Fluorescence Quenching Kinetics of Anthracene Covalently Bound to Poly(Methacrylic Acid): Midchain Labeling

Article abstract of DOI:10.1021/jp984343j

Poly(methacrylic acid) labeled by anthracene in the middle of the chain (A-m-PMA) has been synthesized by anionic polymerization of tert-butyl methacrylate followed by introduction of the difunctional terminating agent 9,10-bis(bromomethyl)anthracene. Deprotection to yield the polyacid using trifluoroacetic acid in the presence of thioanisole results in complete removal of the tert-butyl moieties without destruction of the anthracene chromophore. Steady-state and time-resolved fluorescence quenching by Tl⁺ in aqueous solution at pH 11 at different ionic strengths has been measured. The data have been analyzed and compared to data previously obtained for 9-ethanolanthracene pendant-labeled poly(methacrylic acid) (9EA-PMA) using the following kinetic models: Stern - Volmer; hindered-access; combined Stern - Volmer - Perrin model developed by Morishima et al. (Macromolecules 1993, 26, 4293). These two polymers exhibit very similar quenching behavior, with the most significant difference being that static quenching is more important for A-m-PMA than 9EA-PMA. The Morishima model has been modified to include a preferential binding constant (K_b) for Tl⁺ in the vicinity of the chromophore. The estimated values of K_b are significant, on the order of 12-21, depending on the polymer and ionic strength.

Full text of DOI:10.1021/jp984343j

J. Phys. Chem. A 1999, 103, 2513-2523
2513
Fluorescence Quenching Kinetics of Anthracene Covalently Bound to Poly(Methacrylic
Acid): Midchain Labeling
John H. Clements and S. E. Webber*
Department of Chemistry and Biochemistry, The UniVersity of Texas at Austin, Austin, Texas 78712-1167
ReceiVed: NoVember 6, 1998
Poly(methacrylic acid) labeled by anthracene in the middle of the chain (A-m-PMA) has been synthesized by
anionic polymerization of tert-butyl methacrylate followed by introduction of the difunctional terminating
agent 9,10-bis(bromomethyl)anthracene. Deprotection to yield the polyacid using trifluoroacetic acid in the
presence of thioanisole results in complete removal of the tert-butyl moieties without destruction of the
+
anthracene chromophore. Steady-state and time-resolved fluorescence quenching by Tl in aqueous solution
at pH 11 at different ionic strengths has been measured. The data have been analyzed and compared to data
previously obtained for 9-ethanolanthracene pendant-labeled poly(methacrylic acid) (9EA-PMA) using the
following kinetic models: Stern-Volmer; hindered-access; combined Stern-Volmer-Perrin model developed
by Morishima et al. (Macromolecules 1993, 26, 4293). These two polymers exhibit very similar quenching
behavior, with the most significant difference being that static quenching is more important for A-m-PMA
than 9EA-PMA. The Morishima model has been modified to include a preferential binding constant (K
b
) for
+
Tl in the vicinity of the chromophore. The estimated values of K
b
are significant, on the order of 12-21,
depending on the polymer and ionic strength.
Introduction
at pH 11. This pH is chosen to completely deprotonate the
polymer so that we may concentrate on polyelectrolyte effects
rather than the complex behavior of PMA as a function of pH.7
Poly(meth)acrylates as precursors for polyelectrolytes are
attractive due to an extensive base of synthetic experience for
these types of polymers. However, we have found that the
subsequent hydrolysis or deprotection of the protected (meth)-
acrylates (e.g. methyl (meth)acrylate, tert-butyl (meth)acrylate)
results in destruction of the polymer-bound 9,10-dimethylan-
thracene. We find that interception of the carbocation intermedi-
ate by thioanisole allows a successful deprotection scheme with
little or no loss of this chromophore.
Polyelectrolyte-mediated chemical reactions have been the
subject of considerable research.1-3 An experimentally conve-
nient example is fluorescence quenching of a polyelectrolyte-
bound chromophore by ionic quenchers. It is well-known that
the electrostatic potential of the polyelectrolyte plays a vital
role in the enhancement or retardation of the quenching process.
It is reasonable to expect that the efficiency of quenching will
be strongly influenced by the precise location of fluorescence
probes on the polymer. Morishima and co-workers have
published numerous studies dealing with copolymers of sodium
acrylate and acrylamide with pendant attachment of such probes
as phenanthrene and 4-(4-hydroxyphenyl)ethenylpyridinium
4
Experimental Section
bromide. In earlier work from these laboratories Morrison et
al. have focused their attention on phenanthrene, anthracene,
and diphenylanthracene pendant-tagged to poly(methacrylic
acid).2
I. Synopsis of Synthetic Scheme. Poly(tert-butyl methacry-
late) was prepared by anionic polymerization in THF at -78
°C using cumylpotassium as the initiating species. The polymer
was terminated by reaction with 9,10-bis(bromomethyl)an-
thracene resulting in the joining of two polymer chains through
each dimethylanthracene junction. Low molecular weight mate-
rial was removed by dialysis in acetone and the tert-butyl group
removed by exposure to a trifluoroacetic acid/thioanisole mixture
in the absence of oxygen.
,3
The typical synthesis of these polyelectrolytes utilized free
radical copolymerization of the polyelectrolyte monomer with
a low mole fraction of a monomer with the probe attached as
a pendant group. The molar ratio of these components is chosen
such that, on average, only a few chromophore-containing units
are incorporated into each polymer chain, and the spatial
distribution of chromophores on the polyelectrolyte chain is
expected to be random. Anionic polymerization techniques have
proven invaluable for the development of well-defined polymer-
fluorophore systems. For instance, Hruska et al. found that this
method reliably incorporated a single 1-(2-anthryl)-1-phenyl-
ethylene unit at the junction of a block copolymer of styrene
and tert-butyl acrylate giving a well-defined pendant-tagged
II. Polymerization Reagents. a. SolVent. THF solvent was
stirred over calcium hydride for 24 h and then twice treated
with a sodium mirror to completely dry the solvent.
b. Monomer. tert-Butyl methacrylate monomer (Scientific
Polymer Products) was passed through inactive alumina to
remove inhibitor and predried by stirring over calcium hydride.
After several hours of stirring, the monomer was cryodistilled
and treated with diisobutylaluminum hydride followed by
triethylaluminum. The light green color indicated the absence
of moisture. The dry monomer was cryodistilled into a dry,
5
system. Polymers tagged in this fashion have been used to
elucidate the macromolecular environment in polymer micellar
6
structures. In this work, we report the anionic synthesis of poly-
8
(methacrylic acid) (PMA) labeled in the middle of the chain
sealed ampule and stored at 0 °C until needed.
with a 9,10-dimethylanthracene probe and a study of the
fluorescence quenching by Tl as a function of ionic strength
c. Initiator. Cumyl methyl ether was synthesized by reaction
of 72.9 g (80.2 mL) of R-methylstyrene (1 equiv) with 50.0
+
1
0.1021/jp984343j CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/24/1999

2514 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Clements and Webber
mL of methanol (HPLC grade, 2 equiv) in the presence of 562
mg (338 µL) of 70% HClO4. The mixture was allowed to stir
at 50 °C under nitrogen for 48 h according to the procedure
SCHEME 1
9
developed by Ziegler and Dislich. To prevent polymerization
during the course of the reaction, the p-tert-butylcatechol
inhibitor present in the R-methylstyrene (Aldrich) was not
removed. Chromatographic purification of the ether product was
performed in pure hexane eluent in order to remove starting
material and other impurities. The eluent was then gradually
changed to pure ethyl acetate in order to recover the ether from
the column. The collected solution was rotary evaporated to
obtain the pure cumyl methyl ether product, which is a liquid.
