Fluorescent signaling provides deeper insight into aromatic anion uptake by metal-ion activated molecular receptors

Article abstract of DOI:10.1039/b719183a

Two new, octadentate, fluorescent, macrocyclic ligands, 1-(2-(9-anthrylmethylamino)ethyl)-4,7,10-tris((2S)-2-hydroxy-3-phenoxypropyl)-1, 4,7,10-tetraazacyclododecane (L¹) and 1-(2-(9-anthrylmethylamino) ethyl)-4,7,10-tris((2S)-2-hydroxy-3-[4′-(methyl)phenoxy]propyl)-1,4,7, 10-tetraazacyclododecane (L²), have been prepared with a view to using them to study aromatic anion sequestration. The eight-coordinate Cd(II) complexes of L¹ and L², [CdL¹](ClO ₄)₂·2H₂O and [CdL²](ClO ₄)₂·4H₂O, have both been shown capable of acting as receptors for a range of aromatic oxoanions. This has been demonstrated by perturbation of both ¹H NMR chemical shift values and the anthracene derived fluorescence emission intensity as the potential guest anion and the receptor are combined. Non-linear least squares regression analysis of the resulting titration curves leads to the determination of binding constants in 20% aqueous 1,4-dioxane which lie in the range 10^2.3 M^-1 (benzoate) to 10^7.5 M^-1 (2,6-dihydroxybenzoate). By reference to earlier, X-ray determined structures of related, but non-fluorescent, inclusion complexes the primary anion retention force is known to arise from hydrogen bonding between the anion and four convergent hydroxy groups that exist at the base of a cavity that develops in L¹ and L² as their aromatic groups juxtapose upon metal ion coordination. This work reveals significant stability enhancement when hydroxy groups are positioned on the anion at points where O-H...π hydrogen bonding to the aromatic rings that constitute the walls of the cavity becomes geometrically possible. The Royal Society of Chemistry and the Centre National de la Recherche Scientifique.

Full text of DOI:10.1039/b719183a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Fluorescent signaling provides deeper insight into aromatic anion uptake
by metal-ion activated molecular receptorsw
Adam J. Bradbury,a Stephen F. Lincolnb and Kevin P. Wainwright*a
Received (in Durham, UK) 12th December 2007, Accepted 13th March 2008
First published as an Advance Article on the web 28th April 2008
DOI: 10.1039/b719183a
Two new, octadentate, fluorescent, macrocyclic ligands, 1-(2-(9-anthrylmethylamino)ethyl)-4,7,10-
tris((2S)-2-hydroxy-3-phenoxypropyl)-1,4,7,10-tetraazacyclododecane (L1) and 1-(2-(9-
anthrylmethylamino)ethyl)-4,7,10-tris((2S)-2-hydroxy-3-[40-(methyl)phenoxy]propyl)-1,4,7,10-
tetraazacyclododecane (L2), have been prepared with a view to using them to study aromatic
anion sequestration. The eight-coordinate Cd(II) complexes of L1 and L2, [CdL1](ClO4)2ꢀ2H2O
and [CdL2](ClO4)2ꢀ4H2O, have both been shown capable of acting as receptors for a range of
1
aromatic oxoanions. This has been demonstrated by perturbation of both H NMR chemical shift
values and the anthracene derived fluorescence emission intensity as the potential guest anion and
the receptor are combined. Non-linear least squares regression analysis of the resulting titration
curves leads to the determination of binding constants in 20% aqueous 1,4-dioxane which lie in
the range 102.3 Mꢁ1 (benzoate) to 107.5 Mꢁ1 (2,6-dihydroxybenzoate). By reference to earlier,
X-ray determined structures of related, but non-fluorescent, inclusion complexes the primary
anion retention force is known to arise from hydrogen bonding between the anion and four
convergent hydroxy groups that exist at the base of a cavity that develops in L1 and L2 as their
aromatic groups juxtapose upon metal ion coordination. This work reveals significant stability
enhancement when hydroxy groups are positioned on the anion at points where O–Hꢀ ꢀ ꢀp
hydrogen bonding to the aromatic rings that constitute the walls of the cavity becomes
geometrically possible.
the receptor ligand 1, Fig. 1(b), into the appropriate conical
configuration with the O–H bonds directed into the anion
binding cavity. Up to this point we have studied anion
Introduction
In earlier work we have synthesised anion receptors that are
derived from the macrocyclic ligand cyclen (1,4,7,10-tetraaza-
inclusion in this class of receptor using, principally, NMR
cyclododecane) by equipping it with four pendant arms, each
and X-ray crystallography to probe the anion inclusion beha-
having a hydroxy group and an aromatic group attached at or
viour in solution and the solid state, respectively. We now turn
through the 2-position.1 On coordination to a metal ion with
our attention to fluorescence perturbation as a means of
signaling anion inclusion,2 in solution, within the binding
the potential for eight-coordination, this receptor ligand pre-
assembles into a receptor complex where the four aromatic
cavity of these receptor complexes. Since this technique is of
groups come together to form a partially enclosed space, or
higher sensitivity than routine NMR spectroscopy it allows
cavity. This cavity is well suited to aromatic oxoanion reten-
work down to lower host and guest concentrations and
tion since it has at its base four convergent hydroxy groups
consequently the acquisition of binding constant data for
that are activated towards hydrogen bond donation to an
host–guest complexes over a wider stability range. This in
incoming anion by simultaneous coordination to the metal
turn presents an opportunity to gain a fuller understanding of
ion. A number of these receptor complexes containing bound
the principles of molecular recognition that are operative for
anions have been structurally characterized by X-ray
crystallography,1a–d with the one displayed schematically in
Fig. 1(a) being typical.
Here p-aminobenzoate is the guest anion and Cd(II) is the
eight-coordinating metal ion responsible for pre-assembling
a School of Chemistry, Physics and Earth Sciences, Flinders
University, GPO Box 2100, Adelaide, SA 5001, Australia.
E-mail: Kevin.Wainwright@flinders.edu.au
b School of Chemistry and Physics, University of Adelaide, Adelaide,
SA 5005, Australia
w Electronic supplementary information (ESI) available: Syntheses for
the hydrohalide derivatives of compounds 5, 6, L1 and L2 produced for
the purpose of acquiring microanalytical data relevant to the parent
compounds; Table S1, giving binding constant determinations made
using UV monitored titrations. See DOI: 10.1039/b719183a
Fig. 1 (a) The structural characteristics of p-aminobenzoate inclusion
within the host–guest complex formed from the Cd(II) complex of the
receptor ligand 1, shown in (b).
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1500 | New J. Chem., 2008, 32, 1500–1508

View Article Online
these hosts and of the selectivity of the receptors over a wider
range of guest molecules. The design of our receptor ligands is
such that it is relatively easy to replace one of the 2-hydroxy-
ethyl pendant arms in 1, for example, with a similar arm
containing a photoinduced electron transfer (PeT) responsive
fluorophore.3 To preserve the aromatic anion retaining cap-
abilities of [Cd(1)]2+ the replacement arm, besides being
potentially fluorescent, must also contain an electron donor
group that is capable of binding to the metal ion and simulta-
neously acting as a hydrogen bond donor for the guest. It must
also have an aromatic moiety capable of forming a part of the
cavity wall. A pendant group that meets these requirements is
the 2-(9-anthrylmethylamino)ethyl moiety. In previous work
this has been attached as a pendant arm to cyclen by Kimura
et al. to give 2,4 and to cyclam (1,4,8,11-tetraazacyclotetrade-
cane) by Fabbrizzi et al. to give 3.5 Accordingly, we describe in
this report the means by which ligand 1, and its analogue in
which the phenyl groups are replaced by p-tolyl, can be
modified to incorporate the 2-(9-anthrylmethylamino)ethyl
fluorophore as one of the pendant arms. We demonstrate that
the Cd(II) complexes of the modified ligands retain the anion
including capabilities of [Cd(1)]2+, and we report on their
anion including properties together with the associated fluor-
escence perturbations.
