Sulfonated poly (ether ether ketone)/poly(vinyl alcohol) sensitizing system for solution photogeneration of small Ag, Au, and Cu crystallites

Article abstract of DOI:10.1021/jp045047b

Illumination of air-free aqueous solutions containing sulfonated poly(ether ether ketone) and poly(vinyl alcohol) with 350 nm light results in benzophenone ketyl radicals of the polyketone. The polymer radicals form with a quantum yield 0.02 and decay with a second-order rate constant 6 orders of magnitude lower than that of typical ?±-hydroxy radicals. Evidence is presented that the polymeric benzophenone ketyl radicals reduce Ag⁺, Cu ²⁺, and AuCl₄^- to metal particles of nanometer dimensions. Decreases in the reduction rates with increasing Ag^I, Cu^II, and Au^III concentrations are explained using a kinetic model in which the metal ions quench the excited state of the polymeric benzophenone groups, which forms the macromolecular radicals. Quenching is fastest for Ag⁺, whereas Cu²⁺ and AuCl₄ ^- exhibit similar rate constants. Particle formation becomes more complex as the number of equivalents needed to reduce the metal ions increases; the Au^III system is an extreme case where the radical reactions operate in parallel with secondary light-initiated and thermal reduction channels. For each metal ion, the polymer-initiated photoreactions produce crystallites possessing distinct properties, such as a very strong plasmon in the Ag case or the narrow size distribution exhibited by Au particles. ? 2005 American Chemical Society.

Full text of DOI:10.1021/jp045047b

J. Phys. Chem. B 2005, 109, 7733-7745
7733
Sulfonated Poly(Ether Ether Ketone)/Poly(Vinyl Alcohol) Sensitizing System for Solution
Photogeneration of Small Ag, Au, and Cu Crystallites
A. S. Korchev,† T. S. Shulyak,† B. L. Slaten,*,‡ W. F. Gale,§ and G. Mills*,†
Department of Chemistry, Department of Consumer Affairs, and Materials Research and Educational Center,
Auburn UniVersity, Alabama 36849
ReceiVed: October 29, 2004; In Final Form: February 10, 2005
Illumination of air-free aqueous solutions containing sulfonated poly(ether ether ketone) and poly(vinyl alcohol)
with 350 nm light results in benzophenone ketyl radicals of the polyketone. The polymer radicals form with
a quantum yield 0.02 and decay with a second-order rate constant 6 orders of magnitude lower than that of
typical R-hydroxy radicals. Evidence is presented that the polymeric benzophenone ketyl radicals reduce
-
Ag+, Cu2+, and AuCl4 to metal particles of nanometer dimensions. Decreases in the reduction rates with
increasing AgI, CuII, and AuIII concentrations are explained using a kinetic model in which the metal ions
quench the excited state of the polymeric benzophenone groups, which forms the macromolecular radicals.
-
Quenching is fastest for Ag+, whereas Cu2+ and AuCl4 exhibit similar rate constants. Particle formation
becomes more complex as the number of equivalents needed to reduce the metal ions increases; the AuIII
system is an extreme case where the radical reactions operate in parallel with secondary light-initiated and
thermal reduction channels. For each metal ion, the polymer-initiated photoreactions produce crystallites
possessing distinct properties, such as a very strong plasmon in the Ag case or the narrow size distribution
exhibited by Au particles.
Introduction
are fairly transparent at longer wavelengths, but only a limited
number of metallic species experiencing efficient photoreduction
Synthesis of metal nanoparticles is an area of constant interest
since these particles possess unusual properties and also because
of their possible use in the fabrication of arrays with different
dimensionalities.1-6 Light-initiated reduction of metal ions in
the presence of electron donors is a known method for the
preparation of small metal particles that has found a wide range
of applications including synthesis of metallic colloids,7 imag-
ing,8 and metallization and patterning of polymer films.9,10 Given
the growing technological significance of metallized polymer
films,11 development of simple photochemical routes for the
preparation of metallic structures in these matrixes remains an
important goal. Direct excitation by hard UV photons has been
employed often to induce photoreduction of metal ions or
complexes that absorb strongly below 300 nm,7,9,12,13 allowing
preparation of a variety of metal particles either in solution or
in solids. Limitations of this method are the need for matrixes
exhibiting some transparency to high-energy light and stability
against excessive photochemical degradation. The energetic
photons may also induce undesired oxidation of the products,
as was found in the case of Ag crystallites.14
have been identified. Metal particles sometimes form via
multiphotonic reactions;17-21 one of their distinctive features is
the possibility of inducing such reactions by excitation wave-
lengths at which the metal ions exhibit no absorption of light.
While the feasibility of such processes is hard to anticipate,
reductions derived from mutiphoton absorption by photosensi-
tizers offer a more predictable method for producing metal
crystallites.10b Potential problematic issues of mutiphotonic
reactions are nontrivial effects induced by some matrixes.17,19
Another strategy is based on photoreactions of ketones to initiate
formation of metal crystallites. Illumination of ketones results
in excited states of the carbonyl group able to abstract H atoms
from compounds acting as donors; this process can produce
reducing R-hydroxy (or ketyl) radicals.22 Reduction of Ag, Au,
and Cu complexes to metal particles have been achieved through
excitation of acetone,23 acetophenone,24 benzophenone (BP),25,26
and anthraquinone derivatives in the presence of alcohols as H
atom donors.27 Benzoin has also been employed to photoge-
nerate metal crystallites,28 although this compound forms
reducing radicals via a different process. An advantage of
ketone-initiated photoreactions is their versatility since they are
not dependent on the photochemical properties of the metal
precursors.
An alternative procedure involves utilization of metal com-
plexes that absorb photons with λ > 300 nm. Silver halides
-
and AuCl4 ions are the best known examples of species that
experience photoreduction initiated by near-UV or visible
light;8,15 some Cu2+ complexes undergo similar photoreactions.16
This method is less restrictive since a wide range of matrixes
Stable nanometer-sized particles are frequently prepared in
solutions containing polymers that can hinder uncontrolled
growth of the crystallites. Hence, an alternative method for
generating small metallic crystallites in solution is to use
macromolecules that function as “particles stabilizers” and that
also participate in the photoreduction of metal precursors as
either sensitizers or reductants. An example of such a system
* Authors to whom correspondence should be addressed. E-mail:
slatebl@auburn.edu; millsge@auburn.edu.
† Department of Chemistry.
‡ Department of Consumer Affairs.
§ Materials Research and Educational Center.
10.1021/jp045047b CCC: $30.25 © 2005 American Chemical Society
Published on Web 03/31/2005

