Thiophene-based fluorescent mercury-sensors

Abstract Coordination chemistry of thiophene rings is poorly understood, despite their common use in organic electronic materials. The absorption and emission responses to transition metal ions of three thiophene-based ligands containing pyridine chelating groups are examined. These ligands, 2-(2′-thienyl)-pyridine (L1), 2,5-bis(2-pyridyl)thiophene (L2), and 2,6-bis(2-thienyl)pyridine (L3), show a ratiometric fluorescence response in the presence of Hg(ClO4)2 with reasonable selectivity against many transition metal ions in acetonitrile. 1H NMR data support S,N chelation of L1 and L2 to the Hg(II) center, while L3 coordinates through cyclometalation via one carbon of thiophene. DFT calculations suggest thiophene coordinates to Hg(II) in a bent geometry. Our results indicate that thiophene could offer selectivity for Hg(II) towards the design of fluorescent sensors.


Introduction
The coordination chemistry of heterocyclic aromatic ligands has been studied for a wide variety of applications including catalysts, photosensitizers, and molecular sensors [1][2][3][4][5]. Ligands such as pyridine or N-heterocyclic carbenes have been studied extensively [6][7][8][9]. In contrast, thiophene coordination chemistry is poorly understood. The few examples of thiophene coordination in the literature focus on metal-catalyzed hydrodesulfurization [10]. The photo-electronic properties of polythiophenes and thiophene containing small molecules are extensively studied for use in organic photovoltaic devices (OPVs) [11][12][13][14], yet the ability of sulfur in thiophene to coordinate to metals has not been fully explored. To increase efficiency in OPV applications it is necessary to increase device absorption in the red to near-infrared (NIR) region [11,[15][16][17][18][19][20]. Metal coordination provides a potential avenue to red-shift absorption and enhance the electronic properties of these materials. Also, as a molecular sensor, sulfur of thiophene could offer selectivity for the detection of soft heavy metals such as mercury(II) or lead(II).
Herein, we report several thiophene-based ligands substituted with pyridine groups to facilitate sulfur-coordination (figure 1). We speculate that pyridine should enhance sulfur coordination both through chelation and increasing the electron density on the thiophene. Metal coordination has been studied using absorption, emission and NMR spectroscopy. Based on density functional theory (DFT) calculations, we propose that mercury coordinates through S-N chelation for these ligands. Additionally, the ligands show selectivity towards mercury as evidenced by their absorption and emission responses, making them potentially viable as fluorescent mercury sensors.

Materials and methods
Reagents were obtained from Sigma Aldrich and TCI America Chemicals and used without purification. The ligand 2-(2′-thienyl)-pyridine (L1) was obtained from TCI Chemicals and further purified by recrystallization in hexanes. Acetonitrile (CH 3 CN), acetone, hexanes, tetrahydrofuran (THF), dichloromethane, diethyl ether (et 2 O), chloroform, ethyl acetate, and metal salts were obtained from Sigma Aldrich. 1,4-Dioxane was obtained from TCI America Chemicals. The identity of organic compounds was confirmed by 1 H NMR spectroscopy performed with a Bruker-Avance III 400 MHz spectrometer with a 9.4 Tesla magnet and a 5 mm BBO probe. Mass spectra were recorded on a Thermo electron LTQ-Orbitrap Discovery high-resolution mass spectrometer. emission spectra and quantum yields were obtained using a PTI Quantamax 300 Phosphorimeter with a Xe-flash lamp and PMT detector. Quantum yields were obtained using a PTI K-sphere "petite" integrating sphere. Absorption spectra were collected on a UV-3600 Shimadzu UV-vis NIR spectrophotometer.

Synthesis of 2,6-bis(2-thienyl)pyridine (L3)
A flask containing 2,6-dibromopyridine (0.482 g, 2.03 mmol), thiophene-2-boronic pinacol ester (0.882 g, 4.19 mmol), potassium fluoride dihydrate (0.351 g, 6.04 mmol) and tetrakis(triphenylphosphine)palladium(0) (0.082 g, 0.064 mmol) was purged with N 2 . A solution of 1,2-dimethoxyethane (20 mL) and water (15 mL) was degassed by three freeze-pump-thaw cycles and added. The flask was sealed and heated to reflux for 4 h. The organic layer was then removed in vacuo and 50 mL of water was added. The product was extracted in et 2 O (3 × 50 mL), the organic layers combined, and then washed with water (1 × 50 mL). The organic layer was dried over Na 2 SO 4 and the solvent was evaporated to afford an orange oil. The crude product was then purified by column chromatography, eluted with 50 : 50 CHCl 3 : hexanes. The final product was obtained as a white solid (0.0814 g, 16.4% yield) and characterized by 1 H and 13 C NMR spectroscopy. 1   After an initial spectrum of the ligand was obtained, increasing molar equivalents of the metal salts were added to the solution, via micropipette, until no change in the spectra were observed. 1 H NMR titrations were performed by preparing 4 × 10 −3 M ligand solutions in CD 3 CN. Initial 1 H and COSy NMR spectra of the ligand were obtained prior to addition of a metal salt. Increasing equivalents of the metal salt were added to the solution. 1 H NMR spectra were obtained. COSy spectra were used to assign the peaks of the complexes.

