Towards an efficient microsystem for the Real-Time detection and quantification of mercury in water based on a specifically designed fluorogenic binary Task-Specific ionic liquid

Article abstract of DOI:10.1002/anie.200905037

Chemical Equation Presented Quick test for quicksilver: A microdevice for liquid-liquid extraction and sensing of mercury ions in water uses an ionic liquid containing a task-specific ionic fluorogenic salt (L1) as the extracting liquid phase. The off-on sensing response decreases the likelihood of false positives and allows detection of Hg²⁺ in aqueous solution down to 50 ppb.

Full text of DOI:10.1002/anie.200905037

Communications
DOI: 10.1002/anie.200905037
Analytical Microsystems
Towards an Efficient Microsystem for the Real-Time Detection and
Quantification of Mercury in Water Based on a Specifically Designed
Fluorogenic Binary Task-Specific Ionic Liquid**
Faꢀdjiba Loe-Mie, Gilles Marchand,* Jean Berthier, Nicolas Sarrut, Mathieu Pucheault,
Mireille Blanchard-Desce, Franꢁoise Vinet, and Michel Vaultier*
The control of water resources and the natural environment is
a major current societal challenge that involves various
partners at different levels (local government, water agencies,
water treatment stations, citizens, consumer protection
organizations, etc.). In this general context, facing the
ecological hazards originating from the presence of heavy-
metal ions released in the environment (including water
resources) is one of the important issues to address. Heavy-
metal ions such as mercury cause adverse environmental and
health problems such as bioaccumulation by numerous
organisms and severe physiological problems, including
neurological, neuromuscular, or nephritic disorders. Indeed,
mercury is considered the most toxic nonradioactive metal.
Among the different forms (including the toxic organic
derivative methylmercury), mercuric ions (Hg2+) are not
only toxic but also highly water-soluble, making them
bioavailable for humans and animals by ingestion of water.
The toxicity associated with the bioavailability of Hg2+ ions
entails an increasing number of analyses to be carried out
both on tap water and on water resources. Although
laboratory analyses such as inductively coupled plasma mass
spectrometry (ICP-MS) and atomic absorption or atomic
emission spectroscopy are appropriate for precise quantita-
tive measurements, they suffer from several drawbacks,
including high cost and duration. Moreover the “in-lab”
techniques cannot be readily used for real-time in situ
analysis, which is required to monitor transient events such
as release from thunderstorm overflow. Alternative selective
Hg2+ sensors have been reported based upon optical,[1]
electrochemical,[2] electrical,[3] or biological detection meth-
ods.[4] These sensors allow in situ mercury monitoring, thus
avoiding the use of laboratory analyses. However, most
sensors are limited with respect to sensitivity[5] or are not
adapted for real-time monitoring of transient phenomena
requiring high measurement repetition rates.
This situation clearly calls for the development of efficient
alternatives for in situ sensitive monitoring of water quality,
which is subject to several official norms.[6] To overcome this
limitation, we herein propose a high-performance micro-total
analysis system (mTAS), which is based on both the use of a
novel selective and sensitive molecular fluorescent sensor in
an ionic liquid phase and on liquid–liquid extraction, thus
allowing concentrating extraction of the heavy-metal cation
from a water phase. Several authors[7] have reported the
extraction of heavy metals with ionic liquids containing crown
ethers with very high Nernst coefficients. Furthermore, the
nonvolatility of ionic liquids enables their use in very low
volume in open systems.[8] One of the most recent develop-
ments in the field of ionic liquids is the use of task-specific
ionic liquids (TSILs) or task-specific onium salts (TSOSs) as
soluble supports, therefore expanding their potential appli-
cations far beyond those of conventional ionic liquids. TSILs
retain all of the advantages of ionic liquids and can be used for
liquid–liquid extraction[9] with important benefits. Further-
more, TSILs (or TSOSs) can be dissolved in room-temper-
ature ILs to give binary solutions combining all of the above-
mentioned benefits.
