Probing the surface of transition-metal nanocrystals by chemiluminesence

Article abstract of DOI:10.1021/ja102413k

We propose a simple chemiluminescence (CL) method for investigation of the surface of Co-based nanocrystals (NCs). Using a combination of CL and spin-trap electron paramagnetic resonance techniques, we systematically studied the generation of reactive oxygen species (ROS) at the surface of differently sized CoPt₃ spherical NCs and CoPt₃/Au nanodumbbells. We have shown that differently sized CoPt₃ NCs can promote the formation of ROS and as a result can lead to the oxidation of luminol accompanied by the emission of the light. CL allows monitoring the stability of transition-metal-based NCs against oxidation and dissolution. We found by CL that cobalt ions slowly leach from the surface of CoPt₃ NCs even under very mild conditions; however, the amount of the leached cobalt ions does not exceed the maximal concentration of cobalt at the NC surface indicating that only surface atoms can go into solution.

Full text of DOI:10.1021/ja102413k

Published on Web 06/15/2010
Probing the Surface of Transition-Metal Nanocrystals by
Chemiluminesence
†
†,‡
†,§
†
Galyna Krylova, Nada M. Dimitrijevic, Dmitri V. Talapin, Jeffrey R. Guest,
|
⊥
†
,†
Holger Borchert, Arun Lobo, Tijana Rajh, and Elena V. Shevchenko*
Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439,
Chemical Sciences and Engineering DiVision, Argonne National Laboratory, Argonne, Illinois
6
0439, Department of Chemistry, UniVersity of Chicago, Chicago, Illinois 60637, Energy and
Semiconductor Research laboratory, Department of Physics, UniVersity of Oldenburg,
Oldenburg, 26111, Germany, and HASYLAB at DESY, Hamburg, D-22607, Germany
Received March 23, 2010; E-mail: eshevchenko@anl.gov
Abstract: We propose a simple chemiluminescence (CL) method for investigation of the surface of Co-
based nanocrystals (NCs). Using a combination of CL and spin-trap electron paramagnetic resonance
techniques, we systematically studied the generation of reactive oxygen species (ROS) at the surface of
differently sized CoPt
CoPt NCs can promote the formation of ROS and as a result can lead to the oxidation of luminol
accompanied by the emission of the light. CL allows monitoring the stability of transition-metal-based NCs
against oxidation and dissolution. We found by CL that cobalt ions slowly leach from the surface of CoPt
3 3
spherical NCs and CoPt /Au nanodumbbells. We have shown that differently sized
3
3
NCs even under very mild conditions; however, the amount of the leached cobalt ions does not exceed the
maximal concentration of cobalt at the NC surface indicating that only surface atoms can go into solution.
Introduction
Recent progress in colloidal synthesis provided access to many
types of materials in the form of monodisperse particles with
precisely controlled size, shape, structure, and composition.
Transition-metal-based nanocrystals (NCs) are of great inter-
est because of their interesting magnetic, magnetoelectronic,
and catalytic properties. They are also very promising for
biomedical applications in cancer treatment. Alloys of Pt with
14-18
1
,2
3-5
6
,7
Due to their small size, the surface of NCs plays a substantial
role in the properties of NCs. Some of the application areas are
extremely sensitive to the electronic states and composition of
the surface of NCs. For example, the electronic structure and
chemical composition of the inorganic core and organic shell
affect the ability of NC to serve as a seed material for the growth
of complex multicomponent systems such as binary NCs of
8,9
3
d transition metals have demonstrated high catalytic activity
1
0-13
in the oxygen reduction
improvement of fuel cells.
that is very important for further
†
Center for Nanoscale Materials, Argonne National Laboratory.
Chemical Sciences and Engineering Division, Argonne National
‡
1
9
20
21
Laboratory.
FePt/Fe O
x y
,
3 x y
CoPt /Au, and FePt/CdS and Fe O NCs
§
Department of Chemistry, University of Chicago.
Department of Physics, University of Oldenburg.
HASYLAB at DESY.
|
⊥
(
1) Lacroix, L. M.; Malaki, R. B.; Carrey, J.; Lachaize, S.; Respaud, M.;
Goya, G. F.; Chaudret, B. J. Appl. Phys. 2009, 105, 023911.
2) Zafiropoulou, I.; Devlin, E.; Boukos, N.; Niarchos, D.; Petridis, D.;
Tzitzios, V. Chem. Mater. 2007, 19, 1898–1900.
(12) Chen, W.; Kim, J. M.; Sun, S. H.; Chen, S. W. J. Phys. Chem. C
2008, 112, 3891–3898.
(
(13) Wang, C.; van der Vilet, D.; Chang, K. C.; You, H. D.; Strmcnik, D.;
Schlueter, J. A.; Markovic, N. M.; Stamenkovic, V. R. J. Phys. Chem.
C 2009, 113, 19365–19368.
(
(
(
3) Liu, C.; Wu, X. W.; Klemmer, T.; Shukla, N.; Weller, D. Chem. Mater.
2
005, 17, 620–625.
(14) Murray, C. B.; Sun, S. H.; Doyle, H.; Betley, T. MRS Bull. 2001, 26,
985–991.
4) Salgueirino-Maceira, V.; Liz-Marzan, L. M.; Farle, M. Langmuir 2004,
2
0, 6946–6950.
(15) Shevchenko, E. V.; Talapin, D. V.; Rogach, A. L.; Kornowski, A.;
Haase, M.; Weller, H. J. Am. Chem. Soc. 2002, 124, 11480–11485.
(16) Redl, F. X.; Black, C. T.; Papaefthymiou, G. C.; Sandstrom, R. L.;
Yin, M.; Zeng, H.; Murray, C. B.; O’Brien, S. P. J. Am. Chem. Soc.
2004, 126, 14583–14599.
5) Sun, S. H.; Anders, S.; Thomson, T.; Baglin, J. E. E.; Toney, M. F.;
Hamann, H. F.; Murray, C. B.; Terris, B. D. J. Phys. Chem. B 2003,
1
07, 5419–5425.
(
(
(
(
6) Chen, W.; Kim, J. M.; Xu, L. P.; Sun, S. H.; Chen, S. W. J. Phys.
Chem. C 2007, 111, 13452–13459.
(17) Nandwana, V.; Elkins, K. E.; Poudyal, N.; Chaubey, G. S.; Yano, K.;
Liu, J. P. J. Phys. Chem. C 2007, 111, 4185–4189.
7) Jacinto, M. J.; Kiyohara, P. K.; Masunaga, S. H.; Jardim, R. F.; Rossi,
L. M. Appl. Catal. A: Gen. 2008, 338, 52–57.
(18) Monnier, V.; Delalande, M.; Bayle-Guillemaud, P.; Samson, Y.; Reiss,
P. Small 2008, 4, 1139–1142.
8) Choi, J. S.; Choi, H. J.; Jung, D. C.; Lee, J. H.; Cheon, J. Chem.
Commun. 2008, 2197–2199.
(19) Figuerola, A.; Fiore, A.; Di Corato, R.; Falqui, A.; Giannini, C.;
Micotti, E.; Lascialfari, A.; Corti, M.; Cingolani, R.; Pellegrino, T.;
Cozzoli, P. D.; Manna, L. J. Am. Chem. Soc. 2008, 130, 1477–1487.
