pH-Induced Fluorescence Quenching of Anthracene-Labeled Poly(2-vinylpyridine)

Article abstract of DOI:10.1021/ma035749j

The fluorescence intensity of mid- or end-tagged anthracene-labeled poly(2-vinylpyridine) in acidic aqueous solution is very dependent on the degree of protonation (α). In the absence of protonation (e.g., in methanol solution) there is no appreciable intrapolymer fluorescence quenching. We presume that the quenching is the result of electron transfer from the excited anthracene to neighboring pyridinium units. Even a qualitative fit to a simple Bernoullian statistical model based on the state of nearest-neighbor protonation was not possible unless we propose that the protonation of pyridine units that are adjacent to the chromophore is significantly favored compared to pyridines located elsewhere on the chain. There is a change in the slope of the relative fluorescence vs α curve in the vicinity of 0.42-0.45 that suggests some kind of conformational transition at this degree of protonation.

Full text of DOI:10.1021/ma035749j

Macromolecules 2004, 37, 1531-1536
1531
pH-Induced Fluorescence Quenching of Anthracene-Labeled
Poly(2-vinylpyridine)
J oh n H. Clem en ts a n d S. E. Webber *
Department of Chemistry and Biochemistry, The University of Texas at Austin, Austin, Texas 78712
Received November 21, 2003
ABSTRACT: The fluorescence intensity of mid- or end-tagged anthracene-labeled poly(2-vinylpyridine)
in acidic aqueous solution is very dependent on the degree of protonation (R). In the absence of protonation
(e.g., in methanol solution) there is no appreciable intrapolymer fluorescence quenching. We presume
that the quenching is the result of electron transfer from the excited anthracene to neighboring pyridinium
units. Even a qualitative fit to a simple Bernoullian statistical model based on the state of nearest-
neighbor protonation was not possible unless we propose that the protonation of pyridine units that are
adjacent to the chromophore is significantly favored compared to pyridines located elsewhere on the chain.
There is a change in the slope of the relative fluorescence vs R curve in the vicinity of 0.42-0.45 that
suggests some kind of conformational transition at this degree of protonation.
Sch em e 1
In tr od u ction
There have been many studies of polyelectrolytes with
chemically attached chromophores,1 including our pre-
vious studies on poly(methacrylic acid) with mid- or end-
tagged anthracene moieties.2 In general, for weak
polyacids the fluorescence quantum yield and the access
of quenching species to the chromophore are a strong
function of the solution pH, reflecting the expansion or
contraction of the polyelectrolyte coil and the changes
in the linear charge density. We undertook the present
experiments with the same objective, but of course for
a polypyridine chain the coil expansion will increase as
the pH of the solution is lowered. It was found that the
fluorescence quantum yield changed dramatically as a
function of pH, becoming almost unmeasurable as the
pH approached 2. (The minimum pH required to dis-
solve poly(2-vinylpyridine is approximately 4.) The
fluorescence intensity of the two polymers, A-m-PVP
and BA-e-PVP (see Scheme 1), was studied as a function
of the extent of protonation, R. The fluorescence inten-
sity is qualitatively proportional to (1 - R)2 and (1 - R)
for A-m-PVP and BA-e-PVP, respectively, as would be
expected for a quenching mechanism that is related to
the probability that the adjacent pyridine moieties are
not protonated. However, to achieve even a semiquan-
titative fit to the data, it is necessary to propose that
the pyridines adjacent to the anthracene groups are
preferentially protonated compared to those located in
the main chain. Unfortunately, it is not possible to
obtain a very precise value of the preferential protona-
tion constant from these data fits, but we estimate an
equilibrium constant on the order of 10 for this proton
exchange.
of the more usual sodium mirror. This allows ample time for
purified monomer to be transferred to the receiving ampule
before significant polymerization has occurred. Because of the
high boiling point of 2-VP (20 °C at 1 Torr), a warm water
bath (∼45 °C) was necessary to cryodistill the monomer. Also,
it was often necessary to warm the manifold with a heat gun
to prevent condensation of the monomer in the line. Once in
the presence of sodium metal the monomer was allowed to stir
for ∼45-60 min under vacuum until the appearance of a
purple/gray color, the scavenging anion, indicated purity. At
this point, the monomer was cooled to -78 °C and opened to
vacuum to remove the small quantities of hydrogen gas formed
by the reaction of sodium and water. The monomer was then
warmed to ∼ 45 °C and cryodistilled into an ampule and stored
at 0 °C.