Syntheses
1-(Benzyloxycarbonyl)-4, 7,10-tris((2S)-2-hydroxy-3-phenoxy-
propyl)-1,4,7,10-tetraazacyclododecane (5, R = phenyl). A solu-
tion of (2S)-(+)-3-phenoxy-1,2-epoxypropane (623 mg,
4.15 mmol) in dry EtOH (30 cm3) was added to a refluxing
solution of 1-(benzyloxycarbonyl)-1,4,7,10-tetraazacyclodode-
cane (424 mg, 1.38 mmol) in dry ethanol (10 cm3). The reaction
was monitored by TLC on silica (hexane–CH2Cl2, 9 : 1) and
upon complete disappearance of the epoxide, which occurred
after 10 days, the reaction mixture was cooled and the solvent
was evaporated under vacuum to give the pure product as a
viscous yellow oil (1.05 g, quantitative), dH(CDCl3) 7.29 (10 H,
br m, ArH), 6.94 (10 H, br m, ArH), 5.13 (2 H, s, BnCH2),
4.3–2.0 (34 H, br m, CH, CH2 and OH); dC(CDCl3) 158.57 (3 C,
PhO, ipso), 156.20 (1 C, CQO), 136.61 (1 C, Bn, ipso), 129.28
(6 C PhO), 128.52 (1 C, Bn), 128.37 (2 C, Bn), 127.90 (2 C,Bn),
120.72 (3 C, PhO), 114.43 (6 C, PhO), 69.74 (2 C, OCH2) , 69.62
(1 C, OCH2), 67.04 (1 C, CH2Bn), 66.02 (2 C, CHOH), 65.34
(1 C CHOH) 59.46 (2 C, exo-CH2N), 58.00 (1 C, exo-CH2N),
55.06 (2 C, cyclenCH2), 52.76 (2 C, cyclenCH2), 49.77 (2 C,
cyclenCH2), 47.60 (2 C, cyclenCH2).
1, 4, 7-Tris((2S)-2-hydroxy-3-phenoxypropyl)-1, 4, 7,10-tetra-
azacyclododecane (6, R = phenyl). Cyclohexene (300 mg,
3.65 mmol) was added to a stirred solution of 1-(benzyloxy-
carbonyl)-4,7,10-tetraazacyclododecane (523 mg, 0.69 mmol)
dissolved in absolute ethanol (10 cm3). 10% Palladium-on-
carbon catalyst (500 mg) was then added. The reaction
mixture was refluxed at 80 1C for 5 h, filtered through Celite
and the filtrate concentrated in vacuo to give the free base as a
brown oil (507 mg, 97%); dH(CDCl3) 7.27 (6 H, m, PhH); 6.92
(9 H, m, PhH); 4.8–2.0 (35 H, br m, CH, CH2, OH, NH);
dC(CDCl3) 158.67 (2 C, Ph, ipso), 158.60 (1 C, Ph, ipso), 129.29
(2 C, Ph), 129.23 (4 C, Ph), 120.72 (1 C, Ph), 120.60 (2 C, Ph),
114.45 (4 C, Ph), 114.40 (2 C, Ph), 69.91 (2 C, CH2O), 69.28
(1 C, CH2O), 66.44 (1 C, CHOH), 65.46 (2 C, CHOH), 60.20
(3 C, exo-CH2N), 51.41 (4 C, cyclenCH2), 44.36 (4 C, cyclen-
CH2).
Experimental
General information
All syntheses were performed under dry nitrogen. Solvents were
purified using literature methods.6 Microanalyses were con-
ducted at the University of Otago, New Zealand. NMR data
were collected using a Varian Gemini 300 spectrometer operat-
1-(2-(9-Anthrylmethylamino)ethyl)-4, 7, 10-tris((2S)-2-hydroxy-
3-phenoxypropyl)-1,4,7,10-tetraazacyclododecane (L1). 1,4,
7-Tris((2S)-2-hydroxy-3-phenoxypropyl)-1, 4, 7,10-tetraazacy-
clododecane (2.71 g, 4.35 mmol) was dissolved in dry MeCN
(200 cm3) and 9-(2-bromoethyliminomethyl)anthracene (1.36
g, 4.35 mmol) was added as a solid, along with NaHCO3 (540
mg) and molecular sieves. The reaction was wrapped in foil to
keep out light and then refluxed for 10 days before cooling it to
room temperature, filtering it and removing the solvent under
reduced pressure. This gave the intermediate imine as a red oil
(3.23 g, 87%). The imine was redissolved in EtOH (40 cm3)
and NaBH4 (216 mg, 5.71 mmol) was added. The reaction
mixture was stirred overnight, and then diluted with water
(40 cm3), and extracted with CH2Cl2 (4 ꢃ 30 cm3). The
combined organic layers were dried over Na2SO4, then filtered
and evaporated to yield the crude product as a red oil (1.82 g,
76%). Purification was achieved either by column chromato-
graphy on basic alumina with 1% MeOH–CH2Cl2 as eluent
(yield 1.0 g, 41%) or by converting it to its pentahydrobromide
salt, (1.09 g, 45%), as described in the ESI.w The free ligand
1
ing at 300.075 MHz for protons and 75.462 MHz for 13C. H
NMR chemical shifts were referenced to the residual protonated
solvent peak taken as 7.26 ppm for CDCl3, 2.60 ppm for
DMSO-d6, 3.31 ppm for CD3OD and 1.96 ppm for CD3CN.
For 13C NMR spectra, chemical shifts were referenced to the
central resonance of the solvent multiplet peak taken as
77.00 ppm for CDCl3, 49.00 ppm for CD3OD, 39.52 for
DMSO-d6, and d 118.10 for CD3CN (CN resonance). UV-
Visible absorbance spectra were measured on a Varian Cary 50
SCAN UV-Visible spectrophotometer. Flash chromatography
was carried out using the method of Still et al.,7 with silica gel
from Merck Kieselgel (particle size 230–400 mesh) as the
stationary phase. (2S)-(+)-3-Phenoxy-1,2-epoxypropane,1a
(2S)-(+)-3-[40-(methyl)phenoxy]-1,2-epoxypropane,8 1-(benzyl-
oxycarbonyl)-1,4,7,10-tetraazacyclododecane,9 and 9-(2-bromo-
ethyliminomethyl)anthracene5 were prepared by literature
methods.
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1500–1508 | 1501

View Article Online
was recovered from the HBr salt by dissolving the salt in 1 : 1
water–ethanol and basifying to pH 13. Extraction with
CH2Cl2, drying with Na2SO4, filtration and evaporation of
the solvent under reduced pressure gave the free base as a
reddish brown oil (quantitative), lmax/nm (20% aqueous 1,4-
dioxane) 388.5 nm (e/dm3 molꢁ1 cmꢁ1 6896), 368.4 (7347),
350.6 (4647), 334.4 (2170), 321.1 (sh) (805); dH(DMSO-d6) 8.6
(3 H, m AnthH), 8.1 (2 H, m, AnthH), 7.6 (4 H, m, AnthH),
7.30 (6 H, m, PhH), 6.95 (9 H, m, PhH), 5.3–2.0 (41 H, br m,
-CH, -CH2-, OH, NH); dC(CDCl3) 158.0 (3 C, Ph, ipso),
132.4 (2 C, Anth), 132.2 (2 C, Anth), 130.4
(6 C, Ph), 130.0 (1 C, Anth), 126.9 (2 C, Anth), 126.1 (1 C,
Anth), 126.0 (2 C, Anth), 121.6 (2 C, Anth), 121.3 (2 C, Anth),
120.3 (3 C, Ph), 115.2 (6 C, Ph), 71.4 (1 C, CH2O), 71.1
(1 C, CH2O), 68.7 (1 C, CH2O), 68.4 (1 C, CHOH), 67.5 (1 C,
CHOH), 67.1 (1 C, CHOH), 61.4 (1 C, exo-CH2N), 58.9 (1 C,
exo-CH2N), 58.3 (1 C, exo-CH2N), 57.4 (1 C, exo-CH2N), 56.4
(1 C, CH2NH), 53.1 (2 C, cyclenCH2), 52.5 (2 C, cyclenCH2),
51.4 (2 C, cyclenCH2), 50.4 (2 C, cyclenCH2), 46.0 (1 C,
AnthCH2).
