Rational design of d-PeT phenylethynylated-carbazole monoboronic acid fluorescent sensors for the selective detection of α-hydroxyl carboxylic acids and monosaccharides

Article abstract of DOI:10.1021/ja9060646

We have synthesized three new phenylethynylated carbazole boronic acid sensors, which were predicted to display novel d-PeT fluorescence transduction (PeT, photoinduced electron transfer; fluorophore as the electron donor of the electron transfer, ET) by DFT/TDDFT calculations. The d-PeT effect is characterized by a lower background fluorescence at acidic pH than at neutral pH, which is in stark contrast to the normal a-PeT effect (fluorophore as the electron acceptor of the ET) that shows a strong and undesired background fluorescence at acidic pH. Our experimental results confirmed the theoretical predictions and d-PeT was observed for two of the sensors (with p-dimethylaminophenylethynyl substitution at 6- position of the carbazole core). For the third sensor (with phenylethynyl substitution at 6- position of the carbazole core), however, not d-PeT but rather the normal a-PeT was observed. The discrepancy between the DFT/TDDFT calculations and the experimental observations can be rationalized using free energy changes (Rehm-Weller equations) and the rate constants for the ET (k_ET, Marcus equation). These new d-PeT boronic acid sensors show improved photophysical properties compared to the known d-PeT sensor reported previously by us. In particular, the fluorescence transduction efficiency of the new sensors was improved 8-fold when compared to the known d-PeT boronic acid sensors. Novel fluorescence enhancement/reduction was observed for one of the sensors upon binding with mandelic acid or tartaric acid at pH 5.6. The effect of pH as well as the bonding with analytes on the emission of the sensors were rationalized using DFT/TDDFT calculations. We believe that rational sensor design aided by DFT/TDDFT calculations as well as using free energy changes and electron transfer rate constants to study the emission properties of PeT sensors will become an essential tool in the design of new fluorophores or fluorescent sensors with predetermined photophysical properties.

Full text of DOI:10.1021/ja9060646

Published on Web 11/04/2009
Rational Design of d-PeT Phenylethynylated-Carbazole
Monoboronic Acid Fluorescent Sensors for the Selective
Detection of r-Hydroxyl Carboxylic Acids and
Monosaccharides
Xin Zhang,† Lina Chi,† Shaomin Ji,† Yubo Wu,† Peng Song,‡ Keli Han,‡
Huimin Guo,§ Tony D. James,| and Jianzhang Zhao*,†
State Key Laboratory of Fine Chemicals, School of Chemical Engineering, P.O. Box 40, 158
Zhongshan Road, Dalian UniVersity of Technology, Dalian 116012, P. R. China, State Key
Laboratory of Molecular Reaction Dynamics, Dalian Institute of Chemical Physics, Chinese
Academy of Sciences, 457 Zhongshan Road, Dalian 116023, P. R. China, Department of
Chemistry, School of Chemical Engineering, Dalian UniVersity of Technology, Dalian 116012,
P. R. China, and Department of Chemistry, UniVersity of Bath, Bath BA2 7AY, United Kingdom
Received July 20, 2009; E-mail: zhaojzh@dlut.edu.cn
Abstract: We have synthesized three new phenylethynylated carbazole boronic acid sensors, which were
predicted to display novel d-PeT fluorescence transduction (PeT, photoinduced electron transfer; fluorophore
as the electron donor of the electron transfer, ET) by DFT/TDDFT calculations. The d-PeT effect is
characterized by a lower background fluorescence at acidic pH than at neutral pH, which is in stark contrast
to the normal a-PeT effect (fluorophore as the electron acceptor of the ET) that shows a strong and undesired
background fluorescence at acidic pH. Our experimental results confirmed the theoretical predictions and
d-PeT was observed for two of the sensors (with p-dimethylaminophenylethynyl substitution at 6- position
of the carbazole core). For the third sensor (with phenylethynyl substitution at 6- position of the carbazole
core), however, not d-PeT but rather the normal a-PeT was observed. The discrepancy between the DFT/
TDDFT calculations and the experimental observations can be rationalized using free energy changes
(Rehm-Weller equations) and the rate constants for the ET (kET, Marcus equation). These new d-PeT
boronic acid sensors show improved photophysical properties compared to the known d-PeT sensor reported
previously by us. In particular, the fluorescence transduction efficiency of the new sensors was improved
8-fold when compared to the known d-PeT boronic acid sensors. Novel fluorescence enhancement/reduction
was observed for one of the sensors upon binding with mandelic acid or tartaric acid at pH 5.6. The effect
of pH as well as the bonding with analytes on the emission of the sensors were rationalized using DFT/
TDDFT calculations. We believe that rational sensor design aided by DFT/TDDFT calculations as well as
using free energy changes and electron transfer rate constants to study the emission properties of PeT
sensors will become an essential tool in the design of new fluorophores or fluorescent sensors with
predetermined photophysical properties.
1. Introduction
properties still represents a significant challenge. The three key
components for the successful design of fluorescent sensors are
anappropriatefluorophore,bindingunit,andsensingmechanism.3,10
Popular fluorescence transduction mechanisms are based on
manipulation of electron transfer (ET), for example, photo-
induced electron transfer (PeT),3,11 intramolecular charge trans-
fer (ICT), etc.5,6,12-17 For PeT sensors, a nitrogen atom usually
Fluorescent molecular sensors have attracted much attention
due to their versatile applications in environmental, biological,
and chemical science.1-9 However, the rational design of
fluorescent molecular sensors with predetermined photophysical
† State Key Laboratory of Fine Chemicals, Dalian University of
Technology.
‡ State Key Laboratory of Molecular Reaction Dynamics, Dalian Institute
of Chemical Physics.
(5) James, T. D. Top. Curr. Chem. 2007, 277, 107–152.
(6) Mohr, G. J. Anal. Bioanal. Chem. 2006, 386, 1201–1214.
(7) Gonc¸alves, M. S. T. Chem. ReV. 2009, 109, 190–212.
(8) Caltagirone, C.; Gale, P. A. Chem. Soc. ReV. 2009, 38, 520–563.
(9) Oshovsky, G. V.; Reinhoudt, D. N.; Verboom, W. Angew. Chem.,
Int. Ed. 2007, 46, 2366–2393.
§ Department of Chemistry, Dalian University of Technology.
| Department of Chemistry, University of Bath, UK.
(1) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.;
Kluwer Academic: New York, 1999.
(2) Valeur, B. Molecular Fluorescence: Principles and Applications;
Wiley-VCH Verlag GmbH: New York, 2001.
(10) Ji, S.; Yang, J.; Yang, Q.; Liu, S.; Chen, M.; Zhao, J. J. Org. Chem.
2009, 74, 4855–4865.
(3) De Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley,
A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. ReV.
1997, 97, 1515–1566.
(11) Kavarnos, G.; Turro, N. J. Chem. ReV. 1986, 86, 401–449.
(12) Stibor, I.; Zlatuskova, P. Top. Curr. Chem. 2005, 255, 31–63.
(13) Pu, L. Chem. ReV. 1998, 98, 2405–2494.
(4) Pu, L. Chem. ReV. 2004, 104, 1687–1716.
(14) Lin, J.; Pu, L. Angew. Chem., Int. Ed. 2005, 44, 1690–1693.
9
17452 J. AM. CHEM. SOC. 2009, 131, 17452–17463
10.1021/ja9060646 CCC: $40.75  2009 American Chemical Society