In itia tor . Cumyl potassium was used as the initiating
species although it is more difficult to prepare than reagents
such as n-butyllithium. Because of the much larger size of the
K+ counterion relative to Li+, the ion pair readily dissociates
to yield the highly reactive free cumyl anion, resulting in
extremely rapid initiation. This results in the production of
fairly monodisperse substances (polydispersity ∼ 1.05-1.20).3
Cumylpotassium also possesses a characteristic deep red color
that can easily be titrated to determine precise initiator
concentration.
9,10-Dibr om om eth yla n th r a cen e. 9,10-Dichloromethylan-
thracene (Aldrich) was used without further purification. 3.0
g (1 equiv) of 9,10-dichloromethylanthracene solid was ex-
tracted into a solution of 6.0 g (5.3 equiv) of sodium bromide
in 220 mL of acetone using a Soxhlet apparatus and allowed
to reflux under nitrogen for 18 h according to the procedure
developed by Golden.4 Because of the photoreactive nature of
the product, the material was protected from room light
exposure during the reaction and subsequent purification
process. The reaction mixture was cooled to room temperature
Exp er im en ta l Section
Mon om er s. 2-Vinylpyridine (2-VP) (Scientific Polymer
Products) were passed through inactive alumina to remove
inhibitor and predried by stirring over CaH2 for at least 24 h
and stored in the presence of CaH2 at 0 °C until needed. To
remove any traces of water, 2-VP was exposed to sodium metal.
To slow the inevitable polymerization of monomer by exposure
to the reactive metal, thin shavings of metal were used instead
10.1021/ma035749j CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/24/2004

1532 Clements and Webber
Macromolecules, Vol. 37, No. 4, 2004
and filtered. The yellow residue, which was mostly monobro-
minated material, was collected and extracted a second time
in the presence of sodium bromide using toluene solvent to
yield the desired dibrominated species. The final residue was
collected, recrystallized from toluene, and freeze-dried as an
emulsion in benzene to yield the pure, dry product. Note that
purification via chromatography using alumina or silica
stationary phases results in decomposition of the product into
a variety of species as detected by TLC. MS: m/z 367 (M+),
366, 365, 364, 363, 285, 283, 205. H NMR: δ 5.50 (s, 2 X
-CH2-), 7.63-7.70 (m, 2,3,6,7-H), 8.32-8.40 (m, 1,4,5,8-H);
mp >300 °C in agreement with Golden.4
9-Br om om eth yl-10-br om oa n th r a cen e. 9-Methylanthra-
cene (Aldrich) was reacted with 1.70 mL of bromine (6.2 equiv)
in the presence of triphenylphosphine in acetonitrile. After 1
h of reaction at room temperature the mixture was cold filtered
and allowed to air-dry in a fume hood to remove excess
bromine vapors. The residue was dissolved in dioxane and
rotary-evaporated to further remove excess bromine. This dark
yellow residue was twice recrystallized from chloroform (40
mL of chloroform per 750 mg of material) to obtain pure
product. The material was freeze-dried from dioxane to obtain
the product in a dry powder form. MS: (M+) m/z 191, 269, 270,
271, 272 (parent molecule signals 349, 350, 351 not present
in significant quantities). H NMR: δ 5.46 (s, -CH2-), 7.56-
7.70 (m, 2,3,6,7-H), 8.22-8.32 (d, 4,5-H), 8.56-8.64 (d, 1,8-H)
in agreement with Wang.5 MP: 178 °C (char onset), 193-195
°C (melt) (Wang: 200-202 °C).
P r ep a r a tion a n d Ter m in a tion of P olym er s. The polym-
erization reactor is described in detail elsewhere.6 THF was
cryodistilled into the vessel at -78 °C followed by pressuriza-
tion with high-purity nitrogen. The cold THF was allowed to
warm to near room temperature at which time existing
impurities in the solvent were eliminated by titration with the
initiator solution. The persistence of a light peach color was
indicative of a dilute solution of unreacted cumylpotassium
and signified the end point of the titration. The vessel was
again cooled to -78 °C, and the appropriate amount of initiator
solution was added and allowed to mix thoroughly. Since no
known termination pathway exists for the polymerization of
2-VP, this monomer was added quickly so as to ensure that
all chains initiate simultaneously. After 1 h of stirring at -78
°C a small aliquot of the polymer solution was isolated for
subsequent study via a siphon and terminated with methanol.
