Dalton Transactions View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
PAPER
Cite this: Dalton Trans., 2014, 43, 11557
View Journal | View Issue
Fluorescent Au(I)@Ag2/Ag3 giant cluster for selective sensing of mercury(II) ion† Mainak Ganguly,a Chanchal Mondal,b Jaya Pal,a Anjali Pal,b Yuchi Negishic and Tarasankar Pal*a Highly stable Au(I)core–Ag(0)shell particles have been synthesized in aqueous solution via a green chemistry pathway utilising sunlight irradiation. The shell of the particles is composed of fluorescent Ag2 and Ag3 clusters which make the large core–shell particles highly fluorescent. The Au(I) core of the particles offers long-term stability to the silver clusters, which are otherwise unstable in solution at room temperature, by the transfer of electron density from the shell. Successive additions of Hg(II) ions to the fluorescent solution cause efficient and selective quenching of the fluorescence with gradual red shifting of the emission peak. The metallophilic 5d10(Hg2+)–4d10(Agδ+) interaction as well as Hg(II) stimulated aggregation have been ascribed to causing the fluorescence quenching and red shift. The fluorescent Au(I)core–Ag(0)shell particles are a highly selective and sensitive sensing platform for the detection of Hg(II) down to 6 nM in the presence of various metal ions. The detection limit is far below the permissible level as determined by
Received 21st April 2014, Accepted 6th May 2014
the EPA. Interferences due to Cu(II) and Fe(III) have been eliminated using Na2-EDTA and NH4HF2, respect-
DOI: 10.1039/c4dt01158a
ively. The fluorescent particles are successfully transferred to various solvent systems making Hg(II) determination also possible in non-aqueous media. Finally, the temperature dependent fluorescence change
www.rsc.org/dalton
with and without Hg(II) provides information about the metallophilic interaction.
Introduction Noble metal nanoclusters with a size smaller than 2 nm are gaining importance day by day in various fields like sensing,1,2 labeling,3 catalysis,4 antibacterial activity5 etc. The clusters possess molecule-like properties leading to the generation of fluorescence emission.6 Gold clusters exhibit interesting luminescent properties and have been extensively studied for their exciting optical and structural properties.7,8 In spite of poor photostability towards oxidation, many reports are found showing a stronger fluorescence signal of silver clusters compared to gold confined in the same matrix.9,10 To date, several compounds have been reported for synthesizing Ag clusters.10–22 Polyelectrolytes such as poly(amidoamine) dendrimer, poly(methacrylic acid), and poly(acrylic acid) have been widely employed for obtaining fluorescent Ag clusters in aqueous medium.16,17 Thiolate passivated Ag clusters a Department of Chemistry, Indian Institute of Technology, Kharagpur-721302, India. E-mail: [email protected]; Fax: (+) 91-03222-255303 b Department of Civil Engineering, Indian Institute of Technology, Kharagpur-721302, India c Department of Applied Chemistry, Tokyo University of Science, Tokyo-1628601, Japan † Electronic supplementary information (ESI) available: MALDI-TOF mass analysis, LCMS, elemental mapping, fluorescence spectra. See DOI: 10.1039/ c4dt01158a
This journal is © The Royal Society of Chemistry 2014
have been reported only recently.11–14 The inherent biocompatibility with biomolecules has encouraged researchers nowadays in this context for in vivo applications.18–22 Highly fluorescent and emission tunable Ag clusters encapsulated by DNA have been prepared by Dickson and coworkers.19,20 However, these DNA-encapsulated Ag clusters are not stable under high salt concentration preventing their extensive and extended applications. Although mercury poisoning is a serious threat to human health, global mercury emissions continue to rise with an alarming pace.23 Solid waste incineration and the combustion of fossil fuels are significant sources of mercury emissions.24 The long atmospheric lifetime of mercury is responsible for contamination across vast quantities of land and water.25 The problem becomes worse as bacteria convert elemental and ionic mercury to methyl mercury, which is a potent neurotoxin to the food chain. Serious sensory, motor, and cognitive disorders in human beings are the most severe outcomes of mercury poisoning.26–28 In environmental and health monitoring, detection of Hg(II) is undoubtedly important. As a consequence, different sensing platforms for Hg(II) detection such as small fluorescent organic molecules, various polymers, liposomes, DNAzymes, proteins, oligonucleotides, inorganic materials etc. have been developed.29–34 Noble metal fluorescent clusters of gold and silver to detect Hg(II) have become a hot field of