Ligand synthesis and structure
Three ligands with thiophene and pyridine coordination sites were studied. L1 has one thiophene and one pyridine (figure 1). We hypothesize that pyridine will coordinate to transition metals and hold the thiophene in close proximity to the metal, thereby facilitating sulfur coordination. L2 (central thiophene with two flanking pyridine groups) and L3 (central pyridine with two flanking thiophene groups) were synthesized using modified literature procedures [44,45] and were designed to have several coordination sites in a larger π-conjugated system to increase sensitivity and red-shift the absorption and emission properties (figure 1). Previous attempts to coordinate L3 to chromium(II), cobalt(II), nickel(II), copper(II), silver(II), and zinc(II) were unsuccessful [44]. To the best of our knowledge these ligands have not been shown to coordinate to mercury(II).
The structures of all three ligands were verified by 1 H, 13

Photophysical properties of ligands
Absorption and emission spectra of L1-L3 were obtained with 1 × 10 −5 M solutions in CH 3 CN (figure 3). Absorption spectra of both L2 and L3 are red-shifted relative to L1 due to extended conjugation from the additional ring. The lowest energy absorption maximum of L1 is at 298 nm; L2 and L3 have peaks at 342 and 335 nm, respectively (table 1). Likewise the emission of both L2 (λ max = 394, Φ = 0.56) and L3 (λ max = 376, Φ = 0.34) are red-shifted compared to that of L1 (λ max = 356 nm, Φ = 0.43). extending the conjugation to three rings pushes the emission into the visible region, providing easy detection of metal binding through visual inspection. The lower energy absorption and emission of L2 relative to L3 is likely due to the more bent structure of L3 caused by the geometry of the central pyridine ( figure 2).
Measured absorbance spectra are consistent with excitation energies calculated by TD-DFT (see Supporting Information). The calculated lowest energy absorption for L1 is at 338 nm and the calculated lowest energy absorption for L2 is red shifted to 355 nm. For both L1 and L2 the lowest energy excitation is HOMO to LUMO in nature. The lowest energy peak in calculated absorption spectra of L3 comprises three transitions at 332, 314, and 290 nm. All three excitations are combinations of the HOMO−1 and HOMO to the LUMO and LUMO + 1 (depicted in the Supporting Information). Likewise, three excitations are seen in the measured absorption spectra of L3.

Photophysical response to mercury
In the presence of mercury ions (Hg(ClO 4 ) 2 ), the absorption and emission spectra of all three ligands red-shift. With increasing equivalents of Hg(II), the ligand-based absorption decreases and a new, redshifted absorption grows in. For L1, the absorbance at 298 nm decreases while a peak at 322 nm grows until 50 equivalents of Hg(II) have been added ( figure 4(a)). An isosbestic point is observed at 314 nm. Likewise, the emission of L1 at 356 nm is quenched with addition of Hg(II) and a new peak at 411 nm (Φ = 0.79) appears ( figure 4(b)). Although L2 displays a similar red-shift in both absorption and emission, several distinct features are noted. With increasing amounts of Hg(II), λ max of the absorption at 342 nm decreases as a new peak at 368 nm increases until 20.0 eq of Hg(II) have been added ( figure 4(c)). Similarly, a new emission at 446 nm (Φ = 0.55) is observed ( figure 4(d)). Further additions of Hg(II) past 20.0 eq cause a blue shift in absorption to 360 nm and in emission to 421 nm (Φ = 0.22) (see Supporting Information). A clear isosbestic point is not observed for L2 suggesting the formation of more than one new species in solution.
Addition of Hg(II) to L3 also causes a red-shift in both the absorption and emission, however 100 eq of metal are required to reach saturation ( figure 4(e), (f )). An isosbestic point at 344 nm indicates that no intermediate forms during the reaction. For all three compounds, addition of excess Hg(II) is required to completely quench ligand emission. The emission of the complexes also have higher quantum yield than the free ligand. Likely, this results from suppression of nonradiative decay pathways caused by decreased rotation around the single bond connecting the rings due to metal chelation. The bathochromic shift is likely due to a conformational change of the ligand to a more rigid structure.