7734 J. Phys. Chem. B, Vol. 109, No. 16, 2005
Korchev et al.
would comprise a polymer acting as a sensitizer and another
that performs the donor and particle stabilizer functions at the
same time. Because the macromolecules contain a large number
of functions that can participate in the photoreactions, changes
of a few of such groups during the reactions are not expected
to alter significantly the ability of the polymers to perform their
tasks. Such systems could offer a high degree of processing
flexibility as they may generate not only colloidal solutions but
also metallized polymeric films, membranes, or gels upon
solvent evaporation from the colloids. Furthermore, solution
studies were anticipated to help identify polymer combinations
yielding efficient particle generation in insoluble films derived
from such macromolecular materials. This expectation was based
on earlier results showing that solution experiments were useful
as fast tests for optimizing the polymeric compositions of
films.18c
cold liquids. Product characterization via FTIR, NMR, and
elementary analysis (Atlantic Microlab) yielded results in good
agreement with published SPEEK data.30b
Separate solutions of PVA and of the sodium salt of SPEEK
were prepared by dissolving each of the polymers in water at
70 °C and combining them under stirring, keeping a PVA/
SPEEK molar ratio of 20:1 in all experiments. The mixture was
placed in a glovebag filled with nitrogen and then bubbled
slowly with N2 for 1 h to avoid severe foaming of PVA.
Aqueous solutions of the Ag, Au, and Cu metal ions were
bubbled with nitrogen for 20 min inside the glovebag. Finally,
the polymer solution was combined with one containing a metal
precursor, yielding final concentrations of 1.6 wt % PVA and
0.7 wt % SPEEK. In terms of monomer units, these concentra-
tions correspond to 0.36 M in the case of PVA and 0.018 M
for SPEEK. Such a polymer combination was identified in
preliminary experiments as that enabling the most efficient
photogeneration of Ag crystallites. The kinematic viscosity of
aqueous solutions containing the 1:20 SPEEK/PVA molar ratio
was determined at room temperature by means of a Cannon-
Fenske CFRC 9721-1359 glass viscometer. Detection of poly-
mer crosslinking during illumination was not feasible with this
technique because of the relatively high solution viscosity and
the short illuminations required for completion of the photo-
chemical reactions.
The strategy based on polymeric materials as key components
for the initiation of particle formation was validated in recent
investigations.29 Efficient generation of nanometer-sized Ag
crystallites was demonstrated upon exposure to 350 nm light
of insoluble films made from poly(vinyl alcohol) (PVA),
sulfonated poly(ether ether ketone) (SPEEK), and Ag+ ions.
Selection of such a combination aimed at mimicking the solution
photoreactions of the BP/2-propanol system, which yields
reducing R-hydroxy radicals commonly known as benzophenyl
ketyl (BPK) radicals.22 The benzophenone groups of SPEEK
were envisioned to act in a way analogous to BP molecules,
whereas PVA was intended to fulfill the role of 2-propanol as
an H atom donor. Illumination produced reducing species with
an absorption band centered at 565 nm, thought to be BPK
radicals of the polyketone (hereafter referred to as SPEEK
radicals).29 The long lifetime of the polymer radicals allowed
confirmation of their participation in the reduction of Ag+. A
preliminary report showed that extension of the polymer
photoreactions to solutions was possible since PVA and SPEEK
are water-soluble.29b Examined here are the kinetic features of
the SPEEK-sensitized generation of stable Ag, Au, and Cu
crystallites in aqueous media. Evidence is presented that the
metal ions affect significantly the photoprocesses of SPEEK,
altering the rates of particle formation.
Milton Roy #33-17-80 optical test tubes (path length ) 1
cm) that had flattened bottoms served as irradiation vessels. The
tubes were sealed with rubber septa after introduction of the
solutions, which were stirred throughout the photolysis. Mixtures
containing HAuCl4 were kept at 4-5 °C until shortly before
irradiation to prevent the thermal reduction of Au ions by PVA.31
Uniform illumination of the optical tubes was achieved by
positioning them at midheight inside a Rayonet 100 circular
illuminator that generated photons with λ ) 350 ( 15 nm by
means of 16 RPR-3500A lamps. The temperature inside the
illuminated Rayonet was 29 °C; an Aberchrome 540 was used
to determine the light intensity (I0).32 Optical densities were
measured on a Shimadzu UV-vis PC2501 spectrophotometer.
Kinetic data derived from optical determinations were typified
by a maximum deviation of 30%.
Transmission electron microscopy (TEM) measurements were
carried out by employing a JEOL JEM-2010 microscope with
an accelerating voltage of 200 kV or on a Zeiss EM 10CR
operating at 60 kV. Specimens were prepared on Cu grids
covered with holey-carbon films for Ag and Au colloids. The
grids were dipped into the colloidal solutions, excess liquid was
removed, and the samples were dried under vacuum. Ni grids
were used to image copper colloids; specimen preparation was
conducted inside a glovebag to avoid oxidation of Cu particles
by air. Characterization of the air-stable Ag and Au crystallites
employed X-ray diffraction (XRD) because electron diffraction
measurements were hindered by films of the polymers deposited
on the TEM grids during sample preparation. Samples for XRD
determinations were prepared via colloid centrifugation and
dispersion of the precipitates in water. This procedure was
repeated twice, and the resulting solids were dried under
vacuum, coated onto glass fibers, and fixed with an epoxy resin.
The fibers were mounted on a Bruker SMART APEX charge
coupled device (CCD) diffractometer that employed Mo KR
radiation. Comparisons of average particle diameters obtained
from TEM and XRD determinations were not meaningful given
that the method of sample preparation for diffraction experi-
ments induced particle growth.
Experimental Section
Poly(vinyl alcohol), 99+% hydrolyzed (average molar mass
) 8.9-9.8 × 104 g/mol), was obtained from Aldrich. Poly-
(ether ether ketone) (PEEK) with an average molar mass of 4.5
× 104 g/mol was a gift from Victrex, Inc. AgClO4‚H2O
(Aldrich), HAuCl4‚3H2O (Alfa), CuSO4‚5H2O, HCl, HClO4,
NaOH, and H2SO4 (Fisher) were used as received. Glassware
employed for colloid synthesis was treated with aqua regia;
deionized water from an ion-exchange resin deionizer (U. S.
Filter Service) was used in all procedures. Water-soluble SPEEK
was obtained by complete sulfonation of PEEK using literature
procedures,30 which will be described only briefly. The polyke-
tone was ground to a fine powder and dried under vacuum at
150 °C. A suspension consisting of 150 g of PEEK per L of
96% H2SO4 was heated at 50 °C for several days under constant
stirring, inducing sulfonation followed by dissolution of the
resulting SPEEK. Combination of the resulting solution with
an H2O/HCl mixture (5:1 v/v) precipitated the product. Occluded
H2SO4 was extracted from the sulfonic acid derivative via three
water washings, followed by neutralization with a 0.02 M NaOH
solution, rinsing with H2O, and drying at 150 °C under vacuum.
Losses of the soluble product were minimized by using ice-

Photogeneration of Small Ag and Au Crystallites
J. Phys. Chem. B, Vol. 109, No. 16, 2005 7735
represents the H atom abstraction reaction leading to SPEEK
radicals
where the first structure on the left side corresponds the triplet
excited state of a BP unit in an SPEEK chain. Abstraction of H
atoms from PVA in solution is known to predominantly produce
R-hydroxy radicals that behave as reducing agents.34 For this
reason, only such types of PVA radicals were included as
products in reaction 2.
Figure 1. UV-vis spectra of air-free SPEEK/PVA solution: before
irradiation, (0); after 1 min of irradiation, (i); and after 3, 5, 10, 20,
and 33 min in the dark; λ ) 350 nm and I0 ) 3.4 × 10-5 M(hV)/s.
Inset: Second-order plot for the dark decay of the polymer radical
absorption at 565 nm.
Hydrogen atom abstraction from 2-propanol mainly generates
(CH3)2C•OH radicals that reduce benzophenone molecules into
BPK radicals, (Ph)2C•OH, with a rate constant of k ) 4 × 105
M-1 s-1 35
The similarities in the structures of the R-hydroxy
.
Results and Discussion
radicals derived from 2-propanol and from PVA suggest that
reduction of the BP groups of SPEEK by the polyol radicals is
feasible. This process is represented by reaction 3
Sulfonation of PEEK allowed to use the BP groups of the
soluble polymer derivative as initiators of photoreactions in
aqueous solutions. Slow transformations accompanied by weak
optical signals resulted from illumination of air-free SPEEK
solutions, but large absorptions developed fast in the presence
of PVA. Figure 1 depicts the optical spectra of such a SPEEK/
PVA solution photolyzed for 1 min, showing a band with λmax
) 565 nm. Illumination of films composed of these polymers
produced an identical optical signal that was attributed to SPEEK
radicals.30 Short-lived absorptions with similar shape and λmax
value (560 nm) have been detected during the photoreduction
in the presence of 2-propanol of water-soluble benzophenone
derivatives containing alkyl sulfonate functions in the para
position.33 The photogenerated transient species are believed
to be BPK radicals of the sulfonated BP derivatives. Interest-
ingly, allowing completion of the radical decay process followed
by another illumination step re-forms the absorption with λmax
) 565 nm depicted in Figure 1. No such signal results when
air-saturated solutions of the polymers are irradiated. These
observations and the spectral results from experiments with alkyl
sulfonate derivatives of BP support the interpretation that
SPEEK radicals are responsible for the band centered at 565
nm.
Photogeneration of BPK radicals in SPEEK/PVA systems can
be understood on the basis of the known light-induced trans-
formations of benzophenone.22 Light absorption by BP yields
first a singlet (n, π*) excited state, which converts fast through
intersystem crossing into a triplet (n, π*) excited state able to
abstract an H atom from 2-propanol. A similar path for the
formation of polymer radicals seems reasonable given the
structural analogies between benzophenone molecules and the
BP groups of SPEEK as well as those existing between
2-propanol and PVA. Reaction 1 represents the proposed overall
electronic transition to the triplet excited state
{(CH2)C•OH}y + {R′RCdO}z f
{R′RC•OH}z + {(CH2)CO}y (3)
The efficiency of this secondary formation path of SPEEK
radicals is difficult to estimate because PVA radicals decay in
water with rate constants ranging between 1 and 7 × 109 M-1
s-1 depending on the radical concentration.34 Also, the partici-
pating macromolecular species in the SPEEK/PVA solutions
are anticipated to diffuse slower than benzophenone molecules
and (CH3)2C•OH. However, the high concentration of BP groups
from SPEEK (0.018 M) is expected to facilitate the reduction
of the polymeric carbonyl groups through reaction 3. Conse-
quently, the optical signal shown in Figure 1 after 1 min of
photolysis reflects the total amount of SPEEK radicals produced
via reactions 2 and 3.
The extinction coefficient (ꢀ) of SPEEK at 350 nm in aqueous
solution is 900 M-1 cm-1. This means that most of the incident
photons are absorbed by the polyketone at the employed SPEEK
concentration because the solutions exhibit an absorbance much
larger than 1 at 350 nm. Thus, an initial quantum yield for the
formation of SPEEK radicals (øi(SPEEK•) ) ri(SPEEK•)/I0) can
be evaluated from I0 ) 3.4 × 10-5 M(hV)/s and the initial rate
of radical generation, ri(SPEEK•). Correcting the radical signal
from the absorbance recorded prior to photolysis yields an ini-
tial change in optical density of ∆iOD equal to 2.5 × 10-3 s-1
at 565 nm. Considering that ri(SPEEK•) ) ∆iOD/ꢀ and assum-
ing that the extinction coefficient of SPEEK radicals is close
to the value reported for (Ph)2C•OH radicals (ꢀ ) 3.5 × 103
M-1 cm-1),35 the calculation results in øi(SPEEK•) ) 0.02. This
estimate is only a lower limit because, as illustrated in Fig-
ure 1, spectral data recorded after photolysis is stopped indicate
that the radicals decay in the time range of several minutes.
Also, øi(SPEEK•) will be affected by the cross reaction be-
tween polymer radicals similar to the one noticed for (Ph)2C•-
OH and (CH3)2C•OH that occurs with k ) 8 × 108 M-1
{R′RCdO}z + hV f {3R′RCdO}z*
(1)
where R and R′ are the groups attached to the carbonyl function
within the SPEEK. The structures of the repeating units of each
polymer that undergo changes are presented in reaction 2, which

