Selective sensing of Hg2+ by a proton-ionizable calix[4]arene fluoroionophore

1 23

Analytical and BioanalyticalChemistry ISSN 1618-2642Volume 405Combined 2-3 Anal Bioanal Chem (2013)405:1133-1137DOI 10.1007/s00216-012-6378-8

Selective sensing of Hg2+ by a proton-ionizable calix[4]arene fluoroionophore

Giuseppe Arena, Francesco Attanasio,Dongmei Zhang, Yanfei Yang, RichardA. Bartsch & Carmelo Sgarlata

1 23

Your article is protected by copyright and

all rights are held exclusively by Springer-

Verlag. This e-offprint is for personal use only

and shall not be self-archived in electronic

repositories. If you wish to self-archive your

work, please use the accepted author’s

version for posting to your own website or

your institution’s repository. You may further

deposit the accepted author’s version on a

funder’s repository at a funder’s request,

provided it is not made publicly available until

12 months after publication.

TECHNICAL NOTE

Selective sensing of Hg2+ by a proton-ionizable calix[4]arene fluoroionophore

Giuseppe Arena & Francesco Attanasio & Dongmei Zhang &

Yanfei Yang & Richard A. Bartsch & Carmelo Sgarlata

Received: 27 June 2012 /Revised: 13 August 2012 /Accepted: 21 August 2012 /Published online: 12 September 2012# Springer-Verlag 2012

Abstract A fluorogenic derivative of a calix[4]arene withtwo proton-ionizable N-(phenyl)sulfonyl carboxamide-containing side arms in the 1,3-positions on the lower rim isemployed for the selective sensing of Hg2+ at low concentra-tion levels in water/MeCN (1:1, v/v) solutions containing Pb2+

and Cd2+. All three metal ions quench the fluorescence of theligand in pure MeCN. However, in water/MeCN mixed sol-vent, the recognition of such cations occurs differently as onlyHg2+ complexation quenches the fluorescence of the calixar-ene. Experiments carried out in the presence of an acid and abulky non-complexing cation shows that the quenching of thecalixarene fluorescence upon Hg2+ addition is likely due toproton displacement from the proton-ionizable side arms ofthe ligand. The system may be employed as a simple tool forthe selective and efficient mercury sensing in mixed water/organic solvent.

Keywords Fluorescent sensor . Selective detection .

Mercury . Heavy metals . Calixarenes

Introduction

Heavy metals (such as mercury, cadmium, and lead) areextremely toxic and may have a negative impact on theenvironment and be harmful to humans [1, 2]. The varietyof symptoms due to cardiac, digestive, kidney, and neuro-logical diseases implies that these metals may affect multi-ple target organs. Consequently, the detection and removalof these poisonous species has been a topic of paramountinterest over the past years in analytical, biological, andenvironmental chemistry, as well as in toxicology [3]. Hg2+

is one of the most relevant cations in the ecosystem as itstoxicity has long been recognized as a widespread environ-mental issue owing to the conversion of inorganic mercuryto neurotoxic methylmercury which bioaccumulates throughthe food chain [4, 5]. The toxicity of Hg2+, even at very lowconcentrations, has been acknowledged as a primary con-cern and, accordingly, the development of novel or im-proved analytical methods for the sensitive and selectivedetermination of Hg2+ is highly desirable [6].

Traditional quantitative approaches to Hg2+ and other heavymetals determination employ analytical techniques includingatomic absorption, emission spectroscopy, ICP-MS, electro-chemical measurements and gas chromatography. However,many of these methods require time consuming multistepsample preparation and/or sophisticated instrumentation andtherefore are not well-suited for the quick detection of Hg2+ insitu or for in vivo studies. The use of mercury-responsive smallmolecule ligands, which are relatively easy to synthesize andprovide immediate optical feedback, may overcome the abovelimitations. In particular, fluorescence-based methods offerseveral advantages in terms of sensitivity, selectivity, and cost

Published in the special issue Analytical Science in Italy with guesteditor Aldo Roda.

Electronic supplementary material The online version of this article(doi:10.1007/s00216-012-6378-8) contains supplementary material,which is available to authorized users.