The desired anthracene terminating solution was then added
to the vessel and allowed to react with the living chain ends
for 1.5-2 h at -78 °C. The mixture was then slowly warmed
to room temperature for an additional hour as per the
procedure developed by Valeur and co-workers to form the
labeled polymer products.7
An a lysis a n d P u r ifica tion of La beled P olym er s. GPC
of all labeled polymers was performed before and after any
purification method used. The use of fluorescence and absorp-
tion detection chromatograms allowed us to evaluate just the
labeled population and were used to determine tagging ef-
ficiency (within experimental error, 98-100% efficient for both
polymers). Note that for the doubly terminated polymer A-m-
PVP a smaller than stoichiometric amount of anthracene
terminating agent was added, so there remains in the final
mixture some PVP homopolymer which makes no contribution
to the photophysical results. The GPC number-average mo-
lecular weight and PDI for the A-m-PVP polymer were 28 600
(target 30 000) and 1.26 (determined from the UV detection
signal), and for BA-e-PVP these quantities were 19 100 (target
20 000) and 1.23.
longer be detected by UV-vis absorption spectroscopy. This
procedure usually involved 8-10 outer solution changes. Note
that, prior to use, the membrane tubing was immersed in 5%
acetic acid solution for 10 min to remove trace metals. The
tubing was then conditioned to methanol environment by a
30-45 min soaking in pure water followed by similar soakings
in water/methanol mixtures of progressively higher methanol
content. The polymer solution retained inside the membrane
tubing was evaporated, dissolved in dioxane, filtered, and
freeze-dried to obtain the pure polymer in dry form.
In str u m en ta tion a n d Equ ip m en t. Steady-state and time-
resolved measurements were performed on a SPEX DM3000
Fluorolog-τ2 spectrofluorometer that has been described in our
earlier publications.2 The quantum yields of A-m-PVP and BA-
e-PVP were obtained in methanol and aqueous solution by
referencing to a 9-methylanthracene standard in cyclohexane
taking into account differences in solution refractive index (φref
) 0.29).8 Solutions were prepared with chromophore concen-
trations of ∼10-5 M. Self-absorption, which can result in
erroneously low quantum yields, is minimal at these concen-
trations. Both the labeled polymers and the standard were
excited at 365 nm.
Stock Solu tion s a n d F lu or escen ce Qu en ch in g by HCl
Ad d ition . A known amount of labeled polymer was placed into
a clean, dry volumetric flask. To this was added a predeter-
mined amount of standardized 0.1 M HCl to give a final
solution with pH ∼4 assuming that, at these concentrations,
∼25% of 2-vinylpyridine repeating units are protonated. The
polymer solution was then diluted with deionized water to
yield a 2-vinylpyridine repeating unit concentration within the
range 1.8-2.4 mM. The measured pH value was always within
the range 3.90-4.10, which corresponds to the minimum
amount of HCl required to dissolve the labeled polymers.
Solutions of A-m-PVP and BA-e-PVP in mixed methanol
(MeOH)/water solvents were prepared by dissolving a known
amount of the labeled polymer in MeOH followed by slow
addition of deionized water while stirring to give the final
MeOH/H2O volume ratio. The 2-vinylpyridine repeating unit
concentrations for these solutions were similar to that pre-
pared using dilute HCl.
Because of the observation of systematic changes in the
fluorescence spectrum of the labeled materials as a conse-
quence of prolonged room light exposure, the solutions were
protected from light during storage, which was not allowed to
exceed 1 week. The sensitivity of A-m-PVP fluorescence to
atmospheric oxygen was found to be very similar to that for
A-m-PMA.2 For this reason, solutions of A-m-PVP were
bubbled with nitrogen for at least 20 min immediately after
solution preparation, and each time the solution stock was
opened. Such precautions were not necessary for BA-e-PVP
since, like BA-e-PMA, this material appears to be much more
stable.