7736 J. Phys. Chem. B, Vol. 109, No. 16, 2005
Korchev et al.
s-1 in acetonitrile.35 Equation 4 represents the cross reaction of
SPEEK radicals with PVA radicals
{(CH2)C•OH}y + {R′RC•OH}z f products
(4)
for which the rate constant may be even higher as was
determined for the decay of PVA radicals in water.34 Hence,
the reaction between SPEEK radicals and PVA radicals is
expected to lower the yield of the former species under steady-
state photolysis particularly because, as shown below, the
polyketone radicals exhibit a rather long lifetime.
Included in the inset of Figure 1 is a second-order plot derived
from the corrected OD at 565 nm during the first 20 min of the
decay step (3 half-lives). A straight-line relationship is obtained
with a slope of 5.02 OD-1 min-1 (r2 ) 0.995). The use of the
above-mentioned extinction coefficient for BPK radicals yields
a rate constant for the decay of SPEEK radicals equal to k(decay)
) 290 M-1 s-1. Essentially the same decay rate constant (within
experimental error) results for photolytic periods between 10
and 60 s, followed by monitoring only OD at 565 nm as a
function of time. These results indicate that species absorbing
at 565 nm decay via radical-radical disproportionation and/or
dimerization typical of R-hydroxy radicals
Figure 2. Evolution of normalized optical spectra during photolysis
of a solution containing 6 × 10-5 M AgClO4, PVA, and SPEEK; I0 )
2.4 × 10-5 M(hV)/s. The irradiation time increases according to 5, 10,
15, 20, 25, 30, 35, 40, 45, 50, 60, and 70 s.
{R′RC•OH}z + {R′RC•OH}z f
dimerization/disproportionation (5)
Moderate stirring of the solutions after photolysis accelerates
the radical decay noticeably, yielding in a rate constant 2.5 times
higher than the k5 value obtained in unstirred solutions. This
effect arises from the diffusional restrictions that the macro-
molecules experience since the viscosity of the concentrated
polymer solutions is 4 times the value of H2O. Nevertheless,
the k5 values for SPEEK radicals are several orders of magnitude
smaller than the decay rate constant in water for (Ph)2C•OH
and (CH3)2C•OH (1 × 109 M-1 s-1 in both cases).36,37
Negatively charged radicals of poly(acrylic acid) (PAA) formed
at pH > pKa exhibit a much longer lifetime than that of their
neutral counterparts generated in acid solution.38 As for poly-
carboxylate radicals, electrostatic repulsions between anionic
SPEEK chains increase the radical lifetime by hindering reactive
collisions between the macromolecules.
Photoreduction of Ag+ Ions. Illustrated in Figure 2 is the
evolution with illumination time of the optical spectra from a
SPEEK/PVA solution also containing 60 µM Ag+ ions. The
spectra are corrected from the spectrum acquired prior to
illumination. A broad absorption with λmax ) 415 nm is detected
initially, but the optical signal increases in strength and blue-
shifts with irradiation time to reach 395 nm at the end of the
photoreaction. This signal corresponds to the surface plasmon
of Ag particles with nanometer dimensions.4 The evolution of
this absorption with reaction time has been well-docu-
mented14,18,25,37 and originates from slight changes in the λmax
and strength of the plasmon with increasing particle size.4
Presented in Figure 3a is a TEM image confirming that fairly
spherical Ag particles form during the photochemical process.
Depicted in Figure 3b is a histogram of crystallite diameters
showing that the most frequent particle size is 6 nm whereas
the average diameter amounts to 7.4 nm. XRD measurements
verified the metallic nature of the particles; the relevant data is
presented in the Supporting Information.
Figure 3. (a) TEM image of a colloid made by photolysis of a SPEEK/
PVA solution containing 6 × 10-5 M AgClO4 and the (b) corresponding
histogram of particle diameters.
As shown in Figure 2, the metal particles present at the end
of the photoreaction possess a very strong and narrow plasmon
band. Reduction of silver ions in the range of 18 µM e [Ag+]
e 250 µM yields signal intensities varying linearly with metal
concentration and an ꢀ value (at 400 nm and per mole of Ag
atoms) of 2.4 × 104 M-1 cm-1. This extinction coefficient is
about 5 times higher than that of particles photogenerated in

Photogeneration of Small Ag and Au Crystallites
J. Phys. Chem. B, Vol. 109, No. 16, 2005 7737
PAA/PVA solutions18c and more than 3 times the ꢀ value of
Ag crystallites formed in 2-propanol.14 Particles exhibiting
unusually strong plasmons have been made in aqueous solutions
of several polymeric stabilizers by means of the acetone/2-
propanol sensitizing system,23b in micelles using benzophenone
as photosensitizer,25a or via radiolysis in methanol containing
PVA.39 Colloids produced radiolytically consist of crystallites
spanning the size range of 2-15 nm, which resembles the
diameter distribution depicted in Figure 3b. Thus, the strong
plasmon is not a consequence of a specific method of prepara-
tion or of a particularly narrow size distribution. Common
features of colloids with an intense optical signal are their
average particle diameter of about 7 nm and their sensitivity to
oxidation by air. The reaction with O2 diminishes the strength
and red-shifts the plasmon, but the signal intensity decreases
only by 20% in the case of the SPEEK/PVA system as compared
with at least 50% for the colloids studied earlier.23b,39
Photochemical generation of Ag crystallites also takes place
by irradiation with 350 nm light of solutions containing Ag+
and PVA in the presence or absence of PAA.18b Such a process
is not an important contributor to the Ag+ photoreduction in
the presence of SPEEK since the polyketone absorbs most of
the impinging photons. This conclusion is supported by the large
differences between the plasmon depicted in Figure 2 and those
of Ag particles photogenerated in the absence of SPEEK.18c
Another interesting result shown in Figure 2 is that the band at
565 nm starts being noticed when about 80% of the Ag+ ions
have been reduced. The signal due to SPEEK radicals becomes
particularly discernible after particle formation stops, serving
as an indicator that the photoreduction of Ag+ is completed.
Further insight about the role of the macromolecular radicals
in the reduction process is provided by experiments where the
radicals were created (without Ag+ ions) by illumination of
either SPEEK/PVA films or of solutions containing the two
polymers.29 Metal colloids are produced by mixing a photolyzed
SPEEK/PVA solution with a nonirradiated solution containing
AgClO4 in the absence of air. Also, Ag particles result inside
irradiated films after immersing the polymeric solids into
solutions of Ag+. In contrast, illumination of solutions contain-
ing silver ions and only PVA generates metal crystallites 200
times slower than in the presence of SPEEK. These findings,
together with the data of Figures 1 and 2, are indisputable
evidence that SPEEK radicals act as reductants toward Ag+ ions.
The simplest representation of the reduction process is
Figure 4. Extent of reaction as a function of illumination time for
SPEEK/PVA solutions containing different AgClO4 concentrations: 1.8
× 10-5 M (9); 3.6 × 10-5 M (b); 6 × 10-5 M (2); 9 × 10-5 M ([);
I0 ) 2.4 × 10-5 M(hV)/s.
a rate constant of k ) 4 × 109 M-1 s-1 for the reduction of
Ag+ by BPK radicals in micelles was extracted via multiexpo-
nential fitting of data from a recent laser photolysis study.25a
Such high rate constant is hard to reconcile with earlier
unsuccessful efforts to detect the reaction of Ag+ with the
stronger reductant (CH3)2C•OH.
Inclusion of the H+ concentration effect on E°[(CH3)2C•OH/
(CH3)2CdO, H+] yields an oxidation potential of 1.8 V at pH
) 7,5 demonstrating that reduction of Ag+ by the radical is not
necessarily endergonic. Considering that such a pH is typical
of degassed SPEEK/PVA solutions, the corrected standard
potential of •CH2OH leads to an oxidation potential for SPEEK
radicals smaller than 1.8 V but larger than 1.6 V. Although
reaction 4 is slightly endergonic, complexation of Ag atoms
with Ag+ to form Ag2+ can provide an additional driving force
for the reduction of silver ions by SPEEK radicals. Coupling
of the fast complexation step (k ) 8 × 109 M-1 s-1) to the
Ag+ reduction by (CH3)2C•OH is thought to enable the redox
process.5 In fact, the data of Figure 1 suggest a kinetic rather
than thermodynamic origin for the failed efforts to detect the
reaction between Ag+ and (CH3)2C•OH. Time-resolved mea-
surements require high radical concentrations, but such condi-
tions accelerate the second-order radical decay. Since k(Ag+ +
{R′RC•OH}z + Ag+ f {R′RCdO}z + Ag + H+ (6)
(CH3)2C•OH) < 1 × 106 M-1 s-1 23b
,
the radical-radical decay
At this point, a relevant issue concerns the oxidation potential
of SPEEK radicals, which is expected to be similar to the value
for (Ph)2C•OH radicals. An estimate of the reducing power
of (Ph)2C•OH can be derived after considering that BP is re-
duced in solution by (CH3)2C•OH whereas the methanol radical
(•CH2OH) is unable to induce such reduction.40 The standard
oxidation potentials of the two alcohol radicals are E°[(CH3)2C•-
OH/(CH3)2CdO, H+] ) 1.4 V and E°[•CH2OH/(H2CdO, H+]
) 1.2 V.37 Consequently, E°[(Ph)2C•OH/(Ph)2CdO, H+] is
lower than 1.4 V but higher than 1.2 V, which agrees well with
the E° value of 1.31 V estimated from half-wave potentials.41
Failed attempts to detect the reaction between (CH3)2C•OH and
Ag+ to form Ag atoms with time-resolved methods were
considered evidence that such process is not possible since the
estimated reduction potential for the [Ag+/Ag] couple is very
negative, E ) -1.8 V.23b This conclusion is at odds with our
findings and with the reported photogeneration of Ag crystallites
using (CH3)2CO and (Ph)2CO as sensitizers.23,25 Surprisingly,
channel is significantly faster than the reaction with silver ions.
However, the much longer lifetime of SPEEK radicals allows
them to reduce silver ions efficiently.
Formation of Ag crystallites in water proceeds via a complex
mechanism comprising fast reactions of cationic as well as of
neutral silver clusters.42 Analogous reactions seem to be involved
in the SPEEK/PVA system because the spectra exhibits none
of the long-lived absorptions from polymer-stabilized Ag
clusters found in PAA solutions.42a Presented in Figure 4 is the
advance of reaction (x ) ODt/OD∞) as a function of illumination
time for several metal ion concentrations. Evaluation of x used
the optical data at 400 nm and assumed a nearly constant ꢀ
value for the particles throughout the reaction. In all cases, the
initial stage is characterized by a slow variation of x with time,
followed by a period of increasing acceleration and a final step
of declining rates. The initial rate of particle formation (ri)
decreases steadily when the Ag+ concentration increases
between 36 and 500 µM. Figure 5 demonstrates that the