G. Arena :C. Sgarlata (*)Dipartimento di Scienze Chimiche,Università degli Studi di Catania,Viale Andrea Doria 6,95125 Catania, Italye-mail: [email protected]

F. AttanasioIstituto di Biostrutture e Bioimmagini, UOS Catania, CNR,Viale A. Doria 6,95125 Catania, Italy

D. Zhang :Y. Yang :R. A. BartschDepartment of Chemistry and Biochemistry,Texas Tech University,Lubbock, TX 79409, USA

Anal Bioanal Chem (2013) 405:1133–1137DOI 10.1007/s00216-012-6378-8

Author's personal copy

[7]. Considerable effort has been devoted to the developmentof selective fluorescent molecular sensors for the detection oftoxic metal ions in organic and mixed organic/water solvents.Such small molecules are often called fluoroionophores as theyare composed of a recognition unit along with a signalingmoiety. Calixarene derivatives have been recently proposedas molecular scaffolds for the synthesis of fluorescent sensorsowing to their selectivity and relative ease of functionalizationwith suitable complexing units and/or fluorophores [8, 9].

In earlier work, we showed that macrocyclic ligands obtainedby attaching two proton-ionizable N-(X)sulfonyl carboxamide-containing side arms to a calix[4]arene scaffold provide excel-lent selectivity for Hg2+ and good selectivity for Pb2+ over manyother metal ion species in solvent extraction [10, 11]. We laterreported that Hg2+ is transported with high selectivity fromacidic aqueous source phase solutions of Cd2+, Hg2+, and Pb2+

into aqueous receiving solutions, using the N-(phenyl)sulfonylcarboxamide calix[4]arene 1 (Fig. 1) as a carrier in polymericinclusion membranes (PIM) [12]. Preliminary experiments in-dicated that the above ligand may also act as a promisingfluorogenic probe for Hg2+ recognition. Based on these findings,we now report on the effective and selective detection of Hg2+

over Cd2+ and Pb2+ in mixed water/MeCN (1:1, v/v) solvent bythe versatile molecular fluorescent sensor 1.

Materials and methods

Chemicals

Compound 1 was prepared as previously reported [13]. Sincecalixarenes may contain crystallization and/or absorbed water,a thermogravimetric analysis of 1 was performed with a Perkin

Elmer TGS-2 instrument. TGA (10 °C/min under nitrogen)showed that the ligand contains no crystallization water andonly absorbs water (humidity) in limited amount (1–2%). Leadand mercury perchlorates and perchloric acid were obtainedfrom Aldrich and used as received. Cadmium oxide was pur-chased from Carlo Erba. Tetrabutylammonium perchlorate(TBAP, electrochemical grade) and tetrabutylammonium hy-droxide (TBAOH) were obtained from Fluka. Safer counterions (e.g., tetrafluoroborates) could not be employed owing totheir interfering absorption. Caution: perchlorate salts areshock and heat sensitive and must be handled with care. MeCN(Uvasol for spectrophotometry) was purchased from Merck.High purity water (Millipore, Milli-Q Element A 10 ultrapurewater) and grade A glassware were used throughout.

For the spectroscopic experiments, metal perchlorate sol-utions in MeCN were formulated as follows. Mercuric per-chlorate solutions were prepared by dissolving the salt inMeCN. The Hg2+ concentration was determined by titratingthe resulting solution with standardized potassium thiocya-nate in the presence of ammonium iron(III) sulfate dodeca-hydrate [14]. Lead perchlorate solutions were prepared bydissolving the salt in MeCN and titrating with EDTA usingmethylthymol blue [14]. Cadmium perchlorate solutionswere obtained by dissolving cadmium oxide with a stoichio-metric amount of perchloric acid and dilution with MeCN.

Spectroscopic measurement

UV–Vis spectra were recorded at 25 °C in a quartz cell(1 cm path) using a diode-array spectrophotometer (Agilentmodel 8453).