Stea d y-Sta te Mea su r em en ts. 2.600 mL of labeled PVP
stock solution was placed into a quartz cuvette, equipped with
a screw cap with a gas inlet/outlet, and the solution bubbled
was with nitrogen for 15 min. The samples were excited at
365 nm, and the emission was recorded from 390 to 500 nm
with the signal referenced to the excitation intensity (the S/R
mode). Small amounts of a 35 mM HCl stock solution (∼3-25
µL) were added between each subsequent measurement and
allowed to mix for 2-3 min with N2 bubbling prior to the
acquisition of each spectrum. The decrease of fluorescence by
dilution from the addition of quencher was taken into account
in later analysis. It has been reported that at high degrees of
polymer protonation excitation of pyridinium units at 260 nm
yields an emission centered at 390 nm.9 However, with 365
nm excitation the fluorescence observed in our experiments
is strictly from the anthracene label.
Dialysis was employed to remove traces of any unreacted
chromophores. (Obviously, these could considerably compro-
mise the photophysical studies.) A concentrated solution of the
polymer in methanol (∼10 mg/mL) was placed into a Spectra/
Por series 1 cellulose membrane tubing with a molecular
weight cutoff of 6000-8000, which was sufficient to allow low
molecular weight impurities to pass while retaining the high
polymer. The solution was dialyzed against an outer solution
of the same solvent that was changed every 3-4 h until the
presence of impurities migrating into this solution could no
Resu lts a n d Discu ssion
P r oton a tion Equ ilibr iu m of La beled P olym er s.
Measuring the pH of the labeled polymer solutions at
every addition of HCl allows us to calculate the degree

Macromolecules, Vol. 37, No. 4, 2004
Fluorescence Quenching of Poly(2-vinylpyridine) 1533
Ta ble 1. F lu or escen ce Qu a n tu m Yield for A-m -P VP a n d
BA-e-P VP in Differ en t Solven ts
polymer
solvent
MeOH
pH
φfl
A-m-PVP
0.63
80/20 MeOH:H2O
60/40 MeOH:H2O
40/60 MeOH:H2O
20/80 MeOH:H2O
5/95 MeOH:H2O
H2O (HCl)
0.59
0.44
0.40
0.42
0.28
4.08
1.96
0.054
<10-3
0.15
H2O (HCl)
MeOH
BA-e-PVP
H2O (HCl)
H2O (HCl)
4.27
1.94
0.078
<10-3
P h otop h ysics of La beled P olym er s in Mixed
MeOH/H2O Solu tion . The fluorescence quantum yield,
φ, for A-m-PVP and BA-e-PVP was found to be <10-3
at the pH required for complete protonation (<2),
whereas φ values of 0.63 and 0.15 were found for A-m-
PVP and BA-e-PVP in pure MeOH (Table 1). The
quantum yield of A-m-PVP in different MeOH/H2O
mixtures decreases steadily upon the addition of water.
We presume that the decrease in the quantum yield of
A-m-PVP with increasing water content is the result of
fluorescence quenching due to partial protonation of the
polymer by water (see the next subsection).
The solubility of BA-e-PVP is very different than A-
m-PVP. In particular, it is insoluble in MeOH/H2O
mixtures for MeOH volume fractions below 40%. Thus,
we did not determine the quantum yield values in the
limit of nearly pure H2O for this material.
F igu r e 1. Effect of R on pKd,app for A-m-PVP and BA-e-PVP
based on eq 1 of the text.
of protonation of the polymer, R, as follows:
[HCl]init - [H+]
[PH+]
[P] + [PH+]
R )
)
(1)
[P]init
where [P] and [PH+] represent the concentration of
unprotonated and protonated pyridine units, respec-
tively, and [P]init is the total amount of pyridine added
to the solution. [H+] is given by 10-pH, and [HCl]init is
the HCl concentration assuming no consumption of ions
by polymer protonation.