7738 J. Phys. Chem. B, Vol. 109, No. 16, 2005
Korchev et al.
((4-6) × 1010 M-1 s-1).25a The k7 values are more than an order
of magnitude larger than the quenching constants of benzophe-
none triplets by O2 or BP molecules in H2O.36 Such high rate
constants are inconsistent with diffusion-controlled processes
given the considerable viscosity of the solutions imparted by
their high polymer concentrations but can be rationalized in
terms of the ionic nature of SPEEK. The solution behavior of
polyelectrolytes is influenced by strong Coulombic attractions
between charged macromolecular groups and their counterions.43
For SPEEK, the attractions are expected to keep part of the
Ag+ ions in close proximity of the polymeric sulfonic groups.
Fast quenching of BP triplet states by Ag+ ions located next to
the polymeric charges is possible because of the short distance
separating the sulfonic groups from their adjacent carbonyl
functions.
The experimental intercept results in r(3SPEEK*) ) 9.1 ×
10-7 M/s, and division by I0 yields 0.04 as the quantum yield
of excited BP groups, ø(3SPEEK*). A possible reason for such
low quantum yield is that intersystem crossing from a singlet
to triplet (n, π*) state is less effective for SPEEK than for BP,
which is 100% efficient.36a,b For instance, intersystem crossing
may partially lead to triplet (π, π*) states, which in the case of
4,4′ substituted benzophenones participate less effectively in
the generation of BPK radicals via H atom abstraction.36b
Alternatively, decay of the (n, π*) state may be faster in the
case of the polymeric material than for benzophenone molecules,
or ground-state BP groups of SPEEK may quench the excited
states in a way similar to the self-quenching that the aromatic
ketone experiences in H2O.36a However, self-quenching must
be restricted by the considerable entanglement of SPEEK chains
with PVA macromolecules that is anticipated at the high polyol
concentration. Another possibility is the formation of ground-
state complexes of SPEEK analogous to those responsible for
the emission exhibited by PEEK.36c In fact, SPEEK exhibits an
emission very similar to that of PEEK, which hampered attempts
to detect fluorescence signals originating from the Ag particles.
A maximum quantum yield of radical formation equal to 0.08
is predicted from ø(3SPEEK*) after considering reactions 2 and
3, but øi(SPEEK•) is only 0.02 for solutions without Ag+ ions.
The low radical yield is probably a result of reaction 4 since
the cross reaction consumes SPEEK radicals and also prevents
their further formation via reaction 3.
Figure 5. Dependence of the reciprocal initial reaction rates as a
function of the metal concentration: Ag+ (b); Cu2+ (9).
reciprocal of the initial rate varies linearly with metal ion
concentration according to the relationship 1/ri ) 1.1 × 106 +
3.6 × 1010[Ag+] (r2 ) 0.993). These observations are atypical
because the rate of crystallite formation has been frequently
found to increase linearly with the Ag+ concentration.14,18
Quenching of excited BP molecules by Ag+ occurs during
the photochemical formation of Ag particles in micellar solutions
sensitized by benzophenone.25a A similar reaction is expected
to take place in the polymer solutions
{3R′RCdO*}z +8 {R′RCdO}z
(7)
Ag
Both energy and electron-transfer mechanisms can be involved
during quenching of singlet and triplet excited states of aromatic
ketones by metal complexes.13 Reaction 7 is formulated as an
energy-transfer process because the electron-transfer route is
expected to produce metal particles via formation of Ag atoms
and is unable to explain the data of Figure 5. The decrease in
ri with increasing Ag+ concentration is understood in terms of
a competition between reactions 2 and 7. Only those excited
BP groups of the macromolecules not quenched by Ag+ are
able to form SPEEK radicals that, ultimately, produce Ag
crystallites. Equation 8 is obtained from standard competition
analysis13
An accelerated formation of Ag together with a maximum
reaction rate are noticed in Figure 4 at intermediate illumination
times. Indeed, the highest initial quantum yield of particle
generation is 0.013 at [Ag+] ) 18 µM, whereas the quantum
yield calculated from the maximum reduction rate (ø(Agn)) is
0.06 for this system. Fairly similar rate maxima were derived
from the data of Figure 4; in all cases ø(Agn) > øi(SPEEK•),
meaning that reaction 6 is able to compete with reaction 4. Also,
a topmost ø(Agn) value of 0.07 results from the data for [Ag+]
) 90 µM, which is close to the maximum quantum yield of
SPEEK radical formation predicted from ø(3SPEEK*). The
agreement between experimental and extrapolated quantum
yields indicates that reaction 3 plays a substantial role in the
metal formation. Direct Ag+ reduction by PVA radicals is
deemed less important due to their short lifetime in solution34
and also because the concentration of polymeric BP groups is
over 100 times higher than the Ag+ concentration.
ri ) r(3SPEEK*) k2[PVA]/(k2[PVA] + k7[Ag+]i) (8)
where r(3SPEEK*) represents the formation rate of triplet
excited states from polymeric BP groups, while k2 and k7 are
the rate constants for reactions 2 and 7, respectively. Inversion
of eq 8 yields an expression matching the relationship obtained
from Figure 5
1/ri ) 1/r(3SPEEK*){1 + k7[Ag+]i/k2[PVA]}
(9)
Evaluation of k7 from the slope and intercept of the straight
line shown in Figure 5 and the concentration of monomer units
present in PVA (0.36 M) requires a k2 value. In water, the rate
constants for H atom abstraction from ethanol and 2-propanol
by the triplet state of BP range between 2 and 3 × 106 M-1
The rate of reaction 7 gradually decreases as photolysis
proceeds due to the concomitant decline in Ag+ concentration,
which favors particle formation. Interestingly, the reduction of
s-1 25a,36
.
The use of similar k2 values results in k7 ) (2-3) ×
•
1010 M-1 s-1, which agrees reasonably with the quenching
constant of the BP triplet state by Ag+ in micellar solutions
Ag+ by CH2OH radicals is known to be efficient only when
the silver ions are bound to metal crystallites or clusters.39
A