(20) Pellegrino, T.; Fiore, A.; Carlino, E.; Giannini, C.; Cozzoli, P. D.;
Ciccarella, G.; Respaud, M.; Palmirotta, L.; Cingolani, R.; Manna, L.
J. Am. Chem. Soc. 2006, 128, 6690–6698.
9) Xu, C. J.; Yuan, Z. L.; Kohler, N.; Kim, J. M.; Chung, M. A.; Sun,
S. H. J. Am. Chem. Soc. 2009, 131, 15346–15351.
(
(
10) Yang, H.; Alonso-Vante, N.; Lamy, C.; Akins, D. L. J. Electrochem.
Soc. 2005, 152, A704–A709.
11) Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J. J.; Lucas,
C. A.; Wang, G. F.; Ross, P. N.; Markovic, N. M. Nat. Mater. 2007,
(21) Gu, H. W.; Zheng, R. K.; Zhang, X. X.; Xu, B. J. Am. Chem. Soc.
2004, 126, 5664–5665.
6
, 241–247.
9
102 9 J. AM. CHEM. SOC. 2010, 132, 9102–9110
10.1021/ja102413k  2010 American Chemical Society

Probing the Surface of Transition-Metal Nanocrystals
A R T I C L E S
22
x y 1 2
decorated with two differently sized Au NCs (Fe O /Au /Au ).
3 3
(i) differently sized CoPt NCs; (ii) Co NCs and (iii) CoPt /Au
These transition-metal-based NCs have demonstrated enhanced
nanodumbbells.
1
3,23
electrocatalytic properties.
Unfortunately, very little is
known about the mechanism of the nucleation and growth of
such heterostructures, and their synthesis is rather empirical.
The general concept of growth of such materials is based on
the nucleation of the second material on preformed seeds, and
hence, the surface of seeds and its stability play the crucial role.
Materials and Methods
2
Chemicals. Platinum acetylacetonate (Pt (acac) , Acros Organics,
9
8%), 1-adamantanecarboxylic acid (ACA, 99%, Aldrich), hexa-
decylamine (HDA, 90%, Aldrich), diphenyl ether (99%, Aldrich),
cobalt carbonyl (Co (CO) , stabilized with 1-5% of hexane, Strem),
,2-dichlorobenzene (anhydrous, 99%, Aldrich), trioctylphosphine
oxide (TOPO, 99% Aldrich), oleic acid (OA, 90%, Aldrich),
platinum(II) chloride (PtCl , 99.999%, Aldrich), didodecyldi-
methylammonium bromide (DDAB, 98%, Fluka), NaBH (98.5%,
2
8
1
Chemical stability of the surface of transition-metal-based
NCs is a very important issue for both catalytic and biomedical
2
8
,9
applications. For example, cobalt-based NCs have higher
4
24
saturation magnetization as compared to iron oxides currently
used as agents for MRI and hypothermia treatments; however,
the potential leakage of cobalt ions is a serious concern that
should be addressed. Cobalt ions can easily replace Zn and Mg
active sites in coenzymes and lead to the generation of reactive
Aldrich), toluene (99.5%, Aldrich), chloroform (99.9%, Aldrich),
ethanol (anhydrous, ACS grade, Pharmaco-Aaper), 2-propanol
(99.9%, Fisher Scientific), methanol (99.9% Riedel-de Haen), 1,2-
dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(poly-
ethylene glycol)-2000] (PEG-PE) dissolved in chloroform (10 mg/
mL solution, Avanti Polar Lipids, Inc.), 5,5-dimethyl-1-pyrroline
25,26
oxygen species (ROS) that can provoke cancer.
hand, the controlled release of metal ions from the surface of
NCs capable of catalyzing the decomposition of H with the
On the other
n-oxide (DMPO, 97%, Sigma), CoCl
2
·6H
2
O (99.9%, Aldrich),
, 99.99%, Aldrich)
H
2
O
2
(30% Aldrich), and gold chloride (AuCl
3
2
O
2
were used as received without further purification.
formation of ROS can be also beneficial. Thus, FePt NCs
functionalized with cancer-targeting antibodies exhibited target-
enhanced cytotoxicity to the tumor cells by release of iron ions.
Synthesis of CoPt NCs. CoPt NCs of different sizes were
3
3
synthesized according to the modified protocol described in ref 32.
Briefly, Pt (acac) (33 mg) and ACA (0.25 g) were dissolved at
5 °C under nitrogen flow in a mixture of HDA (4 g) and diphenyl
(CO)
9
2
5
In addition, transition-metal-based nanostructures are of great
ether (2 mL). Then the solution was heated to 170 °C, and Co
2
8
2
7
interest for photochemical and electrochemical water split-
dissolved in ∼1 mL of 1,2-dichlorobenzene was injected. The size
28-30
ting.
These and other catalytic applications require control
of CoPt NCs was controlled by the amount of Co (CO) . Thus,
3
2
8
over chemical composition and electronic states of nanocrystals
as well as their high stability.
0.106 g, 0.09 g, 0.068 g, 0.043 g, and 0.027 g of Co
used to synthesize 4 nm, 4.8 nm, 6 nm, 8 nm, and 10.5 nm CoPt
2
8
(CO) were
3
NCs, respectively. No 1,2-hexadecanediol was used as compared with
ref 32. After injection of cobalt carbonyl solution, the reaction mixture
was kept at 170 °C for 1 h. NCs were annealed at ∼246 °C for 15-20
min, and then the reaction solution was cooled to 70 °C and 5 mL of
chloroform was injected. The solution was then cooled to room
temperature under ambient conditions.
Traditional physical methods of surface analysis such as X-ray
photoelectron or X-ray absorption spectroscopy were proven
to provide insights into composition and electronic states of the
nanocrystals. However, these analytic methods do not probe
the surface layers exclusively and usually average the informa-
tion over more than one monolayer. In addition, spectroscopic
methods require certain sample preparation and data analysis
that lead to relatively slow sample turnover and thus cannot be
utilized for routine monitoring of the surface in time.
3
1
Synthesis of CoPt
bells of CoPt -Au were prepared via the modified procedure
described in ref 33: AuCl (8 mg), DDAB (54 mg), and HDA (74
3
-Au Dumbbell-like Heterodimers. Dumb-
3
3
mg) were dissolved in 4 mL of toluene by sonication for 15 min in
the ultrasonic bath until a clear yellow solution was formed. A
toluene solution of CoPt
diluted to 3 mL and heated to 95 °C in a three-neck flask under
nitrogen. AuCl solution was then added dropwise (1 mL/h) via a
syringe pump, and the solution was kept at 95 °C for an additional
h. The size of the Au portion of the nanodumbbells was found to
3
containing 1.15 mg of 8 nm NCs was
We report that the surface of metal nanocrystals can be probed
by a chemiluminesence technique that allows us to monitor the
presence of the transition-metal atoms at the NC surface and to
detect their leaching into the solution. In this work, we also
studied the ability of alloyed NCs to initiate the formation of
ROS. Three different types of cobalt-based NCs were examined:
3
2
be ∼11 nm.