Phenylethynylated Carbazole Boronic Acid Sensors
A R T I C L E S
acts as the fluorescence switch, then chelation with analytes fixes
the otherwise transferable electrons. In this case, a specific pH
range is required to optimize the fluorescence transduction of
the PeT, since protonation of the N atom (at acidic pH) will
increase the background emission of the sensor and this will
substantiallyreducethesensitivityofthefluorescencesensing.3,5,18-22
Thus, neutral or basic pH are required for most of the a-PeT
sensors (fluorophore as the acceptor of ET) to function
efficiently.
Recently we have embarked on a study of boronic acid
sensors,17,18,23-29 which are unique because covalent bonds,
instead of hydrogen bonds, are formed in molecular recognition.
As a result, the recognition of analytes with boronic acid sensors
can be carried out in aqueous media.5,28,30-54 During our
investigations, we discovered that these a-PeT boronic acid
sensors do not work well in the acidic pH region.5,18,54 Some
analytes, however, such as R-hydroxylcarboxylic acids (man-
delic acid, etc), require recognition at acidic pH, where the
binding is much stronger than that at neutral or basic pH.5,17,18,24,54
Unfortunately, the known a-PeT boronic acid sensors can not
address this problem; therefore, a new sensing mechanism is
required, ideally producing a better signal at acidic pH.
Furthermore, many fluorescent molecular sensors are usually
prepared using a trial and error approach, rather than using a
rational design strategy. Although the trial and error approach
has generated some good systems, a rational design of fluores-
cent chemosensors with predetermined photophysical properties
is more desirable.10
Recently, we found that carbazole-based boronic acid sensor
4 shows d-PeT effect (fluorophore as the donor of the ET)
(Scheme 1); with this system, the fluorescence emission intensity
is decreased at acidic pH relative to neutral pH.27 Recognition
of tartaric acid with sensor 4 was observed at pH 4.0, where
the binding is much stronger than that at neutral and basic pH.
This kind of d-PeT sensor is rarely reported,22 and this efficient
fluorescence transduction at acidic pH upon binding was not
observed for the normal a-PeT sensors.18 However, the excita-
tion/emission wavelength of sensor 4 was short (332 nm/372
nm) and the Stokes shift was also small (40 nm), reducing the
potential applicability for measurement of biological samples.
Furthermore, the PeT fluorescence transduction efficiency of
the sensor 4 was low, for example, the emission intensity
increased by only 0.25-fold when the pH was switched from
acidic to basic.27 In stark contrast, the normal a-PeT sensors
show up to 10-fold fluorescence enhancement on switching the
pH from basic to acidic.18 Therefore, the performance of the
d-PeT boronic acid sensor 4 needed to be improved. More
importantly, the d-PeT mechanism required detailed investiga-
tion, to determine the relationship between molecular structure
and the d-PeT effect to aid future sensor design.
(15) Gunnlaughsson, T.; Glynn, M.; Tocci, G. M.; Kruger, P. E.; Pfeffer,
F. M. Coord. Chem. ReV. 2006, 250, 3094–3117.
(16) Mader, H. S.; Wolfbeis, O. S. Microchim. Acta. 2008, 162, 1–34.
(17) Zhao, J.; Fyles, T. M.; James, T. D. Angew. Chem., Int. Ed. 2004, 43,
3461–3464.
(18) Zhao, J.; Davidson, M. G.; Mahon, M. F.; Kociok-Kohn, G.; James,
T. D. J. Am. Chem. Soc. 2004, 126, 16179–16186.
(19) McCarroll, M. E.; Shi, Y.; Harris, S.; Puli, S.; Kimaru, I.; Xu, R.;
Wang, L.; Dyer, D. J. J. Phys. Chem. B. 2006, 110, 22991–22994.
(20) Sunahara, H.; Urano, Y.; Kojima, H.; Nagano, T. J. Am. Chem. Soc.
2007, 129, 5597–5604.
(21) Cody, J.; Mandal, S.; Yang, L.; Fahrni, C. J. J. Am. Chem. Soc. 2008,
130, 13023–13032.
(22) (a) Ueno, T.; Urano, Y.; Setsukinai, K.; Takakusa, H.; Kojima, H.;
Kikuchi, K.; Ohkubo, K.; Fukuzumi, S.; Nagano, T. J. Am. Chem.
Soc. 2004, 126, 14079–14085. (b) Matsumoto, T.; Urano, Y.; Shoda,
T.; Kojima, H.; Nagano, T. Org. Lett. 2007, 9, 3375–3377.
(23) Zhao, J.; James, T. D. J. Mater. Chem. 2005, 15, 2896–2901.
(24) Zhao, J.; James, T. D. Chem. Commun. 2005, 1889–1891.
(25) Liang, X.; James, T. D.; Zhao, J. Tetrahedron 2008, 64, 1309–1315.
(26) Chi, L.; Zhao, J.; James, T. D. J. Org. Chem. 2008, 73, 4684–4687.
(27) Han, F.; Chi, L.; Liang, X.; Ji, S.; Liu, S.; Zhou, F.; Wu, Y.; Han, K.;
Zhao, J.; James, T. D. J. Org. Chem. 2009, 74, 1333–1336.
(28) Nonaka, A.; Horie, S.; James, T. D.; Kubo, Y. Org. Biomol. Chem.
2008, 6, 3621–3625.
Herein we report our results for the d-PeT boronic acid
sensors 1-3 with phenylethynylated carbazole as the fluoro-
phore (Scheme 1). The D-π-A feature of the sensors was varied
to tune the PeT efficiency, through the introduction of electron-
donating or electron-withdrawing groups. DFT/TDDFT calcula-
tions were then carried out prior to the synthesis of these sensors
to verify the anticipated d-PeT effect. These sensors show red-
shifted emission as well as larger Stokes shifts compared to
the reported d-PeT sensor 4 (Scheme 1).27 More importantly,
the new d-PeT sensor demonstrated substantially improved
(29) Scrafton, D. K.; Taylor, J. E.; Mahon, M. F.; Fossey, J. S.; James,
T. D. J. Org. Chem. 2008, 73, 2871–2874.
(30) Fujita, N.; Shinkai, S.; James, T. D. Chem. Asian J. 2008, 3, 1076–
1091.
(31) Heinrichs, G.; Schellentraeger, M.; Kubik, S. Eur. J. Org. Chem. 2006,
18, 4177–4186.
(32) Levonis, S. M.; Kiefel, M. J.; Houston, T. A. Chem. Commun. 2009,
2278–2280.
(33) Roy, D.; Cambre, J. N.; Sumerlin, B. S. Chem. Commun. 2009, 2106–
2108.
(34) Zhang, L.; Kerszulis, J. A.; Clark, R. J.; Ye, T.; Zhu, L. Chem.
Commun. 2009, 2151–2153.
(35) Cao, H.; McGill, T.; Heagy, M. D. J. Org. Chem. 2004, 69, 2959–
2966.
(47) (a) Zhu, L.; Shabbir, S. H.; Gray, M.; Lynch, V. M.; Sorey, S.; Anslyn,
E. V. J. Am. Chem. Soc. 2006, 128, 1222–1232. (b) Wright, A. T.;
Anslyn, E. V. Chem. Soc. ReV. 2006, 35, 14–28.
(36) DiCesare, N.; Adhikari, D. P.; Heynekamp, J. J.; Heagy, M. D.;
Lakowicz, J. R. J. Fluoresc. 2002, 12, 147–154.
(37) Jin, S.; Li, M.; Zhu, C.; Tran, V.; Wang, B. ChemBioChem. 2008, 9,
1431–1438.
(48) Trupp, S.; Schweitzer, A.; Mohr, G. J. Org. Biomol. Chem. 2006, 4,
2965–2968.
(49) Kim, Y.; Hilderbrand, S. A.; Weissleder, R.; Tung, C. Chem. Commun.
2007, 2299–2301.
(50) Yang, X. P.; Lee, M. C.; Sartain, F.; Pan, X. H.; Lowe, C. R.
Chem.sEur. J. 2006, 12, 8491–8497.
(38) Wang, J.; Jin, S.; Akay, S.; Wang, B. Eur. J. Org. Chem. 2007, 2091–
2099.
(39) Powell, M. E.; Kelly, A. M.; Bull, S. D.; James, T. D. Tetrahedron
Lett. 2009, 50, 876–879.
(51) Dowlut, M.; Hall, D. G. J. Am. Chem. Soc. 2006, 128, 4226–4227.
(52) Jiang, S.; Escobedo, J. O.; Kim, K. K.; Alpturk, O.; Samoei, G. K.;
Fakayode, S. O.; Warner, I. M.; Rusin, O.; Strongin, R. M. J. Am.
Chem. Soc. 2006, 128, 12221–12228.
(53) Wang, Z.; Zhang, D.; Zhu, D. J. Org. Chem. 2005, 70, 5729–5732.
(54) James, T. D.; Samankumara Sandanayake, K. R. A.; Iguchi, R.;
Shinkai, S. J. Am. Chem. Soc. 1995, 117, 8982–8987. (b) Arimori,
S.; Bell, M. L.; Oh, C. S.; Frimata, K. A.; James, T. D. Chem.
Commun. 2001, 1836–1837. (c) Arimori, S.; Bell, M. L.; Oh, C. S.;
Frimat, K. A.; James, T. D. J. Chem. Soc., Perkin Trans. 1 2002,
803–808. (d) Katif, N.; Harries, R. A.; Kelly, A. M.; Fossey, J. S.;
James, T. D.; Marken, F. J. Solid State Electrochem. 2009, 13, 1475–
1482. (e) Bromba, C.; Carrie, P.; Chui, J. K. W.; Fyles, T. M.
Supramol. Chem. 2009, 21, 81–88.
(40) Halo, T. L.; Appelbaum, J.; Hobert, E. M.; Balkin, D. M.; Schepartz,
A. J. Am. Chem. Soc. 2009, 131, 438–439.
(41) Lim, S. H.; Musto, C. J.; Park, E.; Zhong, W.; Suslick, K. S. Org.
Lett. 2008, 10, 4405–4408.
(42) Cao, H. S.; Diaz, D. I.; DiCesare, N.; Lakowicz, J. R.; Heagy, M. D.
Org. Lett. 2002, 4, 1503–1505.
(43) Lavigne, J. J.; Anslyn, E. V. Angew. Chem., Int. Ed. 1999, 38, 3666–
3669.
(44) Zhu, L.; Zhong, Z.; Anslyn, E. V. J. Am. Chem. Soc. 2005, 127, 4260–
4269.
(45) Gao, X.; Zhang, Y.; Wang, B. Tetrahedron. 2005, 61, 9111–9117.
(46) Jin, S.; Zhu, C.; Li, M.; Wang, B. Bioorg. Med. Chem. Lett. 2009,
19, 1596–1599.
9
J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009 17453