The liquid was stirred over calcium hydride to remove trace
moisture and filtered. H NMR was used to confirm the absence
of trace R-methylstyrene, indicated by olefinic doublets at δ
5
.08 and 5.35 (CDCl3). H NMR: δ 1.57 (s, -C(CH3)2), 3.10
(s, -OCH3), 7.25-7.48 (complex, 5 × Ar H).
Cumyl potassium was synthesized from 0.70 g of cumyl
methyl ether (1 equiv) under reduced pressure in the presence
of 0.40 g of potassium metal (mirror, 2.2 equiv) and 100 mL
9
of THF solvent. Upon exposure of the ether to the potassium
mirror a deep red color was observed, indicating the presence
of the cumyl anion. The mixture was allowed to stir for 24 h at
room temperature and in the absence of light. Note that this
method does not produce the fluorescent impurities reported
by Winnik and co-workers.10 The concentration of the deep red
initiating species was determined by titration with dry acetanilide
and usually found to be near 40 mM. The cumylpotassium
solution was transferred to an ampule and used as soon as
possible because it tends to form unidentified side products if
stored for more than 96 h even if light exposure is kept to a
minimum.
III. Preparation and Termination of Polymer. The polym-
erization reactor was assembled, evacuated, and flame-treated,
after which approximately 250 mL of THF was cryodistilled
into the vessel at -78 °C followed by pressurization 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 initiator solution. The persis-
tence 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. Monomer was added dropwise to the
vigorously stirring solution in order to suppress the premature
d. Terminating Agent: 9,10-Bis(bromomethyl)anthracene.
9,10-Bis(chloromethyl)anthracene (Aldrich) was used without
further purification. A total of 3.0 g (1 equiv) of 9,10-bis(chlo-
romethyl)anthracene solid was extracted 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
12
termination mechanism discussed by Warzelhan et al. in which
free monomer may undergo nucleophilic attack by the living
chain end, thereby terminating the chain and eliminating a
methoxy anion. After 1 h a small aliquot of the polymer solution
was isolated for subsequent study via a glass siphon and
terminated with methanol. A total of 0.40 equiv of 9,10-DBMA
in THF was then added to the vessel and allowed to react with
the living chain ends for 1.5-2 h at -78 °C (Scheme 1). The
mixture was then slowly warmed to room temperature for an
additional 1 h to ensure completion as described by Valeur and
11
h, following the procedure developed by Golden. Due to 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 and filtered. The yellow residue, which was mostly
monobrominated 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 pure, dry 9,10-bis(bromomethyl)-
anthracene (9,10-DBMA). Note that purification via chromatog-
raphy using alumina or silica stationary phases results in decom-
position of the product into a variety of species as detected by
13
co-workers. At this stage, any chains that might still be living
were quenched with methanol. The resulting anthracene mid-
tagged poly(tert-butyl methacrylate) (A-m-PTBMA) was puri-
fied and analyzed according to the procedures detailed below.
IV. Purification of Polymer. Dialysis was employed as an
effective method to remove any unreacted terminating fluoro-
phore as well as other impurities from the polymer. A
concentrated solution of A-m-PTBMA in acetone (∼10 mg/
mL) was prepared. This 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 polymer. The solution was
dialyzed against an outer acetone solution that was changed
every 3-4 h until the presence of impurities migrating into this
solution could no 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 and then conditioned to acetone environment by a
+
TLC. MS: m/z 367 (M ), 366, 365, 364, 363, 285, 283, 205.
H NMR: δ 5.50 (s, 2 × -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
11
Golden.
e. Preparation of Terminating Solution. 9,10-DBMA was
placed in a dry, clean round-bottom and exposed to vacuum.
High-purity THF was cryodistilled to dissolve the fluorophore
followed by cryodistillation to dryness in order to entrain off
any trace moisture left in the original solid sample. This process
was repeated once followed by a final dissolution in a minimal
amount of THF to achieve the pure terminating solution used
for the polymerization (∼3-4 mg/mL concentration). This
solution was transferred to a dry, sealed ampule via cannula.

Anthracene Covalently Bound to PMA
0-45 min soaking in pure water followed by similar soakings
in water/acetone mixtures of progressively higher acetone
content. This procedure is extremely important since sudden
immersion into organic solvents can warp membrane pores. The
A-m-PTBMA solution retained inside the membrane tubing was
evaporated, the polymer was dissolved in dioxane, and the
solution was filtered and freeze-dried.
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2515
3
detector and was equipped with an A/D board to simultaneously
record the analogue output from the Waters R401 differential
refractometer and the Perkin-Elmer LS-1 fluorescence detector.
THF was used as the mobile phase at a flow rate of 1.5 mL/
min. Polystyrene standards (Scientific Polymer Products) were
used for calibration.
For each chromatogram the optical density at 230 nm
(
polymer absorption) and 380 nm (anthracene absorption) was
V. Deprotection of Polymers. a. Untagged PTBMA. A 125
mg amount of polymer was dissolved in 75 mL of dioxane. To
this was added 250 µL of concentrated (37 wt %/wt) HCl (3.5
equiv/acid residue), and the resulting mixture was allowed to
stir at 85-90 °C for 8 h. Because the acid-catalyzed elimination
reaction produces volatile isobutylene and no other side
products, the mixture was filtered and freeze-dried without
further purification. Complete elimination of the tert-butyl group
was confirmed by H NMR (see next section).
obtained as a function of elution volume. The chromatograms
were saved as Borland C++ v2.0 based executable files for
calculation of molecular weight averages and polydispersities.14
GPC analysis of polyacid required methylation of -COOH
1
5
groups by diazomethane.
c. Fluorescence Measurements. Steady-state and time-
resolved measurements were performed on a SPEX DM3000
Fluorolog-τ2 spectrofluorometer equipped with a 450 W xenon
light source, Czerny-Turner double monochromators for excita-
tion and emission, and Hamamatsu R928-P emission and
reference (rhodamine-B) photomultipliers. A potassium dideu-
terium phosphate crystal was used to modulate the light source
for time-resolved phase modulation experiments. These time-
resolved measurements were referenced with an aqueous
glycogen scattering solution (τref ) 0). Lifetime data were
obtained using a modulation frequency range of 10-150 MHz.
Subsequent analysis was performed using SPEX Globals
Unlimited Version 2.1 software. Decays were fit to a sum of
b. A-m-PTBMA. A 250 mg amount of A-m-PTBMA was dis-
solved in 10 mL of methylene chloride. This solution was placed
in a clean, dry round bottom and degassed on a high vacuum
system. To this was cryodistilled 3 mL of thioanisole (15 equiv)
followed by 10 mL of trifluoroacetic acid (75 equiv) Each had
previously undergone three freeze-pump-thaw cycles in order
to remove the dissolved oxygen. The resulting ∼45% trifluo-
roacetic acid (TFA) mixture was allowed to react on the high
vacuum system in the absence of room light for 1 h at 50 °C
and at reduced pressure. The solvents were then cryodistilled
away. Dioxane (3 freeze-pump-thaw cycles) was cryodistilled
into the round-bottom flask to redissolve the polymer, followed
by cryodistillation to dryness in an effort to remove the tert-
butyl trifluoroacetate which is formed during the course of the
reaction. Note that this method is similar to the entrainment
procedure used to remove trace moisture from the 9,10-DMBA
terminating species.
2
exponentials using the Marquardt minimization method for ø .