Synthesis of Co NCs. Co NCs were synthesized according to
3
4
the method proposed by Puntes et al. TOPO (0.05 g) and OA
0.05 mL) were dissolved in 1,2-dichlorobenzene (7.5 mL) and
heated to ∼182 °C under continuous stirring. Then, the solution
containing Co (CO) (0.27 g) and 1,2-dichlorobenzene (1.5 mL)
(
(
(
(
(
(
22) Wang, C.; Wei, Y. J.; Jiang, H. Y.; Sun, S. H. Nano Lett. 2009, 9,
544–4547.
23) Zhang, J.; Sasaki, K.; Sutter, E.; Adzic, R. R. Science 2007, 315, 220–
22.
24) Jeong, U.; Teng, X. W.; Wang, Y.; Yang, H.; Xia, Y. N. AdV. Mater.
4
2
8
was quickly injected, and the reaction mixture was kept at the reflux
temperature for 5 min and then quickly cooled with water bath.
Synthesis of 4.5 nm Pt NCs. PtCl (0.03 g) and DDAB (0.1 g)
2
2
2
007, 19, 33–60.
25) Hartwig, A. 4th Congress of Toxicology in DeVeloping Countries;
were dissolved in 10 mL of toluene by sonication for 30 min. After
that, OA (0.2 mL) was added, and the reaction mixture was
sonicated for an additional 5 min and quickly heated to 80 °C. A
IUPAC: Antalya, Turkey, 1999; pp 1007-1014.
26) Hartwig, A.; Asmuss, M.; Ehleben, I.; Herzer, U.; Kostelac, D.; Pelzer,
A.; Schwerdtle, T.; Burkle, A. 3rd International Meeting on the
Molecular Mechanisms of Metal Toxicity and Carcinogenicity; US
Dept Health Human Sciences Public Health Science: Stintino, Italy,
4
4
0 µL portion of NaBH solution in water (0.35 g/mL) was injected
into the reaction mixture followed by immediate addition of HDA
2
001; pp 797-799.
(
27) Kay, A.; Cesar, I.; Gratzel, M. J. Am. Chem. Soc. 2006, 128, 15714–
1
5721.
(32) Shevchenko, E. V.; Talapin, D. V.; Schnablegger, H.; Kornowski, A.;
Festin, O.; Svedlindh, P.; Haase, M.; Weller, H. J. Am. Chem. Soc.
2003, 125, 9090–9101.
(
(
28) Kanan, M. W.; Nocera, D. G. Science 2008, 321, 1072–1075.
29) Surendranath, Y.; Dinca, M.; Nocera, D. G. J. Am. Chem. Soc. 2009,
1
31, 2615–2620.
(33) Pellegrino, T.; Fiore, A.; Carlino, E.; Giannini, C.; Cozzoli, P. D.;
Ciccarella, G.; Respaud, M.; Palmirotta, L.; Cingolani, R.; Manna, L.
J. Am. Chem. Soc. 2006, 128, 6690–6698.
(
(
30) Risch, M.; Khare, V.; Zaharieva, I.; Gerencser, L.; Chernev, P.; Dau,
H. J. Am. Chem. Soc. 2009, 131, 6936.
31) Lee, W. R.; Kim, M. G.; Choi, J. R.; Park, J. I.; Ko, S. J.; Oh, S. J.;
Cheon, J. J. Am. Chem. Soc. 2005, 127, 16090–16097.
(34) Puntes, V. F.; Zanchet, D.; Erdonmez, C. K.; Alivisatos, A. P. J. Am.
Chem. Soc. 2002, 124, 12874–12880.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 26, 2010 9103

A R T I C L E S
Krylova et al.
Table 1. PEG-PE Amounts Used for Water Transfer of
Synthesized NCs
633 nm laser wavelength, in the voltage range of 0.5-40 V.
ꢀ
-Potentials of NCs were measured at 30 and 2 V in chloroform
and in water, respectively. The distance between electrodes was
0.4 cm. The Smoluchowski model was chosen for ꢀ-potential
calculations. X-ray photoelectron spectroscopy (XPS) measurements
were performed with synchrotron radiation at the beamline BW3
at HASYLAB/DESY in Hamburg, Germany. Samples were pre-
vol of PEG-PE
solution per 1
mg of NCs (µL)
NC′
diameter, nm
CoPt
3
NCs
4
124
104
62
47
83
153
134
76
4
8
1
6
.8
3
pared by drop-casting CoPt NCs from colloidal solution on cleaned
0.5
Au substrates. XP spectra were recorded with an Omicron EA 125
hemispherical electron analyzer. Peak fitting was done with Voigt
line shapes and a combined polynomial and Shirley-type back-
ground function. The binding energy scale was referenced to the
Au 4f7/2 level of the substrates that was set to 84.0 eV binding
energy according to ref 36. Electron paramagnetic resonance (EPR)
spectra were obtained using a Bruker Elexys E580 spectrometer.
All experiments were conducted at room temperature. DMPO spin
trap was used, and all solutions were prepared under aerobic
conditions. Stock solutions for the EPR reaction mixture preparation
Co NCs
Pt NCs
9.5
4.5
8/11
CoPt
3
/Au dumbbells
(
0.153 g). The solution was kept at 80 °C for 15 min. The inert
atmosphere was not found to be important for synthesis of Pt NCs.
Postpreparative Procedures. As-synthesized particles were
washed to remove the excess of stabilizers and byproducts by
precipitation with alcohols (methanol, ethanol, 2-propanol) and
centrifugation. The precipitates were redispersed in 2-3 mL of
toluene or chloroform and filtered through a 0.2 µm PTFE filter.
Phase Transfer of NCs. NCs were transferred into water
according to the protocol described in ref 35. The solutions of NCs
were washed by multiple washing steps with alcohols. The first
washing was performed using a mixture of methanol and ethanol.
The second washing was performed with ethanol, and any other
subsequent purification steps were performed with nonsolvents such
as 2-propanol or ethanol/2-propanol mixture. The volume ratio of
the NCs solution to the nonsolvent was usually 1:10. Final
centrifugates were redissolved in 1 mL of chloroform, and
PEGylated phospholipid (PEG-PE) dissolved in chloroform (10 mg/
mL) solution was added. The volume of added PEG-PE solution
was adjusted to obtain the ratio of PEG-PE molecules, and the
number of atoms on the NCs’ surface was approximately 1:5 (Table
were 1.2 M aqueous solution of DMPO, 0.4 mM solution of CoCl
.88 M solution of H , aqueous solution of 6 nm CoPt NCs
2
0
2
O
2
3
with surface Co atoms concentration of 2.32 mM ([Cosurf])
calculated assuming that the amount of Co atoms on the surface
corresponds to general alloy NC stoichiometry (75% Pt atoms per
11
2
5% Co atoms) and aqueous solution of Pt NCs with 9.64 × 10
particles in 1 µL ([Ptsurf] ) 1.78 mM). For EPR measurements, 10
µL of DMPO stock solution was mixed with 10 µL of stock
solutions of other reagents in order to scavenge possible ROS and
finally diluted with DI water to a volume of 80 µL. Sample solutions
were placed in a d ∼1 mm quartz tube, and EPR spectra were
measured immediately after sample preparation (9.4 GHz, power
3
3 mW, modulation amplitude 1 G). All sample preparations and
measurements except otherwise stated were conducted under aerobic
atmosphere. Magnetic properties were studied using SQUID
magnetometry (MPMS XL, Quantum Design). The spectra of
chemiluminescence (CL) were measured with a Perkin-Elmer LS
1
). In the case of CoPt
atoms was calculated by summation of surface atoms at the surface
of CoPt and Au counterparts. After addition of phospholipid,
3
/Au nanodumbbells, the amount of surface
5
5 spectrometer. CL signals were collected by a custom-made setup
consisting of a Si-based light flux detector (THOR Laboratories,
DET 100A) coupled with a digital storage oscilloscope (Tektronix,
TSD 2014B). Samples were placed in disposable 1 cm plastic
cuvettes. Cuvettes were positioned 2 cm from the detector. In a
typical experiment, 1 mL of luminol solution (1.13 mM) in
carbonate buffer (pH ) 9) was mixed with NCs solution or/and
cobalt chloride solution. The total volume of solutions containing
3
chloroform was evaporated completely under the nitrogen flow. The
solid residue was heated to 80 °C, and 2 mL of hot deionized water
was added. After cooling, the resulting aqueous solution was cleaned
from the excess of phospholipid by centrifugation at 20000 rpm
for 2 h. Final aqueous solutions were optically clear and stable for
at least 2 weeks.