1-(Benzyloxycarbonyl)-4,7,10-tris((2S)-2-hydroxy-3-[40-(methyl)-
phenoxy]propyl)-1,4,7,10-tetraazacyclododecane (5,
R
=
p-tolyl). A solution of (2S)-(+)-3-[40-(methyl)phenoxy]-1,2-
epoxypropane (2.05 g, 12.5 mmol) in dry EtOH (30 cm3)
was added to a refluxing solution of 1-(benzyloxycarbonyl)-
1,4,7,10-tetraazacyclododecane (1.28 g, 4.17 mmol) in dry
ethanol (30 cm3). The reaction was monitored by TLC on
silica (hexane–CH2Cl2, 9 : 1) and upon complete disappear-
ance of the epoxide, which occurred after ten days, the
reaction mixture was cooled to room temperature and the
solvent was evaporated to give the product as a viscous yellow
oil (3.33 g, quantitative). dH(CDCl3) 7.37 (5 H, m, Bn), 7.09
(6 H, m, Ph), 6.86 (6 H, m, Ph), 5.32 (2 H, br s, BnCH2), 5.12
(3 H, br s, -OH), 4.6–2.2 (31 H, br m, CH, CH2), 2.26 (9 H, s,
CH3); dC(CDCl3) 156.9 (2 C, Ph, ipso), 156.6 (1 C, Ph, ipso),
155.9 (1 C, CQO), 136.4 (1 C, Bn, ipso), 131.7 (3 C, Ph), 131.6
(6 C, Ph), 128.8 (2 C, Bn), 128.2 (3 C, Bn), 115.6 (6 C, Ph), 70.1
(1 C, CH2O), 69.6 (2 C, CH2O), 67.0 (1 C, BnCH2-), 66.2 (1 C,
CHOH), 65.9 (2 C, CHOH), 60.9 (1 C, exo-CH2N), 59.9 (1 C,
exo-CH2N), 58.1 (1 C, exo-CH2N), 53.8 (2 C, cyclenCH2), 51.3
(2 C, cyclenCH2), 48.0 (2 C, cyclenCH2), 46.0 (2 C, cyclen-
CH2), 20.5 (3 C, CH3).
1-(2-(9-Anthrylmethylamino)ethyl)-4, 7,10-tris((2S)-2-hydroxy-
3-phenoxypropyl)-1,4,7,10-tetraazacyclododecanecadmium(II) di-
perchlorate dihydrate ([CdL1](ClO4)2ꢀ2H2O). [Note: Perchlorate
salts are potentially explosive. Although no problems were
encountered, precautions MUST be taken when handling these
substances.] A solution of cadmium(II) perchlorate hexahydrate
(0.117 g, 0.29 mmol) in EtOH (2.3 cm3) was added dropwise
over 5 min to a refluxing solution of 1-(2-(9-anthrylmethylami-
no)ethyl)-4,7,10-tris((2S)-2-hydroxy-3-phenoxypropyl)-1,4,7,10-
tetraazacyclododecane (0.214 g, 0.25 mmol) in EtOH (7 cm3).
A white precipitate formed instantly. The suspension was
maintained at reflux temperature for 1 h before cooling it to
room temperature and evaporating the solvent. Trituration of
the residue with diethyl ether produced a cream colored powder.
This was collected by filtration, washed with ice-cold
water (1 cm3), and dried under vacuum to give the pure
product. Yield: 0.203 g, 70%. (Found C, 51.9; H, 6.0; N, 5.6.
C52H69CdCl2N5O16 requires C, 51.9; H, 5.8; N, 5.9%);
1,4,7-Tris((2S)-2-hydroxy-3-[40-(methyl)phenoxy]propyl)-1,4,
7,10-tetraazacyclododecane (6, R = p-tolyl). Cyclohexene
(1.91g, 23.3 mmol) was added to a stirred solution of 1-
(benzyloxycarbonyl)-4, 7, 10-tris((2S)-2-hydroxy-3-[40-(methyl)-
phenoxy]propyl)-1,4,7,10-tetraazacyclododecane (3.33 g, 4.17
mmol) dissolved in absolute ethanol (65 cm3). 10% Palladium-
on-carbon catalyst (3.2 g) was then added. The reaction
mixture was refluxed at 80 1C for 5 h, and filtered through
Celite. The filtrate was concentrated in vacuo to give the free
base (1.66 g, 60%). dH(CDCl3) 7.35 (4 H, br s, PhH), 7.05
(4 H, d, J 7.2 Hz, PhH); 6.79 (4 H, d, J 7.2 Hz, PhH); 4.3–2.0
(35 H, br, m, OH, NH, CH, CH2); 2.27 (9 H, s, CH3);
dC(CDCl3) 156.6 (3 C, Ph, ipso), 130.0 (3 C, Ph), 129.5 (6 C,
Ph), 114.4 (6 C, Ph), 70.3 (2 C, CH2O), 70.0 (1 C, CH2O), 67.1
(1 C, CHOH), 66.0 (2 C, CHOH), 61.1 (2 C, exo-CH2N), 60.4
(1 C, exo-CH2N), 59.7 (1 C, cyclenCH2), 55.8 (1 C, cyclen-
CH2), 54.6 (1 C, cyclenCH2), 54.2 (1 C, cyclenCH2), 52.9 (1 C,
cyclenCH2), 49.6 (1 C, cyclenCH2), 47.8 (1 C, cyclenCH2), 46.0
(1 C, cyclenCH2), 20.4 (3 C, CH3).
l
max/nm (20% aqueous 1,4-dioxane) 387.7 (e/dm3 molꢁ1 cmꢁ1
6733), 367.7 (7202), 349.8 (4700), 333.3 (2619), 318.3 (sh)
(1190); dH(DMSO-d6) 8.79–8.20 (3 H, br, m, AnthH), 8.14 (2
H, br d, AnthH), 7.58 (4 H, br m, AnthH), 7.30 (6 H, br s, PhH),
6.95 (9 H, br s, PhH), 5.05 (3 H, br m, -OH), 4.60–2.00 (38 H, br
m, CH, CH2, NH); dC(CD3CN) 159.4 (3 C, Ph, ipso), 132.4
(1 C, Anth, ipso), 130.6 (6 C, Ph), 130.3 (2 C, Anth, ipso),
129.6 (2 C, Anth, ipso), 127.7 (2 C, Anth), 126.7 (1 C, Anth),
126.3 (2 C, Anth), 124.9 (2 C, Anth), 124.4 (2 C, Anth), 122.4
(1 C, Ph), 122.3 (2 C, Ph), 115.8 (2 C, Ph), 115.6 (4 C, Ph),
70.6 (1 C, OCH2), 70.2 (1 C, CH2O), 69.6 (1 C, CH2O),
66.1 (1 C, CHOH), 65.6 (1 C, CHOH), 64.8 (1 C, CHOH),
60.0 (1 C, exo-CH2N), 57.0 (1 C, CH2NH), 55.3 (2 C,
exo-CH2N), 54.7 (1 C, exo-CH2N), 53.3 (1 C, cyclenCH2),
52.7 (1 C, cyclenCH2), 51.2 (1 C, cyclenCH2), 51.0 (1 C,
cyclenCH2), 50.4 (1 C, cyclenCH2), 50.2 (1 C, cyclenCH2),
49.8 (1 C, cyclenCH2), 49.1 (1 C, cyclenCH2), 45.8 (1 C, Anth-
CH2).