Photogeneration of Small Ag and Au Crystallites
J. Phys. Chem. B, Vol. 109, No. 16, 2005 7739
distribution of hydroxo copper(II) species remains controver-
sial.46 Thus, the yellow material seems to originate from
reduction of hydroxo complexes present in the neutral polymeric
solutions. According to Figure 6, development of the signal with
λmax ) 460 nm is completed in about 200 s. This band is less
well-defined thereafter as the CuI compound starts transforming
into metallic Cu, because the metal particles also absorb strongly
below 530 nm. Optical data gathered at 460 and 575 nm when
the first formation step ends in solutions with 15 e [Cu2+] e
50 µM yields ꢀ values for the cuprous compound of 2.7 × 103
and 1.5 × 103 M-1 cm-1, respectively. An interesting observa-
tion is that the CuI product remains unchanged when illumina-
tion is stopped, indicating that the transformation of this material
into metal requires the action of light.
Identification of the product from the second step is facilitated
by the emergence in Figure 6 of the characteristic surface
plasmon of nanometer-sized Cu crystallites centered between
570 and 580 nm.4 Copper particles photogenerated in SPEEK/
PVA solutions exhibit an extinction coefficient of 3.2 × 103
M-1 cm-1 at 575 nm. Sequential red and blue shifts of the
plasmon similar to those shown in Figure 6 occur frequently
during formation of Cu crystallites44 and probably originate from
slight changes in the optical properties of particles with different
sizes.4 The long-term instability of Cu particles in the presence
of O2 and water made their characterization by TEM mandatory,
since the absence of air during such measurements prevented
oxidation reactions. Very laborious experiments were needed
given that only a few particles survived the sample transfer into
the microscope and because electron diffraction analysis was
restricted to grid regions relatively free of polymer films
deposited during sample preparation. Figure 7a is a bright-field
image demonstrating that the particles are nearly spherical but
with a very broad size distribution of 5-40 nm. The corre-
sponding ring pattern shown in Figure 7b depicts the typical
reflections of face-centered cubic (fcc) Cu, confirming the
metallic nature of the nanoparticles.
Figure 6. Evolution of the optical spectra during photolysis of a
solution containing 2.6 × 10-4 M CuSO4, PVA, and SPEEK; I0 )6 ×
10-5 M(hV)/s. The irradiation time increases according to 0, 30, 50,
70, 90, 110, 130, 150, 170, 200, 230, 260, 290, 330, 390, 450, 510,
570, 630, and 780 s.
similar increase in the rate of reaction 6 must take place when
metal is present in the SPEEK/PVA solutions because the
photoreaction accelerates after at least 20% of the Ag+ ions
have been reduced. Another unusual kinetic feature noticed in
Figure 4 is that the rate of metal formation reaches a maximum
at x e 0.42 for [Ag+] e 36 µM, which is when the concentration
of unreduced metal ions is between 15 and 21 µM. In contrast,
the maximum photoreduction rate for solutions with [Ag+] g
60 µM occurs at x g 0.91, when about 6 µM of metal ions are
left unreduced. Thus, significant retardation of the metal ion
photoreduction occurs in solutions containing the highest Ag+
concentrations. Protons are known to quench the triplet state
of BP in aqueous solution.36 Quenching of SPEEK carbonyl
triplets by protons seems logical in systems with high Ag+
concentrations, where large H+ concentrations are generated
through reaction 6. Hence, the shapes of the curves shown in
Figure 4 originate from the quenching of polymeric triplet BP
states by Ag+ and H+, in conjunction with the accelerating effect
induced by metallic clusters/particles.
Figure 8 illustrates the evolution of the advance of reaction
as a function of photolysis time for solutions containing three
different initial concentrations of CuSO4. At a first glance, the
meaning of the plots seems questionable given that calculation
of x employs kinetic data gathered at 575 nm together with OD∞
values corresponding to the full formation of the Cu crystallites.
In fact, the plots aim at depicting the temporal evolution of
transformations involving two-electron-transfer processes. Such
plots are feasible because the ꢀ value for CuI at 575 nm is,
fortuitously, approximately half of the extinction coefficient of
Cu particles. As a consequence of this difference in ꢀ values, x
reaches 0.5 in all plots of Figure 8 upon completion of the first
step producing CuI and increases thereafter to a limiting value
of 1 as metal formation proceeds. Monitoring the evolution of
the signal at 460 nm allowed independent verification that the
first step was completed.
Photoreduction of Cu2+ Ions. Exposure of air-free solutions
containing 2.5 × 10-4 M CuSO4, PVA, and SPEEK to 350 nm
photons induces several spectral changes illustrated in Figure
6. Two consecutive steps can be identified from the optical data;
the initial process generates a broad absorption with λmax
between 460 and 470 nm, which is accompanied by a slightly
turbid yellow coloration of the solution. The second transforma-
tion produces a signal centered at 575 nm and a simultaneous
change in the solution color from yellow to red. Similar observa-
tions have been made during the radiolytic reduction of Cu2+
in aqueous PVA solutions.44 The initial product is believed to
consist of suspended copper(I) hydroxide particles, whereas the
material formed during the second process exhibits properties
typical of Cu crystallites. Having in mind that CuI and the
metallic product are sensitive to air in H2O,44 support for these
assignments is provided by the fact that exposure to O2 of
SPEEK/PVA solutions containing either of the chromophores
induces fast disappearance of their characteristic absorptions.
Reduction of CuII to particles of a yellow CuI product,
presumed to be CuOH, has been observed only in strongly basic
solutions.45 A similar material is formed when free radicals
reduce cupric ions in neutral solutions either through photolysis
of SPEEK or via radiolysis.44 Cu2+ ions undergo hydrolysis in
neutral solution forming Cu(OH)+ as the main product, but the
Formation of a CuI compound implies that SPEEK radicals
act as reducing agents toward Cu2+ complexes according to
{R′RC•OH}z + Cu2+ f {R′RCdO}z + Cu+ + H+ (10)
where Cu2+ represents all of the CuII species present in the
polymer solutions and Cu+ corresponds to the “CuOH” species.
Direct evidence for the redox process is provided by the outcome
of injecting a CuSO4 solution into a preirradiated polymeric
solution in the absence of air and light. As a result, the
absorption of SPEEK radicals decays instantly with concomitant
formation of the signal attributed to the CuI compound.
Reduction of Cu2+ by SPEEK radicals is not surprising

7740 J. Phys. Chem. B, Vol. 109, No. 16, 2005
Korchev et al.
species is explained if SPEEK radicals react faster with the
hydroxo complexes than with nonhydroxylated cupric ions.
As for the Ag system, the data of Figure 8 indicates that the
rates of the initial reaction decrease with increasing Cu2+
concentration. Included in Figure 5 is a plot of 1/ri versus [Cu2+],
where ri is the rate of the initial step; the resulting straight line
obeys the equation 1/ri ) 3 × 105 + 7.6 × 108[Cu2+]. Hence,
the triplet states of polymeric BP groups are quenched by the
CuII complexes via an energy-transfer process similar to reaction
7. Competition between reaction 2 and the quenching process
accounts for the inverse dependence of ri on Cu2+ concentration,
allowing derivation of expressions analogous to eqs 8 and 9.
Quenching constants of kQ ) (2-3) × 109 M-1 s-1 are obtained
using k2 values of (2-3) × 106 M-1 s-1 together with the slope
of the straight-line equation. These constants are 10 times lower
than those of Ag+, probably reflecting the different contributions
from the various CuII species present in the solutions. In
agreement with the Ag system, the experimental intercept
yields ø(3SPEEK*) ) 0.056, and the highest quantum yield of
CuI generation (ø(CuI) ) 0.04) not only results at the lowest
Cu2+ concentration but is within the maximum predicted from
ø(3SPEEK*). Also, ø(CuI) > øi(SPEEK•) at all Cu2+ concentra-
tions, meaning that reaction 10 competes effectively with
reaction 4. However, the data of Figure 8 indicates that the
photoreaction follows an apparent zero-order rate law, which
is confirmed by the linear increase of OD at 460 nm with time.
Thus, decreases in Cu2+ concentration during photolysis affects
both the quenching process and reaction 10 in a similar fashion
because the reduction process is not influenced by the presence
of the CuI product.
According to Figure 8, no simple relationship exists between
the rate of Cu particle formation and metal ion concentration
during the second step. For instance, similar quantum yields of
metal formation (0.007) are obtained at [Cu2+] e 300 µM, but
the efficiency drops at the highest metal ion concentration. Also,
the kinetics of the metal photogeneration are not entirely
reproducible in solutions containing [Cu2+] g 300 µM, hinting
that heterogeneous processes participate in this transformation.
Evidence exists that Cu+ ions are not reduced to metal atoms
by R-hydroxyalkyl radicals,48,49 because of the unfavorable redox
potential, E°[Cu+/Cu(atom)] ) -2.7 V.5 Hence, reduction of
the CuI product by SPEEK radicals appears unfeasible. This
conclusion is supported by the fact that mixing a solution
containing SPEEK radicals with a CuSO4 solution generates
exclusively the “CuOH” product. Cu+ ions disproportionate in
acidic aqueous solutions
Figure 7. (a) TEM image of a photochemically obtained copper colloid
according to the conditions of Figure 6. (b) Electron diffraction pattern
of copper crystallites. The reflections are in good agreement with the
data listed in the Joint Committee on Powder Diffraction Standards
(card 4-0836).
2Cu+ / Cu + Cu2+
(11)
and K ) 5.4 × 105 M-1 was determined from measurements
of the equilibrium on Cu electrodes,47 but a constant about 14
times smaller was obtained in the absence of bulk metal.48b The
results of the present investigation and of previous studies
indicate that the yellow CuI product is stable toward dispro-
portionation.44,49 Nevertheless, fast metal formation was ob-
served upon treatment of “CuOH” with a dilute acid solution.44
Figure 9 illustrates the spectral changes occurring in the absence
of air when a solution containing the yellow CuI product is
acidified with HClO4 to yield [H+] ) 7 × 10-4 M. Immediate
decay of the signal at 460 nm takes place, with concomitant
formation of the plasmon due to Cu crystallites. Comparison
of the optical data at 460 and at 575 nm with the spectra shown
in Figure 7 provide confirmation that the process involves a
2:1 molar ratio of CuI to metallic Cu, as required by reaction
11.
Figure 8. Extent of reaction as a function of illumination time for
SPEEK/PVA solutions containing different CuSO4 concentrations: 1.5
× 10-4 M (2); 3 × 10-4 M (9); 5 × 10-4 M (b).
considering that E°[Cu2+/Cu+] ) 0.168 V47 and that R-hy-
droxyalkyl radicals derived from CH3OH, CH3CH2OH, and
(CH3)2CHOH reduce cupric ions with rate constants between
1.6 and 0.5 × 108 M-1 s-1 48
.
These reactions have been
interpreted in terms of complexation of the alcohol radicals by
Cu2+ and formation of CuIII intermediates; reaction 10 is
expected to occur in a similar fashion. Formation of the “CuOH”