A R T I C L E S
Zhang et al.
Scheme 1. Synthesis of the Fluorescent Boronic Acid Sensors 1-3a
a Reported d-PeT boronic acid sensor 4 is also shown. (i) KI, KIO3, CH3COOH, reflux, 10 min, 45%; (ii) NaH, DMF, n-C4H9Br, room temperature, 1 h,
72%; (iii) POCl3, DMF, CH2Cl2, reflux, 8 h, 63%; (iv) Pd(PPh3)4, CuI, NEt3, 4-ethynyl-N,N-dimethylaniline or phenylacetylene, argon atmosphere, 60 °C,
8 h, 83-90%; (v) ethanol, THF, 4-fluorobenzylamine or benzylamine, reflux, 6 h, then methanol, THF, NaBH4, room temperature, 15 min; (vi) acetonitrile,
K2CO3, 2-(2-bromomethylphenyl)-1,3,2-dioxaborinane, reflux, 10 h, 15-35%.
2. Results and Discussion
fluorescence transduction efficiency, compared to a previous
d-PeT boronic acid sensor reported by us.27 Rich fluorescence
transductions were found for sensor 1 toward recognition of
hydroxyl acids and monosaccharides. For example, fluorescence
enhancement/reduction was found for sensor 1 with tartaric acid/
mandelic acid at pH 5.6. To our knowledge, this is a new
fluorescence transduction profile and such chemoselectivity can
be beneficial in the design of new analytical methods.17
Interestingly, sensor 3 was found to be a normal a-PeT sensor,
not the d-PeT sensor predicted by DFT/TDDFT calculations.
We found that the discrepancy between the theoretical predic-
tions and experimental observations can be rationalized by
consideration of the free energy changes of the ET (∆G°,
by Rehm-Weller equation) and the rate constants of ET (kET,
by Marcus equation). The large fluorescence transductions of
these sensors and the effect of structural modification on the
PeT effect coupling with DFT/TDDFT calculation-aided sensor
design will be of great interest for the future design of
fluorescent sensors or fluorophores with predetermined photo-
physical properties.
2.1. Rational Design of d-PeT Fluorescent Sensors with
Predetermined Photophysical Properties. To investigate the
d-PeT fluorescence transduction mechanism in detail and to
improve the performance of the carbazole-based d-PeT boronic
acid sensor 4 reported by us,27 sensors 1-3 were designed.
Desired improvements include longer excitation/emission wave-
lengths, larger Stokes shifts, and more importantly improved
fluorescence transduction efficiencies, for example, enhanced
emission intensity on switching the pH from acidic to neutral
(Scheme 1).27 The π-conjugation framework of the sensors was
extended by attaching phenylethynyl groups to the parent
carbazole fluorophore. The electron-donating group (-NMe2) was
attached at the para position of alkynyl unit. 4-Fluorobenzyl-
amine was also used to strengthen the electron-accepting
capability of the boronic acid/amine moiety.21
We have shown that the fluorescence transduction of the
d-PeT effect can be rationalized with DFT/TDDFT calcula-
tions.27 Thus, the photophysical properties of the designed
sensors were investigated by DFT/TDDFT calculations prior
9
17454 J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009

Phenylethynylated Carbazole Boronic Acid Sensors
A R T I C L E S
the lack of overlap between the HOMO and the LUMO makes
the transition moment very small, thus S1rS0 is a forbidden
transition and the S1 state can not be directly populated by
photoexcitation but will be populated by internal conversion
(IC) from higher excited states, such as the S8 state.2,10,57,58 Thus,
S1 is not an emissive state; it is a dark state.10,27,58 This means
no radiative decay will occur; instead, nonradiative decay will
act as the drain pipe for the energy of the excited state.
Therefore, we propose that the protonated probe 1 will be
weakly fluorescent. A low-lying allowed excited state of S8
shows f ) 0.9759 (340 nm), which is in good agreement with
the UV-vis absorption at 360 nm (Table 2), thus validating
the theoretical calculations.
In contrast to the protonated 1, the S1 state of the neutral
sensor 1 is a locally excited state (LE, Figure 1b). Furthermore,
the oscillator strength (f) is 0.0877, about 20 times higher than
that of the protonated sensor 1. This means the neutral sensor
1 will probably fluorescence stronger than the protonated sensor
1. The calculated excitation energy of the S1 state is 355 nm,
which is in good agreement with the experimental results (360
nm, Table 2).
On the basis of these calculations, we conclude that the
protonated sensor 1 will give weaker emission than the neutral
sensor 1, that is, sensor 1 is a d-PeT sensor, in that the emission
intensity-pH profile is reversed when compared to the normal
PeT effect (a-PeT). We have demonstrated that the d-PeT effect
is complementary to the normal a-PeT effect,27 that is, the
R-hydroxyl acids can be recognized with higher fluorescence
transduction efficiency at acidic pH, where the binding is
stronger than that at neutral or basic pH.17,18,24
Figure 1. Prediction of the d-PeT effect for sensor 1 by DFT/TDDFT
calculations. (a) Main transitions of the singlet excited state of the protonated
sensor 1. S1rS0 (LUMOrHOMO. f ) 0.0045). S1 is a dark state (no
emission or weakly emissive), with ICT character. (b) Main transitions of
the singlet excited state of the neutral sensor 1, S1rS0 (LUMOrHOMO.
f ) 0.0877). S1rS0 of neutral sensor 1 is an allowed electronic transition
(potentially emissive), and S1 can be recognized as a LE state.
to the synthesis of these sensors. The optimized structures
of the sensors show that the phenylethynyl and the carbazole
moieties are in coplanar geometry, inferring that π-conjuga-
tion between the carbazole core and the peripheral phenyl-
alkynyl moiety will be efficient; thus, we expect a red-shifted
emission wavelength for sensor 1.55,56 The HOMO-LUMO
distributions of the S1 state of sensor 1 are presented in Figure
1. The main electronic transitions of sensor 1 (either the
neutral form or the monoprotonated form of the sensor) are
compiled in Table 1.
To interpret the theoretical calculations for the d-PeT effect
more intuitively, we calculated the fluorescence lifetimes of the
sensors, the Einstein transition probabilities of the spontaneous
transitions, which are described by eq 1.10,59
c3
2(E)2f
τ )
(1)
The evaluation of the electronic excited states and the
photophysical properties of the sensor are based on the selection
rule for electronic transitions.2,10,27,57 A transition can be
considered as forbidden if there is no overlap between the initial
and the destination MOs involved in the transition (e.g., S1 of
the protonated sensor 1, Figure 1a). Another parameter to
evaluate the possibility of a transition is the oscillator strength
(f). Usually transitions with f > 0.01 are allowed; conversely, a
small f < 0.01 infers a forbidden transition (f ) 0.01 corre-
sponding to extinction coefficient, ε, ca. 1000 cm-1 mol-1
dm3).58 The same rules are applicable to the emission
processes.2,10,57,58
where τ is the fluorescence lifetime, c stands for the light
velocity, E is the transition energy, and f is the oscillator strength
(Table 1). All of the parameters are in atomic unit (au). The
calculated fluorescence lifetimes for the protonated sensor 1 and
the neutral sensor 1 are 1002 and 21 ns, respectively. The
reasonable lifetime of the neutral sensor 1 (21 ns) infers that it
is probably fluorescent. For the protonated probe 1, however,
the exceptionally long lifetimes are far beyond the τ range for
relaxation of a singlet excited state (the emissive state for most
organic fluorophores),1,2 which leads to nonradiative decay of
the excited state. These results indicate that the protonated probe
1 will be nonfluorescent or weakly fluorescent.
The low-lying excited states of 2 and 3 were also examined
using similar considerations. d-PeT effects were predicted for
these sensors (see Figures S76-S79 and Tables S2, S3 in
Supporting Information). To validate our theoretical predictions,
we subsequently synthesized sensors 1-3. Experimental results
confirmed that sensors 1 and 2 are d-PeT sensors, as predicted
by the DFT/TDDFT calculations. However, experimentally,
sensor 3 is a-PeT sensor, and not the theoretically predicted
d-PeT sensor. The discrepancy between the theoretical calcula-
tions and the experimental observations for sensor 3 can be
ET from the arylalkynyl moiety to boronic acid moiety was
observed for the S1 state of the protonated sensor 1 (protonation
at the alkylamine, Figure 1a, see also Scheme 2), with f )
0.0045 (excitation energy of 548 nm). The small f value and
(55) Adhikari, R. M.; Neckers, D. C.; Shah, B. K. J. Org. Chem. 2009,
74, 3341–3349.
(56) Xing, J.; Chen, W.; Gu, J.; Dong, X.; Takeyasu, N.; Tanaka, T.; Duan,
X.; Kawata, S. J. Mater. Chem. 2007, 17, 1433–1438.
(57) Parson, W. W. Modern Optical Spectroscopy: With Examples from
Biophysics and Biochemistry; Springer-Verlag: Berlin Heidelberg,
2007.
(58) (a) Zhao, G.; Liu, J.; Zhou, L.; Han, K. J. Phys. Chem. B. 2007, 111,
8940–8945. (b) Saita, K.; Nakazono, M.; Zaitsu, K.; Nanbo, S.; Sekiya,
H. J. Phys. Chem. B. 2009, 113, 8213–8220.
(59) Liu, Y.; Feng, J.; Ren, A. J. Phys. Chem. A 2008, 112, 3157–3164.
9
J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009 17455

