Magnetic field control of photoinduced silver nanoparticle formation

Article abstract of DOI:10.1021/jp061723t

The micellar photoreduction of benzophenone in the presence of Ag ⁺ leads to very rapid and efficient formation of silver nanoperticles. External magnetic fields can be used to control the rate of formation and properties of silver nanoparticles generated by reaction of ketyl radicals formed in the photoreduction of benzophenone in surfactant micelles. The effect is attributed to Zeeman splitting of the triplet sublevels of the confined radical pair that causes a reduction in the rate of geminate processes and increases the availability of ketyl radicals (through escape processes) for metal ion reduction.

Full text of DOI:10.1021/jp061723t

12856
J. Phys. Chem. B 2006, 110, 12856-12859
ARTICLES
Magnetic Field Control of Photoinduced Silver Nanoparticle Formation
,†
†
†
‡
J. C. Scaiano,* Carolina Aliaga, Steven Maguire, and Dashan Wang
Department of Chemistry, UniVersity of Ottawa, Ottawa, Ontario K1N 6N5, Canada, and Institute for Chemical
Process and EnVironmental Technology, National Research Council, Ottawa, Ontario K1A 0R6, Canada
ReceiVed: March 20, 2006; In Final Form: April 22, 2006
+
The micellar photoreduction of benzophenone in the presence of Ag leads to very rapid and efficient formation
of silver nanoperticles. External magnetic fields can be used to control the rate of formation and properties
of silver nanoparticles generated by reaction of ketyl radicals formed in the photoreduction of benzophenone
in surfactant micelles. The effect is attributed to Zeeman splitting of the triplet sublevels of the confined
radical pair that causes a reduction in the rate of geminate processes and increases the availability of ketyl
radicals (through escape processes) for metal ion reduction.
Silver nanoparticles can be readily produced by reductive
processes in homogeneous or micellar solution or in polymer
films1 and characterized through the absorption of the plasmon
SCHEME 1: Photoreduction of Benzophenone Followed
by Reduction of Silver Ions
-4
5
band at ∼420 nm. Ketyl radicals, produced in the photoreduc-
6
tion of ketones, are excellent reducing agents and efficiently
+
0 1,2,7,8
convert Ag to Ag .
The photoreduction of benzophenone
is a convenient source of ketyl radicals; the anticipated mech-
+
anism in the presence of Ag involves electron transfer by ketyl
radicals as illustrated in Scheme 1.
Photoreduction of ketones in micelles can provide confine-
ment for radical pairs (RPs), and such confinement can be used
to control the chemistry of the radicals; the RPs are normally
generated as a triplet pair, reflecting the multiplicity of the
Sample Preparation. Samples (3 mL) were prepared in
Luzchem cuvettes made of UV grade fused silica, with a 10
mm optical path. A stock solution was prepared containing
sodium dodecyl sulfate (SDS) and benzophenone at the desired
concentrations. The silver perchlorate stock solution was added
shortly before the experiment and the sample deaerated under
a nitrogen flow. To avoid excessive foaming, the nitrogen flow
was maintained above the solution (rather than bubbling) for
6,9
reactive excited ketones. The use of anionic micelles provides
+
a convenient medium in which Ag salts, as well as hydrophobic
reagents such as benzophenone (1, BP), and suitable hydrogen
donors can be solubilized; further, surfactants can help stabilize
the metal nanoparticles formed.4
,10
1
0-15 min, while the sample was simultaneously stirred by a
In this work, we have prepared micelles of sodium dodecyl
vortex mixer. Once this procedure was complete, the required
volume of pure 1,4-cyclohexadiene (highly volatile) was added
with a microliter syringe and the sample stirred (vortex) for a
few seconds and sealed with Parafilm.
sulfate (SDS), and to promote the rapid generation of ketyl
radicals,1
1,12
we used 1,4-cyclohexadiene (CHD), an excellent
1
3
hydrogen donor whose behavior in micelles is well-under-
1
1,12
stood.
Irradiation of BP in this system in the presence of
Magnetic Field Effects. The UV-visible absorption spec-
trum of the samples, prepared as described above, was recorded
and the sample exposed to UV light (>320 nm) with or without
a magnetic field, using the setup shown in the Supporting
Information. Immediately after exposure (<10 min), the UV-
visible spectrum was recorded again, and the data used for plots
such as that of Figure 2 in this paper (vide infra). Note that the
spectra shown correspond to the difference between the spectrum
after exposure minus that recorded before irradiation.