As is common with polyelectrolytes, the equilibrium
constant for protonation of the pyridine repeating unit
is a function of R and decreases significantly with
increasing R as a result of increasing electrostatic
repulsive forces between protonated pyridine units. The
apparent dissociation constant, Kd, for the pyridinium
ion can be expressed as
F lu or escen ce Qu en ch in g of La beled P olym er s
by P r oton a tion in Aqu eou s Solu tion . We propose
that electron transfer from excited-state anthracene to
neighboring pyridinium units is the mechanism for
fluorescence quenching:
A + hν f 1A*
1A* + PAH+ f A•+ + PAH•
A
•+ + PAH• f A + PAH+
(3)
R
1 - R
pKd,app ) pH + log
(2)
(
)
In eq 3, A represents the anthracene label and PAH+
represents a pyridinium unit adjacent to the anthracene
label. While it is conceivable that unprotonated pyridine
units might also participate in quenching, the fluores-
cence quantum yield value obtained for A-m-PVP in
MeOH (0.63, see previous section) is identical to that
reported by Cherkasov et al. for 9,10-dimethylan-
thracene in the same solvent.12 This suggests that in
the absence of acid pyridine units do not participate in
fluorescence quenching for pure MeOH solutions.
An oxidation potential of 0.665 V has been measured
by Zhang et al. for 9,10-dimethylanthracene in DMSO,
referenced to the Fc+/Fc (ferrocene) electrode.13 Al-
though no reduction potential data were found for the
pyridinium ion, Raghavan and Iwamoto have obtained
a value of -1.32 V for the 1-methylpyridinium ion in
acetonitrile, referenced to the standard calomel elec-
trode (SC).14 Conversion to the standard hydrogen
electrode (SHE) yields a ∆G value of -2.12 V for the
transfer of an electron from the anthracene label to a
neighboring pyridinium unit. For most anthracene
derivatives the 0-0 transition energy for S0-S1 absorp-
tion is within the range 3.10-3.35 eV. Although we do
not have the corresponding redox value for PVP, it
Values of pKd,app vs R are displayed in Figure 1 for both
A-m-PVP and BA-e-PVP in dilute HCl. pKd,app values
of 3.56 and 3.82 were observed for the A-m-PVP and
BA-e-PVP polymer solutions prepared at pH 4, respec-
tively (RA-m-PVP ) 0.229, RBA-e-PVP ) 0.243). Note that
these values are far removed from the pKd value of 6.12
for the model compound 2-ethylpyridine in water, which
is nearly constant over the full range of R. A similar
result was reported by Kirsh et al.10 The pKd,app curves
for the two materials are surprisingly different, with
pKd,app values for BA-e-PVP that are approximately 0.2
pKd units lower than those for A-m-PVP over the entire
R range investigated. One might think that this differ-
ence is the result of differing molecular weights for the
two materials (DPA-m-PVP ) 272; DPBA-e-PVP ) 182).
However, Kirsh and co-workers reported identical po-
tentiometic titration results for poly(2-vinylpyridine)
with molecular weights ranging from 3.0 × 104 (DP )
286) to 1.0 × 106 (DP ) 9524) in 45/55 EtOH/H2O.10 It
is possible that the pKd difference is an effect of the
chromophore and its location. This suggests that the
introduction of fluorescence probes can significantly
perturb the polymer from its unlabeled state.11

1534 Clements and Webber
Macromolecules, Vol. 37, No. 4, 2004
seems plausible that the proposed electron-transfer
quenching mechanism is thermodynamically allowed.
The fluorescence intensity of the 9-methylanthracene
(9-MA) solution described above was then monitored as
pyridine and HCl were added in succession. Addition
of pyridine up to 1.8 mM to the 9-MA solution did not
result in any fluorescence quenching. Addition of HCl
to give a maximum concentration of 2.4 mM resulted
in a 4.8% decrease in the fluorescence intensity of the
9-MA/pyridine solution. The quenching data can be fit
to the linear Stern-Volmer quenching expression, and
assuming a fluorescence lifetime of 5.2 ns for 9-MA (in
pure ethanol),8 the second-order rate constant for py-
ridinium ion quenching, kq, is found to be 5.8 × 109 M-1
s-1, e.g., approximately a diffusion-controlled reaction.
Chu and Thomas have reported that cetylpyridinium
chloride quenches the pyrene excited singlet state at the
diffusion-limited rate (ca. 1.5 × 1010 M-1 s-1).15
To test any physical model to describe the HCl
quenching behavior observed for the labeled polymers,
it would be convenient to determine the initial fluores-
cence intensity values in the absence of HCl (I0).