Fluorescence emission spectra of ligand and metal-ligandsolutions were monitored from 283 to 420 nm in a quartzcell (1 cm path) using a Spex Fluorolog-2 F-111 spectroflu-orimeter (Horiba Jobin Yvon, NJ, USA). An excitationwavelength of 273 nm was used in all experiments. Bothexcitation and emission bandwidths were set to 5 nm.

Solutions for emission spectra were prepared by addingappropriate volumes of metal perchlorate solutions inMeCN (1×10−4–2.6×10−3M) by a precision electronic pi-pette (Rainin) to volumetric flasks (usually 2 ml) containinga MeCN solution of ligand 1 (4.9×10−4–1.1×10−3M) andthen diluting with a water/MeCN (1:1, v/v) solution toobtain a final ligand concentration of 3.8÷4.2×10−5M andthe desired M/L ratio. Titrations with HClO4, TBAP andTBAOH were carried out by adding increasing amounts ofwater/MeCN (1:1, v/v) solutions of the titrant (1.3÷1.8×10−3M) into the measuring cell containing a known volume(usually 2 ml) of a water/MeCN (1:1, v/v) solution of theligand. The solutions were allowed to equilibrate, afterwhich spectra were recorded. Time intervals for equilibra-tion and reading were systematically changed to avoid arti-facts. At least three independent runs were collected for eachFig. 1 Structure of the calix[4]arene ligand 1

1134 G. Arena et al.

Author's personal copy

M-ligand system (M 0 Hg2+, Pb2+, Cd2+, TBAP, TBAOH,or H+).

Potentiometric measurements

Potentiometric measurements were carried out by means oftwo home-assembled, fully automated apparatuses (Met-rohm meter, dispenser, and combined micro pH glass elec-trodes) controlled by a software set up in our laboratory.Values of E°, Ej, Kw, and the Nernstian slope of the electro-dic system were determined in separate experiments bytitrating a water/MeCN (1:1, v/v) solution of nitric acid withCO2-free potassium hydroxide. To determine the pKa val-ues, solutions of the ligand in water/MeCN (1:1, v/v) withconcentrations ranging from 1.3 to 2.0×10−3M were titratedwith 0.1 M potassium hydroxide. Three independent runswere collected, each run including 50–60 points. The initialpH was always adjusted to 2.5. To avoid systematic errorsand verify the reproducibility, the emf values of each exper-iment were read at different time intervals. All measure-ments were carried out at 25.0±0.1 °C. The ionic strengthwas maintained at 0.1 M (potassium nitrate). Potentiometricdata were refined by using the program Hyperquad [15].

Results and discussion

Preliminary experiments carried out inMeCN showed that theemission spectrum of calixarene 1 is remarkably quenchedupon addition of a Hg2+ solution. Even well below one equiv-alent (e.g., 0.3 equivalents), the addition of Hg2+ results in aremarkable quenching (ca. 40 % of the initial value) of thefluorescence of free 1 (see Electronic SupplementaryMaterial,Fig. S1). Quenching of the fluorescence looks excellent (in-tensity decreases by about 70 %) upon addition of equimolaramounts of Hg2+ (Fig. 2, blue line).

Stability constants for the interactions of 1 with Hg2+ inMeCN were previously determined by UV–Vis titration; thedata refinement was consistent with the formation a MLspecies in addition to a M2L species with log β values of3.8 and 7.7, respectively [12]. Multiple species had beenalso detected in Hg2+solvent extraction experiments fromaqueous into chloroform solutions containing ligands simi-lar to 1 [10]. Despite the relatively low stability constant ofthe ML species, and the consequent small amount formedover the concentration range investigated in the fluorescenceexperiments, the chemosensor responded remarkably welleven at M/L00.3. However, when similar experiments arecarried out with Pb2+ and Cd2+ in MeCN, remarkablequenching of the fluorescence of 1 is also observed. Figure 2shows that the fluorescence intensity of ligand 1 decreasesby 45 % and 60 % when one equivalent of a solution of Pb2+

and Cd2+, respectively, was added. These two metal ionsalso form multiple complex species, as with Hg2+. ML wasthe only species that forms within the range of concentra-tions explored in the fluorescence measurements.

