency (19%) vapor-deposited CIGS device)10
and FFs close to 0.25. The diode response was also relatively
poor (see Supporting Information for data and analysis), with
an ideality factor (A) much larger than 1, revealing that the
device has high series and low shunt resistances.39The high
series resistance is partly attributed to high ITO sheet resistances
(>300 Ω/0) and relatively resistive nanocrystal filmssfour-
point probe measurements gave resistivities of approximately
1 kΩ·cm, which are about 3 orders of magnitude more resistive
than conventional CIGS films with good photovoltaic efficien-
cies.41High shunt conductance (or low shunt resistance) in the
(38) Stolt, L.; Hedstrom, J.; Kessler, J.; Ruckh, M.; Velthaus, K.-O.; Schock,
H.-W. Appl. Phys. Lett. 1993, 62, 597–599.
(39) van Dyk, E. E.; Meyer, E. L. Renew. Energy 2004, 29, 333–344.
Figure 12. The thickness determined by profilometry of CuInSe2nano-
crystal films drop-cast from TCE dispersions with different concentrations.
The SEM image shows a cross-section of a 1.7 µm thick nanocrystal film
on Mo-coated soda lime glass.
Figure 13. Photograph of (a) a CuInSe2nanocrystal dispersion and (b)
the deposition of thin films on an array of glass substrates. After depositing
the films, the substrates were placed in a vacuum oven at room temperature
for 12 h.
Figure 14. (a) Current-voltage characteristics and (b) IPCE spectra of a
CuInSe2nanocrystal photovoltaic device (solid line) with absorbance curves
of CuInSe2and ITO/ZnO layers (dashed). The IPCE spectrum was measured
at zero bias. The nanocrystal absorber layer was 700 nm thick, consisting
of oleylamine-capped CuInSe2nanocrystals with an average diameter of
15 nm. The measured short circuit current density in panel a corresponded
to within a few percent of the integrated IPCE spectra in panel b multiplied
by the AM 1.5 solar spectrum (i.e., the total number of photons converted
to electrons by the device), as it should.
Panthani et al.
devices can result from many factors, including holes or cracks Download full-text
in the nanocrystal film and penetration of the CdS or sputtered
ZnO layers to the back contact.
Synthetic methods for CuInS2, CuInSe2, and CIGS nano-
crystals are described, as well as deposition approaches for
achieving uniform, relatively thick (∼1 µm) nanocrystal layers.
Oleylamine was found to be an effective CuInS2 and CIGS
capping ligand. However, the stabilization of the CIGS nano-
crystals with high Ga content still needs to be improved as the
dispersion stability and quality of the nanocrystals decreased
with increased Ga content.
Films of CuInSe2nanocrystals used as the absorber layer in
conventional layered Mo/CuInSe2/CdS/ZnO/ITO PV devices
gave reproducible photovoltaic responses with power conversion
efficiencies up to ∼0.2% and IPCE as high as 22% for photons
with 400-500 nm wavelength. These devices provide a baseline
performance and demonstrate as a proof-of-concept that these
nanocrystals can be used in PVs. Practical devices, however,
require higher efficiencies. There are many ways to try to
improve PV efficiency, including using nanocrystals with shorter
chain capping ligands, incorporating Ga into the films, and using
various chemical or thermal treatments of the nanocrystal layers
to increase their conductivity.42-48New device architectures
that are more suitable to using nanocrystal absorber layers and
low-temperature manufacturing steps may also provide ways
to increase device efficiency and eliminate the need for high
temperature processing. These are all topics for further study.
Acknowledgment. This research was supported in part by
funding from the National Science Foundation through their STC
program (Grant CHE-9876674), the Robert A. Welch Foundation,
and the Air Force Research Laboratory (FA8650-07-2-5061).
Supporting Information Available: CIGS nanocrystal EDS
data, schematic of PV device structure, thickness profile of dip-
coated nanocrystal film, SEM images of drop-cast nanocrystal
films, PV device diode data and analysis. This material is
available free of charge via the Internet at http://pubs.acs.org.
(40) Wei, S.-H.; Zhang, S. B.; Zunger, A. Appl. Phys. Lett. 1998, 72, 3199–
(41) Wang, X.; Li, S. S.; Kim, W. K.; Yoon, S.; Craciun, V.; Howard,
J. M.; Easwaran, S.; Manasreh, O.; Crisalle, O. D.; Anderson, T. J.
Sol. Energy Mater. Sol. Cells 2006, 90, 2855–2866.
(42) Yu, D.; Wang, C.; Guyot-Sionnest, P. Science 2003, 300, 1277–1280.
(43) Talapin, D. V.; Murray, C. B. Science 2005, 310, 86–89.
(44) Urban, J. J.; Talapin, D. V.; Shevchenko, E. V.; Kagan, C. R.; Murray,
C. B. Nat. Mater. 2007, 6, 115–121.
(45) Jarosz, M. V.; Porter, V. J.; Fisher, B. R.; Kastner, M. A.; Bawendi,
M. G. Phys. ReV. B 2004, 70, 195327.
(46) Konstantatos, G.; Howard, I.; Fischer, A.; Hoogland, S.; Clifford, J.;
Klem, E.; Levina, L.; Sargent, E. H. Nature 2006, 442, 180–183.
(47) Murphy, J. E.; Beard, M. C.; Nozik, A. J. J. Phys. Chem. B 2006,
(48) Law, M.; Luther, J. M.; Song, Q.; Hughes, B. K.; Perkins, C. L.; Nozik,
A. J. J. Am. Chem. Soc. 2008, 130, 5974–5985.
J. AM. CHEM. SOC. 9 VOL. 130, NO. 49, 2008
Synthesis of Nanocrystal Inks