Unfortunately, a minimal amount of HCl is required to
prevent polymer precipitation in pure water. As men-
tioned earlier, the quenching data in Figure 2 are
referenced to the fluorescence intensity in the presence
of 0.586 or 0.600 mM HCl for A-m-PVP and BA-e-PVP,
respectively (Iinit). The HCl quenching data for A-m-PVP
and BA-e-PVP are plotted in Figure 2 as the ratio Iinit/I
() R(R)) vs R (Rmin ) 0.229 and 0.243 for A-m-PVP and
BA-e-PVP, respectively).
While the ratio R(R) for both labeled polymers in-
creases smoothly over the entire range of data, changes
of the slope occur in the range R ca. 0.42-0.45. This
may be the result of counterion condensation occurring
as R approaches the critical charge density. According
to Manning’s counterion condensation theory, the criti-
cal degree of protonation, Rc, at which chlorine counter-
ions will condense onto the polyelectrolyte is 0.357,
assuming the average axial spacing between pyridine
units to be 2.55 Å.16 The slope changes for both polymers
occur at values slightly higher than this. We speculate
that the jump in a near Rc corresponds to a conforma-
tional transition associated with achieving a critical
linear charge density along the PVP chain (perhaps
analogous to the so-called “hypercoiling” in PMA). The
only literature we find concerning the effect of charge
density on PVP is the report of Topp et al. on quater-
nized PVP in aqueous solution.17 They find that there
is a substantial change in the intrinsic viscosity for a
degree of quaternization between 0.09 and 0.25. Their
highest value is at approximately the lowest degree of
protonation we are able to study. There is no confor-
mational change suggested by their data for R ∼ 0.42.
Therefore, if there is a conformational transition that
is responsible for the change of the R(R) slope, it must
be localized in the immediate vicinity of the fluorescence
probes.
F igu r e 2. R(R) for A-m-PVP (upper) and BA-e-PVP (lower)
as a function of R and different K values applied to the
Bernoullian quenching model. For A-m-PVP the model curves
are the smooth curves from the lowest curve up, K ) 1, 8.9
(best least-squares fit), 20, 50, and infinity. For BA-e-PVP only
K ) 1 and K ) infinity are shown.
ability that a pyridine is protonated, then we may write
the following for the anthracene fluorescence intensity:
IA-m-PVP(x) ) φflA-m-PVP[(1 - x)2 + 2xf1 + f2x2]
(4)
The first term inside the brackets is the probability that
neither pyridine is protonated, in which case the
fluorescence of this population is unchanged from the
unquenched reference state (with a fluorescence quan-
tum yield of φflA-m-PVP). The next two terms represent
the probability of one or two protonated neighbors, and
f1 and f2 represents the relative quantum yield in this
case. Presuming that the pyridines quench efficiently
and independently, it is reasonable to imagine that f1
) 2f2. If the quenching is very fast (as expected from
our model compound quenching discussed previously),
then both f1 and f2 would be expected to be very small.
Because there is a minimum degree of pyridine proto-
nation that is required for polymer solubility, the data
presented in Figure 2 are for the ratio
A Ber n ou llia n Mod el of F lu or escen ce Qu en ch -
in g. From the discussion in the previous section we
propose that the mechanism of fluorescence quenching
is via electron transfer to a protonated pyridine moiety,
and because this is expected to be a short-range process,
we focus on the state of protonation of the nearest
neighbors of the anthracene chromophores.
First we consider A-m-PVP in which the anthracene
has two nearest pyridine neighbors. If x is the prob-
R(R)A-m-PVP ) IA-m-PVP(Rmin)/IA-m-PVP(R) (5)

Macromolecules, Vol. 37, No. 4, 2004
Fluorescence Quenching of Poly(2-vinylpyridine) 1535
If we simply substitute R for x in eq 4, then the shape
of the predicted R(R) curve is not a good match to the
experimental findings. In particular, the experiment
results resemble the prediction of eqs 4 and 5 only if
the neighboring pyridines are assumed to have a higher
degree of protonation than the average value (R).