Liquid–liquid extraction microsystems[10] have been inten-
sively developed during the last 10 years to extract, concen-
trate, and detect target molecules present in low concentra-
tions in a liquid. Various solutions have been proposed to
stabilize the interfaces, such as coflowing immiscible liquids
separated by microridges, microporous membranes, or micro-
pillars. However, classical organic solvents with poor Nernst
coefficients were used in those microsystems, thus limiting
their efficiency. In contrast, ionic liquids provide unique
remarkable extraction abilities owing to their very high
Nernst coefficient[9] for metal-ion extraction. On the basis of
the above analysis, we herein report the implementation and
investigation of a novel and efficient microdevice for liquid–
liquid extraction and sensing of mercury ions in water. The
concept is based on an ionic liquid containing a TSOS, which
combines fluorogenic and chelating properties to yield
fluorescent-sensor ionic liquids (FSILs). Such immobilization
of the metal binding unit in a hydrophobic IL would greatly
[*] F. Loe-Mie, Dr. G. Marchand, Dr. J. Berthier, N. Sarrut, Dr. F. Vinet
Department of Technology for Biology and Health
CEA LETI-MINATEC, 17 rue des Martyrs, 38054 Grenoble (France)
Fax: (+33)4-3878-5787
E-mail: gilles.marchand@cea.fr
Dr. M. Pucheault, Dr. M. Blanchard-Desce, Dr. M. Vaultier
CNRS, Chimie et Photonique Molꢀculaire (CPM)
35042 Rennes (France)
and
Universitꢀ de Rennes 1, CPM
Campus de Beaulieu, Bꢁtiment 10 A, 35042 Rennes (France)
Fax: (+33)2-2323-6955
E-mail: michel.vaultier@univ-rennes1.fr
[**] We thank Dr. P. Dubois (University of Rennes1/CEA-LETI), M.
Berger, Dr. P. Peltie, M. Dupoy, and Dr. P. Joly (CEA-LETI) for helpful
discussions. This work was supported by the French Agence
Nationale pour la Recherche (Grant ANR-08-CP2D-11).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200905037.
424
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 424 –427

Angewandte
Chemie
decrease the risk of loss into the aqueous phase compared to
classical ligands.
We thus designed and prepared TSOS L1 (Scheme 1), in
which an 8-hydroxyquinoline benzoate complexing/sensing
unit for mercury cations is grafted onto an ionium salt. A very
Figure 1. Principle of the real-time monitoring of a liquid–liquid
extraction using a fluorogenic task-specific ionic liquid. The two
immiscible phases, containing Hg2+ ions and TSIL L1, respectively,
flow in parallel (separated by micropillars). The extraction induces
fluorescence, the intensity of which is directly correlated to the
generation of the complex [L1-Hg2+]. Progressively, the Hg2+ ions are
extracted by the TSIL L1 at the liquid–liquid interface, which becomes
fluorescent. At the end of experiment, the complex [L1-Hg2+] diffuses
throughout the channel, which becomes totally fluorescent.
with an emission band peaking at 380 nm and corresponding
low quantum yields (F0 = 0.002 in acetonitrile). Such low
quantum yields were observed earlier with 8-hydroxyquino-
line (8-HQ) benzoate derivatives in which the carbonyl
oxygen lone pair is brought into close proximity with the 8-
HQ fluorophore.[12]
Next, the ability of L1 to complex Hg2+ ions was checked
in different solvents. As shown in Figure 2a, the addition of
Scheme 1. Synthesis of the TSIL L1 (see the Supporting Information).
a) Trimethylamine, acetonitrile, 808C, 15 h; b) NaOH 1m, 808C, 15 h,
then HBr 2m, room temperature, 15 min; c) LiNTf2, water, room
temperature, 3 h; d) oxalyl chloride, anhydrous THF, 08C, 15 min, then
room temperature, 3 h; e) 8-hydroxyquinoline, freshly distilled triethyl-
amine, anhydrous THF, room temperature, 4 h.
hydrophobic counteranion was selected to confine the L1-
Hg2+ complex in the IL phase. The heavy-metal detection is
based both on its transfer from a flowing carrier aqueous
solution to a storage ionic liquid phase and on the onset of
fluorescence emission of L1 exclusively upon metal binding
(Figure 1). Such an “off–on” sensing response to metal-ion
binding decreases the likelihood of false positive results and is
expected to improve the detection limit owing to the very low
offset signal.