A R T I C L E S
Zhang et al.
Table 1. Selected Electronic Excitation Energies (eV) and Oscillator Strengths (f), Configurations of the Low-Lying Excited States of the
Neutral Sensor 1 and the Monoprotonated Sensor 1 ([1 + H]+)a
TDDFT//B3LYP/6-31G(d)
electronic transition
energy (eV)b
fc
compositiond
CIe
1
S0fS1
3.49 (355 nm)
0.0877
HfL
0.6381
0.2114
0.6202
0.7052
0.7052
0.7027
0.6627
0.2165
0.6869
0.6390
0.5841
0.3477
HfL+1
HfL+1
HfL+2
HfL+3
H fL
Hf L+1
Hf L+2
Hf L+5
Hf L+6
H-3fL
H-1fL+3
S0fS2
S0fS3
S0fS4
S0fS1
S0fS2
3.68 (336 nm)
3.94 (314 nm)
4.14 (299 nm)
2.26 (548 nm)
2.47 (502 nm)
1.4546
0.0025
0.0087
0.0045
0.0401
[1+H]+
S0fS6
S0fS8
S0fS12
3.10 (400 nm)
3.64 (340 nm)
4.08 (303 nm)
0.3586
0.9759
0.0563
a Calculated by TDDFT//B3LYP/6-31G(d), based on the optimized ground state geometries. b Only selected excited states were considered. The
numbers in parentheses are the excitation energy in wavelength. c Oscillator strength. d H stands for HOMO and L stands for LUMO. Only the main
configurations are presented (CI coefficients >0.2). e CI coefficients are in absolute values.
Scheme 2. Simplified Fluorescence Transduction Mechanisms of Sensor 1a
a (a) Free sensor 1 in media with pH > 6.0, is highly fluorescent. (b) Monoprotonated sensor 1 (weakly fluorescent, in media with pH 4.5-6.0). (c)
Bis-protonated species (in media with pH < 4.5, highly fluorescent). (d) Analyte bonded species (a-PeT is terminated). These assumptions are based on the
experimental results (vide infra) and were supported by DFT/TDDFT calculations.
9
17456 J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009

Phenylethynylated Carbazole Boronic Acid Sensors
A R T I C L E S
Table 2. Photophysical Parameters of Sensors 1-4
a
c
d
e
ε (M-1 cm-1
)
λabs (nm)
λem (nm)
Stokes shift/(nm)
Φb(pH 4.0)
Φ
b(pH7.5)
τ (ns)
kr (× 107 s-1
)
knr (× 107 s-1
)
sensor 1
sensor 2
sensor 3
sensor 4
2.83 × 104
4.88 × 104
2.96 × 104
4.69 × 104
360
370
338
332
430
450
393
372
70
80
55
40
0.04
0.07
0.44
0.24
0.06
0.12
0.21
0.31
5.23
4.77
6.80
7.02
1.22
2.57
3.07
4.43
17.9
18.3
11.6
9.8
a In 5.0 × 10-2 mol dm-3 NaCl ionic buffer (52.1% methanol in water), pH 7.5. b Fluorescence quantum yields, with quinine sulfate as the standard
c
(Φ ) 0.54 in 0.5 M H2SO4). Fluorescence lifetimes, with typical error of 0.01 ns. Concentrations of the sensors are 3.0 × 10-5 mol dm-3
.
d Radiative
decay rate constants at pH 7.5, kr ) Φ/τ. e Nonradiative decay rate constants at pH 7.5, knr ) (1-Φ)/τ.
group for sensors 1 and 2.56,62,63 This observation is supported
by the theoretical calculations. The emission of 1 is centered at
430 nm, which is red-shifted by 58 nm compared to the emission
of the known d-PeT boronic acid sensor.27 Another key
photophysical parameter, the Stokes shift, increased to 70 from
40 nm when compared with the known d-PeT sensor.27 A large
Stokes shift is beneficial for potential fluorescent molecular
sensing, especially fluorescent bioimaging.10,66-68 The excita-
tion/emission wavelengths of sensor 2 are red-shifted compared
to those of sensor 1. The structure difference of 1 and 2 is F
substitution in sensor 1 but not in sensor 2. The electron-
withdrawing F atom is not directly attached to fluorophore so
we did not expect it to disturb the excitation/emission wave-
length profoundly. We attribute this spectral change to the
different polarity of the microenvironment in which the fluo-
rophore resides,69 for example, the perturbation of the π-con-
jugation system by the F atom. We noticed that the pKa of sensor
2 is 5.11 ( 0.05 (see Supporting Information), which is different
from the pKa of sensor 1 (4.79 ( 0.14).
Sensor 3 lacks an electron-donating group (-NMe2). The
excitation/emission (338 nm/393 nm) of sensor 3 shows blue-
shifts compared to that of sensors 1 and 2. The Stokes shift of
sensor 3 (55 nm) is also decreased compared to that of sensor
1 (70 nm) and sensor 2 (80 nm). Furthermore, structured
excitation/emission bands were observed, which is in contrast
to the structureless excitation/emission bands of sensors 1 and
2. These results indicate that the photophysical properties of
ethynylated carbazole can be tuned by electron-donating or
electron-withdrawing substitutions, even at the 3,6-position of
the carbazole moiety.55
To prove the predicted d-PeT effect of sensor 1, the pH
dependency of the emission of sensor 1 was studied (Figure
3). Three pH ranges with dramatic different emission
intensity-pH response can be identified. From pH 11.23 to
pH 6.37, the emission hardly shows any changes (see Figure
S26 in Supporting Information). When the pH was decreased
from 6.37 to 4.23, however, the emission intensity decreased
(Figure 3a); this is in stark contrast to the normal a-PeT
sensors.3-5,18,24,25,70 The pKa for the d-PeT effect was
determined as 4.79 ( 0.14 (Figure 4). This abnormal pH-
emission intensity profile clearly demonstrates the d-PeT
effect and is in full agreement with the theoretical calculations
(Table 1). The fluorescence intensity transduction of the new
sensor is greatly improved compared to the previous d-PeT
Figure 2. Normalized excitation and emission spectrum of sensors 1-3;
3.21 × 10-6 mol dm-3 of sensors in 5.0 × 10-2 mol dm-3 NaCl ionic
buffer (52.1% methanol in water), pH 7.4. For sensor 1: λex ) 360 nm, λem
) 430 nm; for sensor 2: λex ) 370 nm, λem ) 450 nm; for sensor 3: λex
340 nm, λem ) 393 nm; 18 °C.
)
rationalized by consideration of the free energy changes and
the rate constants of the ET (vide infra).
2.2. Synthesis of the Alkynyl Carbazole Fluorescent
Boronic Acid Sensors. The design rationale of sensors 1-3 lies
in the notion that carbazole acts as the fluorophore;60,61 the
phenylethynyl groups were introduced to increase the π-con-
jugation and thus the excitation/emission wavelength. Electron-
donating (Me2N-) and electron-withdrawing units (F atom) were
used to tune the d-PeT efficiencies.56,62,63 First the carbazole
moiety was monoiododied,64 then 5 was butylated in the
presence of NaH.65 3-Formalated compound 7 was obtained with
the Vilsmeier reaction. The π-conjugation frameworks of the
sensors were extended with Sonogashira cross-coupling to obtain
fluorophores with emission at a longer wavelength than the
parent fluorophore (sensor 4, Scheme 1). Then with reductive
amination and reaction with 2-(2-bromomethylphenyl)-1,3,2-
dioxaborinane, the monoboronic acid sensors 1-3 were ob-
tained. All of the compounds were obtained in satisfying yields
(14-35%).
2.3. Excitation/Emission Spectra of the Sensors and
Observation of the Predicted d-PeT Effect. The fluorescence
spectra of probes 1-3 were recorded (Figure 2). The excitation/
emission bands of sensors 1-2 are structureless, compared to
the structured excitation/emission spectra of the previous d-PeT
sensor,27 clearly indicating that electronic communication is
efficient between the moieties at both ends of the ethynylene
(60) Galmiche, L.; Mentec, A.; Pondaven, A.; Her, M. L. New J. Chem.
2001, 25, 1148–1151.
(66) Zhou, Y.; Xiao, Y.; Chi, S.; Qian, X. Org. Lett. 2008, 10, 633–636.
(67) Peng, X.; Du, J.; Fan, J.; Wang, J.; Wu, Y.; Zhao, J.; Sun, S.; Xu, T.
J. Am. Chem. Soc. 2007, 129, 1500–1501.
(61) Velasco, D.; Castellanos, S.; Lopez, M.; Lopez-Calahorra, F.; Brillas,
E.; Julia, L. J. Org. Chem. 2007, 72, 7523–7532.
(68) Peng, X.; Song, F.; Lu, E.; Wang, Y.; Zhou, W.; Fan, J.; Gao, Y.
J. Am. Chem. Soc. 2005, 127, 4170–4171.
(62) Jian, H.; Tour, J. M. J. Org. Chem. 2003, 68, 5091–5103.
(63) Zhu, Z.; Moore, J. S. J. Org. Chem. 2000, 65, 116–123.
(64) Tucker, S. H. J. Chem. Soc. 1926, 546–553.
(69) Jin, S.; Wang, J.; Li, M.; Wang, B. Chem.sEur. J. 2008, 14, 2795–
2804.
(65) Xu, T.; Lu, R.; Qiu, X.; Liu, X.; Xue, P.; Tan, C.; Bao, C.; Zhao, Y.
Eur. J. Org. Chem. 2006, 4014–4020.
(70) Zheng, S.; Lin, N.; Reid, S; Wang, B. Tetrahedron 2007, 63, 5427–
5436.
9
J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009 17457