VII. Quenching Experiments. a. Stock Solutions. A known
amount of A-m-PMA was placed into a clean, dry volumetric
flask. To this was added a predetermined amount of standardized
0
.1 M KOH (titrated against potassium hydrogenphthalate) to
give a final solution with pH 11.00, taking into account the
appropriate amount necessary for complete ionization of the
polyacid. After final dilution of the polymer solution with
deionized water to yield a methacrylic acid residue concentration
of approximately 1.5 mM, the pH was measured. The pH value
obtained was almost always higher than that calculated (probably
due to partial ionization of the original polymer powder) but
never exceeded 11.25. Due to the observation of a systematic
change in the fluorescence spectrum as a consequence of
prolonged room light exposure, the solution was protected from
light during storage, which was not allowed to exceed 1 week
due to measurable drops in the pH from day to day.
A second 3 mL aliquot of thioanisole followed by 10 mL of
TFA was cryodistilled into the reaction vessel, and this ∼75%
TFA mixture was allowed to react with the partially deprotected
polymer for an additional 1 h at 50 °C to complete the reaction.
Two successive entrainments with dioxane were performed
(similar to the procedure described above) to remove TFA, tert-
butyl trifluoroacetate, and thioanisole. The polymer residue was
then exposed to vacuum at room temperature to remove
remaining traces of thioanisole and freeze-dried in dioxane to
obtain pure anthracene mid-tagged poly(methacrylic acid) (A-
m-PMA). Note that removal of thioanisole is crucial prior to
freeze-drying because it depresses the freezing point. Even small
amounts of thioanisole will greatly reduce the effectiveness of
the freeze-drying process, which is essential in order to obtain
the final polymer as a dry powder. The extent of elimination of
the tert-butyl group was determined by comparison of the
integral values at δ 1.39 (s, -C(CH3)3) and 1.60-2.20 (complex,
The thallium(I) heavy-atom quencher was chosen for these
studies because of its similarity in size to the potassium
counterion. Solutions of TlNO3 in water were prepared with
3
concentrations ranging from 40 to 250 mM depending on the
quencher concentration region of interest. The dependence of
the quenching efficiency on ionic strength was investigated by
addition of small amounts of a KNO3 stock solution. The
concentrations of these solutions were chosen such that their
addition to give the highest quencher and salt concentrations
studied resulted in no more than a 10% increase in the polymer
solution volume. Note that we have defined the ionic strength
-
CH2-) in methanol-d and was found to be complete within
the limits of NMR detection (1%).
VI. Instrumentation and Analytic Techniques. a. Nuclear
Magnetic Resonance. All NMR measurements were performed
on either a Bruker AC-250 (250 MHz) or a Varian (300 MHz)
spectrometer.
(IS) of the solution in the following manner:
1
-
IS ) [KOH] + [KNO ] - / [COO ]
(1)
3
2
b. Gel Permeation Chromatography. Purity, extent of tagging,
and molecular weight distribution of the polymer products was
assessed on a gel permeation chromatography (GPC) system
consisting of a model 510 Waters pump connected to a series
-
where [COO ] refers to the concentration of methacrylic acid
residues and is included to take into account neutralization of
hydroxide ions during ionization of the polymer. All solutions
were purged with N2 bubbling for 15 min prior to usage.
b. Steady-State Measurements. 2.600 mL volume of the A-m-
PMA stock solution was placed into a quartz cuvette, equipped
with screw cap containing gas inlet/outlet, and the solution
5
4
3
of four Waters Millipore µStyagel columns (10 , 10 , 10 , and
00 Å, in order of flow) using UV-vis absorption, fluorescence,
5
and refractive index detectors, sequentially. A Hewlett-Packard
Vectra 486DX4-100 computer controlled a HP 1050 diode array

2516 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Clements and Webber
SCHEME 2
Figure 1. Comparison of the absorption at 230 nm (a) and 380 nm
(b) GPC trace for A-m-PTBMA with the 230 nm absorption GPC trace
for the untagged aliquot (c).
SCHEME 3
bubbled with nitrogen for 15 min. Excess salt was added at
this time to achieve the desired ionic strength and the solution
allowed to mix. The sample was excited at 383 nm and the
emission recorded from 395 to 520 nm with the signal
referenced to the excitation intensity (the S/R mode). Small
amounts of quencher stock solution (∼3-10 µL) were added
between each subsequent run and allowed to mix for 2-3 min
with N2 bubbling prior to the acquisition of the spectrum.
Dilution of the fluorophore by addition of quencher and salt
solutions was taken into account in later analysis.
Results
I. Tagging of Polymers. It is essential that the mid-tagged
product not contain any end-tagged material. This could be the
case if the chains displace only one bromine of 9,10-DBMA,
which results in an end-tagged polymer with half the molecular
weight of the desired product (Scheme 2). The presence of an
unwanted end-tagged species would necessitate removal by
GPC, resulting in the purification of only small quantities of
HCl/dioxane method described in the Experimental Section,
employed by Proch a´ zka et al., has been our standard method
for deprotection of untagged PTBMA.
Deprotection of A-m-PTBMA proved much more difficult.
Upon exposure of the mid-tagged material to the HCl/dioxane
mixture, a nearly complete (>99%) loss of anthracene fluores-
cence and absorption (>300 nm) was observed. Furthermore,
H NMR found that only 89% of all residues were deprotected.
In light of these findings, the HCl/dioxane method was
abandoned without further analysis.
1
9
1
6
material, as reported by Sasaki et al. for this type of polymer.
Therefore, we added a substoichiometric ratio of 9,10-DBMA
to the living polymer chains. When the chromatograms moni-
tored at 230 and 380 nm are compared, it is evident that the
anthracene is associated only with the doubled molecular weight
polymer (Figure 1a,b). In addition, the 230 nm trace of the
polymer aliquot taken before 9,10-DBMA termination (Figure
It is widely known that TFA in conjunction with mediators
such as anisole or thioanisole is effective for the removal of
tert-butyl and t-butoxycarbonyl (BOC) groups in the synthesis
1c) has only a small overlap with the 380 nm trace of the tagged
polymer further suggesting the absence of any end-tagged
material. Furthermore, close inspection of the 380 nm trace of
the tagged material shows no low molecular weight tail or
shoulder that would normally be associated with the presence
of a small fraction of tagged low molecular weight polymer.
Analysis of the GPC traces demonstrated that the tagged species
molecular weight (Mn) was 37 500 (target: 40 000), PD ) 1.09,
and there was an untagged species of approximately half the
molecular weight. Analysis of the polymer aliquot taken before
2
0
of polypeptides. The function of these mediators has been
illustrated by the reaction scheme proposed by Lundt and co-
2
1
workers. Thioanisole scavenges the tert-butyl or BOC car-
bocations formed during the elimination reaction thus preventing
these active groups from attacking vulnerable aromatic residues
such as tyrosine and tryptophan, which can ultimately lead to
22
unwanted butylation of these residues (Scheme 3). Deprotec-
tion of A-m-PTBMA in the presence of an excess of thioanisole
gave >99% elimination of tert-butyl groups and ∼100% intact
anthracene residues.
9
,10-DBMA termination yields Mn ) 15 300, PD ) 1.10.