First, the absorption characteristics of L1 were investi-
gated in different solvents. The UV/Vis absorption spectra of
L1 are characterized by two maxima at 227 and 277 nm in
acetonitrile, corresponding to a p–p* transition and to a
lower-lying n–p* transition with intramolecular charge-trans-
fer character.[11] Extinction coefficients e227 and e277 were
determined in acetonitrile to be 44400 and 6700mÀ1 cmÀ1,
respectively. The two absorption maxima (labs) were found to
be almost identical in acetonitrile and in the different studied
hydrophobic ionic liquids ([bmim][NTf2], [bmp][NTf2], and
[tmba][NTf2]; bmim = 1-butyl-3-methylimidazolium, bmp =
butylmethylpyrrolidinium, tmba = N-trimethyl-N-butylam-
monium, NTf2 = bis(trifluoromethylsulfonyl)imide). The
fluorescence spectra of the free ligand L1 in acetonitrile and
in the different ionic liquids show very weak fluorescence,
Figure 2. a) Absorption (A) and fluorescence (F) emission spectra
(excitation at 313 nm) of L1 (10 mm) free and in the presence of
increasing concentrations of Hg(ClO4)2 in [bmp][NTf2]. b) Selectivity
tests for L1 in [bmp][NTf2] with Hg2+, CH3Hg+, Cu2+, Pb2+, Zn2+, Cd2+
,
Ni2+, Mg2+, Ca2+
Angew. Chem. Int. Ed. 2010, 49, 424 –427
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
425

Communications
increasing amounts of Hg(ClO4)2 to a solution of L1 in
crown. The two flows were maintained for the duration of
the experiment. Figure 3a shows typical fluorescence images
of the microsystem during the liquid–liquid extraction. The
[bmp][NTf2] induces a bathochromic shift of both the high-
energy absorption band (shifted to 238 nm) and of the low-
energy band (shifted to 313 nm) and the progressive rise of a
new intense emission band peaking at 448 nm. At saturation,
a maximum fluorescence enhancement factor (FEF) of 94 and
a fluorescence quantum yield of 0.41 in [bmp][NTf2] were
determined for the complex L1-Hg2+. These major changes
can be ascribed to the suppression of effective nonradiative
deactivation processes through a 3(n–p*) state by lowering the
energy of the emitting p--p* excited state with intermolecular
charge-transfer character, thus preventing efficient intersys-
tem crossing.[12]
The selectivity of L1 for Hg2+ cations was evaluated by
testing the response to other environmentally relevant metal
ions, which provided evidence of high selectivity towards
mercury. A maximum fluorescence signal 29-fold smaller than
for Hg2+ was obtained for Cu2+ (lem = 448 nm), while no
significant fluorescence was observed with Cd2+, Mg2+, Ni2+,
and Ca2+. The selectivity ranking was determined to be
Hg2+ @ Cu2+ > Pb2+ > Zn2+ > Cd2+ ꢀ Ni2+ ꢀ Mg2+ ꢀ Ca2+ (Fig-
ure 2b). However, Ca2+ and Mg2+ usually are present in
excess over mercury (at least 1000-fold) in water. To evaluate
the potential interference of these two species in relevant
conditions, L1 was tested with 2 ꢀ 10À3 m Ca2+ and Mg2+ ions.