Magnetic field experiments were carried out with a home-
built magnet capable of dc fields up to 0.12 T (1200 G). The
excitation was provided by a 150 W xenon lamp filtered through
a 320 nm cutoff filter; this ensures that benzophenone is the
only absorbing species.
+
1,2
0
Ag leads to its reduction with rapid generation of the Ag
plasmon band, according to the mechanism of Scheme 1,
0
followed by Ag aggregation to yield nanoparticles.
Experimental Section
General. Chemicals were purchased from Aldrich and used
as received. Absorbance was measured using a Cary 50 Bio
UV-visible spectrophotometer from Varian. Laser flash pho-
tolysis experiments were carried out in a customized Luzchem
system using the third harmonic (355 nm) from Continuum
Surelite laser for excitation.
*
tito@photo.chem.uottawa.ca.
University of Ottawa.
National Research Council.
†
Transmission Electron Microscopy (TEM). TEM speci-
mens were prepared immediately after the nanoparticles were
‡
1
0.1021/jp061723t CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/09/2006

Photoinduced Ag Nanoparticle Formation
J. Phys. Chem. B, Vol. 110, No. 26, 2006 12857
SCHEME 2: Reaction of Benzophenone Triplets with
1
,4-Cyclohexadiene.
SCHEME 3: Separation of the RP Is Competitive with
Intersystem Crossing; Application of an External DC
Magnetic Field (inset box) Separates the Triplet Sublevels
and Reduces the Probability for Geminate Reactions.
Figure 1. Quenching of the benzophenone triplet state in 0.1 M SDS
+
(
monitored at 600 nm to avoid ketyl radical interference) by Ag . The
slope is the quenching rate constant and the intercept the reciprocal of
the triplet lifetime. The insert shows the trace recorded for 0.77 mM
+
Ag .
synthesized: 10 µL of the solution was delivered with a
microliter syringe onto a 300 mesh of copper grid coated with
carbon film; the sample was washed with 25 µL of methanol
and allowed to dry in air. Washing reduces the amount of
organic material and unreacted silver salts and improves TEM
picture quality. TEM examination was carried out using a Philips
CM20 STEM equipped with a Gatan UltraScan 1000 CCD
camera and an energy-dispersive X-ray spectrometer (EDS),
INCA Energy TEM 200. TEM was operated at 200 kV.
Bright field (BF) and high-resolution TEM (HRTEM) images
were taken for particle size measurement and crystallographic
determination. EDS was also used for composition analysis of
the particles.
of the benzophenone triplets decay by reaction with CHD under
our experimental conditions.
The initial radical pair (RP) produced by photoreduction of
benzophenone is well-known to have triplet character (Schemes
2
and 3). Micelles provide RP confinement and increased
opportunities (time) for geminate processes to occur, in the
1
1,12
benzophenone system,
as well as in many other micellar
The RPs decay by a competition of
1
4,15
systems involving RPs.
geminate radical-radical reactions and escape from the micellar
cage, as illustrated in Scheme 3 where the boxes represent
micelles. Geminate reactions require a spin-flip, or intersystem
crossing in the RP from the triplet (that cannot yield products),
to the singlet RP that yields cage or geminate molecular products
Results and Discussion
In order for the mechanism of Scheme 1 to be efficient, it is
essential that the benzophenone triplets be quenched by the
11,12,15-17
rapidly.
This competition can be controlled by applica-
+
hydrogen donor, rather than by other species, notably, Ag . The
tion of external dc magnetic fields. The field (H) causes Zeeman
splitting of the triplet sublevels of the RP, as shown in the inset
in Scheme 3.