The strength of the ML complexes decreases in the orderPb2+>Cd2+>Hg2+ and does not match the sequence ob-served in fluorescence quenching. The decrease of the emis-sion intensity of 1 due to cation complexation does notfollow the cation size order (Pb2+>Hg2+>Cd2+) [16], butrather the hard–soft features of the three metal ions (softnessfollows the order: Hg2+>Cd2+>Pb2+) [17]. However, re-gardless of the number and stability of the species formingin solution for each cation, ligand 1 appeared to be unsuit-able for the selective sensing of a single target ion in MeCN.

When analogous experiments were carried out in water/MeCN (1:1, v/v) mixed solvent, the picture dramaticallychanged as only Hg2+ quenched the fluorescence of 1(Fig. 3). Upon addition of one equivalent of each of thethree metal ions, the intensity of 1 drops by more than 70 %in the case of Hg2+, but remained basically unaltered in thepresence of either Pb2+ or Cd2+. This suggested that theversatile fluoroionophore 1, effectively used so far for theremoval of Hg2+ in solvent extraction and PIM transportexperiments, may be also employed for the selective sensingof Hg2+ at low concentration levels in water/MeCN solu-tions containing Pb2+and Cd2+. The detection limit, calcu-lated as three times the standard deviation of the backgroundnoise from a calibration curve (Fig. S4, Electronic Supple-mentary Material) [18] was found to be 1.56(3)μM. Thisvalue is within the “low micromolar range” reported forother fluorescent chemosensors in mixed solvent [7, 9] andis sufficient for the detection of the submillimolar concen-trations of Hg2+ found in many chemical and biochemicalsystems.

The acidic nature of the sulfonylcarboxamide NH-groupsin 1 results in proton dissociation of the ligand in solutionthus suggesting that complex formation of 1 proceeds via

280 300 320 340 360 380 400 420

1x1011

2x1011

3x1011

4x1011

5x1011

6x1011

Inte

nsity

(cp

s)

wavelength (nm)

Fig. 2 Change in the emission intensity of calixarene 1 in MeCNupon addition of equimolar amounts of different metal ion solutions.CL03.95×10

−5M. Excitation wavelength0273 nm.Black line free ligand1; green line Pb2+-1; red line Cd2+-1; blue line Hg2+-1

Selective sensing of Hg2+ 1135

Author's personal copy

proton displacement by a metal cation [10, 11]. Potentio-metric titrations were carried out to examine the acid–baseproperties of the free ligand in mixed water/MeCN solvent.Preliminary refinement of the data gave pKa values of 6.3(4)and 10.4(2) for the dissociation of the NH group on the firstand the second sulfonylcarboxamide side arm of the calix-arene, respectively. Noticeably, the pKa value determined inmixed solvent is larger than that reported for similar com-pounds in pure water [19] and smaller than the value foundfor analogous sulfonylcarboxamide units in solvents havinglower polarity such as methanol [20]. The extra-protondisplacement resulting from the complexation of the metalion supports the view that the proton may actually play akey role on the fluorescence response of the sulfonylamidemoieties of the ligand. To gain greater insight into thepotential correlation between the fluorescence quenchingand the proton displacement, we monitored the change influorescence resulting from the addition of protons to asolution of 1 in water/MeCN (Fig. 4).

The spectra showed in Fig. 4 provide evidence that pro-tons quench fluorescence; the emission intensity of 1 wasreduced by 60 % when one equivalent of acid was added(red line) and was almost totally extinguished after theaddition of six equivalents of HClO4 (blue line; the fulltitration is reported in the Electronic Supplementary Mate-rial, Fig. S2).

If the protons displaced upon metal ion complexation areresponsible for the quenching, the addition of a cation thatcannot be complexed by 1 should have no effect on thefluorescence of the free ligand. We screened a few suitablecations and found that bulky positively charged TBAP doesnot form any complex with 1; the inset of Fig. S3 (seeElectronic Supplementary Material) shows that the UVspectrum of a 3.5×10−4M solution of the ligand is unaffect-ed by the presence of a 0.01 M solution of TBAP. Asexpected, the addition of an excess of TBAP to a solutionof 1 in water/MeCN leaves the fluorescence of the ligandpractically unaltered (Fig. S3, Electronic SupplementaryMaterial) thus indicating that a cation that does not interactwith the ligand (and therefore cannot cause proton displace-ments from the calixarene ionizable side arms) does nothave effect on the fluorescence of 1.