No signal was observed, showing that these ions cannot fake
the presence of Hg2+. Furthermore, the addition of Ca2+ and
Mg2+ ions in 1000-fold excess over Hg2+ showed a decrease by
only 10% of the Hg2+-induced fluorescence signal obtained in
the absence of these two ions. Next, the response of L1 to
CH3Hg+, a highly toxic form[13] that is only weakly soluble in
water, was determined. A mercury-induced fluorescence
signal 12-fold smaller than that obtained for Hg2+ under
comparable conditions was obtained.
Figure 3. a) Fluorescence intensity monitored at different positions in
the microsystem during a liquid–liquid extraction operation (after one
hour). The aqueous solution contained 500 ppb Hg(ClO4)2, and the
ionic liquid [bmp][NTf2] contained 10À4 m L1. lexc =310 nm and
lem =450 nm; b) Fluorescence intensity (F) of L1-Hg2+ at the micro-
device outlet in the ionic liquid phase as a function of time. texp =1 s,
lexc =310 nm, lem =450 nm. The first measurement (t=1 min) gave a
relatively high value (37 a.u.), as extraction starts during the filling.
When the acquisition was started at the end of the filling, a portion of
the L1-Hg2+ complexes formed during the filling are observed at the
microdevice outlet.
Finally, an experimental detection limit of 10À6 m was
measured for Hg2+, relatively close to the theoretical detec-
tion limit, which is evaluated to be about 2 ꢀ 10À7 m, taking
into account a signal-to-noise ratio of 3.
fluorescence is observed exclusively in the channel containing
the ionic liquid phase with intensity decreasing from the
pillars to the opposite wall and increasing from the inlet to the
outlet of the microsystem. This observation illustrates the
complexation of Hg2+ ions by the nonfluorescent ligand L1
leading to the formation of the fluorescent complex L1-Hg2+
and its very slow diffusion (10À11 m2 sÀ1)—owing to the high
viscosity of [bmp][NTf2] (85 cP at 258C)—away from the
interface into the bulk ionic liquid phase. Indeed, the Peclet
number (Pe) of the ionic liquid flow being relatively small
(Pe = 222), the fluorescent complex diffuses with a mass flux
only slightly affected by convective transport. The homoge-
neous fluorescence obtained at the outlet of the microsystem
allowed us to monitor the extraction as a function of time
(Figure 3b) and to quantify it using a calibration curve. By
progressively decreasing the Hg2+ concentration in the
initially injected aqueous solution, an experimental detection
limit of 50 ppb was obtained.
After this study conducted in pure IL solvents, macro-
scopic-scale liquid–liquid extraction experiments were per-
formed using L1 in [bmp][NTf2], chosen because of its
negligible absorption and fluorescence at operating excitation
and emission wavelengths. The extraction of Hg2+ ions from
aqueous solution was easily performed by shaking, resulting
in the formation of a fluorescent ionic liquid phase. The
Nernst distribution coefficient was determined to be 13.6 and
the extraction yield was 93%. The result was confirmed by
ICP-MS.
A proof-of-principle test of liquid–liquid extraction was
then conducted in microsystems using TSOS L1 (10À4 m) in
[bmp][NTf2] solution as the binary IL storage phase (see
detailed description and fabrication of the microdevice in the
Supporting Information). The aqueous solution containing
the Hg2+ ions was introduced at 4 mLminÀ1 into the 108 mm
wide channel. Next, the solution of L1 in [bmp][NTf2] was
introduced at 0.02 mLminÀ1 into the 150 mm wide channel.
When the filling was completed, optical acquisitions started at
450 nm (lexc = 310 nm) through a Pyrex plate with an
In conclusion, we have described a microsystem prototype
allowing for the real-time detection of mercury in water with a
very high sensitivity reaching a detection limit of 50 ppb. The
proposed technology relies on the combination of several
innovations, including a microdevice specifically designed for
epifluorescence microscope equipped with
a UV-LED
426
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 424 –427

Angewandte
Chemie
[4] Z. S. Zhou, S. J. Wang, Z. M. Yang, Chemosphere 2008, 70, 1500 –
liquid–liquid extraction and a binary task-specific ionic liquid
including a TSOS bearing a 8-hydroxyquinoline benzoate unit
as the extracting/sensing moiety and having the capacity to
concentrate the mercury(II) cation in the hydrophobic ionic
liquid phase, thus increasing the detection sensitivity. This
fluorescent probe is highly selective towards mercury versus
other heavy metals such as copper, lead, and cadmium
(Figure 2b). Another interesting feature is that the ligand
emits fluorescence exclusively when complexed with mercury,
therefore providing an “off–on” sensing response. These
results clearly open the way towards efficient devices for the
real-time tracking of heavy metals in water. Such devices can
also be extended to the detection of other pollutants.