+
rate constant for Ag quenching of benzophenone triplets was
determined under our experimental conditions using laser flash
photolysis techniques. The samples (BP in 0.1 M SDS solution
under nitrogen) were excited at 355 nm to give the characteristic
signal of the triplet state with λmax at ∼530 nm. The decay of
the BP triplet was monitored at 600 nm, where the signal is
weaker than at 530 nm, but the absorption from ketyl radicals
does not interfere;1 a plot of the pseudo-first-order rate constant
In the presence of a magnetic field, the mechanism of Scheme
3 leads to a decrease in the efficiency of geminate RP processes
from T+ and T- that are no longer near-degenerate with the
singlet RP (S); RPs in T+ and T- undergo intersystem crossing
inefficiently or not at all, thus providing an opportunity for
increased micellar escape, since the rate constant for micellar
exit is not affected by external magnetic fields. Increased radical
exit from the micelle favors competitive chemical processes,
such as trapping by ions in the aqueous phase; the MF affects
the radical decay, but not its absorption spectrum. The fields
required for these effects are usually small (e0.1 T), since their
1
+
for triplet decay as a function of Ag concentration yields, from
+
the slope, a triplet quenching rate constant by Ag of 4.6 ×
9
-1 -1
+
1
(
0 M
s
and a triplet lifetime (at [Ag ] ) 0) of ca. 290 ns
Figure 1); this short lifetime is controlled by triplet reaction
11
0
with the surfactant. The quenching by Ag may be assisted
by Coulombic interactions (given the negative charge of the
surfactant) and appears to be dynamic in nature, as judged by
the fact that the decrease in lifetime upon addition of silver
ions resulted in little or no change in the absorbance of triplet
benzophenone immediately after the laser pulse, indicative of
the absence of static quenching.
1
1,16
role is simply to break the S-T degeneracy.
Upon UV (>320 nm, xenon lamp and filter) irradiation of
BP/CHD/SDS containing 0.01 M Ag , one can readily detect
+
the formation of the characteristic plasmon band, as shown in
Figure 2. The band at ca. 330 nm is mostly due to the n,π*
absorption from BP. The formation of the plasmon band is linear
with time, and peaks at 414 nm (vide infra). Formation of this
absorption occurs in a remarkably short time; for comparison,
earlier work under similar irradiation conditions involved
To facilitate hydrogen atom abstraction (rather than Ag+
quenching), we used 0.04 M CHD as a hydrogen donor; under
these conditions, the lifetime of benzophenone triplets in SDS
micelles is ∼12 ns, and hydrogen abstraction leading to ketyl
1
,8
+
exposure times about 100 times longer. In our studies, Ag
+
radical formation is favored even in the presence of Ag . We
triplet quenching is minimized by a careful choice of the
+
11,12
estimate that in the presence of 0.01 M Ag about two-thirds
hydrogen donor.

12858 J. Phys. Chem. B, Vol. 110, No. 26, 2006
Scaiano et al.
Figure 4. Magnetic field effect on the formation of silver nanoparticles
during UV irradiation. (BP 0.005 M, SDS 0.1 M, Ag+ ) 0.01 and
Figure 2. Formation of Ag nanoparticles after 30 s irradiation (BP
0
.04 M CHD, 30 s exposure). Approximately 50% of the maximum
+
0
.005 M, SDS 0.1 M, Ag ) 0.01 and 0.04 M CHD) in the absence
effect is achieved with ca. 27 mT.
(blue) and presence (red) of a 50 mT magnetic field. The thick lines
are the subtraction of the actual spectrum recorded (thin lines) minus
the preirradiation spectrum (green). The insert shows samples irradiated
for 30 s with and without an applied magnetic field.
Figure 5. TEM photos recorded in samples containing CHD. The TEM
sample was prepared within 10 min following UV exposure. The sample
on the left was exposed for 90 s with no field, and the right one at 50
mT. The abundance of particles does not reflect the true abundance in
the samples, but rather the selection of a TEM display area, where
size is representative of the bulk.
Figure 3. Formation of silver nanoparticles during UV irradiation in
few hours) to a much broader band, indicative of further
aggregation. The results reflect that in the absence of CHD the
the absence (red) and presence (blue) of a 50 mT magnetic field (BP
+
0
.005 M, SDS 0.1 M, Ag ) 0.01 and 0.04 M CHD). Absorbance
11
lifetime of excited triplet benzophenone is ca. 300 ns, making
values above 3 have considerable (>10%) error even if coincidentally
they fit perfectly on the line.
+
it an easy target for quenching by Ag ions (see Figure 1).
Interestingly, the growth of the plasmon band under these
conditions is almost twice as fast with 0.005 M Ag , compared
+
+
When Ag is present, the effect of the external field increases
the aqueous availability of ketyl radicals to reduce the metal
ion and induce a large increase in the rate of formation of
nanoparticles as deduced by the observation of the plasmon band
with 0.01 M (note reversal); thus, at both concentrations,
scavenging of ketyl radicals is efficient, and increased Ag
+
concentration simply quenches more benzophenone triplet
18
8
at 414 nm, typical of spherical nanoparticles. The changes
caused by magnetic fields are readily observable, as seen in
the photographs inserted in Figure 2 and in Figure 3. The results
states, thus limiting the ketyl radical generation.