To further prove that the fluorescence quenching resultsfrom the proton displacement due to mercury complexationin water/MeCN, an equimolar amount of Hg2+ was added toa solution of 1 and the change in the fluorescence was thenmonitored upon addition of increasing amounts of a base,tetrabutylammonium hydroxide, whose cation has beenshown not to form complexes with 1 (Fig. 5).

The spectra presented in Fig. 5 show that, when a 4×10−5

M calixarene solution was titrated with an excess ofTBAOH, no change in the emission spectra of 1 was ob-served (the black and the green lines basically overlap). Thisclearly demonstrates that neither the TBA+ cation (see Fig.

280 300 320 340 360 380 400 420

2.0x1011

4.0x1011

6.0x1011

8.0x1011

1.0x1012

1.2x1012

1.4x1012

Inte

nsity

(cp

s)

wavelength (nm)

Fig. 3 Change in the emission intensity of calixarene 1 in water/MeCN (1:1, v/v) upon addition of equimolar amounts of different metalion solutions. CL03.95×10

−5M. Excitation wavelength0273 nm.Black line free ligand 1; green line Pb2+-1; red line Cd2+-1; blue lineHg2+-1

280 300 320 340 360 380 400 420

2.0x1011

4.0x1011

6.0x1011

8.0x1011

1.0x1012

1.2x1012

1.4x1012

Inte

nsity

(cp

s)

wavelength (nm)

Fig. 4 Change in the emission intensity of calixarene 1 in water/MeCN (1:1, v/v) upon addition of increasing amounts of perchloricacid. CL03.95×10

−5M. Excitation wavelength0273 nm. Black linefree ligand 1; green line CH/CL00.09; red line CH/CL01.00; blue lineCH/CL06.05

280 300 320 340 360 380 400 420

2.0x1011

4.0x1011

6.0x1011

8.0x1011

1.0x1012

1.2x1012

Free calixarene Calix + TBAOH

Calix + 1 equiv. Hg2+

Calix + 1 equiv. Hg2+ + TBAOH

Inte

nsity

(cp

s)

wavelength (nm)

Fig. 5 Change in the emission intensity of calixarene 1 (black line)upon addition of equimolar amounts of Hg2+ in MeCN/H2O (1:1, v/v)(blue line); effect of the addition of tetrabutylammonium hydroxide tothe Hg2+-1 solution (red line). CL03.90×10

−5M. Excitation wave-length0273 nm. The green line shows the emission spectrum of a freecalixarene solution containing an excess of TBAOH

1136 G. Arena et al.

Author's personal copy

S3) nor the counterion, OH−, interact with 1. Conversely,the addition of Hg2+ to free 1 results in a pronouncedquenching of fluorescence (blue line). If a base (TBAOH)is added to the solution of 1 containing one equivalent ofHg2+ (blue line), the emission intensity increased remark-ably, up to 80 % of the initial fluorescence of the freecalixarene (red line). Since (1) both Hg2+ and protonsquench the calixarene fluorescence, (2) TBAOH does notinteract with 1 and cannot compete with Hg2+, (3) theaddition of TBAOH does not restore the original fluores-cence, we may conclude that the fluorescence quenching isdue for the most part to the proton displacement from theproton-ionizable side arms of the ligand 1 resulting fromHg2+ complexation. Similar conclusions have been reportedby other authors on ligands functionalized with analogousfluorogenic carboxamide units which undergo deprotona-tion upon complexation of either Hg2+ [21] or anionicspecies [22] and cause fluorescence quenching of the ligand.Though the study of the photophysical properties of thesystem is beyond the scope of the present work, the failureto fully restore the initial fluorescence of 1 (black line)suggests that Hg2+ contributes to fluorescence quenchingas well, although to a smaller extent. Studies on similarfluorophores bearing sulfonyl carboxamide groups supportthis explanation; in fact, based on electrochemical potentialsand spectrofluorimetric experiments in frozen solutions,Leray and Valeur attributed quenching of the fluorescenceof their compounds to an electron-transfer process from theexcited fluorophore to the Hg2+ [8, 21].