II. Deprotection of Polymers. Deprotection of poly(meth)-
acrylates has been discussed by many researchers. Dissolution
of PTBMA in toluene followed by mild heating (∼80 °C) in
the presence of p-toluenesulfonic acid has been shown to yield
the deprotected product,17 while Kamachi and co-workers have
shown that prolonged exposure to H2SO4 at ambient temperature
III. Steady-State Quenching. a. Stern-Volmer Analysis of
Quenching. The steady-state fluorescence spectra of A-m-PMA
as a function of added Tl is shown in Figure 2a. A distinct
overall red-shifting of the spectrum as well as a partial loss of
vibrational structure with addition of quencher is observed. This
spectral shift can be clearly seen in the normalized spectra at
+
1
8
is effective in deprotecting a variety of (meth)acrylates. The

Anthracene Covalently Bound to PMA
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2517
Figure 3. (a) I
0
/I plot vs Tl+ concentration up to 12 mM for the
Figure 2. (a) (upper) Steady-state fluorescence spectra of A-m-PMA
at [Tl ] values as follows: (a) 0; (b) 0.131 mM; (c) 0.392 mM; (d)
+
following initial ionic strengths (top to bottom): 1.83 mM (squares);
6.36 mM (circles); 14.83 mM (triangles). (b) Like (a) except maximum
1
.604 mM. (b) (lower) Normalized steady-state fluorescence spectra
+
+
of A-m-PMA for [Tl ] ) 0 (a) and 11.2 mM (b) (I
respectively).
0
/I ) 1 and 14.97,
Tl concentration 90 mM with the following initial ionic strengths (top
to bottom): 1.19 mM (squares); 10.85 mM (circles); 18.28 mM
(
triangles). In both cases the smooth line is the best fit to the modified
two different Tl+ concentrations (Figure 2b). The intensity
Morishima model (eq 10). See Table 3 for the derived parameters.
values used in the SV quenching analysis were obtained from
the integral of the spectrum from 395 to 520 nm. An analogous
spectral shift was not observed by Morrison et al. for 9-etha-
nolanthracene pendant-tagged to poly(methacrylic acid) (9EA-
centration, i.e., the well-known Stern-Volmer (SV) plot. The
nonlinear SV expression for fluorescence quenching is given as
2
2
I /I ) 1 + K [Q] + A [Q]
(2)
PMA). According to the fit of our data to the hindered access
0
sv
1
model (see later), ca. 5% of the chromophores are “inaccessible”
+
to the Tl quencher. We speculate that if this inaccessibility is
where [Q] is the quencher concentration. In our use of eq 2 A1
real, and not just an artifact of the model, then it may arise as
a result of a particular polymer tacticity in the vicinity of the
chromophore. We will argue later that Tl binds preferentially
is an empirical coefficient representing the deviation from SV
2
3
linearity and Ksv is the SV constant. At low quencher
concentrations, the higher order SV term is negligible and the
SV constant can be calculated from the initial slope of the data.
The apparent second-order quenching rate constant, kq, can be
extracted from the defining relationship Ksv ) kq<τ0>, where
<τ0> is the fluorescence lifetime in the absence of quencher.
Figure 3a shows the dependence of the SV data on the ionic
+
in the vicinity of the anthracene chromophore. If so, the spectral
+
shift may be induced by the Tl ion for a subset of chro-
mophores that are not quenched completely.
The quenching data are expressed as the ratio of the fluores-
cence intensity in the absence of quencher (I0) to that in the
presence of quencher (I) plotted as a function of quencher con-
+
strength of the solution for [Tl ] up to 12 mM. From the initial

2518 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Clements and Webber
TABLE 1: Stern-Volmer Analysis of Initial Slope of I
0
/I
potential of the polyelectrolyte in concentrating the thallium
quencher ion around the probe.
10-3
sv (M-1
10-11
(M-1 s-1
polymer
IS (mM)
K
)
k
q
)
The Coulombic force between point charges is accounted for
a
a
a
b
14.2c
6.92
3.33
9.41d
A-m-PMA
A-m-PMA
A-m-PMA
1.83
6.63
14.83
1.91
12.2
7.40
3.56
11.3
2
5
by the multiplication of kDiff by the Debye term.
W
9
EA-PMA
W/(e - 1)
(4a)
(4b)
a
Data from Figure 3a for A-m-PMA. b Data from ref 2. c
τ
0
) 10.7
where
ns. d
τ
) 12 ns.
0
2
W ) z z e /(4πꢀ ꢀkT[r + r ])
slope of the quenching data in the absence of added salt
B C
0
B
C
-
1
(
background IS ) 1.83), Ksv was found to be 12 220 M . Using
In eq 4b zB and zC are the formal charges on species B and C,
the value <τ0> ) 10.7 ns (see later section), the quenching
12
-1 -1
e is the proton charge, ꢀ is the permittivity of free space, ꢀ is
rate constant was found to be 1.42 × 10
M
s . This value
0
the solvent dielectric constant, k is Boltzmann’s constant, and
compares reasonably well with that of 9EA-PMA at pH 11.00,
1
1
-1
T is temperature. For monovalent ions of opposite charge the
electrostatic correction is a factor of 1.478. The formal charge
on the anthracene reactant must equal 113 in order to yield the
quenching rate constant observed for A-m-PMA at low quencher
concentrations, which corresponds to approximately half the total
formal charge of the polyanion.
IS ) 1.91 mM, for which kq was found to be 9.41 × 10
M
-1
s . The values of Ksv and kq for different ionic strengths are
presented in Table 1.
As in the previous work, addition of salt to the solution
decreases the magnitude of quenching at all quencher concentra-
tions. This phenomenon can be interpreted in two ways. First,
the excess counterions reduce the effective electrostatic attraction
of the polyanion for the quencher ion through Debye shielding.
Second, the addition of excess counterions may displace
quencher ions within the vicinity of the polyelectrolyte. We have
At higher quencher concentration (Figure 3b) the I0/I data
-
1
roll over and approach a constant slope on the order of 48 M
for all IS values studied, giving an apparent quenching constant
9
-1 -1
on the order of 4.5 × 10 M
s
in good agreement with the
9
-1 -1
2
2
value of 2.92 × 10 M
s
for 9EA-PMA. These values are
referred to this as “statistical dilution” in our earlier paper.
very similar to the diffusion-controlled limit discussed above.
This is a point we will return to later in the text.
The results for very high quencher concentrations (up to 90
mM) are presented in Figure 3b. As in the earlier work, there
+
b. Inaccessible Fluorophores: Hindered-Access Model. The
negative deviation from linearity in the SV plot (Figure 3)
suggests that a fraction of probes are inaccessible to quencher
ions. In such cases a two-state model is often employed in which
the population of probes is divided into an accessible fraction,
fa, and an inaccessible fraction, fb, (fb ) 1 - fa). In the usual
application of this model it is assumed that as quencher is added
to the system the fluorescence intensity of the accessible probes
is very strong negative curvature for [Tl ] > 10 mM. For the
+
highest Tl concentration there is essentially no effect of added
9
-1
salt and the limiting slope corresponds to kq ) 4.4 × 10 M
-
1
s , very close to the expected value for a diffusion-limited
reaction (see below). In all plots the solid line is the best fit to
the modified Morishima model discussed later.
Previous work has shown that increasing the ionic strength
of the solution has no effect on quenching by the neutral species
3
decreases according to the linear SV relation (e.g. eq 2 with A1
acrylamide. Similarly we find for acrylamide quenching of
-1
) 0; we denote the SV constant for the accessible probes by
A-m-PMA Ksv values of 16.0, 15.4, and 14.7 M for IS values
of 1.65 (background), 9.28, and 16.6 mM, respectively (kq )
a
Ksv ) while the fluorescence intensity of the b probes remains
2
3
9
-1 -1
unchanged. This yields the following expression:
1
.50-1.37 × 10 M s ). Although the SV constant does
decrease slightly with ionic strength, the effect is small compared
a
+
I /(I - I) ) 1/f + 1/(K f [Q])
(5)
to that for the Tl quencher suggesting that the polyanion does
0
0
a
sv a
not undergo any significant local conformational changes as a
consequence of high IS that would provide protection for the
anthracene chromophore.