Collaborative research along these lines is currently in
progress.
1509.
[5] J.-S. Lee, M. S. Han, C. A. Mirkin, Angew. Chem. 2007, 119,
4171 – 4174; Angew. Chem. Int. Ed. 2007, 46, 4093 – 4096.
[6] Guidelines for drinking water, 3rd edition. World Heath
Organization, Geneva, 2004.
[7] a) S. Dai, Y. H. Ju, C. E. Barnes, J. Chem. Soc. Dalton Trans.
1999, 1201 – 1202; b) H. Luo, S. Dai, P. V. Bonnesen, A. C.
Buchanan III, J. D. Holbrey, N. J. Bridges, R. D. Rogers, Anal.
Chem. 2004, 76, 3078 – 3083; c) K. Shimojo, M. Goto, Anal.
Chem. 2004, 76, 5039 – 5044.
[8] a) G. Marchand, P. Dubois, C. Delattre, F. Vinet, M. Blanchard-
Desce, M. Vaultier, Anal. Chem. 2008, 80, 6051 – 6055; b) P.
Dubois, G. Marchand, Y. Fouillet, J. Berthier, T. Douki, F.
Hassine, S. Gmouh, M. Vaultier, Anal. Chem. 2006, 78, 4909 –
4917.
[9] a) A. E. Visser, R. P. Swatloski, W. M. Reichert, R. Mayton, S.
Sheff, A. Wierzbicki, J. H. Davis, R. D. Rogers, Environ. Sci.
Technol. 2002, 36, 2523 – 2529; b) A. Ouadi, B. Gadenne, P.
Hesemann, J. J. E. Moreau, I. Billard, C. Gaillard, S. Mekki, G.
Moutiers, Chem. Eur. J. 2006, 12, 3074 – 3081.
[10] a) M. Tokeshi, T. Minagawa, K. Uchiyama, A. Hibara, K. Sato,
H. Hisamoto, T. Kitamori, Anal. Chem. 2002, 74, 1565 – 1571;
b) J. G. Kralj, H. R. Sahoo, K. F. Jensen, Lab Chip 2007, 7, 256 –
263.
[11] C. H. Chen, J. Shi, Coord. Chem. Rev. 1998, 171, 161 – 174.
[12] H. Zhang, L.-F. Han, K. A. Zachariasse, Y.-B. Jiang, Org. Lett.
2005, 7, 4217 – 4220.
[13] E. Climent, M. D. Marcos, R. M. Manez, F. Sancenon, J. Sato, K.
Rurack, P. Amoros, Angew. Chem. 2009, 121, 8671 – 8674;
Angew. Chem. Int. Ed. 2009, 48, 8519 – 8522.
Received: September 8, 2009
Revised: October 21, 2009
Published online: December 3, 2009
Keywords: analytical methods · environmental chemistry ·
.
fluorescent probes · ionic liquids · microsystems
[1] E. M. Nolan, S. J. Lippard, Chem. Rev. 2008, 108, 3443.
[2] B. Kumar Jena, C. Retna Raj, Anal. Chem. 2008, 80, 4836 – 4844.
[3] S. Casilli, C. Malitesta, S. Conoci, S. Petralia, S. Sortino, L. Valli,
Biosens. Bioelectron. 2004, 20, 1190 – 1195.
Angew. Chem. Int. Ed. 2010, 49, 424 –427
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
427