Photogeneration of Small Ag and Au Crystallites
J. Phys. Chem. B, Vol. 109, No. 16, 2005 7741
Figure 9. (1) UV-vis spectrum of a solution containing 2.6 × 10-4
M CuSO4, PVA and, SPEEK photolyzed for 200 s. (2) After injection
of 20 µL of 0.1 M HClO4.
Figure 10. Dependence of the induction period on HAuCl4 concentra-
tion in the presence of both polymers (9), I0 ) 5.6 × 10-5 M(hV)/s
and in the presence of only PVA (b), I0 ) 3.3 × 10-5 M(hV)/s.
These observations indicate that bound Cu+ is unable to
disproportionate but that reaction 11 takes place when cuprous
ions become free in solution after dissolution of the hydroxide
particles by protons. The dynamics of the particle photogenera-
tion depicted in Figure 8 can be rationalized on the basis of
these results since protons formed via reaction 10 accumulate
in the polymer solutions, and when x approaches 0.5 their
concentrations are close to the initial Cu2+ concentration.
Dissolution of the cuprous hydroxide becomes then appreciable,
and the resulting Cu+ ions start disproportionating to Cu2+ and
metal particles. SPEEK radicals continue to reduce cupric ions
through reaction 10, although such ions are now those resulting
from reaction 11. Coupling of reactions 10 and 11 explains the
similar slopes noticed in Figure 8 prior to and after x reaches a
value of 0.5; the slope changes noticed thereafter suggest that
another process participates in the generation of crystallites. Cu
particles are known to catalyze the reduction of CuI species by
R-hydroxyalkyl radicals.49 Having in mind that the concentration
of copper ions left unreduced is very small at x values close to
1 (which makes reaction 11 improbable), a metal-mediated
reduction of cuprous ions by SPEEK radicals seems a possible
explanation for the last reaction step.
-
Reduction of AuCl4 to the metal usually involves two
consecutive processes; the first step forms AuCl2- (λmax ) 246
nm) in homogeneous solution51 or species believed to be Au
clusters exhibiting signals centered at 350-380 nm in polymeric
systems.15a,17,50,52 The first step occurs with bleaching of the
-
yellow color of AuCl4 (λmax ) 319 nm),9a,31 followed by
photoreduction of the initial products to metallic Au. A similar
two-step transformation takes place during illumination of
SPEEK/PVA solutions in the presence of HAuCl4. Au crystal-
lites are not detected during the first step (referred to as the
induction period), which occurs without any color change for
solutions with [HAuCl4] < 300 µM since SPEEK absorbs most
of the light under such conditions. However, bleaching is noticed
at higher AuIII concentrations and in SPEEK-free solutions,
where a yellow coloration due to the auric ions can be detected.
Metal generation starts once the bleaching process is over;
particles are not produced when photolysis is stopped at the
end of the induction period, even when irradiated solutions are
left in the dark for several hours. These findings indicate that
the second step is a light-initiated process that begins after most
of the AuIII ions have been consumed and that this phototrans-
formation involves products from the first reaction.
-
Photoreduction of AuCl4 Ions. Elucidation of the trans-
formations of HAuCl4 induced by light in SPEEK/PVA solutions
faces challenges not encountered in the cases of Ag+ or Cu2+
.
Illustrated in the curve labeled with squares in Figure 10 is
the evolution of the induction period as a function of HAuCl4
concentration in photolyzed SPEEK/PVA solutions, whereas the
circles correspond to data from experiments performed with-
out SPEEK. Considerable uncertainty remains about the true
induction period for SPEEK/PVA solutions with 90 µM HAuCl4
given that the measured value (5 s) is within the time required
for the optical determinations. However, the longer induc-
tion period (65 s) for the lowest HAuCl4 concentration (100
µM) in the absence of SPEEK is less affected by the experi-
mental constraints. Nevertheless, the data of Figure 10 support
the conclusion that direct photoreduction of HAuCl4 is not
significant in SPEEK/PVA solutions because the induction
periods of these systems are, in general, considerably shorter
than those obtained in the absence of the polyketone. Figure
11 depicts the corresponding quantum yields of photoreduction
(ø(-HAuCl4)) as a function of AuIII concentration in the
presence of SPEEK and PVA. Evaluation of the quantum yields
assumed that the induction periods correspond to the times
required to photoreduce all of the auric ions initially present.
This assumption is supported by the fact that the induction
Indeed, radiation with λ e 350 nm initiates the reduction of
AuCl4 ions in the presence of reducing agents, including
-
alcohols.9a,15,17,50 Direct photoreduction of HAuCl4 seems
improbable in our system since the 350 nm photons are
overwhelmingly absorbed by SPEEK. Nevertheless, some
testing of the photoprocesses taking place in the absence of the
polyketone was conducted to assess their impact on the SPEEK-
initiated photoreaction. Specifically, the experiments aimed at
detecting possible chain reductions of the AuIII ions similar to
-
those induced by direct excitation of AuCl4 in polymer gels
containing methanol.17 AuCl4- ions are also reduced thermally
by PVA molecules with a molar mass similar to that of the
polyol used in the present study.31 Metal formation lasts more
than 6 days when 100 µM AuCl4- reacts with 3 mM PVA but
only a few hours in the SPEEK/PVA solutions where the PVA
concentration is approximately 100 times higher than the
concentration used earlier. Although the dark reaction is
suppressed before photolysis via the method described in the
Experimental Section, parallel occurrence of the thermal reduc-
tion during illumination is possible.

7742 J. Phys. Chem. B, Vol. 109, No. 16, 2005
Korchev et al.
Figure 12. Evolution of the optical spectra during photolysis of a
solution containing 2.7 × 10-4 M HAuCl4, PVA, and SPEEK; I0 )
5.6 × 10-5 M(hV)/s. The irradiation time increases according to 0, 3,
4, 5, 6, 7, 8, 9, 11, 12, 13.5, 15, 17, and 20 min.
Figure 11. Variation of the quantum yield for AuIII reduction as a
function of HAuCl4 concentration, I0 ) 5.6 × 10-5 M(hV)/s. Inset:
Concentration dependence of the reciprocal rate of photoreduction for
AuIII (b) and AuI (O).
signal centered at 565 nm is detected throughout the first step,
and mixing a solution containing photogenerated SPEEK
radicals with another containing HAuCl4 results in an instan-
taneous decay of the radical absorption. The large ø(-HAuCl4)
values can be explained if the SPEEK-initiated photoreaction
occurs in parallel with the known thermal reduction of AuIII by
PVA.31 Such an interpretation is supported by the slow room-
temperature bleaching of auric ions detected in solutions that
contain only PVA or at high HAuCl4 concentrations in SPEEK/
PVA mixtures. Another consequence of the negative intercept
extracted from Figure 11 is that calculation of kQ for AuIII via
an expression analogous to eq 9 is not possible. Utilization of
the k2 and ø(3SPEEK*) values used for Cu2+ yields a rough
estimate of kQ ≈ 7 × 109 M-1 s-1. This estimate is similar to
the kQ value of the cupric species, probably reflecting the
participation of several products from the HAuCl4 hydrolysis
in the quenching process.54
Figure 12 depicts the evolution of the optical spectra recorded
during the second step for a SPEEK/PVA solution containing
initially 270 µM HAuCl4. In this plot, the origin of time
corresponds to the end of the induction period; the initial
spectrum demonstrates that the first reaction occurs without
significant spectral changes above 400 nm. Photolysis beyond
the induction period induces a blue coloration of the solutions
and a broad absorption centered at about 550 nm. During the
first 11 min, the signal increases in strength, followed by a
substantial narrowing at longer times as well as a blue-shifting
of the maximum to 520 nm. The solutions turn increasingly
red, exhibiting the typical surface plasmon of spherical Au
particles.4 Plasmon intensities that vary linearly with Au
concentration are obtained upon complete metal ion reduction
in the range of 90 e [AuIII] e 550 µM. The corresponding
extinction coefficient (4 × 103 M-1 cm-1) is in good agreement
with ꢀ values reported earlier.53 Figure 13a is a typical TEM
image of a gold colloid, confirming that, apart from a few
aggregates of crystallites, the photoreaction forms spherical
particles. The histogram of crystallite diameters illustrated in
Figure 13b is evidence of a very narrow distribution since the
most frequent particle size coincides with the average diameter
of 3 nm. XRD data verifying the metallic nature of the
crystallites are included in the Supporting Information.
periods coincide with the times needed for complete bleaching
of the AuIII color. Not included in the figure is the quantum
efficiency (0.36) estimated for 90 µM HAuCl4 in the presence
of SPEEK due to the uncertainties mentioned above. Although
the quantum yields are higher than the efficiencies obtained for
the Ag+ and Cu2+ systems, the data are not consistent with a
chain reaction.
-
Detection of the optical signals from AuCl2 or clusters
produced via the AuIII reduction is impaired by the absorption
due to SPEEK. Photoreduction of AuIII in polymer films or gels
produces stable Au clusters,15,17 but these species exist only as
transients when auric ions are reduced in solution.31 Interest-
ingly, the products of the first step remain unchanged in the
dark for times longer than those needed to complete the
photogeneration of metal particles. This observation is consistent
with the fact that AuI disproportionates very slowly to form
Au even in the presence of metal.17,52,53 For simplicity, we have
assumed that aurous ions are photogenerated during the first
reaction, which means that the quantum yields of Figure 11
correspond to a two-electron transformation. According to
this interpretation, the photoreduction of HAuCl4 is at least 2
times more efficient than those for Ag+ and Cu2+, implying
that ø(3SPEEK*) and the quantum yield of radical formation
are higher for the AuIII system. Such conclusions are hard to
justify since the quantum yields of triplet states and of polymer
radicals found in Ag+ and Cu2+ solutions are mutually
consistent. Figure 11 shows that ø(-HAuCl4) decreases with
increasing AuIII concentration, and the inset demonstrates that
the reciprocal of the photoreduction rate (r(-HAuCl4)) is a linear
function of the metal ion concentration with 1/r(-HAuCl4) )
-1.8 × 105 + 2.6 × 109[HAuCl4]. Obviously, auric ions quench
the triplet states of polymeric BP groups in a way similar to
reaction 7. The negative value of the intercept means that ø(3-
SPEEK*) < 0, which has no physical meaning, but supports
the notion that the quantum yields depicted in Figure 11 are
abnormally high.
Reduction of AuCl4- to AuCl2- is characterized by a positive
potential, E° ) 0.92 V.54 In cationic polymer gels, the reduction
of AuIII by •CH2OH (a weaker reductant than SPEEK radicals)
generates clusters containing metal atoms bound to AuI.17 Hence,
the polymer radicals are anticipated to reduce AuIII ions, and
several findings corroborate such expectation. For example, no
According to the spectral changes depicted in Figure 12,
photoreduction of AuI first produces the chromophores respon-