1-(2-(9-Anthrylmethylamino)ethyl)-4, 7, 10-tris((2S)-2-hydroxy-
3-[40-(methyl) phenoxy]propyl)-1, 4, 7, 10-tetraazacyclododecane
(L2). 1,4,7-Tris((2S)-2-hydroxy-3-[40-(methyl)phenoxy]propyl)-
1,4,7,10-tetraazacyclododecane (0.87 g, 1.23 mmol) was
dissolved in dry MeCN (60 cm3). 9-(2-Bromoethylimino-
methyl)anthracene (0.39 g, 1.23 mmol) was added as a solid,
along with NaHCO3 (0.13 g) and molecular sieves. The
reaction was wrapped in foil to keep out light and refluxed
for 10 d before cooling it to room temperature, filtering it and
removing the solvent under reduced pressure. This gave the
intermediate imine as a red oil. The imine was redissolved in
EtOH (25 cm3) and NaBH4 (100 mg, 2.64 mmol) was added as
a solid. The reaction mixture was stirred overnight then
diluted with water (30 cm3) and extracted with CH2Cl2 (4 ꢃ
30 cm3). The combined organic layers were dried over
Na2SO4, then filtered and evaporated to yield the crude
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1502 | New J. Chem., 2008, 32, 1500–1508

View Article Online
1
Binding constant measurements using H NMR
product as a red oil (0.74 g, 67%). Purification was achieved
by formation of the pentahydrobromide salt as described in
the ESI.w The free ligand was recovered by dissolving the salt
in water–ethanol 1 : 1 and basifying to pH 13 with 1 M NaOH.
Extraction with CH2Cl2, drying with Na2SO4, filtration and
evaporation of the solvent under reduced pressure gave the
pure product as a reddish brown oil (quantitative); lmax/nm
(20% aqueous 1,4-dioxane) 386.4 (e/dm3 molꢁ1 cmꢁ1 6226),
366.5 (6673), 348.8 (4345), 332.6 (2141), 320.1 (sh) (920);
dH(CDCl3) 8.44 (1 H, s, AnthH), 8.02 (2 H, d, J = 8.0 Hz,
AnthH), 7.46 (2 H, d, J = 9.0 Hz, AnthH), 7.33 (4 H, m,
AnthH), 7.07 (6 H, m, PhH), 6.80 (6 H, br s, PhH), 5.3–2.5
(41 H, br m, NH, OH, CH and CH2), 2.23 (9 H, s, CH3);
dC(CDCl3) 156.5 (1 C, Ph, ipso), 156.3 (2 C, Ph, ipso), 131.4
(2 C, Anth), 130.0 (6 C, Ph), 129.2 (3 C, Ph), 129.0 (1 C, Anth),
128.7 (2 C, Anth), 128.4 (1 C, Anth), 128.15 (2 C, Anth),
128.05 (2 C, Anth), 126.3 (2 C, Anth), 125.3 (2 C, Anth), 115.3
(2 C, Ph), 114.3 (4 C, Ph), 70.3 (2 C, CH2O), 69.2 (1 C, CH2O),
66.4 (1 C, CHOH), 65.6 (2 C, CHOH), 62.9 (1 C, exo-CH2N),
60.4 (2 C, exo-CH2N), 59.7 (1 C, exo-CH2N), 57.2 (1 C,
CH2NH), 53.0–50.0 (8 C, m, cyclenCH2), 45.5 (1 C, Anth-
CH2), 20.4 (3 C, CH3).
Binding constants for host–guest associations were determined
by monitoring changes in the chemical shift of a convenient
1
guest H NMR resonance whilst known aliquots of the host
complex, in DMSO-d6, were added to 1 ꢃ 10ꢁ3 mol dmꢁ3
samples of the sodium salt of the guest in DMSO-d6. A series
1
of H NMR spectra were recorded at 294 ꢄ 0.5 K for each
host–guest combination from a host : guest ratio of 0 : 1 to
10 : 1. Approximately fifteen different host : guest ratios were
used to construct each titration curve which was then fitted
using a non-linear curve fitting procedure to yield the value of
the binding constant.
Binding constant measurements using fluorescence emission
In a similar way to the 1H NMR method, titration of the host
complex with the sodium salt of the various different guest
anions, during which the fluorescence emission intensity of the
host at 416 nm was monitored, gave a titration curve that was
analysed by non-linear least squares regression to yield the
host–guest binding constants and the molar fluorescence
(e0HG) of the host–guest complexes. These titrations were
conducted in 20% aqueous 1,4-dioxane (I = 0.1 M Et4NClO4)
at pH 7.00 (0.02 M lutidine). The fluorescence emission spectra
at 298 ꢄ 0.1 K were recorded on a Varian Cary Eclipse
fluorescence spectrophotometer, using stoppered 10 mm
quartz cells, over a wavelength range of 370–550 nm at
0.15 nm intervals, with a scan rate of 40 nm minꢁ1. Both the
excitation and emission monochromator slit widths were set at
5 nm, and due to the highly fluorescent nature of the com-
pounds, a 1.5 absorbance attenuator was used for 10ꢁ3 and
10ꢁ4 mol dmꢁ3 solutions, while 10ꢁ6 mol dmꢁ3 solutions
required no attenuation. All samples were purged with N2
prior to use. Quantum yields were calculated with respect to
quinine in 0.5 mol dmꢁ3 H2SO4 according to the method of
Demas and Crosby,10 and are corrected for solvent refractive
index.
1-(2-(9-Anthrylmethylamino)ethyl)-4, 7, 10-tris((2S)-2-hydroxy-
3-[40- (methyl) phenoxy]propyl)- 1, 4, 7, 10- tetraazacyclododecane-
cadmium(II) diperchlorate tetrahydrate ([CdL2](ClO4)2ꢀ4H2O).
[Note: Perchlorate salts are potentially explosive. Although no
problems were encountered, precautions MUST be taken
when handling these substances.] A solution of cadmium(II)
perchlorate hexahydrate (0.183 g, 0.43 mmol) in EtOH (3.5
cm3) was added dropwise over 5 min to a refluxing solution of
1-(2-(9-anthrylmethylamino)ethyl)-4,7,10-tris((2S)-2-hydroxy-
3-[40- (methyl)phenoxy]propyl)- 1, 4, 7, 10-tetraazacyclodode-
cane (0.302 g, 0.39 mmol) in EtOH (10 cm3). A white
precipitate formed instantly. The suspension was maintained
at reflux temperature for 1 h before cooling it to room
temperature and evaporating the solvent. Trituration of the
residue with diethyl ether produced a cream colored powder.
This was collected by filtration, washed with ice-cold water
(1 cm3), and dried under vacuum to give the pure product.
Yield: 0.261 g, 50%. (Found: C, 51.9; H, 6.5; N, 5.6.