If there is a preferential protonation of neighboring
pyridines, we can write
PH+ + PA ) P + PAH+
(6)
with an equilibrium constant K. This yields the follow-
ing relationship for x in terms of R:
RK
x )
(7)
1 + R(K - 1)
According to this approach, the fluorescence quenching
ratio, R(R), is a function of f1, f2, and K (and we expect
f1 and f2 to be very small). Substituting for x in eq 4,
one obtains the rather unwieldy expression
1/2
F igu r e 3. Plot of (R(R)A-m-PVP
R(R)BA-e-PVP (open symbols) vs R.
)
(solid symbols) and
and f2 can be taken to be zero within experimental
error.18 A nonlinear least-squares fit to the R(R)A-m-PVP
data for all R values using eq 8 (with f1 and f2
constrained to zero) yields KA-m-PVP ) 8.9 ( 1.1, but
fitting just the points between Rmin (0.229) and 0.42 (e.g.,
before the “break” in the R(R)A-m-PVP curve) yields very
large and nonphysical values for KA-m-PVP. The reason
for this is that if K is much larger than unity, eq 9
becomes
IA-m-PVP(R) )
(1 - R)2 + 2f1KR + (2f1K(K - 1) + f2K2)R2
A-m-PVP
φfl
(1 + R(K - 1))2)
(8)
If we ignore f1 and f2, then
2
(1 - Rmin
)
1 + R(K - 1)
1 - R
R(R)A-m-PVP
)
(9)
2
1 - Rmin
Rmin 1 - R
{
}
R
1 + Rmin(K - 1)
R(R)A-m-PVP
)
(13)
{
}
A similar approach predicts the following for BA-e-
PVP:
e.g., independent of K. As can be seen from Figure 2,
this limiting case fits the initial data reasonably well,
while using K ) 8.9 forces a fit through the experimen-
tal values for R > 0.42. Also illustrated in Figure 2 is
the predicted R(R)A-m-PVP curve for K ) 1, which clearly
does not fit the data at all. The other K values presented
(20 and 50) illustrate that obtaining a value of K from
fitting the experimental data will have very poor preci-
sion, but we can certainly state that for the Bernoullian
model to be applicable K must be equal to or larger than
10. For BA-e-PVP (Rmin ) 0.243) the large K limiting
form is a reasonably good fit to all the data (Figure 2),
but once again there is a small “break” in the curve
between R ) 0.45 and 0.55.
IBA-e-PVP(x) ) φflBA-e-PVP[(1 - x) + xf1] (10)
where the value of f1 is not necessarily the same as for
A-m-PVP because of different electronic factors, the
differing lifetime of the excited state, etc. In any case,
f1 is expected to be small. Substituting for x using eq
10 and ignoring f1, we obtain
1 - Rmin
1 + R(K - 1)
1 - R
R(R)BA-e-PVP
)
(11)
1 + Rmin(K - 1)
From eqs 9 and 11 we see that if the preferential
protonation constant K is equal for these two polymers
that
It is also possible to estimate the value of K from the
absolute value of the fluorescence intensity using eqs 8
and 10. If the fluorescence quantum yield in pure
BA-e-PVP
methanol is taken to be equal to φflA-m-PVP or φfl
,
1/2
R(R)BA-e-PVP ) (R(R)A-m-PVP
)
(12)
then we can write (ignoring the f1 and f2 terms)
Equation 12 suggests that a square-root relationship
should exist between R(R)BA-e-PVP and R(R)A-m-PVP. In
Figure 3 we present this comparison and this expecta-
tion is realized, especially for R values between Rmin and
ca. 0.4.
Fitting the full Bernoullian model to the experimental
data is not so straightforward because there are two or
three independent parameters that affect the calculated
R(R) curve in different ways. Roughly the values of f1
and f2 determine the dynamic range of the quenching
while K shifts the x values to values larger than R
(assuming K > 1). By far the preferential protonation
constant K has the largest influence on the fit of the
experimental data to the models, and we find that f1
n
R
minK
IMeOH/I(Rmin) ) 1 +
(14)
{
}
1 - Rmin
where n ) 1 and 2 for BA-e-PVP and A-m-PVP,
A-m-PVP
respectively. Using the values in Table 1 and Rmin
BA-e-PVP
) 0.229 and Rmin
) 0.243, we obtain the
estimate KA-m-PVP ) 8.0 and KBA-e-PVP ) 2.9. The
former is reasonably close to the value obtained from
the least-squares fit, but the latter is much smaller than
expected from Figure 2.
