Enhanced luminescence of Cu-In-S nanocrystals

by surface modification
Young-Kuk Kim, Young-Sang Cho, Kookchae Chung, Chul-Jin Choi
1Division

of Functional Materials,Korea Institute of Materials Science (KIMS), Changwon 641-010, Korea

We have synthesized highly luminescent Cu-In-S(CIS) nanocrystals (NCs) by heating the

mixture of metal carboxylates and alkylthiol under inert atmosphere. We modified the surface
of CIS NCs with zinc carboxylate and subsequent injection of alkylthiol. As a result of the
surface modification, highly luminescent CIS@ZnS core/shell nanocrystals were synthesized.
The luminescence quantum yield (QY) of best CIS@ZnS NCs was above 50%, which is 10
times higher than the initial QY of CIS NCs before surface modification (QY=3%). Detailed
study on the luminescence mechanism implies that etching of the surface of NCs by
dissociated carboxylate group (CH3COO-) and formation of epitaxial shell by Zn with sulfur
from alkylthiol efficiently removed the surface defect which are known to be major nonradiative recombination sites in semiconductor nanocrystals. In this study, we developed a
novel surface modification route for monodispersed highly luminescent Cu-In-S NCs with less
toxic and highly stable precursors.

Keywords: CuInS2, nanocrystals, ZnS, surface

1. INTRODUCTION
Colloidal semiconductor nanocrystals (NCs) or
quantum dots (QDs) based on semiconductors open
possibilities for applications in new areas such as light
emitting diodes, solar cells, and biological imaging,
etc.1-3 As alternatives to II-VI semiconductors, several
materials system including III-V semiconductor NCs
(e.g. InP)4 and impurity-doped ZnS NCs5 have been
developed. However, each materials systems have
some drawbacks such as requirement of expensive
pyrophoric compounds as precursors, chemical stability,
tunability of emission wavelength, etc.6 Ternary
chalcopyrite-type I-III-VI2 group semiconductor NCs
were also synthesized as less-toxic new candidates to
II-VI semiconductor NCs.6-10 I-III-VI2 group
semiconductors with appropriate band-gap energies
(e.g. 1.5eV for CuInS2 and 1.68eV for CuGaSe2) are
direct transition semiconductors favoring ionic
bondings similar to II-VI semiconductors. However, in
contrast to II-VI group semiconductors which make
luminescence by direct exciton recombination, the
dominant
luminescence
mechanism
for
the
chalcopyrite-type I-III-VI2 group semiconductor NCs
in the visible spectrum is known to be donoracceptor pair recombination (DAP) which requires
defects to play an important role in the luminescence

process.6 Recently, several results have been reported

for synthesis of non-aggregated I-III-VI2 group
semiconductor NCs of a few nanometers with
fluorescence quantum yield of a few percent.7-9 Many
studies have shown that surface overcoating of I-IIIVI2 group semiconductor NCs by epitaxial growth of
higher band gap semiconductor layers can enhance
their emission efficiency.6,8 Such core/shell structure
can effectively improve photoluminescence quantum
yield
by
passivating
surface
non-radiative
recombination sites. The recent report by Li et al. on
synthesis of highly luminescent CuInS2/ZnS NCs is
quite impressive.10 They improve the fluorescence
quantum yield by forming ZnS layer on the surface of
CuInS2 cores using zinc ethylxanthate. As an
alternative to the coreshell approach, Micic et al. have
proposed the etching of InP nanocrystals with HF as a
way to improve their luminescence efficiency11 and
Talapin et al. have shown that HF treatment combined
with
size-selective
photoetching resulted
in
monodisperse InP colloids with room temperature
quantum yields up to 40%.12 In particular, Ryu et al.
reported a novel shell formation approach using zinc
acetates for high luminescence of InP nanocrystals.13
The addition of zinc acetate was proved to be effective
on enhancement of luminescence by etching the

surface of InP cores and by passivating surface

dangling bonds through forming ZnS shell.
In this work, we intended to report a one-pot
synthesis route to highly luminescent CuInS2/ZnS coreshell NCs with high quantum yield through defect
control and shell formation by refluxing with zinc
acetate and palmitic acid. Zinc acetate played an
important role in formation of inorganic shell and size
regularization of CuInS2 (CIS) NCs. We showed that
Cu-deficiency in precursor solution is essential to
enhancing quantum yield above 60% and surface
modification by refluxing with zinc acetate and
palmitic acid enables a reproducible synthesis of
monodispersed highly luminescent CuInS2 NCs with
less toxic and highly stable precursors.