Dalton Trans., 2014, 43, 11557–11565 | 11557
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Paper
research.2,35,36 Guo and Irudayaraj identified the 5d10(Hg2+)– 4d10(Ag+) metallophilic interaction to be responsible for selective fluorescence quenching of silver clusters by Hg(II).37 In our present study, we have synthesized Au(I) supported fluorescent Ag clusters (AuAgF) in solution via a green chemical approach under solar light irradiation. The ultra small fluorescent Ag2 and Ag3 clusters, responsible for the fluorescence, cover the Au(I) moiety rendering highly fluorescent Au(I)core–Ag(0)shell giant clusters. The fluorescence intensity is greatly reduced specifically after the introduction of Hg(II) ions, causing successive red shifting of the emission maxima. The release of gold upon removal of fluorescent silver clusters by Hg(II) from the Au(I) surface is documented here. The rupture of the core–shell particle and subsequent fluorescence quenching make the core–shell particle a highly selective as well as sensitive Hg(II) sensor down to the detection limit of 6 nM, well below the EPA permissible level.38
Experimental section Materials and instruments All the reagents were of AR grade. Triple distilled water was used throughout the experiments. Chloroauric acid (HAuCl4), silver nitrate (AgNO3), glutathione (GSH), (R)-S-lactoylglutathione, ethylenediaminetetraacetic acid disodium salt dehydrate (Na2-EDTA), ammonium hydrogen difluoride (NH4HF2) and all the metal salts were purchased from Sigma-Aldrich. Dioctyl sodium sulfosuccinate (AOT) was obtained from Wako Pure Chemical Industries Ltd. All the solvents were purchased from Merck. All glassware was cleaned with freshly prepared aqua regia, subsequently rinsed with copious amount of distilled water and thoroughly dried before use. UV-vis absorption spectra were recorded on a SPECTRASCAN UV 2600 digital spectrophotometer (Chemito, India). The fluorescence measurements were performed at room temperature using a LS55 fluorescence spectrometer (Perkin Elmer, USA). A fluorescence microscope (Olympus DP72) was used to take fluorescence images. Particle morphology was observed using a field emission scanning electron microscope (Supra 40, Carl ZEISS Pvt. Ltd). X-ray photoelectron spectroscopy (XPS) analysis was done with a VG Scientific ESCALAB MK II spectrometer (UK) equipped with Mg Kα excitation source (1253.6 eV) and a five-channeltron detection system. For the measurement of ζ potential, a Malvern Nano ZS instrument employing a 4 mW He–Ne laser operating at a wavelength of 633 nm was used. TEM analyses were carried out with a H-9000 NAR, Hitachi instrument, using an accelerating voltage of 300 kV. Atomic fluorescence measurements were performed on an atomic fluorescence spectrometer (AFS-9700, Beijing, China). Synthesis of fluorescent solution A milky white suspension was observed as soon as 2.0 mL 10−2 M HAuCl4 was added to 60 mL 16 × 10−4 M aqueous glutathione. After solar light irradiation along with vigorous stirring
11558 | Dalton Trans., 2014, 43, 11557–11565
Dalton Transactions
for ∼6 hours, 3.4 mL 10−2 M AgNO3 was added. Then the mixture was kept under the sun for a further ∼3 hours with vigorous stirring. A straw yellow solution (AuAgF) is obtained with strong fluorescence (λem = 564 nm). Aliquots of fluorescing 3 mL AuAgF and HgAuAgF [partially fluorescence quenched by 10 × 10−6 M Hg(II) after ageing for 9 hours] solutions were used thoroughout the course of the study. Synthesis of fluorescent particles by AOT mediated route For the preparation of the fluorescing species in water immiscible organic solvents, 0.1 M AOT in heptane was introduced. In the 6 mL AOT–heptane solution, GSH, aqueous chloroauric acid and silver nitrate were added to obtain the same final concentrations as mentioned for AuAgF. w0 i.e., the ratio of the number of water and AOT molecules was maintained to be 10. The solution was exposed to the sun for ∼15 hours with vigorous stirring to achieve a fluorescent yellow solution. The fluorescent solution was vacuum dried and redispersed in various water immiscible organic solvents.