Dialysis. Aqueous solutions of NCs (1.5 mL) placed into the
.5-3 mL Slide-A-Lyser dialysis cassette with molecular weight
500 MWCO were dialyzed versus 300 mL of DI water under
stirring. After the dialysis was stopped, the solution was evaporated
slowly at 50 °C until its volume became 1.5 mL (equal to the
volume of the NCs solution inside the cassette). Both dialysis water
and NCs solution taken out from the cassette were tested in the
CL experiments.
The X-ray diffraction data (XRD) were collected using a Bruker
D8 Discover diffractometer (Cu KR radiation, 0.1 mm detector slit).
Samples for transmission electron microscopy (TEM) were prepared
by droping and drying of 1-2 µL of toluene solution of NCs on a
carbon-coated copper grid (Ted Pella).
Fourier transform infrared (FTIR) measurements were conducted
using a Bruker (Billerica, MA) Vertex 70 spectrometer. Films
containing of ∼100 µg of NCs were directly placed on top of the
sample stage of the switchable attenuated total reflection (ATR)
device. Thermogravimetric analysis (TGA) was performed using
Mettler Toledo 851 apparatus. The initial amount of the sample
for TGA measurements was not less than 3 mg; the samples were
loaded into a light aluminum crucible by dropping and drying the
concentrated NC solution. All measurements were done under
continuous nitrogen flow. The ꢀ-potential measurements of NCs
were performed on ZetaSizer (Mavern Instruments) operating at
2
+
NCs or Co ions was within the range of 0.5-600 µL. DI water
was used to dilute the final volume of the sample up to 1.7 mL to
0
3
simplify the normalization. After that, 0.3 mL of H
44 mM) were quickly injected into luminol solution under vigorous
stirring. Calibration was performed with blank solutions of CoCl
with seven different concentrations: 2, 4, 20, 40, 200, 400, and
600 µM. The volume of the standard solutions taken for the CL
2 2
O solution
(
2
1
2
+
reaction was 250 µL. Thus, final Co concentrations were 0.25,
0
.5, 2.5, 5, 25, 50, and 200 µM, respectively. The CL signal was
registered with an oscilloscope. Quantitative information of the CL
intensity was obtained by integration of CL signal over time (Figure
S1, Supporting Information).
Results and Discussion
3
CoPt NCs of five different sizes were synthesized in high-
boiling solvent (Figure 1). All particles have a chemically
disordered face-centered cubic (fcc) phase as previously
13,32
reported.
3
CoPt NCs were transferred into water using the
phospholipids (see Phase Transfer of NCs in the Materials and
Methods). We have found that multistep washing of NCs is
critical for the phase transfer. According to the TGA data, there
is ∼45 wt % of organic molecules in the sample of 8 nm CoPt
3
NCs after basic purification of NCs from the excess of precursors
(
35) Dubertret, B. M/S: Medecine Sci. 2003, 19, 532–534.
104 J. AM. CHEM. SOC. 9 VOL. 132, NO. 26, 2010
(36) Powell, C. J. Appl. Surf. Sci. 1995, 89, 141–149.
9

Probing the Surface of Transition-Metal Nanocrystals
A R T I C L E S
Figure 1. TEM overview images of 4.0, 4.8, 6.0, 8.0, and 10.5 nm CoPt
3
3
NCs and CoPt /Au (8 nm/11 nm), respectively (a-f). Scale bars are 10 nm.
and byproduct (Figure S2, Supporting Information). The solu-
tions of such samples in toluene or chloroform were stable
against aggregation and oxidation for months; however, NCs
could not be transferred into water. The subsequent purification
procedures reduced the concentration of the organic component
down to ∼5-6% that was optimal for phase transfer from
chloroform into water. To calculate the weight of ligand
monolayer on the NC surface, we assumed a 1:3 ratio of ACA
to HDA assuming a 1:3 ratio of Co to Pt. The footprint of HDA
2
,37,38
is 18.6-20 Å,
and the cross-section of ACA is calculated
2
to be 23-28 Å on the basis of the effective diameter of the
molecule reported in ref 39. According to the calculations, 5-6
wt % of organic molecules cannot provide the full passivation
3
of the surface of 8 nm CoPt NCs that obviously can facilitate
the phase transfer. FTIR measurements (Figure S3, Supporting
Information) indicate the higher concentration of chlorine
containing organics in the properly washed samples. This signal
can originate from the absorbed molecules of chloroform that
was used as a solvent, confirming the lack of HDA and ACA
molecules at the surface of NCs.
3
Figure 2. XPS spectra of 8 nm CoPt NCs in the energy regions of the Co
2p bands at two different photon energies. Experimental data (dotted curves)
are plotted together with a deconvolution into contributing peaks (solid
curves).
4
5-50
51,52
(
CL).
This highly sensitive reaction
is routinely used
in forensic medicine to detect trace amounts of blood since blood
2+
It is well-known that cations of transition metals (e.g., Co ,
2
+
2+
Fe , Ni , etc.) can induce Fenton-type chemical reactions and
(42) Baj, S.; Krawczyk, T. J. Photochem. Photobiol. A: Chem. 2006, 183,
•
-
•
•
111–120.
lead to the formation of free radicals ꢀ
are called reactive oxygen species.
is an important stage in oxidation of luminol (Scheme S1).
2
, OH , and OOH that
The generation of ROS
(
43) Mercnyi, G.; Lind, J.; Eriksen, T. E. J. Phys. Chem. 1984, 88, 2320–
4
0,41
2
323.
(44) Rose, A. L.; Waite, T. D. Anal. Chem. 2001, 73, 5909–5920.
45) Lind, J.; MerBnyi, G.; Eriksen, T. E. J. Am. Chem. Soc. 1983, 105,
655–7661.
(46) Burdo, T. G.; Seitz, W. R. Anal. Chem. 1975, 47, 1639–1643.