2. EXPERIMENTAL
Cu1-xInS2 NCs were prepared with the method
described in ref. 7 based on Cu(II) acetate, In(III)
acetate, 1-dodecanethiol (DDT) as precursors for Cu,
In, S. The amount of In(III) acetate and 1dodecanethiol (DDT) were fixed to 1mmol and 5ml (ca.
21mmol), respectively. The concentration of DDT
was intentionally maintained to be much larger than
indium and copper precursor to minimize formation of
CuxS.6 The amount of Cu(II) acetate was varied from
0.2mmol to 1mmol. Those precursors were dispersed in
30ml of 1-octadecene under inert atmosphere at room
temperature and subsequently heated to 240oC. After
reaction for 15 minutes, the solution was cooled to
room temperature.
To construct core-shell structure, 1mmol of zinc acetate
was added to the CuInS2-based core solution after
cooling to room temperature without purification and
separation of synthesized NCs. In some experiment, we
substitute cadmium acetate for small part of zinc
acetate maintaining total amount of metal acetate
constant to be 1mmol. For colloidal stability of NCs,
3mmol of palmitic acid mixed with 10ml of 1octadecene was also added to the solution.
Subsequently, the mixture was heated to 230oC under
inert N2 atmosphere and reacted without further
addition of precursors. The mixed solution became
uniform at above 120oC and further dwelling at 230oC
to enhance the luminescence. After refluxing for 5
hours, DDT with quantity equivalent to zinc acetate
was injected to the solution and maintained for an hour
to form ZnS shell at the same temperature. Aliquots
were taken at different time interval and rapidly cooled

to room temperature. Before further characterization,

polar solvent such as acetone, methyl alcohol, etc. was
added to precipitate Cu-In-S2-based NCs and the
precipitates were purified by repetition of
centrifugation and re-dispersion to n-hexane.

3. RESULTS AND DISCUSSION

3.1. Effect of Cu-deficiency in the starting solution
Reports have shown that the main fluorescence process
in CIS NCs is closely related to lattice defects such as
copper vacancy, indium substitute anti-site defect,
etc.6,8,14-16 This implies that controlling crystal defect
can enhance the fluorescence of CuInS2 nanocrystals
and intentional formation of some crystal defects was
proved to be efficient to high fluorescence quantum
yield.14 Fig. 1(a) shows various fluorescence spectra
obtained from Cu1-xInS2 (x = 0 ~ 0.8) NCs with various
different starting compositions.

Figure 1. Photoluminescence (PL) and (b) Raman spectra of Cu 1xInS2 nanocrystals prepared with different degree of copper
deficiency in the starting solutions. (c-d) TEM images of Cu1-xInS2
nanocrystals with different starting composition of x=0 and x=0.8,
respectively.

The luminescence intensity of CIS NC was highly

enhanced and the wavelength of emission spectra
showed blue-shift with minimizing copper content in
the starting solution. As mentioned earlier,
incorporation of crystal defect in CIS NCs by highly
Cu-deficient composition in the starting solution can be
ascribed to the source of improved luminescence.
Uehara et al. explained the blue-shift of fluorescent
spectra by the lowering of valence band due to Cudeficiency in CIS NCs.14 The composition of CIS NCs
prepared from precursor solution with Cu:In=0.2:1 was
measured to be Cu:In=0.41:1 from induction-coupled

plasma-mass spectrometry (ICP-MS).