+
It should be noted that Ag is well-known to form complexes
1
9
20
with many alkenes, including CHD. At the concentrations
+
+
suggest that Ag ions trap predominantly ketyl radicals after
used in this work, only a small fraction of the Ag will be
they escape from the micelle. It should be noted that it is
unimportant which radical leaves the micelle first (the cyclo-
hexadienyl radical probably does); once the geminate character
of the RP has been lost, the remaining radical has abundant
opportunity (time) to escape, since there is little that these
radicals can do, other than exit, when they are single micelle
occupants.
complexed with CHD (given the modest equilibrium constant),20
and thus, we do not expect a significant effect on the rate or
mechanism for silver atom formation. However, we see
increased particle stability in the presence of CHD, suggesting
that the alkene may play a role in assisting nucleation and
nanoparticle formation. Stabilization of silver clusters by alkenes
21,22
is known,
although their role at the concentration used (0.04
2
2
The effect of the magnetic field occurs at relatively low fields
and appears to saturate at around 0.1 T, as shown in Figure 4.
In the plateau region, we see a 67% signal increase; for
M) is surprising; it is possible that the SDS micelles home
the growing nanoparticles and also provide a high local
concentration for CHD.
+
comparison, when the RP separation is measured (without Ag )
TEM analysis of the nanoparticles (Figure 5) reveal particles
with an average size 4-5 nm in the absence of a field; larger
particles are multicrystalline and probably reflect aggregation
of the smaller nanoparticles. In the presence of a field, the
particles appear to be somewhat larger, typically ∼8 nm. Aged
samples (e.g., 2 h under nitrogen) show clear signs of particle
aggregation, with complex networks (see Supporting Informa-
tion) readily detectable by TEM and reminiscent of those
the increase is 69%,11 supporting the idea that both phenomena
are based on the same primary processes, i.e., that magnetic
fields promote radical separation as a result of Zeeman splitting
of the triplet sublevels.
In the absence of CHD, Ag nanoparticles are also formed,
but their formation is slower, with the plasmon band somewhat
broader and less stable, as indicated by its evolution (over a

Photoinduced Ag Nanoparticle Formation
J. Phys. Chem. B, Vol. 110, No. 26, 2006 12859
4
observed elsewhere. This initial (ca. 2 h) aggregation does not
Supporting Information Available: Experimental methods,
TEM aggregation data, nanoparticle formation kinetics, and
stability information. This material is available free of charge
via the Internet at http://pubs.acs.org.
lead to significant change in the shape of the plasmon band;
eventually (days) the visible absorption shifts and broadens (see
Supporting Information); this process is accelerated if air is
allowed in the samples.
The reaction and the field effects are best observed when the
samples are irradiated under a nitrogen atmosphere. However,
even under air, the effect is readily observable, albeit the
magnitude of the changes is reduced, and the particles are less
stable. Oxygen is a well-known quencher of excited triplet states,
References and Notes
(
1) Eustis, S.; Krylova, G.; Eremenko, A.; Smirnova, N.; Schilla, A.
W.; El-Sayed, M. Photochem. Photobiol. Sci. 2005, 4, 154.
2) Yonezawa, Y.; Sato, T.; Kuroda, S.; Kuge, K. J. Chem. Soc.,
Faraday Trans. 1991, 87 (12), 1905.
3) Korchev, A. S.; Bozack, M. J.; Slaten, B. L.; Mills, G. J. Am. Chem.
(
23
as well as a free radical scavenger. In the system under study,
the short lifetime of the benzophenone triplet, the efficient
(
Soc. 2004, 126 (1), 10. Evanoff, D. D., Jr.; Chumanov, G. J. Phys. Chem.
B 2004, 108, 13948. Korchev, A. S.; Shulyak, T. S.; Slaten, B. L.; Gale,
W. F.; Mills, G. J. Phys. Chem. B 2005, 109 (16), 7733.
+
scavenging of the ketyl radicals by Ag , and the rather low
solubility of oxygen in water all combine to allow a good
fraction of the reaction to proceed even under air.
(4) Pileni, M. P. J. Phys. Chem. B 2001, 105 (17), 3358.
(
5) Burda, C.; Chen, X.; Narayanan, R.; El-Sayed, M. A. Chem. ReV.
Given that particles formed in the presence of a field are
somewhat larger, and these are known to have a larger
2
005, 105 (4), 1025.