47) Kitajima, N.; Fukuzumi, S.; Ono, Y. J. Phys. Chem. 1978, 82, 1505–
42-44
(
The oxidation of luminol is accompanied by intense emission
of a blue light in the process known as chemiluminesence
7
(
1
509.
(
(
37) Plaut, D. J.; Lund, K. M.; Ward, M. D. Chem. Commun. 2000, 769–
(48) Lin, J. M.; Shan, X. Q.; Hanaoka, S.; Yamada, M. Anal. Chem. 2001,
73, 5043–5051.
7
70.
38) Espina, A.; Garcia, J. R.; Guil, J. M.; Jaimez, E.; Parra, J. B.;
Rodriguez, J. J. Phys. Chem. B 1998, 102, 1713–1716.
(49) Ren, J. C.; Huang, X. Y. Anal. Chem. 2001, 73, 2663–2668.
(50) Du, J. X.; Li, J. J.; Yang, L. L.; Lu, J. R. Anal. Chim. Acta 2003, 481,
239–244.
(
(
(
39) Valyon, J.; Engelhardt, J. React. Kinet. Catal. Lett. 1998, 63, 27–32.
40) Gutteridge, J. M. C. Free Radical Res. Commun. 1993, 19, 141–158.
41) Konovalova, T. A.; Lawrence, J.; Kispert, L. D. J. Photochem.
Photobiol. A: Chem. 2004, 162, 1–8.
(51) Rongen, H. A. H.; Hoetelmans, R. M. W.; Bult, A.; Vanbennekom,
W. P. J. Pharm. Biomed. Anal. 1994, 12, 433–462.
(52) Dodeigne, C.; Thunus, L.; Lejeune, R. Talanta 2000, 51, 415–439.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 26, 2010 9105

A R T I C L E S
Krylova et al.
contains iron-based oxygen-transport metalloprotein (hemoglo-
5
3,54
bin) that can promote the generation of ROS.
Other
2
+
transition-metal ions in the oxidation state +2 (e.g., Co , see
the reactions 1-5 below) can also lead to the formation of ROS
that further oxidize the luminol into aminophthalate. Note that
2
+
3+
during this reaction M ions converts irreversibly into M ,
and hence, M2 ions are not recycled as a true catalyst. The
+
2
+
intensity of CL is proportional to the concentration of M ions
at their high dilutions that is used to determine quantitatively
the low (nM) concentrations of these ions.
•
-
Co(II) + ꢀ f Co(III) + ꢀ
(1)
(2)
2
2
2
H+
Co(II) + ꢀ2•-
2
2
Figure 3. (a) Dependence of the integrated surface area (ISA) on the
concentration of cobalt atoms at the surface([Cosurf]) of differently sized
8 Co(ΙΙΙ) + H ꢀ
CoPt
different concentrations; (b-d) optical images of chemiluminesence made
at the decay time of 0.2 s in of the solutions of 53 µM CoCl , 4.8 nm
CoPt NCs with [Cosurf] ) 65 µM and CoPt /Au dumbbells, respectively;
e) summary of different contributing factors (scavenging of radicals at the
3 2
NCs and the calibration plot obtained with solutions of CoCl of
•
-
Co(ΙΙ) + H ꢀ f Co(III) + ꢀH + ꢀH
(3)
2
2
2
3
3
(
•
-
Co(ΙΙ) + ꢀH f Co(III) + ꢀH
(4)
(5)
surface of 4.8 nm CoPt NCs and losses of signal due to absorption of
3
emitted light by NCs) that describe the lower CL signal in case of NCs as
compared with calibration CoCl solution.
2
•
+
Co(III) + H O f Co(II) + OOH + H
2
2
Since Co2 ions can be oxidized by oxygen and peroxide in
+
We employed the chemiluminescence technique based on the
oxidation of luminol to analyze the surface of NCs containing
4
0,41
alkaline solutions,
we expected that cobalt atoms at the
surface of NCs can also promote the formation of ROS and
induce the oxidation of luminol. Indeed, Figure 3 provides the
visualization of the CL promoted by cobalt ions in solution and
atoms of transition metals. CoPt
NCs and CoPt /Au dumbbells) were chosen as the model system
i) because these NCs can be synthesized in a wide size range
3 3
-based NCs (individual CoPt
3
(
3
by the surface of CoPt NCs.
with narrow size distribution (Figure 1) and (ii) because of the
relatively long decay times of the CL promoted by cobalt ions
The spectrum of CL exhibits a maximum at 420 nm (Figure
S1a, Supporting Information). The integrated signal area (ISA)
of the CL intensity was chosen as a parameter minimizing the
effect of fluctuations in mixing of injected peroxide solution
(
Figure S1).
3
XPS analysis of CoPt NCs revealed the presence of different
types of Co species in the NCs (Figure 2). The doublet labeled
V” in Figure 2 has a binding energy of 778.8 eV for the Co
3/2 level and a spin-orbit splitting of 15.2 eV. These values
(
Figure S1b, Supporting Information). The calibration curve of
“
2+
ISA versus [Co ] shows linear dependence in the concentration
2
p
5
5
range of 0-200 µM (Figure S4, Supporting Information).
are in good agreement with reference data for metallic Co.
2+
Further increase of [Co ] results in the formation of Co(OH)
2
We assigned this doublet to Co atoms in the volume of the
bimetallic particles. The second doublet, labeled “S”, was
observed at ∼781 eV (2p3/2) with a larger spin-orbit splitting
of ∼15.8 eV. These values are typical for cobalt in the oxidation
that precipitates from the solution leading to the drop in CL
signal (Figure S4, Supporting Information).
We measured the ISA of CL as a function of concentration
of cobalt atoms [Cosurf] at the surface of CoPt NCs of different
3
sizes. We used the following assumptions: (i) the ratio of Co
to Pt atoms at the surface of NCs is equal 1:3 (the elemental
2+
56
state Co , as it was previously observed in spectra of CoO
The third doublet, labeled “Sat” can be identified as a shakeup
56
2+
satellite feature. We assign the component “S” to Co surface
sites that are coordinated by the ACA ligands.
3
analysis is in a good agreement with CoPt stoichiometry (Table
S1, Supporting Information), and (ii) we assumed that all surface
atoms of cobalt are chemically active. The weighting of the
dried samples and subtracting the amount of organic molecules
determined by TGA allows us to calculate the number of
individual NCs in studied samples based on their sizes deter-
mined from TEM. Figure 3a shows the plots of ISA of CL
versus the concentration of surface atoms of cobalt [Cosurf] for
It was indeed assumed previously that ACA molecules bind
preferably to the Co surface sites and Pt surface atoms are
15,57,58
coordinated by HDA.
Coordinated oxidized cobalt surface
sites may contribute to the component “S” as well. Note also
that this component is very broad, suggesting inhomogeneous
broadening due to the presence of Co2+ species in slightly
different chemical environments that cannot be further resolved,
e.g., oxidized cobalt surface sites and surface sites capped with
the ACA ligands.