Thus, Cudeficiency in starting solution resulted in large Cu
deficiency in CIS NCs. In addition, the Raman spectra
of Cu-deficient CIS NCs displayed in Fig. 1(b) show a
characteristic peak of CuAu-type ordering (CA) at
340cm-1 as well as chalcopyrite (CH) peak at 300cm1
. Since the CA peak indicates the clustering of defect
pair at high concentration of defects,14,17-19 CA peak in
the Raman spectrum of Cu-deficient Cu-In-S NCs also
supports a large number of crystal defects induced by
copper deficiency in the starting solutions. In particular,
the more intense CA peaks compared with CH
peaks in the Raman spectra implied that the highly Cudeficient composition in the starting solution gave rise
to higher concentration of lattice defects in the
synthesized NCs. As observed in the TEM image in Fig.
1(c) & (d), the grown NCs showed quite irregular
shapes and broad size distribution irrespective of Cudeficiency in the starting solutions. According to Zhong
et al.s work, CuInS2 NCs were synthesized from the
thermal decomposition of intermediate compound,
CuIn(SC12H25)x complex and prepared NCs were
highly aggregated due to the oriented attachment of
small particles.7 As a result of irregular shape and polydispersed size distribution, the absorption spectrum of
NCs showed broad shoulder with long wavelength trail
and photoluminescence excitation spectrum (PLE)
showed broad intensity distribution shown in Fig. 2(a).
3.2. Surface modification of copper indium sulfide
nanocrystals
The non-radiative recombination of electron-hole pairs
by the surface states has been attributed to the major
origin of the low quantum yield in semiconductor NCs.
In addition, the passivation of surface defects by
further shell growth on the semiconductor NCs was
proved to be effective for high luminescence properties.
6,10
Ryu et al. have suggested that zinc acetate acts as a
source of ZnS shell on InP and acetic acid produced
from dissociation of zinc acetate etches the defective
surface of as-synthesized InP cores.13 In this study,
impressed by Ryu et al.s work, we tried to improve the
photoluminescence (PL) of CIS NCs by modifying
their surface through refluxing with zinc acetate and
palmitic acid. Fig. 2(b) shows optical spectra of CIS
NCs after treatment with zinc acetate and palmitic acid
for 5 hours at 230oC. The long wavelength trail in the
absorption spectrum was efficiently suppressed and
photoluminescence excitation spectrum (PLE) showed
narrow intensity distribution.

Figure 2. Absorption, PL and PLE of (a) as-synthesized Cu1-xInS2

nanocrystals (x=0.2) and (b) Cu1-xInS2 nanocrystals (x=0.2) after
refluxing with zinc acetate and palmitic acid for 5 hours at 230oC.

Since the long wavelength trail in the absorption

spectra and broad PLE is mainly caused by
inhomogeneous size distribution, more uniform size
distribution of NCs was obtained after refluxing with
zinc acetate and palmitic acid as inferred from the
narrow absorption band in the optical spectra. In
addition, TEM images of NCs shown in Fig. 3(a) also
support narrowing of the size distribution after
refluxing with zinc acetate and palmitic acid. The low
resolution image showed equally spaced dots, and
oblate nanocrystals with crystalline lattice was
observed from the high resolution image. Thus, the
refluxing of CIS core solution with zinc acetate and
palmitic acid resulted in a digestive ripening of CIS
NCs. The aggregated CIS NCs were supposed to be
broken into quasi-monodispersed NCs by etching with
acetic acid produced from the reaction between zinc
acetate and palmitic acid.13
Fig. 2(b) shows temporal evolution of the
photoluminescence spectra of CIS NCs after refluxing
with zinc acetate and palmitic acid at 230 oC. The
fluorescence quantum yield of NCs was highly
enhanced to be above 40%, which was initially 3% for
pristine CIS NCs. In particular, the quantum yield was
improved up to 56% by further injection of DDT and
subsequent refluxing for an hour at the same
temperature. Since CIS core solution contained large
amount of sulfur to suppress CuxS formation [6,7] and
as-synthesized CIS cores were refluxed without
removing unreacted precursors, formation of ZnS shell

is expected even without further injection of sulfur

precursor such as DDT during the refluxing. In
addition, the crystal structure of CuInS2 and ZnS is
quite similar and lattice mismatch between them is
about 2%. Thus, as a result of passivation of surface
defects and confinement of charge carriers by coherent
shell formation, fluorescence of the CIS NCs was
intensified by simple refluxing with zinc acetate and
palmitic acid.13 Moreover, we suppose that etching of
defect-rich surface with acetic acid also contributed to
high quantum yield in CIS NCs refluxed with zinc
acetate and palmitic acid.

acid and defect control through highly offstoichiometric starting composition of precursor
solution. The best fluorescence quantum yield was 56%,
which was obtained from core shell nanocrystals
consisting of Cu-deficient core and ZnS shell formed
by prolonged refluxing with zinc acetate and palmitic
acid and subsequent injection of DDT. In addition,
formation of ZnS shell layer on CIS NCs change the
thermal dependence of the integrated PL intensity from
thermal quenching to thermal antiquenching with
increasing temperatures. In particular, formation of
thick shell by refluxing with zinc acetate and palmitic
acid resulted in the drastic change in thermal behavior
of PL intensity for CIS NCs.