A R T I C L E S
Zhang et al.
Figure 3. Fluorescence spectral changes of sensor 1 (3.21 × 10-6 mol dm-3) at different pH; 5.0 × 10-2 mol dm-3 NaCl ionic buffer (52.1% methanol in
water). (a) From pH 6.37 to 4.23; (b) from pH 4.23 to 2.00. λex ) 360 nm; 18 °C.
in Supporting Information). For probe 3, however, the normal
a-PeT effect was observed (see later discussion). We propose
that the a-PeT effect for probe 3 is due to the lack of the electron
donating group, -NMe2. The d-PeT as well as the a-PeT effect
of the sensors can be rationalized by considering the free energy
changes of the sensors.2,19-22,72 We noticed that sensors 1 and
2 show relatively strong emissions at basic pH; this is different
from the normal a-PeT sensors.18 Fluorescence quantum yields
also support the analysis (Table 2). For example, sensor 1 gives
Φ of 0.04 at pH 4.0 (protonated). Conversely, Φ ) 0.06 was
observed at pH 7.5. This is in contrast to sensor 3, which show
the a-PeT effect with reversed Φ profiles (Table 2).
2.4. pH and Analyte Titration of Sensor 1. The recognition
of analytes by sensor 1 show rich fluorescence transduction
motifs. The pH titration of sensor 1 with and without analytes
is presented in Figure 4. In the presence of tartaric acid, the
Figure 4. Fluorescence intensity-pH profile of probe 1 in the presence of
D-tartaric acid, D-mandelic acid, and D-lactic acid. The concentration of
hydroxyl acids is 2.5 × 10-2 mol dm-3; 3.21 × 10-6 mol dm-3 of sensor
1 in 5.0 × 10-2 mol dm-3 NaCl ionic buffer (52.1% methanol in water).
λex ) 360 nm, λem ) 430 nm; 18 °C.
emission intensity of sensor 1 increased at pH 4.0 compared to
the emission of blank sensor but decreased at pH 6.0 and
increased again in the pH range 8-10 (Figure 4). Notably, the
sensor.27 For example, the fluorescence enhancement is ca.
2-fold for sensor 1, on switching the pH from 4.23 to 6.37,
while the reported d-PeT boronic sensor shows only a 0.25-
fold enhancement.27 We attribute this improvement to the
stronger electron-donating ability of the modified fluorophore.
By decreasing the pH from 4.23 to 2.00, a new emission
band at 390 nm developed and the emission intensity at 475
nm concomitantly decreased (Figure 3b). The pKa value for
this change is 3.31 ( 0.10 (Figure 4). We propose that this
pH-dependency of the emission at pH 2.0-4.5 is due to the
switch from an ICT to LE emission state, caused by
protonation of the dimethylaminophenyl moiety. With pro-
tonation of the aryl nitrogen atom (Scheme 2c), the ICT
excited state will be replaced by a LE excited state.3,10,71 At
the same time the electron-donating capability of the phe-
nylethynylene moieties decreased, so we also propose that
the d-PeT is terminated at low pH, which enhances the
fluorescence emission.
The above postulation about the fluorescence changes at pH
4.5-2.0, and the transition of the S1 from ICT to LE state was
also rationalized by DFT/TDDFT calculations (Figure S75 and
Table S1 in Supporting Information). The f value for S1 of the
bis-protonated sensor 1 is as high as 1.0008 and indicates S1 as
an emissive state.10,58 This is fully supported by experimental
observations (Figure 3).
sensor gave large fluorescence enhancement at basic pH in the
presence of tartaric acid and mandelic acid (for the binding
curves, see Figure S28 and Figure S29 in Supporting Informa-
tion); this is in stark contrast to the sensors with normal a-PeT
effect.18 Such a rich fluorescence transduction was not observed
for normal PeT boronic sensors.18,24,73-75 The apparent pKa
values of sensor 1 increased to 6.83 ( 0.11 (r2 ) 0.9356) in
the presence of tartaric acid, from 4.79 ( 0.14.
Good chemoselectivity was found for sensor 1 toward the
recognition of tartaric acid and mandelic acid. For example, at
pH 5.6, the emission of the sensor decreased in the presence of
tartaric acid but increased in the presence of mandelic acid
(Figures 4 and 5). The mechanism for this enhancement/
reduction is unclear.27
The emission spectra changes of sensor 1 in the presence of
tartaric acid and mandelic acid at pH 4.0, 5.6, and 7.5 are
presented in Figure 5. At pH 4.0, fluorescence enhancement
was observed in the presence of either tartaric acid or mandelic
acid (Figure 5a,d). Thus, the recognition of hydroxyl acid with
significant fluorescence transduction was achieved at acidic pH
with the d-PeT effect.18,27
At pH 5.6, the emission intensity decreased in the presence
of tartaric acid but increased in the presence of mandelic
The emission intensity-pH profiles of probes 2 and 3 were
also studied and d-PeT was found for probe 2 (see Figure S35
(72) Mineno, T.; Ueno, T.; Urano, Y.; Kojima, H.; Nagano, T. Org. Lett.
2006, 8, 5963–5966.
(73) Kawanishi, T.; Romey, M. A.; Zhu, P. C.; Holody, M. Z.; Shinkai, S.
J. Fluoresc. 2004, 14, 499–512.
(71) Zhao, G.; Chen, R.; Sun, M.; Liu, J.; Li, G.; Gao, Y.; Han, K.; Yang,
(74) Phillips, M. D.; James, T. D. J. Fluoresc. 2004, 14, 549–559.
(75) Cao, H.; Heagy, M. D. J. Fluoresc. 2004, 14, 569–584.
X.; Sun, L. Chem.sEur. J. 2008, 14, 6935–6947.
9
17458 J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009

Phenylethynylated Carbazole Boronic Acid Sensors
A R T I C L E S
Figure 5. Emission spectra of sensor 1 in the presence of D-tartaric acid at (a) pH 4.0; (b) pH 5.6; (c) pH 7.5 and the emission spectra in the presence of
D-mandelic acid at (d) pH 4.0; (e) pH 5.6; (f) pH 7.5. 3.21 × 10-6 mol dm-3 of sensor 1 in 5.0 × 10-2 mol dm-3 NaCl ionic buffer (52.1% methanol in
water), 18 °C. λex ) 360 nm.
Figure 6. Termination of the d-PeT effect of the protonated sensor 1 upon binding with analyte (the analyte was simplified as ethylene glycol, OHCH2CH2OH,
to reduce the calculation time (MeOH inserted zwitterionic structure). The transitions for the low-lying excited states of S1rS0 (LUMOrHOMO) and
S2rS0 (LUMO+1rHOMO) are presented. Note S1rS0 is an allowed electronic transition with oscillator strength f ) 0.0433. The optimized structure of
the binding complex is shown.
Table 3. Selected Electronic Excitation Energies (eV) and
Corresponding Oscillator Strengths (f), Main Configurations and CI
acid (Figure 5). Moreover, red-shift of the emission peaks
were observed in the presence of analytes; this observation
isdifferentfromthenormala-PeTboronicacidsensors,17,18,23,24
which show no shift of the emission band in the presence of
analytes.
We propose that the fluorescence enhancement in the presence
of tartaric acid at pH 4.0 is due to the transformation of boron
from sp2 to sp3 hybridized, with release of the steric constrain
caused by bonding with tartaric acid, and finally the termination
of the d-PeT effect (quenching effect, see Scheme 2).
Coefficients of the Low-Lying Electronically Excited States of
Sensor 1 Bonded with Ethylene Glycol (see Figure 6 for the
Structure of the Binding Complex)a
TDDFT//B3LYP/6-31G(d)
electronic transition
energy (eV)b
fc
compositiond
CIe
S0fS1
S0fS2
S0fS3
S0fS4
3.34 (371 nm)
3.62 (342 nm)
3.75 (330 nm)
3.83 (323 nm)
0.0433
1.3642
0.0015
0.1183
HfL
0.6784
0.1984
0.6980
0.6941
HfL+1
HfL+2
HfL+3
2.5. Rationalization of the Binding Induced Emission
Enhancement at Acidic pH and the Solvent-Inserted
Zwitterionic Structure of the Binding Complex. The above
postulation on the enhancement of the emission of sensor 1 upon
binding with analytes at acidic pH, that is, the fluorescence
transduction mechanism with the d-PeT effect, can be rational-
ized using DFT/TDFFT calculations. The binding complex of
1 with analytes (simplified as ethylene glycol) was optimized,
and the MOs of the low-lying electronic transitions of the
complex are presented in Figure 6 and Table 3. It is accepted
a Methanol inserted zwitterionic structure was used for the
calculation. Calculated by TDDFT//B3LYP/6-31G(d), based on the
optimized ground state geometries. b Only selected low-lying excited
states were considered. The numbers in parentheses are the excitation
energy in wavelength. c Oscillator strength. d H stands for HOMO and L
stands for LUMO. Only the main configurations are presented. e CI
coefficients are in absolute values.
that upon bonding with an analyte, a solvent molecule inserted
complex will result; that is, a methoxy group will bond to boron
atom to release the steric strain, by transformation of sp2 boron
9
J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009 17459