3
five different sizes of CoPt NCs. The dependence of ISA on
the [Cosurf] is linear only at very high dilutions of NCs. The
deviation of the slope of linear part of the curves from the slope
of the calibration curve can originate from faceting of NCs that
can provide higher or lower concentration of surface atoms of
cobalt. In addition, not all surface atoms of cobalt can participate
in the generation of ROS. Further increase of NC concentrations
leads to the saturation of the ISA curves in all five cases of NC
sizes. In order to understand this observation, we studied the
role of the factors discussed below. First, we examined the effect
of PEG-PE used to transfer NCs into water. However, the
addition of PEG-PE to the Co2 solution in the same concentra-
(
53) Zhou, S. L.; Wang, J. H.; Huang, W. H.; Lu, X.; Cheng, H. K.
J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2007, 850, 343–
3
47.
(
(
54) Zhi, Q.; Xie, C.; Huang, X. Y.; Ren, J. C. Anal. Chim. Acta 2007,
5
83, 217–222.
55) Mandale, A.; B, B. S.; Date, S. K J. J. Electron Spectrosc. Relat.
Phenom. 1984, 33, 61–72.
(
(
56) Chuang, T. J. B. C. R.; Rice, D. W. Surf. Sci. 1976, 59, 413–429.
57) Samia, A. C. S.; Schlueter, J. A.; Jiang, J. S.; Bader, S. D.; Qin, C. J.;
Lin, X. M. Chem. Mater. 2006, 18, 5203–5212.
+
(58) Bagaria, H. G.; Ada, E. T.; Shamsuzzoha, M.; Nikles, D. E.; Johnson,
D. T. Langmuir 2006, 22, 7732–7737.
3
tion of lipids per cobalt atom as in the case of CoPt samples
9
106 J. AM. CHEM. SOC. 9 VOL. 132, NO. 26, 2010

Probing the Surface of Transition-Metal Nanocrystals
A R T I C L E S
Scheme 1. Depiction of the DMPO Adducts with OH and O
•-/
•
did not affect the intensity of the CL signal. To figure out the
2
•
65
HOO Radicals
contribution of light absorption by CoPt
3
NCs into losses of
CL we examined the ISA of Co solution under different
concentrations of “inactive” CoPt NCs (during the oxidation
of luminol cobalt(II) is irreversibly converted into Co(III) and
thus, the addition of extra portions of H and luminol
suppresses the activity of CoPt NCs in generation of ROS (see
the Supporting Information for details)). The absorption of light
by CoPt NCs can lead to the drop in the intensity of CL (Figure
2
+
3
2
O
2
3
3
S5, Supporting Information); however, it cannot explain all
losses of ISA, and other factors have to be taken into
consideration.
It is known that platinum can catalyze the H
tionation reaction.
2
O
2
dispropor-
5
9
counterintuitive. We assume that platinum at the surface of
CoPt NCs can scavenge the ROS more efficiently than Pt NCs.
Previously, it was shown experimentally and explained theoreti-
3
Pt
2
H O 98 2H O + O
2 2 2 2
(6)
cally that small Pt clusters and ad-islands were better catalysts
63,64
than Pt NCs.
In addition, the better performance of Pt-based
alloys as compared with pure platinum was found to occur due
to the change of electronic structure of platinum. All three
The proposed mechanism of this reaction is based on the
1
1
formation of Pt-(OHads) and Pt-(OOHads) complexes adsorbed
60
scenarios can affect the calculation accuracy in the case of
at the surface of NCs. In addition to that, the platinum surface
NCs by addition of Pt NCs to Co2 solution.
+
6
1
modeling CoPt
3
is known to be a good scavenger for the ROS. We decided to
examine the possibility of radical scavenging by Pt and CoPt
NCs. For that, we added different amounts of CoPt and Pt NCs
to the solutions of cobalt(II) chloride. We normalized the
concentration of both CoPt and Pt NCs to the concentration of
surface platinum assuming the spherical shape of NCs. Addition
of both CoPt and Pt NCs to the solutions of CoCl led to a
3
To prove the ROS scavenging ability of Pt and CoPt NCs,
3
we carried out the room-temperature EPR experiments. The
DMPO was used as a spin-trap agent (Scheme 1). The
concentrations of the reaction species used in EPR experiments
were increased as compared to those in the CL experiments
due to the lower sensitivity of the spin-trapping method. Blank
3
3
3
2
experiments with H
radical product (Figure 4). Addition of Pt NCs to H
2
O
2
and DMPO do not show any adduct
and
significant decrease of CL (Figure S6, Supporting Information).
Using experimental data presented in Figures S5 and S6
Supporting Information), we can estimate the ISA/[Cosurf]
2
O
2
DMPO solution does not result in radical production as well.
However, the formation of gaseous products can be observed
even by the naked eye. The catalytic decomposition of H O
2 2
(
dependences after subtraction of absorption and radical scaveng-
ing losses form calibration curve. As we can see, light absorption
and scavenging of radicals at the surface of NCs explains ∼70%
on platinum involves dissociation of peroxide on the Pt surface
•
with formation of strongly adsorbed OH radicals that further
of the deviation of the experimental results obtained on CoPt
3
6
0
convert into oxygen and water.
NCs from the calibration curve (Figure 3e). In addition, it was
shown that noble metal NCs are able to accumulate negative
In the case when Co2 and H
under anaerobic conditions, the EPR revealed a four-line
spectrum with g ) 2.0057 and hyperfine splitting a ) 14.9 G
Figure 4). According to the literature, the spectrum can be
attributed to the DMPO-OH adduct. The same measurements
performed in the presence of air revealed superposition of a
+
2
2
O were mixed with DMPO
6
2
charge during their synthesis. Thus, we have to consider the
possibility of the chemisorption of the metal cations from
the solution at the surface of NCs due to Coulomb attraction.
N
(
65
The measurements of the ꢀ-potential on CoPt
indicate that both types of NCs can adsorb Co ions (Figure
3
and Pt NCs
2
+
new signal with g ) 2.0058 and a
N
) 13.9 G that corresponds
S6, insert, Supporting Information) decreasing the effective
65,66
concentration of Co2+ ions in the solution. Based on the same
to the DMPO-OOH adduct
(Figure 4). In Figure 4, the lines
corresponding to DMPO-OH radicals are labeled by circles,
while ones corresponding to DMPO-OOH radical adducts are
labeled with stars. The DMPO-OOH radical adduct is not stable,
and its transformation to DMPO-OH, according to Scheme 1,
was completed in 20 min (Figure 4).
3
trend in change of ꢀ-potential, we can assume that CoPt and
Pt NCs adsorb similar amount of cobalt ions. This adsorption
may result in the decrease of the amount of active Pt sites on
the NCs’ surface, potentially causing underestimation of the
scavenging effect of surface Pt atoms. In addition to that, we
cannot exclude the effect of passivation of Pt surface by PEG-
PE that can also depends on size and faceting of NCs. In our
experiment, we used only 4.5 nm Pt NCs; the size of those is
•
•
•-
The formation of OH and OOH/O
2
radicals can lead
to the oxidation of luminol (Scheme S1, Supporting Informa-
2+
2 2
tion). The addition of Pt NCs to the Co /H O /DMPO mixture
results in much lower peak intensities (Figure 4), confirming
efficient scavenging of the radicals by Pt NCs. The formation
close to the size of 4.8 nm CoPt
As shown in Figure S6 (Supporting Information), higher
dilutions of CoPt NCs demonstrate CL quenching almost two
times stronger as compared to Pt NCs, which is rather
3
NCs.