(6) Scaiano, J. C. J. Photochem. 1973/74, 2, 81.
24
absorbance cross-section, one may wonder if the difference
in absorbance caused by a magnetic field (Figures 2 and 3) can
be solely the result of having different nanoparticle sizes (rather
(
7) Sato, T.; Onaka, H.; Yonezawa, Y. J. Photochem. Photobiol., A
1
999, 127, 83.
(
8) Kometani, N.; Doi, H.; Asami, K.; Yonezawa, Y. Phys. Chem.
0
than different Ag yields). However, particle absorbance
Chem. Phys. 2002, 4, 5142.
increases roughly linearly24 with their diameter, while the
number of atoms grows with the cube of the particle size;
therefore, the absorbance per gram-atom of silver actually
decreases with increasing particle size. Thus, in terms of the
yield of atomic silver, the nanoparticle size effect (larger
particles in the presence of a magnetic field) would be to
attenuate to some extent the observable magnetic field effect.
Attempts to detect the time-resolved growth of the plasmon
band by laser flash photolysis failed, probably because growth
at nucleation centers occurs beyond the millisecond window
observable in our instrument. Formation of the plasmon band
was readily observable in longer time scales (seconds and several
laser pulses) under 355 nm laser irradiation.
(9) Turro, N. J. Modern Molecular Photochemistry; Benjamin/Cum-
mings Publishing Co.: Menlo Park, 1978; p 628.
(
10) Maillard, M.; Pileni, M. P.; Link, S.; El-Sayed, M. A. J. Phys. Chem.
B 2004, 108 (17), 5230.
11) Scaiano, J. C.; Abuin, E. B.; Stewart, L. C. J. Am. Chem. Soc.
982, 104 (21), 5673.
(
1
(
12) Scaiano, J. C.; Abuin, E. B. Chem. Phys. Lett. 1981, 81, 209.
(
13) Paul, H.; Small, R. D., Jr.; Scaiano, J. C. J. Am. Chem. Soc. 1978,
100, 4520.
(14) Turro, N. J. Chem. Commun. 2002, 2279. Turro, N. J.; Garcia-
Garibay, M. Thinking Topologically about Photochemistry in Constrained
Spaces. In Photochemistry in Organized and Constrained Media; Rama-
murthy, V., Ed.; VCH Publishers: New York, 1991; p 1.
(
15) Turro, N. J.; Gr a¨ tzel, M.; Braun, A. M. Angew. Chem., Int. Ed.
Engl. 1980, 19, 675.
In conclusion, magnetic field effects on ketone photoreduction
in micelles can be used to control silver nanoparticle synthesis
under mild conditions. The method provides spatial and temporal
control for particle formation. External weak magnetic fields,
add a new methodology for the manipulation of nanoparticle
synthesis. Similar magnetic field control will probably apply
to other metals and alloys; interestingly, significant effects occur
in the range of commonly employed magnetic stirring bars (e10
mT).
(16) Turro, N. J.; Kraeutler, B. J. Am. Chem. Soc. 1978, 100, 7432.
Nagakura, S.; Hayashi, H. Int. J. Quantum Chem. 1984, 18, 571.
(
17) Kleinman, M. H.; Shevchenko, T.; Bohne, C. Photochem. Photobiol.
998, 67 (2), 198.
18) Mock, J. J.; Barbic, M.; Smith, D. R.; Schultz, D. A.; Schultz, S.
J. Chem. Phys. 2002, 116 (15), 6755.
1
(
(
(
(
(
(
19) Hartley, F. R. Chem. ReV. 1973, 73 (2), 163.
20) Muhs, M. A.; Weiss, F. T. J. Am. Chem. Soc. 1962, 84 (24), 4697.
21) Andrews, M. P.; Ozin, G. A. J. Phys. Chem. 1986, 90 (13), 2923.
22) Andrews, M. P.; Ozin, G. A. J. Phys. Chem. 1986, 90 (13), 2929.
23) Maillard, B.; Ingold, K. U.; Scaiano, J. C. J. Am. Chem. Soc. 1983,
Acknowledgment. J.C.S. is grateful to Natural Sciences and
Engineering Research Council of Canada and the Province of
Ontario Premier’s program for generous support.
1
05 (15), 5095.
24) Evanoff, D. D., Jr.; Chumanov, G. J. Phys. Chem. B 2004, 108
(37), 13957.
(