NMR titration for structure determination
In order to elucidate the structure of the L1-L3/Hg(II) complexes, formation of the complexes were monitored using 1 H NMR during the metal titrations into solutions of each ligand ( figure 5(a)-(c)). COSy spectra were obtained at the end of each titration to assign peaks. Downfield shifts for ligand peaks  were attributed to deshielding of the protons due to metal coordination [46]. The positive charge is delocalized on the rings such that the protons para to the Hg(II) coordination site are most deshielded. Initial addition of Hg(II) to all ligand solutions caused downfield shifts for all peaks in each spectrum.
In the case of L1 the para proton on the pyridyl group (H e ) ( figure 5(a)) shifts 0.717 ppm with addition of 2 eq of Hg(II), suggesting nitrogen coordination. H C in the three-position of thiophene was shifted 0.457 ppm relative to the free ligand suggesting possible sulfur-coordination to Hg(II). These results are consistent with S,N chelation of L1 to Hg(II).
Since two new products are formed with addition of Hg(II) to L2, emission spectroscopy was used in conjunction with 1 H NMR to differentiate the two products. The symmetry of the molecule was retained during the titration, as evidenced by retention of the singlet peak from H e on the thiophene ( figure 5(b)). After 1.00 eq of Hg(II) was added, emission spectroscopy confirmed the identity of the first product (λ max = 435 nm). A low downfield shift (Δppm = 0.083 ppm) is observed for the protons at the 3,4-positions of thiophene (H e ), while a 0.471 ppm shift was observed for the proton para to the pyridyl nitrogen (H C ) (figure 5(b)). After 5 eq an emission at 420 nm indicated the second product had formed in solution. A greater downfield shift for all peaks was observed. Though it is evident that pyridine chelation occurs, the identity of the products formed in solution between L2 and Hg(II) is unclear. The minimal shift observed for the singlet from the proton on the thiophene would indicate sulfur coordination does not occur until the second product is formed in solution. Further investigation into the structures of the two products formed for L2 are currently underway.
During the titration of L3 with Hg(II), a broadening of all peaks was observed until 5.00 eq of Hg(II) were added and the product was formed. Unlike titrations with L1 and L2, the total number of aromatic protons decreased by one after the addition of Hg(II). Another distinct feature is an apparent break in symmetry observed by an increase in the number of proton environments ( figure 5(c)). Following addition of 5 eq of Hg(II), the 1 H NMR spectrum suggests a proton at the three-position of one of the thiophenes had been removed from the ligand. Removal of this proton indicates S,N,C-chelation by L3 to Hg(II). This is the only ligand that appears to coordinate through a carbon rather than sulfur.

Calculated structure
DFT structure optimization was used to explore the coordination geometry of Hg(II) with L1. The optimized structure shows S,N-chelation to Hg(II) with the sulfur on the thiophene coordinating in a side on manner (see Supporting Information). This causes L1 to twist so that the N and S are on the same side with a N-C-C-S dihedral angle of 41.2°. The lowest energy excitation for the calculated absorption spectrum from the optimized geometry is red shifted to 349 nm compared to L1 (λ max = 338 nm), consistent with experiment. The orbitals involved in the absorption of the complex formed between L1 and Hg(ClO 4 ) 2 involve a mix of metal centered and ligand centered orbitals (see Supporting Information).

Selectivity towards mercury
The selectivity of the fluorescent response was tested against other metal perchlorate salts: Hg(II), Ag(I), Cd(II), Co(II), Cu(II), Fe(II), Fe(III), Mg(II), Mn(II), Na(I), Ni(II), Pb(II), and Zn(II). The emissions of solutions with 5 eq of each metal salt were measured. For L1, only addition of Hg(II), Cu(II), and Fe(III) resulted in an emission response. The presence of other metal ions in solution did not affect the emission response to Hg(II), with the exception of Fe(III) which showed an enhanced emission (figure 6(a)). Similar selectivity was observed for L2. There is a bathochromic shift for 5 eq of Fe(III) (λ max = 420 nm) relative to the free L2. However, addition of Hg(II) alone results in a different emission (λ max = 446 nm) ( figure 6(b)). Notably, L3 emission was only minorly quenched in response to 5 eq of all metal salts, including Hg(II). Fe(III) addition caused a new emission at 457 nm (see Supporting Information) (figure 6(c)). These results indicate that thiophene containing fluorescent sensors can display selectivity for Hg(II).

Conclusion
We have developed three new thiophene-based ligands with a fluorescence response to Hg(II) via chelation to the metal center. The addition of pyridyl groups enhances coordination of thiophene via chelation, as is evident through the decreased amount of Hg(II) needed to reach equilibrium with L2, which contains two pyridines, relative to the amount needed to reach equilibrium with L1 and L3, which both contain only one pyridine. This suggests that the sulfur of thiophene is not a strong coordinating  group. 1 H NMR suggests a S,N,C-chelation of L3 to Hg(II). Our results indicate that thiophene ligands offers selectivity for Hg(II). A unique coordination chemistry is observed for L2 and Hg(II), forming two distinct products dependent on the amount of Hg(II) in solution. Currently our lab is researching the identity of these products towards the design of thiophene-based molecular sensors. In conclusion, we have shown that thiophene containing ligands display selectivity for Hg(II) giving a visible, ratiometric fluorescence response.