Photogeneration of Small Ag and Au Crystallites
J. Phys. Chem. B, Vol. 109, No. 16, 2005 7743
Highlighted by arrows are discontinuities corresponding to the
starting points of the blue to red change. Evaluation of x used
the optical data at 520 nm together with the extinction coefficient
of the Au particles. Such a procedure underestimates the advance
of reaction at times shorter than those marked by the arrows
because the species responsible for the broad signals are known
to be weaker light absorbers at that wavelength than Au
particles.55 Despite this shortcoming, the results clearly show a
progressively less efficient evolution of x with an increasing
AuI concentration, signifying that aurous ions produced in the
first reaction quench the triplet states of polymeric BP groups.
Only lower limits of the initial photoreduction rates can be
obtained from Figure 14 in the absence of the extinction
coefficient at 520 nm for the species originating the broad
signals. Nevertheless, these rough estimates confirm the validity
of a competition expression similar to eq 8 since the plot
included in the inset of Figure 11 shows that the inverse of the
rate is linearly dependent on the AuI concentration with 1/ri )
4 × 104 + 6 × 109[AuI]. The intercept yields ø(3SPEEK*) )
0.45, which is at least 8 times larger than the quantum
efficiencies derived from the Ag+ and Cu2+ systems. In fact,
the actual ø(3SPEEK*) must be even larger since the real initial
rates are higher than those used for the plot shown in the inset
of Figure 11. For the same reason, the corresponding slope is
only a lower limit of the true value, making questionable any
attempt to evaluate kQ. As in the case of the initial step, the
most probable reason for the large photoreduction rates is the
occurrence of a process in parallel with the reaction involving
SPEEK radicals. However, this secondary reduction channel is
light-initiated because the products of the first step remain
unchanged in the dark within the time frame of the photochemi-
cal experiments. The exact nature of such reaction is not known,
but possible pathways include quenching of the singlet state of
excited polymeric BP groups via electron transfer or a chain
photoreaction similar to the process noticed in polymer gels.17
Metal crystallites form during photolysis from a reaction
between SPEEK radicals and the chromophores responsible for
the broad signals since the radical absorption at 565 nm is not
observed in Figure 12. Generation of Au particles from
precursors displaying a blue color has been noticed when
AuCl4- is reduced either thermally in some solutions56,57 or via
photolysis inside the channels of porous glasses.55 The blue
coloration originates from scattering of light by very large
aggregates containing weakly interacting Au species that do not
exhibit metallic diffraction signals. Broad spectra centered at
550 nm are also obtained when a HAuCl4 solution is mixed
with another containing SPEEK radicals. Irrespective of the
method of preparation (through direct photolysis or solution
mixing), the chromophores remain unchanged for about 100
min, followed by a sluggish emergence of the plasmon from
Au crystallites. Identification of the chromophores by TEM was
not feasible because of their short lifetime as well as the
interfering effects induced by the polymeric components of the
solutions. The dark transformation of the species responsible
for the blue color resembles the postirradiation reaction found
Figure 13. (a) TEM image of gold colloid obtained as indicated in
Figure 12 and the (b) corresponding histogram of particle diameters.
during radiolysis of AuCl4 solutions, where AuI ions bound
-
Figure 14. Extent of reaction as a function of illumination time for
to small metal particles are slowly converted into gold atoms
via disproportionation as well as through reduction by PVA.53
Occurrence of the latter reaction is not unreasonable given that
aurous ions are known to react with solvent molecules in water/
acetonitrile mixtures with formation of metal.58
SPEEK/PVA solutions containing different HAuCl4 concentrations: 9
× 10-5 M (O); 1.8 × 10-4 M (b); 3 × 10-4 M (2); 5.5 × 10-4
M
([).
sible for the blue color, followed by their subsequent transfor-
mation to Au particles at the later stages of the second step.
This transformation is also evident in the plots of Figure 14,
depicting the evolution of the advance during the second reaction
for solutions containing several initial AuIII concentrations.
The earlier findings suggest that aggregates of small Au
particles containing bound aurous ions are the species respon-
sible the blue coloration in SPEEK/PVA solutions. According

7744 J. Phys. Chem. B, Vol. 109, No. 16, 2005
Korchev et al.
R.; Van Duyne, R. P. J. Phys. Chem. B 1999, 103, 3854. (c) Storhoff, J. J.;
Mucic, R. C.; Mirkin, C. A. J. Cluster Sci. 1997, 8, 179.
(7) (a) Han, M. Y.; Quek, C. H. Langmuir 2000, 16, 362. (b) Esumi,
K.; Suzuki, A.; Aihara, N.; Usui, K.; Torigoe, K. Langmuir 1998, 14, 3157.
(c) Yonezawa, Y.; Sato, T.; Ohno, M.; Hada, H. J. Chem. Soc., Faraday
Trans. 1 1987, 83, 1559.
(8) (a) Belloni, J. In Homogeneous Photocatalysis; Chanon, M., Ed.;
John Wiley: New York, 1997; p 169. (b) Sahyun, M. R. V. J. Imaging
Sci. Technol. 1998, 42, 23.
(9) (a) Boitsova, T. B.; Gorbunova, V. V.; Loginov, A. V. Russ. J.
Gen. Chem. 2002, 72, 642. (b) Rogash, A. L.; Shevchenko, G. P.;
Afanas’eva, Z. M.; Sviridov, V. V. J. Phys. Chem. 1997, 101, 8129. (c)
Yonezawa, Y.; Konishi, Y.; Hada, H.; Yamamoto, K.; Ishida, H. Thin Solid
Films 1992, 218, 109.
(10) (a) Akamatsu, K.; Ikeda, S.; Nawafune, H. Langmuir 2003, 19,
10366. (b) Stellacci, F.; Bauer, C. A.; Meyer-Friedrichen, T.; Wenseleers,
W.; Alain, V.; Kuebler, S. M.; Pond, S. J.; Zhang, Y.; Marder, S. R.; Perry,
J. W. AdV. Mater. 2002, 14, 194.
to this scenario, part of the aurous ions formed in the first
reaction react with SPEEK radicals producing the aggregates.
Once this step is completed, AuI ions bound to the aggregates
are reduced by SPEEK radicals whereas the slower dark metal
formation probably results from their reaction with PVA. The
transformation of aggregates is envisioned to resemble the
process that yields gold particles in citrate solutions.57 Reduction
of the surface-bound AuI species induces fusion of the metal
particles, which transforms the aggregates into larger metal
crystallites with well-defined plasmon bands. Particle formation
processes that exhibit different levels of mechanistic complexity
have been encountered as well during the reduction of Pd and
Pt ions by SPEEK radicals. Kinetic data pertaining to such
systems will be present in a forthcoming report, which will also
demonstrate that metal colloids made photochemically from
SPEEK/PVA solutions can easily be converted into polymer
films containing metallic crystallites with nanometer dimensions.
(11) Metallized Plastics; Mittal, K. L., Ed.; VSP: Utrecht, The
Netherlands, 2001; Vol. 7 and earlier volumes.
(12) Horvath, O.; Stevenson, K. L. Charge-Transfer Photochemistry of
Coordination Compounds; VCH: New York, 1993.
(13) Marciniak, B.; Buono-Core, G. E. J. Photochem. Photobiol., A 1990,
52, 1.
Conclusions
(14) Huang, Z.-Y.; Mills, G.; Hajek, B. J. Phys. Chem. 1993, 87, 11542.
(15) (a) Bornstein, L.; Chernysov, D.; Valetsky, P.; Tkachenko, N.;
Lemmetyinen, H.; Hartmann, J.; Forster, S. Langmuir 1999, 15, 83. (b)
Kartuzhanskii, A. L.; Studzhinskii, O. P.; Plachenov, B. T.; Sokolova, I.
V. Russ. J. Appl. Chem. 1986, 59, 2265. (c) Studzhinskii, O. P.;
Kartuzhanskii, A. L.; Plachenov, B. T. Russ. J. Appl. Chem. 1985, 55, 597.
(16) Avey, A. A.; Hill, R. H. J. Am. Chem. Soc. 1996, 118, 237.
(17) Malone, K.; Weaver, S.; Taylor, D.; Cheng, H.; Sarathy, K. P.;
Mills, G. J. Phys. Chem. B 2002, 106, 7422.
(18) (a) Gaddy, G. A.; McLain, J. L.; Steigerwalt, E. S.; Broughton,
R.; Slaten, B. L.; Mills, G. J. Cluster Sci. 2001, 12, 457. (b) Gaddy, G. A.;
McLain, J. L.; Korchev, A. S.; Slaten, B. L.; Mills, G. J. Phys. Chem. B.
2004, 108, 14858. (c) Gaddy, G. A.; Korchev, A. S.; McLain, J. L.; Slaten,
B. L.; Steigerwalt, E. S.; Mills, G. J. Phys. Chem. B 2004, 108, 14850.
(19) (a) Kondo, Y.; Inouye, H.; Fujiwara, S.; Suzuki, T.; Mitsuyu, T. J.
Appl. Phys. 2000, 88, 1244. (b) Kondo, Y.; Miura, K.; Suzuki, T.; Inouye,
H.; Mitsuyu, T.; Hirao, K. J. Non-Cryst. Solids 1999, 253, 143.
(20) (a) Peyser, L. A.; Lee, T. H.; Dickson, R. M. J. Phys. Chem. B
2002, 106, 7725. (b) Zheng, J.; Dickson, R. M. J. Am. Chem. Soc. 2002,
124, 13982.
(21) (a) Monti, O. L. A.; Fourkas, J. T.; Nesbitt, D. J. J. Phys. Chem.
B 2004, 108, 1604. (b) Muniz-Miranda, M. J. Raman Spectrosc. 2002, 33,
295. (c) Wu, P. W.; Cheng, W.; Martini, I. B.; Dunn, B.; Schwartz, B. J.;
Yablonovitch, E. AdV. Mater. 2000, 12, 1438. (d) Crawford, K. D.; Hughes,
K. D. J. Phys. Chem. B 1998, 102, 2325.
(22) Gilbert, A.; Baggott, J. Essentials of Molecular Photochemistry;
CRC Press: Boca Raton, FL, 1991; Chapter 7.
(23) (a) Yonezawa, Y.; Sato, T.; Kuroda, S.; Kuge, K. J. Chem. Soc.,
Faraday Trans. 1991, 87, 1905. (b) Henglein, A. Chem. Mater. 1998, 10,
444.
The present study complements earlier investigations29 and
demonstrates the versatility of the SPEEK/PVA system for the
generation of small metal particles in different matrixes.
Formation of long-lived BPK radicals derived from SPEEK
occurs via illumination with 350 nm photons of solutions
containing PVA and the polyketone in the absence of air.
Electrostatic repulsions between the anionic SPEEK chains in
conjunction with the considerable viscosity of the polymer
solutions contribute to the long lifetime of the macromolecular
radicals. Their stability at room temperature enables the polymer
radicals to efficiently reduce AgI, CuII, and AuIII ions with the
formation of nanometer-sized metal crystallites. Quenching of
the triplet excited state of BP groups from SPEEK by the metal
ions limits the efficiency of the particle photogeneration
processes. However, crystallites form under optimum conditions
with efficiencies higher than the quantum yield of polymer
radicals measured in the absence of metal ions. This observation
implies that the metal ions compete with processes that restrict
the effectiveness of the radical production. The dissimilar
mechanistic features exhibited by the crystallite formation
processes of the three metals are exemplified by the generation
of kinetically metastable intermediates that are different for each
metal ion.
(24) Sato, T.; Kuroda, S.; Yonezawa, Y. J. Photochem. Photobiol., A
1999, 127, 83.
(25) (a) Kometani, N.; Doi, H.; Asami, K.; Yonezawa, Y. Phys. Chem.
Chem. Phys. 2002, 4, 5142. (b) Sato, T.; Maeda, N.; Ohkoshi, H.; Yonezawa,
Y. Bull Chem. Soc. Jpn. 1994, 67, 3165.
(26) (a) Kapoor, S.; Mukherjee, T. Chem. Phys. Lett. 2003, 370, 83. (b)
Kapoor, S.; Palit, D. K.; Mukherjee, T. Chem. Phys. Lett. 2002, 355, 383.
(c) Kapoor, S. Langmuir 1998, 14, 1021.
(27) Finter, J.; Lohse, F.; Zweifel, H. J. Photochem. 1985, 28, 175.
(28) Itakura, T.; Torigoe, K.; Esumi, K. Langmuir 1995, 11, 4129.
(29) (a) Korchev, A. S.; Bozak, M. J.; Slaten, B. L.; Mills, G. J. Am.
Chem. Soc. 2004, 126, 10. (b) Korchev, A. S.; Sartin, M.; Mills, G.; Slaten,
B. L.; Gale, W. F. Light Induced Metal Nanoparticle Formation (M ) Ag,
Au, Cu, Pd, Pt) in Aqueous Solutions and in Polymer Films Composed of
SPEEK-PAA. In Proceedings of the ISCANA 2003 Conference.
(30) (a) Bailly, C.; Williams, D. J.; Karasz, F. E.; MacKnight, W. J.
Polymer 1987, 28, 1009. (b) Jin, X.; Bishop, T.; Ellis, T. S.; Karasz, F. E.
Br. Polym. J. 1985, 17, 4. (c) Devaux, J.; Delimov, D.; Daoust, D.; Legras,
R.; Mercier, J. P.; Straziell, C.; Nield, E. Polymer 1985, 26, 1994.
(31) (a) Longenberger, L.; Mills, G. Nanotechnology: Molecularly
Designed Materials; Chow, G.-M., Gonsalves, K. E., Eds; ACS Symposium
Series 622, American Chemical Society: Washington, DC, 1996; p 128.
(b) Longenberger, L.; Mills, G. J. Phys. Chem. 1995, 99, 475.
(32) Heller, H. G.; Langan, J. R. J. Chem. Soc., Perkin Trans. 2 1981,
341.
Acknowledgment. The authors are grateful to R. Leibfried
(Victrex, Inc.) for the generous gift of PEEK. We also thank
M. E. Miller and T. Albrecht-Schmitt for helping us with TEM
and XRD measurements and the Department of Commerce for
partial funding of this project.
Supporting Information Available: XRD traces of pre-
cipitated metal colloids. This material is available free of charge
via the Internet at http://pubs.acs.org.
References and Notes
(1) Metal Nanoparticles Synthesis, Characterization and Applications;
Feldheim, D. L.; Foss, C. A., Jr. Eds.; Marcel Dekker: New York, 2002.
(2) (a) Kamat, P. V. J. Phys. Chem. B 2002, 106, 7729. (b) Link, S.;
El-Sayed M. A. J. Phys. Chem. B 1999, 103, 8410.
(3) Nanoscale Materials in Chemistry; Klabunde, K. J., Ed.; John
Wiley: NewYork, 2001.
(4) Kreibig, U.; Vollmer M. Optical Properties of Metal Clusters;
Springer-Verlag: Berlin, 1995.
(5) Belloni, J., Mostafavi, M. In Metal Clusters in Chemistry; Braun-
stein, P.; Oro, L. A.; Raithby, P. R., Eds.; Wiley-VCH: Weinheim, 1999;
Vol. 2, p 1213.
(6) (a) Xia, Y.; Gates, B.; Yin, Y.; Lu, Y. AdV. Mater. 2000, 12, 693.
(b) Hulteen, J. C.; Treichel, D. A.; Smith, M. T.; Duval, M. L.; Jensen, T.
(33) Allen, N. S.; Chen, W.; Catalina, P.; Green, P. N.; Green, A. J.
Photochem. Photobiol., A 1988, 44, 349.