Results and discussion The water soluble Au(I)@Ag2/Ag3 giant cluster has been synthesized by sunlight irradiation in the presence of AgNO3, HAuCl4 and GSH precursors. Thus, the green chemically evolved yellow solution (AuAgF) exhibits an intense fluorescence (quantum yield 6% using quinine sulfate in 0.1 M H2SO4 as reference and highly stable with unaltered emissive property with respect to intensity and emission maximum for months) with the emission maxima ∼564 nm when the solution is excited at ∼400 nm. Consequently, a large Stokes shift of 164 nm is observed. It is intriguing to note that the presence of both gold and silver are essential to obtain such giant fluorescent Au(I)core–Ag(0)shell clusters. The absence of any one of the metals i.e., gold or silver, evolves a non-fluorescent solution proving the synthetic tactic to be synergistic. The Stokes shift can be further increased by increasing the concentration of Hg(II). In other words, a gradual red shift of the emission maxima is witnessed along with strong and successive quenching of the fluorescence intensity. The fluorescence of the AuAgF solution is quenched by the Hg(II) ion. The Hg(II)–AuAgF mixed solution reaches a stable fluorescence condition after a waiting time of 9 hours. The quenched fluorescence is associated with the concentration dependent red shifting of the emission maxima. Under the experimental conditions employed, we introduced 10 × 10−6 M Hg(II) into 3 mL AuAgF solution and the solution exhibits a 55 nm red shifted peak after 9 hours. This becomes the representative test solution (HgAuAgF) for investigation. Here the quenching is 73%. In the case of other metal ions, virtually no alteration of the fluorescence intensity is found except for Cu(II) and Fe(III). Fig. 1A and 1A1 display the comparative effect of different metal ions on the fluorescence intensity of the as produced fluorescent solution [I0 = fluorescence intensity of AuAgF,
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Dalton Transactions
Fig. 1 (A) Fluorescence spectral profile and (A1) bar diagram showing the effect of different metal ions on the fluorescence of the AuAgF solution. (B) Fluorescence spectral profile and (B1) bar diagram showing the effect of different metal ions on the fluorescence of the HgAuAgF solution. Conditions: AuAgF solution = 3 mL; [Mn+] = 10 × 10−6 M; λex = 400 nm.