3
(
63) Vajda, S.; Pellin, M. J.; Greeley, J. P.; Marshall, C. L.; Curtiss, L. A.;
Ballentine, G. A.; Elam, J. W.; Catillon-Mucherie, S.; Redfern, P. C.;
Mehmood, F.; Zapol, P. Nat. Mater. 2009, 8, 213–216.
(
59) Qin, L. D.; Banholzer, M. J.; Xu, X. Y.; Huang, L.; Mirkin, C. A.
J. Am. Chem. Soc. 2007, 129, 14870.
(64) Strmcnik, D. S.; Tripkovic, D. V.; van der Vliet, D.; Chang, K. C.;
Komanicky, V.; You, H.; Karapetrov, G.; Greeley, J.; Stamenkovic,
V. R.; Markovic, N. M. J. Am. Chem. Soc. 2008, 130, 15332–15339.
(65) Mojovic, M.; Spasojevic, I.; Vuletic, M.; Vucinic, Z.; Bacic, G. J.
Serb.Chem. Soc. 2005, 70, 177–186.
(
(
60) Mededovic, S.; Locke, B. R. Appl. Catal. B 2006, 67, 149–159.
61) Hamasaki, T.; Kashiwagi, T.; Imada, T.; Nakamichi, N.; Aramaki,
S.; Toh, K.; Morisawa, S.; Shimakoshi, H.; Hisaeda, Y.; Shirahata, S.
Langmuir 2008, 24, 7354–7364.
(66) Kadiiska, M. B.; Maples, K. R.; Mason, R. P. Arch. Biochem. Biophys.
1989, 275, 98–111.
(
62) Henglein, A. J. Phys. Chem. 1993, 97, 5457–5471.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 26, 2010 9107

A R T I C L E S
Krylova et al.
•
•
Figure 4. EPR spectra of DMPO adduct radicals formed in the presence of OH and OOH radicals. DMPO-OH and DMPO-OOH peaks are depicted by
circles and stars, respectively. In all experiments, the concentrations of H
O
2 2
and DMPO were 0.11 and 0.15 M, respectively.
of small amounts of DMPO-OH and DMPO-OOH was observed
even though the sample was prepared and kept under nitrogen
atmosphere during the measurements. This observation indicates
the generation of oxygen catalyzed by Pt (eq 6) that can be
•
•-
2+
further converted into OH and O
Addition of CoPt NCs to the Co /H
demonstrates a substantial decrease in the radical production
2
in the presence of Co .
2
+
3
2 2
O /DMPO mixture
2
+
2+
efficiency comparative to Co similar to the case of Co ions
and Pt NCs mixture.
For comparison purposes, we also performed experiments
with pure 9.5 ( 1.5 nm Co NCs (Figure S7, Supporting
Information). Co NCs were also found to promote the oxidation
of luminol. However, addition of cobalt NCs into an alkaline
solution of luminol was accompanied by a color change from
black to greenish, indicating possible dissolution of cobalt NCs.
Indeed, ISA-[Co] plots calculated with the assumption that there
was no nanocrystal dissolutions are far above the calibration
plot, confirming the leaching of cobalt ions from the nanocrystals
Figure 5. Kinetic of dialysis for 6 nm CoPt
from Co leached into water (blue line with circles); from CoPt
3
NCs: the signal measured
NCs (black
2+
3
2+
line with squares) and summation of signals measured from leached Co
and from CoPt NCs in dialysis cassette (black dashed line). Calibration
curve (red line) corresponds to the [Cosurf] ) 70 µM calculated for 6 nm
CoPt NCs assuming their spherical shape and 25% and 75% of Co and Pt
3
3
NCs, respectively.
(
Figure S7, Supporting Information). The ease of dissolution
of pure Co NCs prompted the idea that cobalt atoms can also
leave the surface of CoPt NCs, leaving a Pt-enriched surface
layer that could be very beneficial for further catalytic
higher numbers approaching the CL signal for [Cosurf] calculated
for 6 nm large CoPt
concentration of Co in dialysis water almost reaches the
concentration corresponding to the maximum concentration of
3
NCs. After ∼11 days of dialysis, the
3
2+
6
7,68
applications.
In addition to that, washing out the interfacial
3
cobalt atoms at the surface of 6 nm CoPt NCs assuming the
spherical shape of NCs and 1:3 stoichiometry.
This observation allows us to assume that cobalt atoms leach
atoms of transition metals can lower the toxicity of transition-
metal-based magnetic NCs and potentially allow their biomedi-
cal applications (e.g., magnetic drug delivery, MRI agents)
without the risk of deleterious side effects.
only from the surface of CoPt
ments performed on 4.8 nm CoPt
3
NCs. The magnetic measure-
NCs before and after 11 days
3
3
We found that continuous dialysis of CoPt NCs in deionized
of dialysis show that these nanocrystals do not exhibit significant
differences (Figure 6). There is about a 14% decrease in the
coercivity and a minor decrease in the saturation magnetization.
These changes can be attributed to the removal of the cobalt
atoms from the surface of NCs. Since the impact of the canted
water results in the leaching of cobalt atoms from the surface
of NCs (Figure 5). We monitored the rate of dissolution of the
surface cobalt using the chemiluminescence technique. Both
dialysis water and solution of NCs were measured at different
intervals of dialysis. The dialysis water was concentrated up to
spins of surface atoms into the total magnetization of the particle
1
.5 mL. Note that the summation of ISA signals of dialysis
69,70
is lower than the impact of “bulk” spins,
the small losses
water and solution of NCs at later stages of dialysis gives the
in saturation magnetization are in good agreement with the
removal of surface atoms, even though the concentration of total
(
67) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.;
Lucas, C. A.; Markovic, N. M. Science 2007, 315, 493–497.
68) Stamenkovic, V.; Mun, B. S.; Mayrhofer, K. J. J.; Ross, P. N.;
Markovic, N. M.; Rossmeisl, J.; Greeley, J.; Norskov, J. K. Angew.
Chem., Int. Ed. Engl. 2006, 45, 2897–2901.
3
number of cobalt atoms at the surface of CoPt NCs is relatively
(
(69) Kachkachi, H.; Ezzir, A.; Nogues, M.; Tronc, E. Eur. Phys. J. B 2000,
14, 681–689.
9
108 J. AM. CHEM. SOC. 9 VOL. 132, NO. 26, 2010

Probing the Surface of Transition-Metal Nanocrystals
A R T I C L E S
assume that the adsorption of Co2 takes place because there is
+
a significant shift (∼25 nm) of the plasmon resonance peaks to
62
the red (Figure S8, Supporting Information), and no aggrega-
tion of Au NCs took place as it was confirmed by light scattering
experiments. We suggest that the adsorbed Co2 ions can still
participate in the generation of ROS since even cobalt atoms
located at the surface of NCs can promote their formation as it
was shown above. Two times stronger quenching of the CL in
+
3
the case of CoPt NCs by Au NCs as compared to the solution
of Co2 ions can originate from the difference in the kinetics
+
of ROS generation by free ions in the solution and atoms at the
surface of CoPt
of ROS generated at the surface of CoPt
compared to the case of the solution of ions. The second reason
of zero activity of CoPt /Au nanodumbbells in the promotion
of ROS formation can be that there are no cobalt atoms in the
appropriate form at the surface of CoPt part in heterodimers.