The diffusion-controlled reaction rate (kDiff) between neutral
species B and C, with radii rB and rC and diffusion rates DA
and DC, is given by the Smoluchowski expression:
where I0 ) I0a + I0b. Therefore, a plot of I0/(I0 - I) vs 1/[Q]
should yield a linear function in which fa can be obtained from
the intercept and Ksv from the ratio of the intercept and slope.
a
By fitting the data to this model (Figure 4), we obtain quenching
rate constants in fair agreement with the values obtained from
the initial slope of the SV plot. The fraction of accessible probes
is 0.93-0.95 for three different ionic strengths (Table 2). Once
again, this result is similar to 9EA-PMA for which fa was
kDiff ) 4πN (r + r )(D + D )
(3)
A
B
C
B
C
Here NA is Avogadro’s number. For Tl+ and anthracene
calculated to be 0.95. These findings are consistent with the
2
reactants we have previously estimated from space-filling
insensitivity of the neutral quencher acrylamide to ionic strength
(see earlier discussion). The persistence length of fully neutral-
2
+
models that rB + rC ) 8.52 Å. From the mobility of Tl it is
+
-5
2
24
+
found that DTl ) 2.00 × 10 cm /s, and if we assume that
ized PMA in aqueous solution (Na counterion) has been found
+
26
only Tl contributes to D because anthracene is bound to the
to be 420 Å through electric permittivity measurements. It is
polyelectrolyte, then the diffusion-controlled rate constant for
difficult to envision the complete encapsulation of a probe
species by flexible segments of the chain, thereby rendering
them inaccessible to quencher ions. Upon closer inspection of
the fit, we find that this model linearizes the data fairly well in
the case of excess salt but fails in the absence of excess salt.
Therefore, we believe this model is not really appropriate for
the present system.
c. Combined Stern-Volmer and Perrin (Morishima) Analysis.
For quenching by counterions that are condensed within the
vicinity of a polyelectrolyte, the concept of a sphere-of-action
1
0
quenching in water at 25 °C is calculated to be 1.29 × 10
-1
-1
M
s . For spherical species the diffusion rate is often
estimated from the Stokes-Einstein expression, D ) kT/6πηr,
where η is the solvent viscosity. Using this relation, kDiff for
9
-1 -1
water at 25 °C is 3.5 × 10 M
s
for one immobile species.
Therefore we expect the diffusion-limited rate to be in the range
10
-1 -1
(0.35-1.3) × 10
M
s . These values are on the order of 2
orders of magnitude lower than the observed rate constant,
demonstrating the overwhelming effect of the electrostatic

Anthracene Covalently Bound to PMA
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2519
TABLE 3: Analysis of Quenching Data Using the Modified
Morishima Model (eq 10)
polymer
IS (mM)
K
sv,app (M-1)
〈N〉
K
b
A-m-PMA (Figure 3a)
1.83
26.2a
9.38
2.50 12.6
2.52 15.0
2.42 20.8
2.48 15.5
a
6
4.8
.63
1
17.4a
17.6
3.49
3.42
4.39
3.06
A-m-PMA (Figure 5a)
A-m-PMA (Figure 3b)
5.37
1.94
1.6
2.72
7.75
1
2
2.71 13.9
2.66 17.4
0.9
9
EA-PMAb
1.91
2.76
7.72
a
K
sv,app values from Figure 3b are considered to be more reliable
because there are more data points in the concentration region where
diffusive quenching dominates. b Data from Figure 3 of ref 2. Note
that the values quoted in this reference are average ionic strengths,
including the addition of quencher.
In our application of the Morishima model we assume that
the number of ions that reside near the chromophore is fixed
(〈N〉) and when quencher ions are added the average number of
quenchers is given by 〈N〉xQ, where xQ is the mole fraction of
quencher ions. This assumption is in the spirit of Manning
Figure 4. Fit of the data of Figure 3a to the hindered access model.
The following identify the initial ionic strength: 1.83 mM (a); 6.36
mM (b); 14.83 mM (c).
2
9
condensation theory and also was the basis of a simplified
3
Poisson-Boltzmann model applied earlier. xQ is computed on
the basis of the total concentration of all cations, but if the
quencher ion is preferentially bound to the polyelectrolyte or
the region near the chromophore, the apparent value of the
quencher mole fraction may be higher that the value obtained
from using the bulk concentrations of all ions. We will encounter
this situation in our data fits. Thus we obtain our modification
of the Perrin expression in the absence of preferential binding:
TABLE 2: Parameters from Fitting Data of Figure 3a to
Hindered Access Modela
IS (mM) 10-3Ksv (M-1
)
f
IS (mM) 10-3
K
sv (M-1
)
f
a
a
1
6
.83
.63
13.0
7.61
0.945
0.929
14.83
4.30
0.930
a
Equation 5.
+
quenching (Perrin model) was invoked by Morishima et al.4b
For the Perrin model quenchers residing within a quenching
sphere of action immediately surrounding the probe have a
quenching probability equal to unity.23,27 It is assumed that the
probability of finding n quenchers within this spatial region can
I0/I ) (1 + K
[Tl ]atm) exp[〈N〉(x
Tl
+
)]
(8)
sv,app
Here
+
+
+
(xTl
+
) ) [Tl ] /([Tl ] + [K ] )
(9)
0
0
0
n
be described by a Poisson distribution, P(n) ) exp(-〈nQ〉)〈nQ〉 /
and the subscript 0 denotes bulk concentrations. Equation 8
displays negative curvature in the plot of I0/I vs [Tl ]. If there
is preferential binding of the Tl near the chromophore this
can be accounted for by replacing [K ]0 by [K ]0/Kb (see the
n!, where 〈nQ〉 is the average number of quenchers that reside
within the volume. Therefore, the probability that the probe will
not be quenched is equal to P(0) () exp(-〈nQ〉).28 In contrast,
quenchers that reside outside the condensation zone are assumed
to quench by a dynamic mechanism. The quenching reaction
involving these so-called “atmospheric” quenchers is assumed
to be well described by SV kinetics. In polyelectrolyte systems
both mechanisms are expected to be operative such that
quenching can be described by a combined Perrin and Stern-
Volmer mechanism. An expression incorporating this idea was
originally proposed by Morishima et al.4b
+
+
+
+
+
+
Appendix). Assuming that [Tl ]atm ∼ [Tl ]0, our fitting function
becomes
+
[
Tl ]
+
0
I0/I ) (1 + Ksv,app[Tl ] ) exp 〈N〉
0
{
+
+
}
]
[
[Tl ] + [K ] /K
b
0
0
(
10)
with the fitting parameters 〈N〉, Kb, and Ksv,app. Equation 10 failed
to fit the data satisfactorily unless Kb was allowed to vary. The
Kb values obtained from this fit have an average value of ca.
+
I0/I ) (1 + Ksv,app[Tl ] ) exp[〈n 〉]
(6)
atm
Q
1
6 but vary with the total ionic strength over the range 12-21
+
where Ksv,app is the “apparent” SV constant, [Tl ]atm is the
“
(Table 3). The values of 〈N〉 obtained from this fit are
+
atmospheric concentration” of Tl , and 〈nQ〉 is the average
remarkably constant. The Ksv,app values obtained from the fit to
the data in Figure 3b are considered more reliable because of
the larger number of data points in the saturation region. We
note that all the parameters obtained are in reasonable agreement
with those of 9EA-PMA (Table 3).
number of quencher ions in the active quenching sphere of the
chromophores. In the original model 〈nQ〉 is related to the ratio
of condensed counterions to fluorophores through an adjustable
parameter, R, which represents the fraction of condensed
counterions that are in proximity to the fluorescent probe:
+
d. Tl Displacement Experiment. Morishima et al. have used
a displacement method to assess competitive binding between
〈
n 〉 ) RC /[anth]
(7)
+
+
+
+ 4d
Q
b
0
Tl and other monovalent ions such as Li , Na , and K .