I = fluorescence intensity of AuAgF after individual addition of different metal ions]. It is worth noting that Cu(II) as well as Fe(III) result in fluorescence quenching like Hg(II). But, the phenomenon of quenching by Cu(II) and Fe(III) is not as efficient as Hg(II). Besides, the gradual red shift with increased concentration is unique for mercury, showing promise for designing a solution based sensing platform for mercury. Now, to investigate the effect of various metal ions on the Hg(II) induced quenched fluorescence (HgAuAgF), we have found that it is almost indifferent to other interfering metal ions except for Cu(II) and Fe(III). For both the Cu(II) and Fe(III) cases, the fluorescence of HgAuAgF is further decreased [Fig. 1B and 1B1]. In this context, it can be stated that Fe(III) exhibits a significant quenching capability towards the AuAgF solution. But, the quenching effect with Fe(II) is insignificant. Some efforts39,40 have been devoted to distinguish between Fe(II) and Fe(III). The distinction between Fe(II) and Fe(III) has been achieved fluorometrically even at very low concentration (nM level) exploiting the as-prepared fluorescent giant cluster (AuAgF solution) [Fig. S1, ESI†]. The quenching capability of Hg(II) is clearly much higher than Fe(III) and Cu(II). Nevertheless, we have introduced Na2EDTA and NH4HF2 to eliminate the interference due to Cu(II) and Fe(III), respectively. It may be noted that these two masking agents cause no observable interference to the fluorescence of AuAgF. When we introduced excess Na2-EDTA to the aqueous fluorescent AuAgF solutions containing the three quencher metal ions [Cu(II), Fe(III), Hg(II)] individually, appreciable turn on fluorescence was noticed only for Cu(II) owing to a reasonable stability constant of the Cu-EDTA complex. Surprisingly, the stability constants for Cu(II), Fe(III) and Hg(II) with EDTA are 18.8, 25.7 and 21.5, respectively. This indicates that the affinity of Cu(II) to the fluorescent giant cluster is
This journal is © The Royal Society of Chemistry 2014
Paper
comparatively low unlike Fe(III) and Hg(II). Complexation of Cu(II) with EDTA regenerates the lost fluorescence of the AuAgF solution although the stability constant of the EDTA complex is comparatively low for Cu(II) [Fig. 2A]. Quenching may also be due to free Cu(II) ions in solution exerting spin–orbit coupling.41,42 It may also be due to the formation of a weak complex43 of copper with glutathione decreasing the capping capability for the Au(I)@Ag2/Ag3 giant cluster. The fluorescence re-appears when the copper forms an EDTA complex leaving free GSH. To eliminate the interference of Fe(III), NH4HF2 has been added after metal ion addition [Fig. 2B]. Decent ‘turn on’ fluorescence is observed in the solution in which fluorescence was originally quenched by Fe(III). The HF2− restores the fluorescence due to the formation of highly stable [FeF6]3−.44 Besides, for both the Na2-EDTA and NH4HF2 treatments, Hg(II) stimulated quenched fluorescence remains indifferent. Consequently, the discrimination of the three metal ions [Hg(II), Fe(III), Cu(II)] causing quenching has been achieved with the addition of Na2-EDTA and NH4HF2. So, after separate incubation of different metal ions in AuAgF solution followed by the addition of excess Na2-EDTA and NH4HF2, Hg(II) exclusively quenches (with successive bathochromic shifts) the fluorescence of AuAgF solution qualifying the AuAgF solution as a potential sensing platform [Scheme 1] for Hg(II) [Fig. 2C]. In other words, fluorescence of AuAgF is altered only with Hg(II), Fe(III) and Cu(II) as indicated in Fig. 1. The interference due to Fe(III) and Cu(II) has been eliminated by masking them with NH4HF2 and Na2-EDTA, respectively [Fig. 2]. Other cations, including Cd(II) and Zn(II), have no effect on the fluorescence property of AuAgF. Again, they do not affect the
Fig. 2 Effect of addition of (A) Na2-EDTA and (B) NH4HF2 on the fluorescence of AuAgF solution after the quenching of fluorescence by Cu(II), Fe(III) and Hg(II). (C) Comparative account of fluorescence alteration after incubation of AuAgF solution with different metal ions followed by addition of Na2-EDTA + NH4HF2. Conditions: AuAgF solution = 3 mL; [Mn+] = 10 × 10−6 M; λex = 400 nm.