A R T I C L E S
Zhang et al.
to sp3 boron, and an intramolecular hydrogen bond will
form.18,28,34,41,47,52,76-81 It is only in aprotic solvents that the
dative B-N will be formed.5,47a,76
A zwitterionic structure was used for the DFT calculations,
and the boron atom takes a sp3 tetrahedron geometry. The
optimization displays an intramolecular hydrogen bond (H-bond)
formed with the methoxy group as the H-bond acceptor and
N-H as the H-bond donor. The optimized H-bond length
(N-H· · ·O) is 2.657 Å, which is very close to a crystal structure
of boronate complex with a H-bond length (N-H· · ·O) of 2.655
Å.18,34,76 These results indicate that molecular optimization
based on DFT calculations can precisely predict the geometries
of the boronate complex.
On the basis of the optimized ground state geometries, we
carried out the TDDFT calculation on the low-lying electroni-
cally excited states of the complex (Table 3). The oscillator
strength of the S1 state (f ) 0.0433) of the complex increased
by ca. 10-fold compared to that of the protonated sensor 1 (f )
0.0045, Figure 1 and Table 1). Moreover, there is overlap
between the HOMO and LUMO (Figure 6).10,58 Thus, we
propose that the fluoresce of the complex may be intensified
compared to the protonated sensor (free sensor 1). These
expectations based on the DFT/TDDFT calculations were
verified by the experimental observations (Figure 5); that is,
the emission of protonated sensor 1 intensified in the presence
of analytes (Figure 5). Furthermore, we can see that the d-PeT
effect of protonated sensor 1 (Figure 1) terminated upon binding
an analyte. On the basis of the electronic transitions, we propose
that the protonated amine/boronic acid groups act as ET
acceptors and play an indispensable role for the d-PeT effect
of protonated sensor 1. The pH dependency of the compound
without a boronic acid group, that is, amine 9a, was also
investigated and no d-PeT effect but rather the normal a-PeT
effect was observed (see Figure S32 in Supporting Information).
2.6. Analyte Titration with the Sensors: Chemoselectivity
through Emission Enhancement or Reduction. The binding
curves of sensor 1 with mandelic acid and tartaric acid at pH
5.6 are presented in Figure 7. At pH 5.6, enhancement of the
emission intensity was observed in the presence of mandelic
acid, whereas the intensity was reduced in the presence of
tartaric acid. Binding constants of (3.20 ( 0.28) × 105 M-1
and (1.14 ( 0.17) × 104 M-1 were determined for tartaric acid
and mandelic acid, respectively. To our knowledge, chemose-
lectivity with enhancement/reduction of intensity in the presence
of different analytes has never been reported for fluorescent
chemosensors.4,9,82,83
Figure 7. Relative fluorescence intensity of probe 1 vs concentrations of
D-mandelic acid and tartaric acid at pH 5.6. For D-mandelic acid, λex
)
360 nm, λem ) 457 nm. For d-tartaric acid, λex ) 360 nm, λem ) 454 nm.
The solid lines are the fitting results of 1:1 binding; 3.21 × 10-6 mol dm-3
of sensor 1 in 5.0 × 10-2 mol dm-3 NaCl ionic buffer (52.1% methanol in
water), 18 °C.
Figure 8. Binding isotherms of sensor 1 with monosaccharide at pH 7.5.
The solid lines are the fitting results of 1:1 binding. λex ) 360 nm, λem
)
430 nm; 3.21 × 10-6 mol dm-3 of sensor in 5.0 × 10-2 mol dm-3 NaCl
ionic buffer (52.1% methanol in water); 18 °C.
determined for tartaric acid and mandelic acid, respectively.
Such a recognition ability at acidic pH is impossible for the
normal a-PeT sensors, where the background emission of the
sensor is too strong.3,5,17,23,24
The fluorescence transduction mechanisms for sensor 1 can
be summarized in Scheme 2. The free sensor 1 (in media with
pH > 6.0) is highly fluorescent. With monoprotonation of sensor
1, the d-PeT effect was switched on; thus, the fluorescence
intensity is reduced. On further reduction of the pH to pH <
4.0, the sensor became bis-protonated and the d-PeT effect was
terminated (Figures 3 and 4); thus, the emission peak is blue-
shifted and the intensity is enhanced. Upon binding with
analytes, the d-PeT effect was also terminated; thus, fluorescence
enhancement was observed (Figure 4). These assumptions are
in agreement with the fluorescence measurement and are
supported by the DFT/TDDFT calculations.
The emission intensity-pH profile of sensor 1 in the presence
of glucose, fructose, and galactose were also investigated (Figure
8). Sensor 1 showed reduced emission in the presence of fructose
at pH 7.5. But fluorescence enhancements were observed in the
presence of glucose and galactose at pH 7.5. Binding constants
of (1.52 ( 0.33) × 105 and (1.28 ( 0.53) × 105 M-1 were
determined for glucose and fructose, respectively. The binding
constant was (1.70 ( 0.50) × 105 M-1 for galactose. To our
knowledge, recognition of monosaccharides with such fluores-
The emission intensity of sensor 1 at pH 4.0 was enhanced
in the presence of both tartaric acid and mandelic acid (see
Figure S27 in Supporting Information). Binding constants of
(2.67 ( 0.42) × 104 and (3.99 ( 0.84) × 104 M-1 were
(76) Collins, B. E.; Sorey, S.; Hargrove, A. E.; Shabbir, S. H.; Lynch, V. M.;
Anslyn, E. V. J. Org. Chem. 2009, 74, 4055–4060.
(77) Kaiser, P. F.; White, J. M.; Hutton, C. A. J. Am. Chem. Soc. 2008,
130, 16450–16451.
(78) Sartain, F. K.; Yang, X.; Lowe, C. R. Chem.sEur. J. 2008, 14, 4060–
4067.
(79) Gamsey, S.; Miller, A.; Olmstead, M. M.; Beavers, C. M.; Hirayama,
L. C.; Pradhan, S.; Wessling, R. A.; Singaram, B. J. Am. Chem. Soc.
2007, 129, 1278–1286.
(80) Berube, M.; Dowlut, M.; Hall, D. G. J. Org. Chem. 2008, 73, 6471–
6479.
(81) Kano, N.; Yoshino, J.; Kawashima, T. Org. Lett. 2005, 7, 3909–3911.
(82) Bell, T. W.; Hext, N. M. Chem. Soc. ReV. 2004, 33, 589–598.
(83) Lin, J.; Hu, Q.; Xu, M.; Pu, L. J. Am. Chem. Soc. 2002, 124, 2088–
2089.
9
17460 J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009