+
For the present system the Tl displacement experiment was
In this expression, [anth]0 is the bulk concentration of anthracene
probes and Cb represents the concentration of bound counterions
and is given by simple Manning condensation theory.
performed by first measuring the fluorescence intensity corre-
sponding to 3.75 mM Tl (IS ) 5.37, I0/I ) 11.5) and then
measuring this value as a function of added excess KNO3 up to
+

2520 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Clements and Webber
Figure 6. Steady-state (curve a) and time-resolved (curve b) quenching
of A-m-PMA. The circles and triangles in curve b correspond to a
double and triple exponential fit to the decay curve.
ionic strength, including the quencher ion. In the case of simple
SV quenching the slope (â) should be unity. Figure 5b shows
the data plotted in this fashion, which has a slope of 0.329.
The modified Morishima model with Kb ) 16.1 (the average
value of all determinations) also yields a plot that is essentially
linear and matches experiment very well.
III. Time-Resolved Quenching. The lifetime and steady-
state quenching of A-m-PMA are compared in Figure 6. Due
to signal-to-noise limitations, the lifetime measurements were
limited to relatively low quencher concentrations. The fluores-
cence decays were fit to the sum of exponentials:
I(t) ) R exp(-t/τ )
(12)
∑i
i
i
Here I(t) is the fluorescence intensity as a function of time
(
normalized to unity at t ) 0) and Ri and τi are the preexpo-
nential factor and lifetime of the ith component, respectively.
We define the average lifetime as
+
Figure 5. (a) I
0
/I vs added K+ for [Tl ] ) 3.704 mM, initial ionic
<
τ> ) Riτi
(13)
∑i
strength 5.37. The smooth line is the fit to the modified Morishima
model (eq 10). (b) Data for the displacement experiment plotted
according to eq 11. The linear fit yields â ) 0.329. Also shown is the
calculation from the modified Morishima model (see text).
<
τ> represents the integral of I(t) from 0 to ∞, which is directly
3
0
comparable to the steady-state intensity. The fluorescence
lifetime in the absence of quencher, <τ0>, was determined to
be 10.7 ns (essentially equal to the 12.0 ns for 9EA-PMA).
The lifetime quenching is characterized by the value of <τ0>/
1
4.88 mM (IS ) 15.65, I0/I ) 8) (see Figure 5a). If these data
are fit to eq 10, we obtain fitting parameters in good agreement
with those obtained from a normal quenching experiment (see
Table 3).
<
τ>. As can be seen from Figure 6 (curve b), these results are
unchanged using a two- or three-exponential fit. If there is an
element of “static quenching” (i.e. a quenching process faster
than we can measure), then we expect <τ0>/<τ> to lie below
In our earlier paper in which a displacement experiment was
analyzed, we started with the simple linear Stern-Volmer
expression and wrote [Tl ] ) cxTl , where c is the total
concentration of cations. From this, one can propose the
+
+
3
1
I0/I. We should note that the models we used to fit the data in
the previous sections do not directly calculate the dynamics of
quenching but rather the integral of I(t).
3
following expression as a measure of preferential binding:
+
The Ri, τi data are given in Table 4. We note that the un-
quenched decay is well-described by a single exponential (R1
) 0.991) and in fact in all cases the fluorescence decay is only
modestly nonexponential. The nonexponentiality is more pro-
(
I0/I - 1)
[K ]
0
excess
-
1 ) â
(11)
(I0/I - 1)K
+
IS0
+
Here the subscript 0 refers to the solution before the addition
of excess K . [K ]excess corresponds to the additional K added
during the displacement experiment, and IS0 is the initial total
nounced as quencher is added (R1 ) 0.838 at [Tl ] ) 0.392
+
+
+
mM). After addition of quencher and analysis of the fluorescence
+
lifetime, K was added to the system and the effect on the life-

Anthracene Covalently Bound to PMA
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2521
TABLE 4: Results from Fitting Fluorescence Decay to Two
Exponentials
SCHEME 4
+
Tl ] (mM) IS (mM)a
[
τ
1
(ns)
R
1
τ
2
(ns)
R
2
<τ>
0
0
0
0
0
0
0
0
0
.0
1.83
1.97
10.76 0.991
7.22 0.956
5.04 0.867
4.09 0.812
3.39 0.838
4.60 0.857
5.36 0.893
5.91 0.924
0.37
0.93
1.22
1.15
0.78
1.19
1.26
1.06
0.009 10.70
.066
.175
.284
.392
.388
.384
.379
0.044
0.133
0.188
0.162
0.143
0.107
0.076
6.94
4.53
3.54
2.97
4.11
4.92
5.54
2.21
2.45
2.68
6.69
10.61
14.44
a
In this table the value of IS given is the sum of the Tl+ plus
background and added salt.
time was monitored. It was found that the decay is slightly more
exponential (R1 ) 0.838 and 0.924 for IS ) 2.68 and 14.44
mM, respectively). This can be rationalized as the result of
+
+
excess K displacing Tl from the region of the chain, which
decreases the static-quenching component. This is qualitatively
similar to results obtained for other polymer systems for which
single photon counting methods were used to obtain the lifetime
data.
In contrast, our earlier work on 9EA-PMA shows almost only
dynamic quenching throughout the entire quenching range even
though both steady-state and time-resolved quenching curves
+
both strongly deviate from linearity at higher Tl concentra-
3
2
tions. The relative importance of static quenching for A-m-
PMA is the most significant difference we have found between
these two polymer systems. We presume that this a consequence
of the point of chromophore attachment and not an instrumental
artifact (e.g. the difference between single photon and phase
modulation time-dependent fluorescence measurements).
Discussion
two cases. In fact most aspects of the fluorescence quenching
are essentially identical for these two systems. According to
our estimate of the thickness of the condensation zone (16.3 Å;
see above), it is reasonable that both anthracene probes would
experience similar ion densities and therefore be quenched
equivalently. A similar conclusion was reached by Morishima
et al. on the basis of studies of a pH-dependent merocyanine
dye attached to poly(acrylic acid)-co-acrylamide.34 The only
significant difference we have found is that static quenching is
more important for the in-chain case, which suggests that the
density of quenching ions in the immediate vicinity of the
chromophore is somewhat higher.
To achieve a better understanding of the role of ionic strength
in the quenching process it is helpful to discuss Manning’s
3
3
counterion condensation theory. The effects of polyelectrolytes
on a number of chemical reactions have been explained quite
well within the basic framework of the model. According to
Manning, polyelectrolytes in solution can be characterized by
a charge density parameter ê, given by
2
ê ) ꢀe
(14)
kTb
where e, ꢀ, k, and T have been defined previously and b is the
average axial spacing between charged groups on the polyelec-
trolyte. The critical charge density parameter is defined as êc
+
If most quenching events originate from Tl ions that are
condensed on the polyelectrolyte, and if the concentration of
condensed ions is constant, then in the absence of preferential
binding all quenching processes should depend on the mole
-
1
)
N , where N is the charge of the counterion. When ê > êc,
counterions will condense on the polyelectrolyte until the
effective charge density parameter equals êc. For ê < êc no
condensation occurs. For fully ionized PMA ê ) 2.8 (b ) 2.55)
and it is predicted that condensation of monovalent ions proceeds
+
35
fraction of Tl with respect to all monovalent cations. This is
in contrast to classical homogeneous quenching processes in
which it is the molar concentration of quenchers that determines
the rate of quenching. In the present work the plots of I0/I vs
-
1
until (1 - ê ) ) 0.64 of the total polymer charge has been
effectively neutralized. Excess monovalent ions added to the
system may freely exchange with previously condensed species,
but the net population of condensed counterions remains the
same. The concentration of counterions remains constant from
the surface of the polyelectrolyte to a radius that is a function
of the linear change density and ê. We have previously estimated
+
xTl almost superimpose (Figure 7) as was the case for 9EA-
PMA (ref 2, Figure 8). There is a slight positive displacement
for higher ionic strengths which can be the result of the
+
following: (1) preferential binding of the Tl ion in the vicinity
of the probe; (2) a contribution to the quenching via a
homogeneous mechanism for which the quenching is propor-
+
+
3
,29
tional to [Tl ] and not xTl . We note the relation
the radius of the condensation zone to be 16.3 Å for PMA.