We can assume that the mechanism of the formation of CoPt
Au nanodumbbells can involve either oxidation of surface cobalt
atoms into Co(III) or their leaching from the surface of CoPt
during nucleation and growth of Au counterpart. The mechanism
of the CoPt /Au nanodumbbells formation is currently under
investigation.
The results obtained for CoPt
that the toxicity of NCs associated with their ability to promote
the formation of ROS can be significantly lowered by the proper
design of NCs. Thus functionalization of transition metal based
magnetic NCs with material (for instance gold) that can both
scavenge the radicals and render them ease of further biofunc-
tionalization can be a promising way toward high performance
and safe agents for magnetic hyperthermia and MRI.
3
NCs and by higher probability of the collision
3
NCs and Au NCs as
3
Figure 6. Magnetization curves of 4.8 and 10.5 nm CoPt NCs at 5 K
before and after dialysis. The duration of dialysis was 11 days.
3
high (∼31%). The dialysis of larger 10.5 nm NCs that have
∼
14% Co atoms on their surface did not affect their magnetic
3
3
/
properties. There were no measurable changes in coercivity and
saturation magnetization (Figure 6). These results can be
explained both by higher chemical stability of larger NCs due
3
7
1
to their lower chemical potential and by lower concentration
of cobalt atoms at the surface of NCs. The FC and ZFC
temperature dependencies of the magnetization of the sample
before and after dialysis do not exhibit significant difference as
well (Figure S8, Supporting Information). These data allows
3
3
/Au nanodumbbells indicate
us to assume that in the case of as-transferred into water CoPt
3
nanoparticles CL comes mainly from the surface of NCs since
the leaching of cobalt atoms is a slow process whereas CL
decays very fast (days vs seconds).
The third tested system was CoPt
3
/Au dumbbell-like NCs
(
Figure 1f). The gold containing NCs were proven to be very
8
promising agents for diagnostics and therapeutic treatment
7
2,73
Conclusions
purposes.
on the presynthesized 8 nm CoPt
the CoPt /Au nanodumbbells was examined by EDX analysis,
and the atomic ratio of Co/Pt/Au was about 1/2.95/9.9. No
detectable CL signal was observed when CoPt /Au nanodumb-
bells were added to the solution of luminol instead of CoPt
The 11 ( 1 nm gold counterparts were grown
3
NCs. The composition of
We have proposed a simple chemiluminescence-based method
of investigation of surface properties of Co-containing nano-
crystals. The CL method allows efficient monitoring over time
of the stability of transition-metal NCs against oxidation and
dissolution. This method reveals high sensitivity to only surface
atoms. Using a combination of the chemiluminsence method
and spin-trap EPR techniques, we studied the generation of ROS
at the surface of different types of NCs. In fact, the findings
from both techniques correlate very well. We have shown that
3
3
3
NCs (Figure 3a,d). EPR data showed that the formation of
hydroxyl and/or superoxide radicals does not occur in the case
of dumbbell NCs (Figure 4). There are several factors that can
lead to that observation. First, it can be associated with the
synergetic effect of gold and platinum surfaces that act as the
ROS scavengers. Addition of 15 nm Au NCs to the solution of
differently sized CoPt NCs can promote the formation of ROS
3
and as a result can lead to the oxidation of luminol accompanied
CoPt
Figure S9, Supporting Information). For instance, 175 µM of
Au NCs leads to ∼43% decrease in CL signal of CoPt NCs
that is about two times more than CL losses induced by the
3
NCs results in the significant quenching of the CL signal
by emission of light. On the other hand, CoPt NCs can serve as
3
(
even better scavengers for ROS than Pt NCs. We observed by CL
3
method that cobalt atoms slowly leach from the surface of CoPt3
NCs even under very mild conditions; however, the concentration
of the leached cobalt atoms does not exceed the maximal
concentration of cobalt calculated assuming the spherical shape
and 1:3 stoichiometry indicating that only surface atoms go into
solution. Since the formation of layers of Pt at the surface of alloyed
NCs (Pt skin) results in the superior catalytic performance of such
materials, the leaching of cobalt atoms from the surface can be
potentially beneficial to prepare highly efficient catalysts. Since
only the dissolution of the atoms of transition metals from the
surface of alloy NCs occurs and further leaching of their atoms
has not been found, alloy NCs based on transition metals other
than iron can be considered as alternative for iron-based nano-
structures for biomedical applications.
same concentration of Pt NCs (∼19%). In the case of the CoCl
2
solution, the same concentration of Au NCs causes the decrease
in the intensity of CL signal of ∼23% that is comparable to the
CL losses caused by Pt NCs (∼19%). Note, in this case the
surface of Au NCs can absorb ions of cobalt that can potentially
lower the concentration of cobalt ions in the solution. We
(
70) Tronc, E.; Ezzir, A.; Cherkaoui, R.; Chaneac, C.; Nogues, M.;
Kachkachi, H.; Fiorani, D.; Testa, A. M.; Greneche, J. M.; Jolivet,
J. P. 3rd EuroConference on Magnetic Properties of Fine Particles
and Their ReleVance to Materials Science; Elsevier Science Bv:
Barcelona, 1999; pp 63-79.
(
(
(
71) Talapin, D. V.; Rogach, A. L.; Haase, M.; Weller, H. J. Phys. Chem.
B 2001, 105, 12278–12285.
We believe that our method can be also utilized to monitor
the stability of the shell grown around the transition metal based
nanocrystals as well in the studies of the mechanism of the
72) Choi, J. S.; Jun, Y. W.; Yeon, S. I.; Kim, H. C.; Shin, J. S.; Cheon,
J. J. Am. Chem. Soc. 2006, 128, 15982–15983.
73) Melancon, M. P.; Lu, W.; Li, C. MRS Bull. 2009, 34, 415–421.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 26, 2010 9109

A R T I C L E S
Krylova et al.
nucleation and growth of core/shell or dumbbell-like structures
when the seed material contains the transition metals. This work
surements were supported by the German Science Foundation
(DFG) within the framework of the SFB 508.
3
was focused mostly on CoPt NCs; however, there are many
Supporting Information Available: EDX data; scheme of
oxidation of luminol; spectrum and time decay of chemilu-
reasons to expect that similar trends will be observed in the
case of other cobalt-based alloys as well as nickel- and copper-
based NCs.
minesence; TGA data; FTIR data; calibration plot for CoCl
solutions; details of preparation of “inactive” CoPt NCs;
analysis of the different quenching factors of chemilumin-
ensence; chemiluminsence on Co NCs, FC/ZFC on CoPt NCs.
2
3
Acknowledgment. We acknowledge Prof. Horst Weller (Uni-
versity of Hamburg) and Thomas M o¨ ller (HASYLAB at DESY)
for fruitful discussions. The work at the Center for Nanoscale
Materials (ANL) was supported by the U.S Department of Energy
under Contract No. DE-AC02-06CH11357. D.V.T. acknowledges
the NSF CAREER under Award No. DMR-0847535. XPS mea-
3
This material is available free of charge via the Internet at http://
pubs.acs.org.
JA102413K
9
110 J. AM. CHEM. SOC. 9 VOL. 132, NO. 26, 2010