Phenylethynylated Carbazole Boronic Acid Sensors
A R T I C L E S
constants of the sensors with selected analytes of R-hydroxyl
carboxylic acid and monosaccharide are compiled in Table 4.
Above all, sensor 1 and 2 are found to be d-PeT sensors,
which is consistent with theoretical predictions. However,
experimentally sensor 3 is a normal a-PeT effect system in
disagreement with the theoretical predictions. We also demon-
strated that the fluorescence transductions of the sensors in the
presence of analytes are good (Figures 4, 5, 7, and 8).
2.8. Evaluate the d-PeT Effect Using Thermodynamics
and Kinetics: Application of the Rehm-Weller and the
Marcus Equations. To explain the discrepancy between the
theoretically predicted d-PeT effect by DFT/TDDFT calculations
and the experimentally observed a-PeT effect for sensor 3
(Figure 9), we used a thermodynamic criteria to evaluate the
possibility of the electron transfer process, that is, the free energy
changes (∆G°) of the d-PeT effect. This driving force for the
ET can be obtained by using the Rehm-Weller equation (eq
2).1,2,11,19-22
Figure 9. Fluorescence emission intensity-pH profile of sensor 3 with and
without analytes. The solid lines are the pKa fitting results of pH-intensity
curves. The concentration of hydroxyl acids is 1.25 × 10-2 mol dm-3; 3.0
× 10-6 mol dm-3 of sensor 3 in 5.0 × 10-2 mol dm-3 NaCl ionic buffer
(52.1% methanol in water). λex ) 340 nm, λem ) 389 nm; 18 °C.
∆Go ) Eo(D+/D) - Eo(A/A-) - ∆E0,0 + wp
cence enhancement/reduction has never been reported.5,6,46,48,52,84
Such a phenomenon is rare, because boronic acid based sensors
usually display fructose selectivity over other monosaccharides.
The emission intensity-pH profile of sensor 2 was also
investigated, and the response in the presence of analytes, such
as tartaric acid and mandelic acid, etc., was determined. It was
found that sensor 2 is a d-PeT sensor (see Figure S34 in
Supporting Information). However, the fluorescence transduction
is relatively simple compared to that of sensor 1. No fluores-
cence enhancement/reduction was observed in the presence of
different analytes (see Figure S40, S42 and S44 in Supporting
Information).
2.7. Normal PeT Effect of Sensor 3. Sensor 3 was designed
without the electron-donating group (-NMe2). d-PeT effect was
predicted by DFT/TDDFT calculations (see Figures S78 and
S79 and Table S3 in Supporting Information). The fluorescence
emission spectra of sensor 3 as a function of the pH or in the
presence of analytes was investigated. Interestingly, a normal
a-PeT emission intensity-pH profile was observed, that is, sensor
3 show intensified emission at acidic pH and reduced emission
at basic pH, with pKa values of 5.43 ( 0.05 (Figure 9). This
emission intensity-pH profile is typical for a-PeT system.
In the presence of tartaric acid, mandelic acid, and lactic acid,
the emission intensity of sensor 3 was enhanced in the pH range
6.0-8.0, with apparent pKa values of 7.53 ( 0.03, 8.28 ( 0.05,
and 7.44 ( 0.05, respectively. This emission intensity-pH profile
of the sensor alone or the sensor in the presence of analytes
were similar to the normal a-PeT sensors.18 Herein we can
clearly see that the normal a-PeT mechanism is in operation.
The recognition of tartaric acid or mandelic acid at pH 4.0 is
impossible for sensor 3, because the fluorescence enhancement
in the presence of analytes is too small.85-89 But with d-PeT
sensor, that is, sensor 1, the recognition of mandelic acid/tartaric
acid at pH 4.0 was possible (Figures 4 and 5). The binding
e2
wp
)
(2)
4πεR
where E°(D+/D) is the oxidation potential of the electron donor,
E°(A/A-) is the reduction potential of the electron acceptor;
herein, the values are approximated as the energy difference of
HOMO and LUMO (calculated with DFT). ∆E0,0 is the
zero-zero transition energy (approximated as the crossing point
of the excitation and the emission spectra of the sensors). ꢀ is
the dielectric constant of media (approximated as methanol with
ꢀ value of 32.7). R is the distance between the donor and
acceptor (distance between the two N atoms of the carbazole
and amine moieties, respectively, calculated by the DFT
optimization).
Negative ∆G° means the ET is thermodynamically possible.
Conversely, positive ∆G° value means the ET is impossible.
Free energy changes for the ET of sensor 1 were calculated as
∆G° ) -49.0 kJ mol-1 (Table 5). For sensor 2, ∆G° ) -28.8
kJ mol-1. These negative values infer that ET may occur for
sensor 1 and sensor 2. Experimentally, d-PeT effect was
observed for both sensor 1 and sensor 2 (Figure 4 and
Supporting Information).
Similar consideration was also applied for sensor 3, for which
d-PeT effect was also predicted by DFT/TDDFT calculations
(see Figure S77, Figure S78, and Table S3 in Supporting
Information), but not observed experimentally (Figure 9). For
sensor 3, a positive free energy change of ∆G° ) 14.5 kJ mol-1
was obtained (Table 5), which infers that the ET can not occur
spontaneously.19 Thus the ∆G° values can be used to predict
the d-PeT effect.
Alternatively, the effect of d-PeT on the emission profile of
the sensors can also be evaluated by kinetics of the ET process
of the ET effect. The kinetics of the ET is important because
the two major competing process, the ET (quenching effect)
and the radiative decay of the excited molecules (giving
emission) dictate the fate of the excited states.1-3 With the free
energy changes of the ET (Table 5), the ET rate constants (kET)
can be estimated by the Marcus equation (eq 3).2,19,21
(84) James, T. D.; Shinkai, S. Top. Curr. Chem. 2002, 218, 159–200.
(85) Schertzer, B. M.; Baker, S. N.; Diver, S. T.; Baker, G. A. Aust.
J. Chem. 2006, 59, 633–639.
(86) Tan, W.; Zhang, D.; Wang, Z.; Liu, C.; Zhu, D. J. Mater. Chem. 2007,
17, 1964–1968.
(87) Yu, Y.; Zhang, D.; Tan, W.; Wang, Z.; Zhu, D. Bioorg. Med. Chem.
Lett. 2007, 17, 94–96.
1/2
4π3
(∆Go + λ)2
4λkBT
V2 exp -
(3)
(88) DiCesare, N.; Lakowicz, J. R. Chem. Commun. 2001, 2022–2023.
(89) Luvino, D.; Gasparutto, D.; Reynaud, S.; Smietana, M.; Vasseur, J.
Tetrahedron Lett. 2008, 49, 6075–6078.
kET
)
2
[
]
(
)
p λkBT
9
J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009 17461

A R T I C L E S
Zhang et al.
Table 4. Stability Constants (M-1) of Sensors 1-3 with R-Hydroxyl Acid and Monosaccharides
stability constants
pH
sensor 1
sensor 2
sensor 3
c
tartaric acid
4.0
5.6
7.5
4.0
5.6
7.5
4.0
5.6
7.5
7.5
7.5
7.5
(2.67 ( 0.42) × 104
(3.20 ( 0.28) × 105
(1.74 ( 0.93) × 103
(3.99 ( 0.84) × 104
(1.14 ( 0.17) × 104
(2.28 ( 1.78) × 103
(4.33 ( 0.56) × 103
(6.47 ( 1.62) × 104
(4.67 ( 0.90) × 104
(1.52 ( 0.33) × 105
(1.70 ( 0.50) × 105
(1.28 ( 0.53) × 105
(1.81 ( 0.59) × 104
(3.98 ( 0.49) × 104
(1.49 ( 0.27) × 105
(4.28 ( 1.01) × 104
(1.22 ( 0.16) × 104
(5.46 ( 2.44) × 103
(3.31 ( 1.06) × 103
(5.28 ( 0.11) × 103
(6.32 ( 2.22) × 104
(6.56 ( 1.68) × 104
(2.33 ( 0.89) × 104
(1.73 ( 0.62) × 104
-
b
(4.20 ( 0.51) × 103
c
-
-
c
mandelic acid
lactic acid
b
b
(1.58 ( 0.16) × 104
c
-
-
c
a
(4.16 ( 0.41) × 103
(1.33 ( 0.28) × 104
(2.56 ( 0.73) × 103
(4.36 ( 0.84) × 103
c
-
glucose
galactose
fructose
a Fluorescence intensity decreased in the presence of analytes. b pH 6.0. c At which pH is not appropriate for binding constant determination.
Table 5. Parameters Used in the Calculation of the Free Energy
was observed; instead, the normal a-PeT effect was observed.
This discrepancy between the DFT/TDDFT calculations can be
rationalized by consideration of the kinetics and thermodynamics
of the ET process, that is, by using the Marcus (kET) and
Rehm-Weller equations (free energy changes, ∆G°). The d-PeT
boronic acid sensors 1 and 2 show improved photophysical
properties compared to the reported sensor, such as red-shifted
emission wavelength, larger Stokes shifts, and most importantly
improved fluorescence transduction efficiency of the d-PeT
effect. The fluorescence intensity of 1 and 2 can be modulated
by ca. 2-fold compared to the reported value of 0.25-fold; thus,
the fluorescence modulation is 8× more efficient for the new
d-PeT sensors. Fluorescence transduction of analyte recognition
was observed, for example, novel fluorescence enhancement/
reduction was observed for sensor 1 in the presence of mandelic
acid or tartaric acid at pH 5.6. The effect of pH as well as the
bonding with analytes on the emission of the sensors was
rationalized using DFT/TDDFT calculations. We believe our
successful application of DFT/TDDFT calculations and the free
energy change considerations in the rationalization of the
photophysical properties of the sensors reported herein may
inspire others in the quest for the rational design of fluorophores
and sensors with predetermined photophysical properties.
Changes of the Potential d-PeT Effect With Rehm-Weller
Equation (eq 2)a
b
b
c
d
EHOMO/(eV) ELUMO/(eV) ∆E/(eV) ∆E0,0(eV) wp(eV) ∆G/kJ mol-1/(eV)
sensor 1 -6.18
sensor 2 -6.15
sensor 3 -7.13
-3.67
-3.62
-3.76
2.50
2.53
3.37
3.16
2.98
3.38
0.15 -49.0(-0.51)
0.15 -28.8(-0.30)
0.15 +14.5(+0.15)
a MeOH as the Solvent. b Calculated values. c ∆E ) ELUMO - EHOMO
.
d Crossing point of the excitation and emission spectra.
where p is Planck’s constant, kB is the Boltzmann constant, T
is the absolute temperature, λ is the reorganization energy, and
V is the electronic coupling matrix element. On the basis of the
assumption that the radii of the donor moiety is about 5 Å, the
radii of the ET acceptor is 2.5 Å (DFT optimization value), the
λ values of the sensor 1 was calculated as 2.4 eV. The electronic
coupling matrix of the three sensors molecules was arbitrarily
assumed as 0.5 eV.19 The kET values of the protonated sensors
1, 2, and 3 were calculated as 1.20 × 109 s-1, 4.69 × 107 s-1
,
max
and 7.37 × 104 s-1, respectively (kET
) 2.71 × 1015 s-1).
The ET rate constants of sensors 1 and 2 are comparable to the
rate constants of radiative decay (fluorescence rate constant, kr
on scale of 107 s-1, Table 1);1,2 thus, the ET can compete with
radiative decay, the fluorescence will be reduced, and this is
the d-PeT effect. For the protonated sensor 3, however, the
extremely slow ET rate constants (kET ) 7.37 × 104 s-1) implies
that it cannot compete efficiently with the radiative decay
process (kr ) 3.07 × 107 s-1); thus, fluorescence is possible
and a-PeT effect is expected for sensor 3. These considerations
are in agreement with the experimental observations. It should
be pointed out that these considerations are approximations
because some of the parameters are estimated. However, the
results clearly demonstrate the trend of the kinetics of the ET
process for protonated sensors 1-3 and it is useful for evaluation
of the photophysical properties of the sensors from a kinetic
point of view.
3. Experimental Section
3.1. Analytical Measurements. NMR spectra were recorded on
a Bruker 400 MHz spectrometer (CDCl3 or CDCl3/CD3OD as
solvents, TMS as standard, δ ) 0.00 ppm). High resolution mass
spectra (HRMS) were determined on a LC/Q-TOF MS system
(UK). Melting points were measured on a XPR-300C microscopic
instrument. Fluorescence spectra were measured on a F4500
(Hitachi) or a CRT 970 spectrofluorometer. Fluorescence lifetimes
were measured with a Fluoromax-4 spectrofluorometer (Horiba
Jobin Yvon). Absorption spectra were recorded on a Perkin-Elmer-
Lambda-35 UV/vis spectrophotometer. A 0.05 M NaCl (52.1%
methanol in water, w/w) ionic buffer was used in the experiment.
All pH were measured using a Delta 320 Microprocessor pH meter
(Mettler Toledo).
3.2. Synthesis of 2-[[[(N-Butyl-6-[4′-dimethylaminophenyl-
ethynyl]-9H-carbazol-3-yl)methyl]-[[4′-fluorophenylmethyl]ami-
no]methyl]phenyl]boronic Acid (Sensor 1). 9a (100 mg, 0.20
mmol), 2-(2-bromomethylphenyl)-1,3,2-dioxaborinane (65.0 mg,
0.26 mmol), and K2CO3 (118.0 mg, 0.86 mmol) were mixed in
acetonitrile (5 mL); the mixture was refluxed for 10 h under N2.
The mixture was cooled, and dichloromethane was added. The
organic layer was washed with water and dried over anhydrous
Na2SO4. The solvent was removed, and the residue was purified
with column chromatography (Al2O3, ethyl acetate/MeOH, 100:1,
V/V). Yellow powder (18.5 mg) was obtained, yield: 14.6%. Mp
Herein we have demonstrated the potential, as well as the
limitations of the DFT/TDDFT calculation in the rational PeT
sensor design. Our result demonstrated that thermodynamics as
well as kinetics are also useful for evaluation of the photo-
physical properties of the fluorescent sensors. These results will
be helpful for future rational design of fluorescent sensors and
fluorophores with predetermined photophysical properties.
2.9. Conclusions. In conclusion, we synthesized three new
phenylethynylated carbazole boronic acid sensors 1-3. The
sensors were rationally designed using DFT/TDDFT calcula-
tions, which predicted a d-PeT effect for all the sensors.
Experimental results proved the theoretical predictions of a
d-PeT effect for sensor 1 and 2. However, for sensor 3, no d-PeT
9
17462 J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009