One of the motivations for carrying out the present synthetic
approach was to compare the fluorescence quenching of an in-
chain anthracene probe (A-m-PMA) with previous work from
this laboratory using a pendent chromophore (9EA-PMA). In
the context of a charged cylinder model (Scheme 4) one might
expect to see different ion distributions and/or transport for these
xTl
+
+
[
Tl ] ) IS
(15)
0
1 - xTl+
where IS0 is the initial ionic strength of the solution before the
addition of Tl . Therefore a positive displacement is expected
+

2522 J. Phys. Chem. A, Vol. 103, No. 15, 1999
Clements and Webber
At high Tl+ concentration the slope of I0/I is greatly
diminished. Note that the maximum value of I0/I produced by
〈
N〉
condensed ion quenching is e . The additional quenching is
by noncondensed ions that diffuse into the region of the
+
chromophore and is described by the 1 + KSV,app[Tl ] factor in
eq 10. The KSV,app values extracted from the Morishima model
3
are approximately 10 lower than the initial slope of the Stern-
Volmer plot and correspond approximately to a second-order
reaction rate between neutral species, one of which is immobile.
This is very much in keeping with the physical picture described
above.
Summary
A detailed procedure for the synthesis and tagging of
anthracene mid-tagged poly(methacrylic acid) has been de-
scribed that includes an effective deprotection route that does
not destroy the attached anthracene probe.
The steady-state and time-resolved fluorescence quenching
+
of A-m-PMA by Tl have been obtained and analyzed using
several different models. The data are very similar to that
observed for the anthracene pendant-tagged system 9EA-PMA
studied by Morrison et al. According to the two-state or
Figure 7. Plot of I
0
/I as a function of xTl+ for the following initial
ionic strengths: curve a, 1.83 mM (squares); curve b, 6.36 mM (circles);
curve c, 14.83 mM (triangles) (from the data of Figure 3a).
2
hindered-access model, the fluorophore populations for both
systems appear to be largely accessible by quenchers even at
high ionic strengths. However this model is not believed to be
appropriate because of the relatively poor fit at low ionic
strength.
if homogeneous quenching is important. However, for all fits
to the Morishima model the homogeneous quenching term (1
+
+
Ksv,app[Tl ]0) in eq 10 makes a relatively small contribution
to the overall quenching.
The primary effect of ionic strength on the quenching process
was found to be statistical dilution of Tl , as was observed for
In our previous work we measured the preferential binding
+
+
of Tl to homopolymer PMA and found binding constants
+
+
3
9EA-PMA. However, analysis of several aspects of the data
relative to K or Na on the order of 1.4. This is not necessarily
the same binding constant obtained by fits our modification of
the Morishima model (eq 10) which reflects preferential binding
in the immediate vicinity of the probe. This model fits our data
very well, but of course it has three adjustable parameters. The
physical picture is strongly influenced by the Manning model.
As quencher is added, the fixed number of condensed ions (〈N〉)
+
have suggested that there is preferential binding of Tl near
the chromophore. Our previous attempts to understand the
degree of preferential binding using a more complex model did
3
not lend themselves to fitting experimental data. In contrast,
modification of the model proposed by Morishima does allow
a fitting procedure. Using this model, preferential binding
+
constant values (K ) of ∼12-21 were found for A-m-PMA at
near the fluorophore are replaced by Tl . If Kb in eq 10 is greater
b
+
+
the different ionic strengths studied while a value of ∼7 was
than unity, then Tl replaces the K counterions more efficiently
than expected on the basis of mole fraction alone. Kb is found
to be higher than for bulk PMA (cf. 1.4 and 12-20) so there
found for 9EA-PMA. While the precise numerical value should
not be taken too seriously, we believe that the preponderance
+
+
of evidence is consistent with preferential binding of Tl near
must be a favorable interaction of Tl with the aromatic
the anthracene chromophore.
chromophore or the lower charge density near the chromophore
+
somehow favors Tl .
Time-resolved measurements demonstrated that as quencher
is added to the system, the magnitude of static quenching
increases. Static quenching is diminished by the addition of salt.
Similar observations were made for 9EA-PMA with the
important difference that static quenching is much less pro-
nounced than for A-m-PMA. This was found to be the most
significant difference between the two systems. For A-m-PMA
the nearly single exponential character of the fluorescence decay
persists even in the presence of quencher. This suggests to us
that within the condensation zone of the counterions the
concentration is relatively uniform and that the condensation
zone is significantly larger than the anthracene probe. Therefore
the lifetime quenching is similar to that obtained in homoge-
In our earlier work the physical picture proposed by Mor-
ishima et al. stimulated us to consider a simplified model for
the quenching of a fluorophore covalently bound to a polylelec-
3
trolyte, which we referred to as the “tube model”. For this
model, reaction-diffusion equations were solved for a constant
concentration of ions confined to the immediate vicinity of the
polyelectrolyte. Because one must carry out numerical integra-
tion of the diffusion-reaction equations for specific geometric
parameters, this model is hard to apply for data-fitting purposes.
Many of the features of the tube model are very similar to the
modified Morishima model (e.g. rollover in I0/I at high quencher
concentration, the effect of preferential binding on the ionic
strength dependencies). While a quantitative fit to experimental
data was not attempted in ref 3, the qualitative behavior of the
data suggested that preferential binding had to be evoked. In
the present experiments the value of Kb obtained by fitting
different experimental data sets is in reasonable agreement for
A-m-PMA and approximately a factor of 2 higher than for 9EA-
PMA (Table 3). We believe that the conclusion that there is
+
neous solution at a comparable Tl ion concentration.
The objective of this work has been to understand the
dependence of fluorescence quenching on the position of the
probe with respect to the backbone of the polyelectrolyte. From
this we are beginning to obtain a more sophisticated three-
dimensional view of the distribution of quencher and other
counterions about the polyelectrolyte. Future work will address
+
36
preferential binding of Tl near the chromophore is inescapable.
polyelectrolyte end-effects on quenching reactions.

Anthracene Covalently Bound to PMA
J. Phys. Chem. A, Vol. 103, No. 15, 1999 2523
Acknowledgment. This research has been supported by the
Office of Naval Research (Grant N00014-91-J-1667) and the
Robert A. Welch Foundation (Grant F-356). This support is
gratefully acknowledged. The authors also wish to thank Prof.
G. C. Willson for many helpful discussions regarding the
deprotection procedure including the critical suggestion that
thioanisole could be used to prevent destruction of the an-
thracene moiety.
Y.; Kamachi, M. J. Phys. Chem. 1991, 95, 6027. (f) Sassoon, R.; Hug, G.
J. Phys. Chem. 1993, 97, 7823.
(2) Morrison, M.; Dorfman, R.; Clendening, W.; Kiserow, D.; Rossky,
P.; Webber, S. E. J. Phys. Chem. 1994, 98, 5534.