Photogeneration of Small Ag and Au Crystallites
J. Phys. Chem. B, Vol. 109, No. 16, 2005 7745
(34) Ulanski, P.; Bothe, K.; Rosiak, J. M.; von Sonntag, C. Macromol.
Chem. Phys. 1994, 1443.
(46) (a) Kolosov, I. V.; Tokareva, G. I.; Panteleev, V. M. Russ. J. Inorg.
Chem. 1985, 30, 535. (b) Plyasunova, N. V.; Wang, M.; Zhang, Y.;
Muhammed, M. Hydrometallurgy 1997, 45, 37.
(47) Ciavatta, L.; Ferry, D.; Palombari, R. J. Inorg. Nucl. Chem. 1980,
42, 593.
(48) (a) Freiberg, M.; Meyerstein, D. J. Chem. Soc., Faraday Trans. 1
1980, 76, 1825. (b) Buxton, G. V.; Green, J. C. J. Chem. Soc., Faraday
Trans. 1 1978, 74, 697.
(49) Johnston, F. J. Radiat. Phys. Chem. 1989, 33, 113.
(50) Weaver, S.; Taylor, D.; Gale, W.; Mills, G. Langmuir 1996, 12,
4618.
(51) Kunkely, H.; Vogler, A. Inorg. Chem. 1992, 31, 4539.
(52) Lingane, J. J. J. Electroanal. Chem. 1962, 4, 332.
(53) Gachard, E.; Remita, H.; Khatouri, J.; Keita, B.; Nadjo, L.; Belloni,
J. New J. Chem. 1998, 1257.
(54) Gmelin Handbook of Inorganic and Organometallic Chemistry,
Gold, 8th ed.; Keim, R. Ed.; Springer-Verlag: Berlin, 1992, Supplement
Vol. B1, pp 180-258.
(55) Mu, R.; Ueda, A.; Wu, M. H.; Henderson, D. O.; Malone, K.; Mills,
G.; Meldrum A. Proc. Electrochem. Soc. 1999, 98-19 (Quantum Confine-
ment: Nanostructures), 439.
(56) Quinn, M.; Mills, G. J. Phys. Chem. 1994, 98, 9840.
(57) Chow, M. K.; Zukowski, C. F. J. Colloid Interface Sci. 1994, 165,
97.
(58) Kissner, R.; Welti, G.; Geier, G. J. Chem. Soc., Dalton Trans. 1997,
1773.
(35) Demeter, A.; Berces, T. J. Phys. Chem. 1991, 95, 1228.
(36) (a) Ledger, M. B.; Porter, G. J. Chem. Soc., Faraday Trans. 1.
1972, 68, 539. (b) Beckett, B.; Porter, G. Trans. Faraday Soc. 1963, 59,
2051. (c) Cha, H.-J.; Frank, C. W. Korea Polym. J. 1999, 7, 141.
(37) Schwarz, H. A.; Dobson, R. W. J. Phys. Chem. 1989, 93,
409.
(38) Ulanski, P.; Bothe, E.; Hildenbrand, K.; Rosiak, J. M.; von Sonntag,
C. J. Chem. Soc., Perkin Trans. 2 1996, 13.
(39) Mostafavi, M.; Dey, G. R.; Francois, L.; Belloni, J. J. Phys. Chem.
A 2002, 106, 10184.
(40) Hoshino, M.; Arai, S.; Imamura, M.; Ikehara, K.; Hama, Y. J. Phys.
Chem. 1980, 84, 2576.
(41) Canonica, S.; Hellrung, B.; Wirz, J. J. Phys. Chem. A 2000, 104,
1226.
(42) (a) Gachard, E.; Remita, H.; Khatouri, J.; Keita, B.; Nadjo, L.;
Belloni, J. New J. Chem. 1998, 1257. (b) Janata, E.; Henglein, A.; Ershov,
B. G. J. Phys. Chem. 1994, 98, 10888.
(43) Dautzenberg, H.; Jaeger, W.; Ko¨tz, J.; Philipp, B.; Seidel, Ch.;
Stscherbina, D. Polyelectrolytes: Formation, Characterization and Ap-
plication; Hanser/Gardner: Cincinnati, OH, 1994; Chapters 3 and 6.
(44) Khatouri, J.; Mostafavi, M.; Amblard, J.; Belloni, J. Chem. Phys.
Lett. 1992, 191, 351.
(45) Cheng, K. L. Anal. Chem. 1955, 27, 1165.