We are not aware of any literature precedent that
would explain why pyridines adjacent to anthracene
moieties would be preferentially protonated. The pyri-

1536 Clements and Webber
Macromolecules, Vol. 37, No. 4, 2004
Y. Polymer 1998, 39, 3035. (c) Tiera, M. J .; Neumann, M. G.
J . Braz. Chem. Soc. 1995, 6, 191.
(2) (a) Clements, J . H.; Webber, S. E. J . Phys. Chem. A 1999,
103, 2513. (b) Clements, J . H.; Webber, S. E. J . Phys. Chem.
B 1999, 103, 9366.
dinium ion could be stabilized by the relatively electron-
rich anthracene moiety provided that there is sufficient
mobility in the -CH-CH2- bridge to permit their
juxtaposition.
(3) Martin, T. J . Dissertation, The University of Texas at Austin,
Su m m a r y
1996; Chapter 2, p 26.
(4) Golden, J . H. J . Chem. Soc. London 1961, 3741.
(5) Wang, H.; Zhang, B.; Cao, Y. J . Photochem. Photobiol. A 1995,
92, 29.
The fluorescence intensity of mid- or end-tagged
anthracene-labeled PVP in acidic aqueous solution is
very dependent on the degree of protonation (R). This
result was unexpected, and no precedent could be found
in the literature. We presume that the quenching is the
result of electron transfer from the excited anthracene
to neighboring pyridinium units. (Unprotonated pyri-
dine does not quench the anthracene singlet state.) Even
a qualitative fit to a simple Bernoullian statistical model
based on the state of nearest-neighbor protonation was
not possible unless we propose that the protonation of
pyridine units that are adjacent to the chromophore is
significantly favored compared to pyridines located
elsewhere on the chain. There is a change in the slope
of the relative fluorescence vs R curve in the vicinity of
0.42-0.45 that suggests some kind of conformational
transition at this degree of protonation.
(6) Clements, J . H. Dissertation, The University of Texas at
Austin, 1999; Chapter 2.
(7) Valeur, B.; Rempp, P.; Monnerie, L. C. R. Hebd. Seances Scad.
Sci. Paris, Ser. C. 1974, 279, 1009.
(8) Birks, J . Photophysics of Aromatic Molecules; J ohn Wiley-
Interscience: New York, 1970; Chapter 4.
(9) (a) Pierola, I.; Turro, N.; Kuo, P.-L. Macromolecules 1985,
18, 508. (b) Vyprachticky, D.; Sung, K.; Okamoto, Y. J . Polym.
Sci., Part A 1999, 37, 1341.
(10) Kirsh, Yu. E.; Komarova, O. P.; Lukovkin, G. M. Eur. Polym.
J . 1973, 9, 1405.
(11) Morawetz, H. Macromolecules 1996, 29, 2689.
(12) Cherkasov, A.; Molchanov, V.; Vember, T.; Voldaikina, K. Sov.
Phys. Dokl. 1956, 1, 427.
(13) Zhang, X.; Bordwell, F.; Bares, J .; Cheng, J .; Petrie, B. J .
Org. Chem. 1993, 58, 3051.
(14) Raghavan, R.; Iwamoto, R. J . Electroanal. Chem. 1978, 92,
101.
(15) Chu, D.-Y.; Thomas, J . K. Macromolecules 1984, 17, 2142.
(16) Manning, G. J . Chem. Phys. 1969, 51, 924.
(17) Topp, A.; Belkoura, L.; Woermann, D. Macromolecules 1996,
29, 5392.
(18) Using the full eq 8 in a nonlinear least-squares fitting
procedure seems to lead to an instability in the fitting. The
improvement in the ø2 between ignoring the fi terms or
including them is very minor, but the minimization tends to
push K to very large values and fi to very small ones.
Ack n ow led gm en t. This research was supported by
a grant from the Robert A. Welch Foundation (F-356)
whose support is gratefully acknowledged.
Refer en ces a n d Notes
(1) (a) Anghel, D. F.; Toca-Herrera, J . L.; Winnik, F. M.; Rettig,
W.; von Klitzing, R. Langmuir 2002, 18, 5600. (b) Anghel,
D. F.; Alderson, V.; Winnik, F. M.; Mizusaki, M.; Morishima,
MA035749J