References

Figure 3. (a) TEM images of Cu1-xInS2 nanocrystals (x=0.2) treated

with zinc acetate and palmitic acid for 5 hours at 230oC. (b) Temporal
evolution of PL of Cu1-xInS2 nanocrystals (x=0.2) (c) PL of Cu1-xInS2
NCs with various starting composition (x=0, 0.5, 0.8) after refluxing
with zinc acetate and palmitic acid for 5 hours at 230oC.

We have checked the effect of Cu-deficiency on the

luminescence of CIS/ZnS NCs. Fig. 3(c) shows PL
spectra of CIS/ZnS NCs prepared with different copper
content in the starting solutions. The quantum yield of
Cu1-xInS2/ZnS NCs was measured to be 3%, 16%, 70%
for x=0, x=0.5, x=0.8, respectively. This implies that
refluxing CIS NCs with zinc acetate and palmitic acid
does not remove copper defects located in the interior
of CIS NCs, but passivate surface defects of the
nanocrystals. Thus, it is concluded that Cu-deficiency
is essential to high quantum yield of CIS NCs even
after refluxing with zinc acetate and palmitic acid.

4. CONCLUSIONS
In conclusion, we have synthesized highly luminescent
CIS NCs by refluxing with zinc acetate and palmitic

1. M. Bawendi, M. Steigerwald, L. Brus, Ann. Rev. Phys. Chem. 41

(1990) 477.
2. A.P. Alivisatos, Science 271 (1996) 933.
3. G. Markovich, C. Collier, S. Henrichs, F. Remacle, R. Levine, J. R.
Heath, Acc. Chem.. Res. 32 (1999) 415.
4. S. Xu, J. Ziegler, T. Nann, J. Mater. Chem. 18 (2008) 2653.
5. N. Murase, R. Jagannathan, Y. Kanematsu, M. Watanabe, A. Kurita,
6. K. Hirata, T. Yazawa, T. Kushida, Nature 259 (2009) 686.
6. R. Xie, M. Rutherford, X. Peng, Chem. Mater. 20 (2008) 531.8.
7. H. Zhong, Y. Zhou, Y. Mingfu, Y. He, J. Ye, C. He, C. Yang, Y. Li,
Chem. Mater. 20 (2008) 6434.
8. K. Nose, Y. Soma, T. Omata, S. Otsuka-Yao-Matsuo, Chem. Mater.
21 (2009) 2607.
9. H. Zhong, Zhou, Y. Mingfu, Y. He, J. Ye, C. He, C. Yang, Y. Li,
Nanotechnology 18 (2008) 025602.
10. L. Li, J. Daou, I. Texier, T. T. K. Chi, N. Q. Liem, P. Reiss, Chem.
Mater. 21 (2009) 2422.
11. L. Micic, C. J. Curtis, K. M. Jones, J. R. Sprague, A. J. Nozik, J.
Phys. Chem. 98 (1994) 4966.
12. D. V. Talapin, N. Gaponik, H. Borchert, A. L. Rogach, M. Haase,
H. Weller, J. Phys. Chem. 106 (2002) 12659.
13. E. Ryu, S. Kim, E. Jang, S. Jun, H. Jang, B. Kim, S. Kim, Chem.
Mater. 4 (2009) 573.
14. M. Uehara, K. Watanabe, Y. Tajiri, H. Nakamura, H. Maeda, J.
Chem. Phys. 129 (2008)134709; b) W. Kato, M. Uehara, K. Nose,
T. Omata, S. Otsuka-Yao-Matsuo, M. Miyazaki, H. Maeda, Chem.
Mater. 18 (2006) 3330-3335.
15. K. Watanabe, M. Uehara, H. Nakamura, H. Maeda, Mater. Res.
Soc. Symp. Proc. 1064 (2008) PP03-19.
16. H. Nakamura, W. Kato, M. Uehara, K. Nose, T. Omata, S.
Otsuka-Yao-Matsuo, M. Miyazaki, H. Maeda, Chem. Mater. 18
(2006) 3330.
17. M. Nanu, J. Schoonman, A. Goosens, Thin Solid Films 451-452
(2004) 193.
18. J. Hofhuis, J. Schoonman, A. Goossens, J. Phys. Chem. C 112
(2008) 15052.
19. J. Alvarez-Garcia, B. Barcones, A. Perez-Rodrigues, A. RomanoRodrigues, J. R. Morante, A. Janotti, S. H. Wei, R. Scheer, Phys.
Rev. B 71 (2005) 054303