C55H79CdCl2N5O18 requires C, 51.6; H, 6.2; N, 5.5%);
Results and discussion
Synthesis of the receptor ligands and complexes
The receptor ligands L1 (R = phenyl) and L2 (R = p-tolyl),
which are analogues of 1 where one of the pendant arms is the
fluorophore 2-(9-anthrylmethylamino)ethyl, were synthesised
in three steps from the known compound 1-(benzyloxycarbo-
nyl)cyclen, 4,9 using the procedure shown in Scheme 1. The use
of an enantiomerically pure epoxide in the first step, which
proceeds quantitatively, is necessary to ensure that only the
homochiral diastereomer of the monoprotected product, 5, is
formed. For removal of the benzyloxycarbonyl protecting
group, in the second step, both the catalytic transfer hydro-
genation method,11 using cyclohexene with palladium-on-
charcoal catalyst, and acid hydrolysis, using HBr in acetic
acid,12 were investigated. The former method proved to be the
superior of the two, giving the pure product, 6, directly and in
near quantitative yield. The final step proceeds in ca. 40%
yield, over 10 days in acetonitrile, during which time precau-
tions must be taken to avoid photodegradation of the anthra-
cene sub-unit.13
l
max/nm (20% aqueous 1,4-dioxane) 386.4 (e/dm3 molꢁ1 cmꢁ1
6464), 366.4 (6960), 348.6 (4564), 332.6 (2274), 319.5 (sh)
(1008); dH(DMSO-d6) 8.78–8.30 (3 H, br m, AnthH), 8.10 (2
H, br d, AnthH), 7.53 (4 H, br m, AnthH), 7.26 (6 H, br s,
PhH), 6.90 (6 H, br s, PhH), 5.00 (3 H, br m, -OH), 4.70–2.10
(38 H, br, m, NH, CH and CH2), 2.00 (9 H, CH3); dC(DMSO-
d6) 156.4 (3 C, Ph, ipso), 131.4 (2 C, Anth, ipso), 130.7 (2 C,
Anth, ipso), 130.0 (6 C, Ph), 129.5 (3 C, Ph, ipso), 129.2
(1 C, Anth, ipso), 127.6 (1 C, Anth), 127.2 (2 C, Anth),
125.9 (2 C, Anth), 125.2 (2 C, Anth), 121.8 (2 C, Anth),
115.3 (6 C, Ph), 71.3 (3 C, CH2O), 68.0 (1 C, CHOH), 67.0
(1 C, CHOH), 66.0 (1 C, CHOH), 61.8 (1 C, exo-CH2N), 59.8
(1 C, CH2NH), 58.3 (1 C, exo-CH2N), 57.0 (1 C, exo-CH2N),
56.0 (1 C, exo-CH2N), 53.9 (2 C, cyclenCH2), 54.0 (2 C,
cyclenCH2), 52.8 (2 C, cyclenCH2), 51.3 (2 C, cyclenCH2),
45.9 (1 C, AnthCH2), 20.2 (3 C, CH3).
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1500–1508 | 1503

View Article Online
Table 1 Binding constantsa in DMSO-d6, expressed as log(K/Mꢁ1),
for the inclusion of guest anions within the receptor complexes,
1
determined by H NMR monitored titrations
Receptor complex
[Cd(1)]2+b
[CdL1]2+c
[CdL2]2+c
Guest anion
p-Nitrophenolate
p-Formylphenolate
p-Nitrobenzoate
Phenoxyacetate
(D)-Histidinated
(L)-Histidinated
a
4.2 ꢄ 0.1
44.5c
3.5 ꢄ 0.1
44.5
4.4 ꢄ 0.1
4.8 ꢄ 0.1
4.5 ꢄ 0.2
44.5
3.3 ꢄ 0.1
44.5
4.0 ꢄ 0.1
4.4 ꢄ 0.1
3.6 ꢄ 0.1
4.1 ꢄ 0.2
4.5 ꢄ 0.4
45
4.2 ꢄ 0.2
4.2 ꢄ 0.4
Measured at 294 ꢄ 0.5 K with [guest anion] = 10ꢁ3 M. Uncertainties
b
c
d
are one SD. Taken from ref. 1c. This work. At 313 K in 10% (v/v)
D2O in DMSO-d6.
Scheme 1
pH Dependence of the UV absorption and fluorescence emission
of L1 and L2
Cd(II) complexes of L1 and L2 were readily formed as their
diperchlorate salts by the addition of cadmium(II) perchlorate
to a solution of the ligand in hot ethanol and were charac-
terised by NMR spectroscopy, microanalysis and their fluor-
escence behaviour. The structure of these complexes is
expected to be similar to many other Cd(II) complexes which
have been formed from cyclen derived ligands that have four
appended 2-hydroxyethyl, carboxymethyl, or carbomoyl-
methyl moieties where the structures have been determined
using X-ray diffraction and found to be as shown in Fig.
1(a).1b In the case of L1 and L2 the fluorescence data that
follow confirm that the fluorescent arm is coordinated through
the anthrylamine under the conditions used for anion inclu-
sion experiments. Thus, it seemed reasonable to initiate anion
inclusion experiments on the basis that the anthryl moiety has
the potential to juxtapose with the other three aromatic
moieties to form a cavity suitable for anion retention, but this
first needed to be verified.
Since the fluorescence emission intensity of the 2-(9-anthrylmethyl-
amino)ethyl moiety is known to be modulated by the extent of
protonation of the anthrylamine, through the PeT effect,4 it was
important at the outset of the fluorescence investigations to
establish the pH dependence of both the UV and fluorescence
emission spectra of L1 and L2. This and all subsequent work was
performed in 20% aqueous 1,4-dioxane since the ligands and
complexes under discussion were insufficiently soluble in water. As
the pH of a 10ꢁ4 M solution of the respective ligand was increased
from 1.7 to 10.1 the three major UV spectral bands shown by L1
(Fig. 2) and L2 underwent a hypsochromic shift of 0.4 nm and the
minor band a bathochromic shift of 0.4 nm. This behaviour
produced an isosbestic point at 350 nm which would subsequently
be used as the excitation wavelength for the fluorescence emission
experiments. The intensity of the four well defined UV bands
increased by up to ca. 10% as the pH was increased, which is
consistent with an increase in electron delocalisation brought
about by deprotonation of the anthrylamine.15
The intensity of the fluorescence emission spectra of L1
(Fig. 3) and L2 initially increased slightly as the pH was raised
Study of the anion binding ability of [CdL1]2+ and [CdL2]2+
1
compared to [Cd(1)]2+, using H NMR monitoring
We set out to verify that [CdL1]2+ and [CdL2]2+ show similar
anion including behaviour that seen with [Cd(1)]2+ by using
1
them in the same type of H NMR monitored titrations, with
the same set of aromatic anions, as we used previously to
measure anion binding constants with [Cd(1)]2+ in DMSO-
d6.1c In this way any commonality in the binding constant
values, subject to known differences in the structure of the
receptor, could be taken as evidence of similar anion retaining
behaviour. The binding constants obtained are shown in
Table 1. Those recorded for [CdL1]2+ and [CdL2]2+ were
found to be of similar, but smaller, magnitude to those
recorded with [Cd(1)]2+, which is consistent with the new
anion receptors presenting three convergent O–H groups plus
one weaker N–H hydrogen bond donor group,14 compared to
[Cd(1)]2+, which presents four convergent O–H groups. These
data were taken as confirmation of the anion binding proper-
ties previously discovered with [Cd(1)]2+ being substantially
Fig. 2 Changes in the UV spectrum of protonated L1, 10ꢁ4 mol dmꢁ3
in 20% aqueous 1,4-dioxane (I = 0.1 mol dmꢁ3, Et4NClO4), as the pH
is raised from 1.7 to 10.1 by the addition of aliquots of NEt4OH, at
298 K. The pH interval between spectra is ca. 0.5 pH units.
retained by [CdL1]2+ and [CdL2]2+
.