(3) Morrison, M.; Dorfman, R.; Webber, S. E. J. Phys. Chem. 1996,
00, 15187.
1
(
4) (a) Morishima, Y.; Higuchi, Y.; Kamachi, M. J. Polym. Sci. 1991,
2
9, 677. (b) Morishima, Y.; Ohgi, H.; Kamachi, M. Macromolecules 1993,
26, 4293. (c) Morishima, Y.; Higuchi, Y.; Kamachi, M. J. Polym. Sci. 1993,
31, 373. (d) Morishima, Y.; Sato, T.; Kamachi, M. Macromolecules 1996,
2
2
9, 1633. (e) Morishima, Y.; Sato, T.; Kamachi, M. Macromolecules 1996,
9, 3960.
Appendix
(
5) Hruska, Z.; Vuillemin, B.; Riess, G.; Katz, A.; Winnik, M.
Makromol. Chem. 1992, 193, 1987.
6) (a) Webber, S. E.; J. Phys. Chem. B 1998, 102, 2618. (b) Proch a` zka,
+
+
We consider the Tl and K ions to compete for binding in
the region of space near the chromophore. The preferential
(
K.; Kiserow, D. J.; Webber, S. E. Acta Polym. 1995, 46, 277.
(7) (a) Katchalsky, A. J. Polym. Sci. 1951, 7, 393. (b) Arnold, R. J.
Colloid Sci. 1957, 12, 2689. (c) Leyte, J. C.; Mandel, M. Polym. Sci., Part
A-1 1964, 2, 1879. (d) Morawetz, H. Macromolecules 1996, 29, 2689.
(8) Allen, R.; Long, T.; McGrath, J. Polym. Bull. 1986, 15, 127.
+
binding of the Tl can be expressed by the equilibrium
+
+
+
+
Tl (aq) + K (b) ) Tl (b) + K (aq)
(A-1)
(A-2)
(9) Ziegler, K.; Dislich, H. Chem. Ber. 1957, 90, 1107.
+
+
[
Tl (b)][K (aq)]
(10) Hruska, Z.; Winnik, M. A.; Hurtrez, G.; Riess, G. Polym. Commun.
1990, 31, 402.
(11) Golden, J. H. J. Chem. Soc. London 1961, 3741.
)
Kb
+
+
[
Tl (aq)][K (b)]
(
12) Warzelhan, V.; Hocker, H.; Schulz, G. V. Makromol. Chem. 1978,
79, 2221.
13) Valeur, B.; Rempp, P.; Monnerie, L. C. R. Hebd. Seances Acad.
1
where b refers to an ion bound in the sphere of action and aq
refers to an ion in bulk solution. The region of space around
(
Sci. Paris, Ser. C 1974, 279, 1009.
+
+
the chromophore is small, so the fraction of all Tl or K ions
bound in the region is also small. Therefore we may write
(14) Programs written by Dr. T. J. Martin.
(15) Procedure referenced in Aldrich catalog under the trade name
Diazald.
(
16) Sasaki, T.; Yamamoto, M.; Nishijima, Y. Macromolecules 1988,
21, 610.
(17) (a) Long, T.; DePorter, C.; Patel, E.; Dwight, D.; Wilkes, G.;
+
+
[
Tl (b)]
[Tl (aq)]
0
)
Kb
(A-3)
+
+
[
K (b)]
[K (aq)]0
McGrath, J. Polym. Prepr. 1987, 28, 214. (b) Deporter, C.; Long, T.;
Venkateshwaran, L.; Wilkes, G.; McGrath, J. Polym. Prepr. 1988, 29, 343.
(
18) Kamachi, M.; Kurihara, M.; Stille, J. Macromolecules 1972, 5, 161.
+
where [Tl (aq)]0 refers to the bulk concentration. Note that Kb
is the preferential binding constant specifically for the region
near the chromophore and may differ numerically from pref-
erential binding (if any) on the polyelectrolyte chain itself. The
ratio of concentrations is equal to the ratio of mole fractions,
so
(19) Proch a´ zka, K.; Kiserow, D.; Ramireddy, C.; Tuzar, Z.; Munk, P.;
Webber, S. E. Macromolecules 1992, 25, 454.
(20) (a) Kemp, D.; Fotouhi, N.; Boyd, J.; Carey, R.; Ashton, C.; Hoare,
J. Int. J. Peptide Protein Res. 1988, 31, 359. (b) Pearson, D.; Blanchette,
M.; Baker, M.; Guindon, C. Tetrahedron Lett. 1989, 30, 2739.
(
21) Lundt, B.; Johansen, N.; Volund, A.; Markussen, J. Int. J. Peptide
Protein Res. 1978, 12, 258.
(
22) Sieber, P. Pept., Pro. Eur. Symp., 9th 1968, 236.
(
23) Lakowicz, J. Principles of Fluorescence Spectroscopy; Plenum
(
xTl
+
)b
(xTl
+
)b
(xTl )0
+
Press: New York, 1983.
(24) CRC Handbook of Chemistry and Physics, 66th ed.; Weast, R.,
Ed.; CRC Press: Boca Raton, FL, 1985; p D-167.
)
) Kb
(A-4)
(
xK
+
)b 1 - (xTl
+
)b
1 - (xTl )0
+
(25) Levine, I. Physical Chemistry, 3rd ed.; McGraw-Hill Book Co.:
Since
New York, 1988; p 559.
(
(
26) Van Der Touw, F.; Mandel, M. Biophys. Chem. 1974, 2, 231.
27) Birks, J. B. Photophysics of Aromatic Molecules; Wiley-Inter-
+
[
Tl ]
0
science: New York, 1970; p 441.
(
xTl )0 )
+
(A-5)
(28) This is clearly an oversimplifcation because ions in the sphere of
action are not necessarily 100% effective at quenching. However, the simple
Perrin model captures the main physical features of “contact quenching”.
+
+
[
Tl ] + [K ]
0 0
(
(
(
29) Manning, G. J. Chem. Phys. 1969, 51, 924.
30) Webber, S. E. Photochem. Photobiol. 1997, 65, 33.
31) See the discussion of static quenching in ref 30. Static quenching
we can substitute into eq A-4 and after some simplification
obtain
may be described operationally as quenching which occurs too quickly to
be observed by the time-dependent detection equipment. The phase-
modulation SPEX τ2 system is able to detect lifetimes as short as 0.1-0.2
ns, but fitting such a short component along with much longer-lived
components is problematic.
+
[
Tl ]
0
(xTl
+
)b )
(A-6)
+
+
[
Tl ] + [K ] /K
0 0 b
(
32) See Figure 3 of ref 2.
(33) Anderson, C. F.; Record, M. T. Annu. ReV. Biophys. Biophys. Chem.
References and Notes
1
990, 19, 423.
(
1) (a) Taha, I. A.; Morawetz, H. J. Polym. Sci, Part A-2 1971, 9, 1669.
(34) Morishima, Y.; Higuchi, Y.; Kamachi, M. J. Polym. Sci, Part A:
b) Sassoon, R.; Rabani, J. J. Phys. Chem. 1985, 89, 5500. (c) Delaire, J.;
Sanquer-Barrie, M. J. Phys. Chem. 1988, 92, 1252. (d) Kim, H.; Claude,
B.; Tondre, C. J. Phys. Chem. 1990, 94, 7711. (e) Morishima, Y.; Tominaga,
Polym. Chem. 1991, 29, 677.
+
(35) In ref 4e the displacement of Tl by divalent ions is discussed.
(36) John Clements, manuscript in progress.