Phenylethynylated Carbazole Boronic Acid Sensors
A R T I C L E S
169-171 °C. 1H NMR (400 MHz, CDCl3/CD3OD) δ 8.26 (s, 1H),
7.94 (s, 1H), 7.85 (d, 1H, J ) 6.0 Hz), 7.61 (d, 1H, J ) 8.4 Hz),
7.45 (d, 2H, J ) 8.8 Hz), 7.25-7.38 (m, 7H), 7.17 (d, 1H, J ) 6.4
Hz), 7.02-7.06 (m, 2H), 6.69 (d, 2H, J ) 8.8 Hz), 4.29 (t, 2H, J
) 7.2 Hz), 3.79 (s, 2H), 3.77 (s, 2H), 3.60 (s, 2H), 3.00 (s, 6H),
1.82-1.88 (m, 2H), 1.38-1.43 (m, 2H), 0.96 (t, 3H, J ) 7.6 Hz).
13C NMR (100 MHz, CDCl3/CD3OD) δ 150.0, 141.8, 140.5, 140.1,
136.8, 132.7, 132.1, 131.7, 131.6, 131.5, 130.6, 129.5, 128.0, 127.7,
126.4, 123.8, 122.7, 122.6, 122.3, 115.6, 115.4, 114.5, 112.1, 110.9,
109.2, 108.9, 88.7, 88.6, 61.5, 57.9, 56.5, 43.2, 40.5, 31.3, 20.7,
14.0. ESI-HRMS ([C41H41BFN3O2 + H]+) calcd 638.3354, found
638.3377.
3.3. Synthesis of 2-[[[(N-Butyl-6-[4′-dimethylaminophenyl-
ethynyl]-9H-carbazol-3-yl)methyl]-[[phenylmethyl]amino]me-
thyl]phenyl]boronic Acid (Sensor 2). Similar procedure to the
synthesis of sensor 1 was used. Yellow powder (27.1 mg) was
obtained, yield: 21.3%. Mp 76-78 °C. 1H NMR (400 MHz, CDCl3/
CD3OD) δ 8.26 (s, 1H), 7.96 (s, 1H), 7.83 (d, 1H, J ) 6.4 Hz),
7.61-7.63 (m, 1H), 7.45-7.47 (m, 2H), 7.28-7.39 (m, 10H), 7.19
(br, 1H), 6.70-6.72 (m, 2H), 4.30 (t, 2H, J ) 7.2 Hz), 3.83 (s,
2H), 3.80 (s, 2H), 3.66 (s, 2H), 3.00 (s, 6H), 1.83-1.90 (m, 2H),
1.37-1.46 (m, 2H), 0.94 (t, 3H, J ) 7.2 Hz). 13C NMR (100 MHz,
CDCl3/CD3OD) δ 150.1, 140.5, 140.2, 136.3, 132.7, 131.5, 131.4,
131.3, 130.1, 130.0, 129.4, 128.7, 128.6, 128.2, 127.8, 127.7, 127.5,
123.7, 122.7, 122.5, 122.4, 114.4, 112.2, 111.0, 109.0, 108.9, 88.6,
88.5, 61.2, 57.2, 56.6, 43.2, 40.4, 31.3, 20.7, 14.0. ESI-HRMS
([C41H42BN3O2 + H]+) calcd 620.3448, found 620.3429.
31.3, 20.7, 14.0. ESI-HRMS ([C39H36BFN2O2 + H]+) calcd
595.2932, found 595.2913.
3.5. Computational Details. The ground state structures of
sensors were optimized using density functional theory (DFT) with
B3LYP functional and 6-31G (d) basis set. The excited state related
calculations were carried out with the Time dependent DFT (TD-
DFT) with the optimized structure of the ground state (DFT/6-
31G(d). There are no imaginary frequencies in frequency analysis
of all calculated structures; therefore, each calculated structures are
in local an energy minimum. All of these calculations were
performed with Gaussian 03.90
Acknowledgment. We thank the NSFC (20642003, 20634040,
20833008 and 40806042), Ministry of Education (SRF for ROCS,
SRFDP-200801410004 and NCET-08-0077), PCSIRT(IRT0711),
NKBRSF (2007CB815202), State Key Laboratory of Fine Chemi-
cals (KF0710 and KF0802), State Key Laboratory of Chemo/
Biosensing and Chemometrics (2008009), the Education Depart-
ment of Liaoning Province (2009T015), and Dalian University of
Technology (SFDUT07005 and 1000-893394) for financial support.
T.D.J. and J.Z. are grateful to the Royal Society for fostering their
collaboration through the China-UK Science Networks. The scheme
has helped initiate annual Symposia and Joint Laboratories devoted
to Catalysis & Sensing for our Environment (CASE). We are also
grateful to the reviewers for their critical comments on the
manuscript.
3.4. Synthesis of 2-[[[(N-Butyl-6-[phenylethynyl]-9H-carbazol-
3-yl)methyl]-[[4′-fluorophenylmethyl]amino]methyl]phenyl]-
boronic Acid (Sensor 3). Similar procedure to the synthesis of
sensor 1 was used. Yellow powder (45.8 mg) was obtained, yield:
Supporting Information Available: General experimental
methods, synthesis of compounds, compound characterization
data (1H NMR, 13C NMR and HRMS), fluorescence spectra
(emission intensity-pH profiles and binding curves) and DFT/
TDDFT calculation details of sensors 1-3. This material is
available free of charge via the Internet at http://pubs.acs.org.
1
34.9%. Mp 122-124 °C. H NMR (400 MHz, CDCl3/CD3OD) δ
8.30 (s, 1H), 7.96 (s, 1H), 7.84 (d, 1H, J ) 6.8 Hz), 7.64 (d, 1H,
J ) 7.6 Hz), 7.58 (d, 2H, J ) 6.8 Hz), 7.26-7.40 (m, 11H),
7.02-7.06 (m, 2H), 4.31 (t, 2H, J ) 7.2 Hz), 3.80 (s, 2H), 3.78 (s,
2H), 3.61 (s, 2H), 1.83-1.90 (m, 2H), 1.38-1.44 (m, 2H), 0.96 (t,
3H, J ) 7.6 Hz).13C NMR (100 MHz, CDCl3/CD3OD) δ 163.6,
161.2, 141.8, 140.6, 136.9, 132.2, 131.7, 131.6, 131.4, 130.5, 129.6,
128.5, 128.2, 128.0, 127.7, 126.6, 124.2, 124.1, 122.8, 122.6, 122.3,
115.6, 115.4, 113.6, 109.3, 109.0, 91.0, 87.8, 61.5, 58.0, 56.7, 43.3,
JA9060646
(90) Frisch, M. J.; et al. Gaussian 03, Revision D.01, Gaussian, Inc.:
Wallingford, CT, 2003.
9
J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009 17463