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1504 | New J. Chem., 2008, 32, 1500–1508

View Article Online
Fig. 3 (a) Changes in the fluorescence emission spectra (lex = 350 nm) of protonated L1, 10ꢁ6 mol dmꢁ3 in 20% aqueous 1,4-dioxane (I =
0.1 mol dmꢁ3, Et4NClO4) during titration with Et4NOH at 298 K. Emission maxima are at 398.1, 418.8 and 442.6 nm at pH 3 and 396.0, 416.7 and
441.2 nm at pH 10. Maximum fluorescence emission intensity decreases from pH 3.0 to pH 12.0; (b) fluorescence emission intensity and quantum
yields (F) of L1 at lmax plotted against pH, values derived from (a).
from pH 1.7 to 3.0 using Et4NOH. This behaviour with
anthracene derivatives has been noted previously by Czarnik
et al. and attributed to deprotonation of the aromatic ring.2a
Beyond pH 3 and up to the limit of the titration at pH 12 the
fluorescence decreased significantly. The sigmoidal shape of
this region of the associated titration curve (Fig. 3(b)) clearly
indicated a connection between the fluorescence decrease and
a deprotonation occurring with a pKa of 8.4 ꢄ 0.4 in L1 and
8.2 ꢄ 0.4 in L2 in a similar way to that which was seen with
ligand 2, where the corresponding pKa was measured as 7.2 in
water.4 In that work 1H NMR experiments were used to
associate the deprotonation responsible for the fluorescence
decrease with the anthrylamine and hence to establish PeT
originating from the anthrylamine lone pair electrons as the
reason for the fluorescence quenching. The same explanation
appears to hold for L1 and L2 since the measured pKa values in
20% aqueous dioxane are slightly higher than the value
obtained in water, as would be expected in a solvent of lower
dielectric constant, and quite similar to the value of 8.91
measured for propyl(9-anthrylmethyl)amine by Fabbrizzi
et al. in 20% aqueous acetonitrile.5 The quantum yield values
recorded at pH 12 for L1 and L2 are 0.117 and 0.047,
respectively. Both of these values are in excess of the value
of 0.02 recorded for 2 in water by Kimura et al.4 and of 0.028
measured by ourselves for 2 in 20% aqueous 1,4-dioxane. The
most likely explanation for the higher baseline fluorescence of
L1 and L2 lies with the presence of the hydroxy group on the
pendant arms adjacent to the anthrylamine. If one or more of
these should act as a hydrogen bond donor towards it, the
result would be a withdrawal of lone pair electron density from
the anthrylamine thereby diminishing the effectiveness of the
PeT quenching.3
eight-coordinating metal ions, such as Pb(II) and Hg(II), or
the potentially six-coordinating Zn(II) ion.1 It was necessary,
therefore, to characterise fully the fluorescence emission beha-
viour of [CdL1]2+ and [CdL2]2+ before investigations of its
perturbation upon anion uptake could begin. For comparative
purposes the fluorescence emission of the Pb(II), Hg(II) and
Zn(II) complexes of L1 was also investigated.
Titration of a micromolar solution of L1 or L2 in 20%
aqueous dioxane at pH 7.0 (1 mM HEPES, I = 0.1 M
(Et4NClO4)) with Cd(ClO4)2 produced only very slight increases
to the molar absorptivity of the UV spectral bands, but very
significant increases in the intensity of the fluorescence emission
bands. This is shown for L1 in Fig. 4. With both L1 and L2 these
increases terminate once one equivalent of Cd(II) has been
added. Well established stability constant data leave no doubt
that the cyclen component of each ligand molecule will be fully
bound to the metal ion, under these conditions, once one
equivalent of Cd(II) has been added;16 the fact that the
The effect of metal ion complexation on the fluorescence of L1
and L2
Metal complexes of L1 and L2 formed with Cd(II) ion were of
principal interest in this study because of the observations
made in earlier work that had demonstrated their superiority
in anion sequestration over those of other potentially
Fig. 4 The relative fluorescence emission intensity at lmax for L1,
10ꢁ6 mol dmꢁ3 in 20% aqueous 1,4-dioxane, I = 0.1 mol dmꢁ3
(Et4NClO4), at pH 7.0 (0.01 mol dmꢁ3 HEPES) as it is titrated with
Cd(ClO4)2ꢀ6H2O.
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1500–1508 | 1505

View Article Online
Table 2 Binding constants determined from fluorescence intensity monitored titrations,a expressed as log (K/Mꢁ1), and fluorescence intensity
perturbations at 416 nm (DI)b for the inclusion of guest anions within the receptor complexes [CdL1]2+ and [CdL2]2+
Receptor complex
[CdL1]2+
[CdL2]2+
log K
Guest anion
log K
DI
117%
DI
47%
Benzoate
2.3 ꢄ 0.1d
4.5 ꢄ 0.1
5.3 ꢄ 0.2c
7.1 ꢄ 0.2c
6.1 ꢄ 0.1c
7.5 ꢄ 0.4c
7.1 ꢄ 0.2c
4.9 ꢄ 0.2c
6.5 ꢄ 0.1c
4.1 ꢄ 0.4
5.5 ꢄ 0.1c
—
p-Hydroxybenzoate
m-Hydroxybenzoate
o-Hydroxybenzoate
3,5-Dihydroxybenzoate
2,6-Dihydroxybenzoate
Gallate (3,4,5-trihydroxybenzoate)
p-Nitrobenzoate
p-Aminobenzoate
p-Dimethylaminobenzoate
Phenoxyacetate
(D)-Histidinate
(L)-Histidinate
p-Toluenesulfonate
Benzenesulfonate
a
27%
21%
20%
ꢁ8%
ꢁ4%
10%
ꢁ13%
59%
ꢁ30%
2%
3.0 ꢄ 0.1
4.7 ꢄ 0.4
4.3 ꢄ 0.4c
—
—
—
ꢁ6%
12%
0%
0%
0%
0%
8%
8%
39%
6.9 ꢄ 0.6c
4.6 ꢄ 0.3
3.4 ꢄ 0.1
—
0%
ꢁ11%
4.6 ꢄ 0.2
3.2 ꢄ 0.3
20%
[Receptor complex] = 10ꢁ4 mol dmꢁ3, measured at pH 7.0 (0.02 mol dmꢁ3 lutidine) in 20% aqueous 1,4-dioxane, at 298 K, I = 0.1 mol dmꢁ3
b
(Et4NClO4). Uncertainties are taken as one SD. Values calculated as (IHG ꢁ IH)/IH) ꢃ 100 where IHG is calculated together with K from non-
c
d
linear least squares regression analysis of the titration curve. [Receptor complex] = 10ꢁ6 mol dmꢁ3
.
[Receptor complex] = 10ꢁ3 mol dmꢁ3
.
fluorescence enhancement also terminates at this point indicates
that the PeT induced quenching is fully eliminated and therefore
that the anthrylamine component of each ligand molecule is also
strongly bound. The enhanced fluorescence corresponds to a
45% gain in quantum yield for L1 (from 0.471 to 0.684 for
[CdL1]2+, both at pH 7.0), and to a 34% gain for L2 (from 0.406
to 0.546 for [CdL2]2+, both at pH 7.0). Similar titrations using
Pb(II), Hg(II), and Zn(II) with L1 also demonstrated termination
of the fluorescence change once the metal ion concentration
became equimolar with that of the ligand; the quantum yields,
with changes compared to L1 at pH 7.0 given in brackets, for the
fully formed complexes were found to be 0.518 (+10%), 0.358
(ꢁ24%) and 0.518 (+10%), respectively. It is interesting to note
that L1 shows a reversal in the relative magnitude of the
fluorescence enhancement induced by Zn(II) and Cd(II) com-
pared to that which was seen with ligand 2,4 which gave a higher
fluorescence with Zn(II). This is likely to be due to the inability of
Zn(II) to exceed a coordination number of six with the types of
ligand under consideration here,1d,17 which prevents it from
binding the four pendant arms of L1 simultaneously. On a
time-averaged basis, therefore, the anthrylamine is likely to be
partially non-coordinated in the Zn(II) complex of L1, allowing
some PeT quenching to occur, but fully coordinated in the five-
coordinate Zn(II) complex of 2, where there is no competition
from other chelating pendant donor atoms. Cd(II) on the other
hand will form eight-coordinate complexes with L1 thereby fully
coordinating the anthrylamine. A similar, but less marked, effect
was seen with another cyclen derivative carrying four pendant
arms one of which was the dansyl fluorophore.18
(Et4NClO4)), were titrated with the aromatic anions shown in
Table 2 perturbations of the fluorescence emission intensity at
416 nm were seen. The maximum extent of these for each anion
(DI) is shown in Table 2. In most cases there was also a change in
the molar absorbance of the UV band at lmax (350 nm) ranging
from ꢁ7% (for (L)-histidinate with [CdL2]2+) to +52% (for
benzoate with [CdL1]2+). Since blank titrations using the un-
complexed receptor ligands L1 and L2 failed to produce any
perceptible changes at the monitored absorption and emission
wavelengths under these conditions, both UV absorption and
fluorescence emission perturbations induced by a potential guest
anion on [CdL1]2+ and [CdL2]2+ were taken as indicators of its
inclusion within the binding cavity and thus the resulting
titration curves could be used to determine the guest binding
constants. However, it should be noted that since the fluores-
cence and UV perturbations (as well as 1H NMR perturbations)
embrace both positive and negative displacements of the mon-
itored signal, the observation of zero perturbation does not
necessarily exclude the possibility of anion inclusion.