Conclusions

The fluorogenic calix[4]arene 1 with two N-(phenyl)sul-fonyl carboxamide-containing side arms, while being unableto discriminate Hg2+ from Pb2+ and Cd2+ in MeCN, hasbeen successfully employed for the selective sensing of Hg2+

at low concentration levels in water/MeCN (1:1, v/v) solutionscontaining Pb2+ and Cd2+. The role of the proton-ionizableside arms seems to be crucial for the rationalization of thefluorescence features of the sensor.

The selective recognition/removal of a target metal ionfrom waste aqueous solutions remains an important scien-tific endeavor. To this aim, the versatile molecular sensor 1,which has been shown to recognize Hg2+ both in solvent

extraction and transport through liquid membranes, mightalso be employed as a cheap and “easy to use” tool for theoptical detection of Hg2+ along a pipeline. For example, onemay envision resorting to a bypass of the aqueous stream tobe monitored, mixing the sample flowing through this alter-nate path with an appropriate amount of MeCN in which 1 isdissolved1 and recording the emission spectrum to obtain aquick analysis/check of the streaming solution.

Acknowledgment MIUR (PRIN 2008F5A3AF_005) is gratefullyacknowledged for partial support.

References

1. Hutchinson TC, Meema KM (1987) Lead, mercury, cadmium andarsenic in the environment. Wiley, New York

2. Gadd GM (2000) Heavy metal pollutants: environmental andbiotechnological aspects. In: Lederberg J (ed) Encyclopedia ofmicrobiology, 2nd edn. Academic, San Diego

3. Ferguson JE (1990) The heavy elements: environmental impactand health effects. Pergamon, Oxford

4. Boening DW (2000) Chemosphere 40:1335–13515. Harris HH, Pickering IJ, George GN (2003) Science 301:12036. Selid PD, Xu H, Collins EM, Striped Face-Collins M, Zhao JX

(2009) Sensors 9:5446–54597. Nolan EM, Lippard SJ (2008) Chem Rev 108:3443–34808. Leray I, Valeur B (2009) Eur J Inorg Chem 24:3525–35359. Ocak U, Ocak M, Bartsch RA (2012) Inorg Chim Acta 381:44–57

10. Talanova GG, Hwang HS, Talanov VS, Bartsch RA (1998) J ChemSoc Chem Commun 1329–1330

11. Talanova GG, Hwang HS, Talanov VS, Bartsch RA (1998) J ChemSoc Chem Commun 419–420

12. Sgarlata C, Arena G, Longo E, Zhang D, Yang Y, Bartsch RA(2008) J Membr Sci 323:444–451

13. Talanova GG, Talanov VS, Gorbunova MG, Hwang HS, BartschRA (2002) J Chem Soc Perkin Trans 2 12:2072–2077

14. Vogel AI (1961) Quantitative inorganic analysis. Longmans,London

15. Gans P, Sabatini A, Vacca A (1996) Talanta 43:1739–175316. Lide DR (1990) Handbook of chemistry and physics, 71st edn.

CRC, Boston17. Pearson RG (1963) J Am Chem Soc 85:3533–354318. Harris D (2007) Quantitative chemical analysis, 7th edn. Freeman,

New York, Section 5-219. Babic S, Horvat AJM, Pavlovic DM, Kastelan-Macan M (2007)

Trends Anal Chem 26:1043–106120. Fernandez FA, Bartsch RA, Arnaud-Neu F, Hubscher-Bruder V,

Michel S (2010) ARKIVOC 7:191–20221. Metivier R, Leray I, Valeur B (2004) Chem Eur J 10:4480–449022. Bhalla V, Kumar R, Kumar M, Dhir A (2007) Tetrahedron

63:11153–11159

1 The acidity of the incoming solution would not be an issue as it canbe adjusted to the desired value upstream from the bypass

Selective sensing of Hg2+ 1137

Author's personal copy