Recording the UV spectral perturbations was necessary to
verify that there were no inner filter effects affecting the
fluorescence spectra and also to acquire the absorption data
necessary for the computation of the quantum yields. Also it
enabled measurements of the guest binding constants to be
made that are supplementary to those reported in Table 2.
These are reported in the ESI,w but show no significant
variance from those obtained using fluorescence data.
The fluorescence emission intensity of [CdL1]2+ and
[CdL2]2+ is sufficiently high that perturbations to it can easily
be recorded at 10ꢁ6 M, compared to 10ꢁ4 M for UV and
10ꢁ3 M for 1H NMR perturbations. Because of the greater
level of host–guest dissociation at the lower concentration this
widens the range of anion binding constants that can be
measured to an upper limit of ca. 107.5, compared to 104.5 in
Fluorescence emission and UV absorbance perturbation in
[CdL1]2+ and [CdL2]2+ due to anion sequestration
When [CdL1]2+ and [CdL2]2+ at 10ꢁ4 or 10ꢁ6 M in 20%
aqueous dioxane at pH 7.0 (0.02 M lutidine, I = 0.1 M
1
studies using H NMR monitoring. A typical titration curve
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1506 | New J. Chem., 2008, 32, 1500–1508

View Article Online
basic than benzoate; so there must be another reason. The clue
to this appears to lie with the observed fluorescence emission
intensity perturbations: The interaction of benzoate with
[CdL1]2+ brings about a 117% increase in its fluorescence
emission intensity. This most likely arises from the elimination
of quenching effects due to solvent water molecules initially
hydrogen bonded in the binding cavity (it was seen earlier that
the quantum yield for 2 is depressed in water and this appears
to be a general effect for PeT perturbed anthracene fluores-
cence19) together with some enhancement of PeT quenching
due to hydrogen bond acceptance by the benzoate carboxy
moiety from the anthrylamine, which will weaken the N–H
bond and release electron density for PeT. This 117%
enhancement diminishes to 27% when the p-hydroxybenzoate
is introduced. Since the water elimination effect is common to
all entering guest molecules the diminished enhancement must
be attributable to its higher basicity, which leads to a greater
weakening of the anthrylamine N–H bond and consequently
greater PeT quenching than occurred in response to benzoate
inclusion. With m- and o-hydroxybenzoate the level of fluor-
escence enhancement compared to [CdL1]2+ is similar to that
of p-hydroxybenzoate (21% and 20%, respectively) yet this
cannot be due to similarly strong hydrogen bond acceptance
from the anthrylamine since both are of lower basicity than
p-hydroxybenzoate. The stronger binding, that gives rise to
similar fluorescence enhancement, must, therefore, arise from
another cause associated with o- and m-positioning of the
hydroxy group, but not p-positioning. We suggest that this
cause may be O–Hꢀ ꢀ ꢀp hydrogen bonding to the aromatic
rings that form the walls of the cavity, including the anthra-
cene moiety.20 As shown in Fig. 6, which is based on our
knowledge of the crystal structure of p-aminobenzoate
Fig. 5 Molar fluorescence changes (e0) at 416 nm for 10ꢁ6 mol dmꢁ3
[CdL1](ClO4)2 in 20% aqueous 1,4-dioxane at pH 7.0 (0.02 mol dmꢁ3
lutidine, I = 0.1 M (Et4NClO4)), as it is titrated with sodium
p-hydroxybenzoate. Squares indicate the experimental data points
and the curve indicates the theoretical e0 values for the calculated
values of K and e0HG
.
monitored by fluorescence emission at a [receptor] = 10ꢁ6 M,
that of p-hydroxybenzoate with [CdL1]2+, is shown as Fig. 5.
The wider range of binding constants quantified through the
fluorescence emission studies allows a much better insight into
the manner of anion inclusion than was hitherto possible. The
first seven entries in Table 2 are particularly interesting. These
constitute a set of benzoate anions substituted by hydroxy
groups at one or more additional sites on the aromatic ring.
Acquisition of binding constant data up to 107.5 reveals that
whilst the binding of the unsubstituted benzoate anion is quite
weak (log K = 2.3), the strength of the binding can be
augmented, incrementally, by the presence of additional hy-
droxy groups in a manner dependent on their number and
positioning. Thus, a p-hydroxy group enhances log K to 4.5, a
m-hydroxy to 5.3, and an o-hydroxy to 7.1. The effect is
correspondingly heightened if two m- or o-hydroxy groups
are present. The enhanced stability of the p-hydroxybenzoate
interaction, compared to benzoate, is expected on the grounds
of its greater basicity (pKa = 4.58 compared to 4.20), which
causes it to hydrogen bond at the base of the cavity more
strongly. However, this argument cannot be applied to the
even greater binding strength of m-hydroxybenzoate (pKa =
4.08) or o-hydroxybenzoate (pKa = 2.98) since they are less
included in [Cd(1)]2+ 1c
and of salicylate hydrogen bonded
,
to the analogue of [Cd(1)]2+ that has methyl moieties in place
of the phenoxymethylene groups,21 the O–H group is only
geometrically positioned to do this when located ortho or, to a
lesser extent, meta to the carboxy group. Such an interaction
would not only strengthen the binding of the anion, but would
also be likely to quench the fluorescence of the anthracene
either through de-excitation of the anthracene through energy
transfer to the O–H oscillator,22 if the oscillator is directed
towards it rather than toward one of the other three aromatic
Fig. 6 Showing the potential for –O–Hꢀ ꢀ ꢀp hydrogen bonding as the positioning of the hydroxy substituent on the guest molecule changes. The
ellipse in each case is the approximate locus of points accessible to the hydroxy H atom in response to rotation around the O–H bond. Only in
structure (c) is the hydroxy group constantly in range for –O–Hꢀ ꢀ ꢀp hydrogen bonding to be possible.
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1500–1508 | 1507

View Article Online
groups, or alternatively by through space PeT if p-stacking
with the anthracene occurs.3 In none of the five X-ray deter-
mined structures of this class of host–guest complex that exist
has p-stacking between a guest and an aromatic ring been
seen.1a,c,d In favor of the former suggestion we note that the
binding of 3,5-dihydroxybenzoate (pKa = 4.04) and 2,6-
dihydroxybenzoate (pKa = 1.05) is stronger than for their
mono-substituted counterparts, even though the basicity is still
diminishing, and that the fluorescence is quenched to a greater
extent, which is consistent with two appropriately positioned
O–H groups doubling the probability of the anthracene being
involved in an O–Hꢀ ꢀ ꢀp hydrogen bond whilst the probability
of p-stacking with it remains unchanged. The binding of 3,4,5-
trihydroxybenzoate (gallate, pKa = 4.41) on the grounds of
basicity should be stronger than that of 3,5-dihydroxybenzo-
ate, which it is, but its fluorescence is not diminished to the
same extent, this may in some way be related to intramolecular
hydrogen bonding between the p-hydroxy group and one or
other of the m-hydroxy groups, which diminishes the level of
O–Hꢀ ꢀ ꢀp hydrogen bonding. Intramolecular O–Hꢀ ꢀ ꢀp hydro-
gen bonding between two gallate moieties in p-tert-butyl-
calix[4]arene-1,3-digallate has been observed where the hydro-
xy oxygen atom of one gallate sits 3.247 A above the facing
gallate.23 In this work preliminary ab initio molecular model-
ling of the structure in Fig. 6(c) using B3LYP/LanL2MB
calculations suggests that the hydrogen atom in the O–Hꢀ ꢀ ꢀp
hydrogen bond proposed for that structure is positioned
4.73 A above the nearest carbon atom of the aromatic ring.24
This initially appeared somewhat distant for a strong interac-
tion, but a referee has pointed out that since our binding
constant and fluorescence measurements are conducted in
partially aqueous solution it is quite possible that a bridging
water molecule could interpose between the hydroxy group
and the ring, and that the observed phenomena could be
accounted for in that way.
38, 4986; (b) K. P. Wainwright, Adv. Inorg. Chem., 2001, 52, 293;
(c) C. B. Smith, A. K. W. Stephens, K. S. Wallwork, S. F. Lincoln,
M. R. Taylor and K. P. Wainwright, Inorg. Chem., 2002, 41, 1093;
(d) T. S. Robinson, O. Wyness, S. F. Lincoln, M. R. Taylor, E. R.
T. Tiekink and K. P. Wainwright, Inorg. Chim. Acta, 2006, 359,
1413.
2 See for example: (a) E. U. Akkaya, M. E. Huston and A. W.
Czarnik, J. Am. Chem. Soc., 1990, 112, 3590; (b) L. Fabbrizzi and
A. Poggi, Chem. Soc. Rev., 1995, 24, 197; (c) L. Fabbrizzi, M.
Licchelli, G. Rabaioli and A. Taglietti, Coord. Chem. Rev., 2000,
205, 85; (d) P. D. Beer and E. J. Hayes, Coord. Chem. Rev., 2003,
240, 167; (e) P. V. Bernhardt and E. G. Moore, Aust. J. Chem.,
2003, 56, 329.
3 (a) R. A. Bissell, A. P. de Silva, H. Q. N. Gunaratne, P. L. M. Lynch,
G. E. M. Maguire, C. P. McCoy and K. R. A. S. Sandanayake, Top.
Curr. Chem., 1993, 168, 223; (b) A. P. de Silva, H. Q. N. Gunaratne,
T. Gunnlaugsson, A. J. M. Huxley, C. P. McCoy, J. T. Rademacher
and T. E. Rice, Chem. Rev., 1997, 97, 1515; (c) A. P. de Silva, B.
McCaughan, B. O. F. McKinney and M. Querol, Dalton Trans.,
2003, 1902; (d) G. De Santis, L. Fabbrizzi, M. Licchelli, A. Poggi and
A. Taglietti, Angew. Chem., Int. Ed. Engl., 1996, 35, 202; (e) L.
Fabbrizzi, M. Licchelli, A. Perotti, A. Poggi, G. Rabaioli, S.
Donatella and A. Taglietti, J. Chem. Soc., Perkin Trans. 2, 2001,
2108; (f) L. Bruseghini, L. Fabbrizzi, M. Licchelli and A. Taglietti,
Chem. Commun., 2002, 1348.
4 S. Aoki, S. Kaido, H. Fujioka and E. Kimura, Inorg. Chem., 2003,
42, 1023.
5 L. Fabbrizzi, M. Licchelli, P. Piersandro and L. Parodi, Angew.
Chem., Int. Ed., 1998, 37, 800.
6 D. D. Perrin, W. L. F. Armarego and D. R. Perrin, Purification
of Laboratory Chemicals, Pergamon, Oxford, 3rd edn, 1988.
7 W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43, 2923.
8 A. Damsyik, S. F. Lincoln and K. P. Wainwright, Inorg. Chem.,
2006, 45, 9834.
9 J. Yoo, D. E. Reichert and M. J. Welch, Chem. Commun., 2003,
766.
10 J. N. Demas and G. A. Crosby, J. Phys. Chem., 1971, 75, 991.
11 C. H. Manning, A. Skierawska, J. N. Marx and D. J. Bornhop,
Synth. Commun., 2003, 33, 457.
12 (a) W. F. Gilmore and H. A. McBride, J.Pharm. Sci., 1974, 63,
1087; (b) C. Yuan, G. Wang and S. Chen, Synthesis, 1990, 6
522.
13 R. Dabestani, K. J. Ellis and M. E. Sigman, J. Photochem.
Photobiol., A: Chem., 1995, 86, 231.
14 G. A. Jeffrey, An Introduction to Hydrogen Bonding, Oxford
University Press, Oxford, 1997.
15 A. G. Martinez, J. O. Barcina, A. deFresno Gerezo, A.-D. Schluter
and J. Frahn, Adv. Mater., 1999, 11, 27.
16 A. Bianchi, M. Micheloni and P. Paoletti, Coord. Chem. Rev.,
1991, 110, 17.
17 H. Maumela, R. D. Hancock, L. Carlton, J. H. Reibenspies and K.
P. Wainwright, J. Am. Chem. Soc., 1995, 117, 6698.
18 S. Aoki, H. Kawatani, T. Goto, E. Kimura and M. Shiro, J. Am.
Chem. Soc., 2001, 123, 1123.
The remaining binding constants reported in Table 2 are
broadly in line with those determined by the 1H NMR method
and reported either in Table 1 or elsewhere.1c The fluorescence
diminishments seen with aromatic anions that have electron
withdrawing or donating groups in the para position:
p-nitrobenzoate, p-dimethylaminobenzoate and p-toluene-
sulfonate, compared to their non-substituted parent anion,
follow the pattern seen by Fabbrizzi et al. and are believed to
be associated with through space PeT processes involving the
anthracene moiety.3d
19 (a) H. Ji, R. Dabestani, G. M. Brown and R. L. Hettich, J. Chem.
Soc., Perkin Trans. 2, 2001, 585; (b) H. Xu, X. Xu, R. Dabestani,
G. M. Brown, L. Fan, S. Patton and H. Ji, J. Chem. Soc., Perkin
Trans. 2, 2002, 636.
20 (a) J. Atwood, F. Hamada, K. D. Robinson, G. W. Orr and R. L.
Vincent, Nature, 1991, 349, 683; (b) J. F. Malone, C. M. Murray,
M. H. Charlton, R. Docherty and A. J. Lavery, J. Chem.
Soc., Faraday Trans., 1997, 93, 3429; (c) M. Nishio, M. Hirota
and Y. Umezawa, The CH/p Interaction, Wiley-VCH, New York,
1998.
Acknowledgements
We are grateful to Prof. S. Salehzadeh of Bu-Ali-Sina Uni-
versity, Hamadan, Iran for undertaking the preliminary
ab initio calculations. Funding of this work by the Australian
Research Council is gratefully acknowledged.
21 C. B. Smith, S. F. Lincoln, M. R. Taylor and K. P. Wainwright,
Acta Crystallogr., Sect. E, 2002, 58, m33.
22 J. F. Callan, A. P. de Silva and N. D. McClenaghan, Chem.
Commun., 2004, 2048.
23 E. Nomura, A. Hosada and H. Taniguchi, J. Org. Chem., 2001, 66,
8030.
24 S. Salehzadeh and K. P. Wainwright, unpublished calculations.
References
1 (a) C. B. Smith, K. S. Wallwork, J. M. Weeks, M. A. Buntine, S. F.
Lincoln, M. R. Taylor and K. P. Wainwright, Inorg. Chem., 1999,
ꢂc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1508 | New J. Chem., 2008, 32, 1500–1508