Dalton Trans., 2014, 43, 11557–11565 | 11559
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Paper
Scheme 1 Giant fluorescent Au(I)core–(Ag2/Ag3)shell clusters as a highly selective sensing platform for Hg(II).
fluorescence of HgAuAgF indicating that they are inert in the context of Hg(II) detection. The lifetime measurement of the AuAgF solution implies that three different components 5.3 ns (24%), 0.3 (46%) ns and 75.5 ns (30%) are present with an average lifetime of 24 ns. After the addition of Hg(II), the three different components change to 6.6 ns (26.4%), 0.6 ns (33.7%) and 89 ns (39.8%) with an average lifetime of 71 ns [Fig. S2, ESI†]. So, the particles stay in the excited state for a longer time after being attached to Hg(II) ions which is unusual. Dynamic quenching reduces the lifetime while static quenching keeps the lifetime unaltered. We mention later that Au(I)core–Ag(0)shell particles are changed in the presence of Hg(II). The longer stability in the excited state of particles of the HgAuAgF solution may reduce the radiative decay rate by non-radiative transfer of energy to Hg(II) or solvent as reported by Lakowicz et al.45 As a result, dissipation of energy takes place as heat or lossy surface waves.45,46 The formation and fluorescence of silver clusters are well documented in the literature. The emissive nature of Ag clusters originates from the fact that electrons can travel from the submerged and quasi-continuum 5d band to the lowest unoccupied conduction band of Ag clusters, termed as an interband transition.2,6 The MALDI-tof analysis [Fig. S3, ESI†] detects the formation of Ag2 and Ag3 particles that are responsible for the origin of strong fluorescence. As MALDI is a hard ionization technique and fragmentation is associated with the MALDI laser desorption process, we have also performed LCMS mass analysis in ES+ and ES− mode confirming the presence of such tiny fluorescent Ag2 and Ag3 clusters [Fig. S4(a) and S4(b), ESI†].
11560 | Dalton Trans., 2014, 43, 11557–11565
Dalton Transactions
Not only mass, but also absorption and emission spectra support the formation of silver clusters. No plasmon band is observed for Ag2–Ag3 clusters as reported by different groups.10,47 The emission spectrum matches closely with similar tiny clusters reported by Maretti et al.48 Usually, silver clusters are unstable and their dissociation energy is similar to the excitation energy. The Au+ surface in our case provides a support to stabilize the clusters. Zheng et al.6 reported that Ag2–Ag8 clusters are highly emissive due to their interband transition. In the present case, Ag2–Ag3 clusters are likely to be responsible for the fluorescence of the giant clusters. The Au(I) surface helps to stabilize the Ag clusters48 which are otherwise unstable. Recently, aggregation induced emission of Au@Ag clusters has been reported.49 Gold clusters become more emissive in the presence of silver.49 In our case, no emission is observed for gold if the process is repeated without silver. It indicates that the (Ag2/Ag3)shell on Au(I)core is responsible for the intense fluorescence. XPS analysis [Fig. 3A] of the highly fluorescent AuAgF solution under freeze drying condition indicates the formation of Au(I) as well as Ag(0) [85.81 eV and 89.46 eV binding energy for Au(I) 4f7/2 and Au(I) 4f7/2; 369.20 eV and 375.20 eV binding energy for Ag(0) 3d5/2 and Ag(0) 3d3/2].21 In the HgAuAgF solution, the binding energies of both Au and Ag are shifted towards ∼1 eV lower energy indicating the tendency of reduction and oxidation, respectively. Besides, the binding energy of S(II) 2P3/2 is also decreased by ∼1 eV after the addition of mercury owing to the strong Hg–S affinity. Moreover, 101.32 and 105.14 eV binding energies of Hg in the HgAuAgF solution corrspond to Hg(II) 4f7/2 and Hg(II) 4f5/2, respectively.25 In other words, the +2 oxidation state of
Fig. 3 (A) XPS spectra of AuAgF: wide angle (a) and high resolution for the elements Au (b), Ag (c), S (d); (B) XPS spectra of HgAuAgF: wide angle (a) and high resolution for the elements Au (b), Ag (c), Hg (d), S (e).
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Dalton Transactions
mercury is unaltered in the presence of fluorescent Au(I) @Ag(0)–thiolate core–shell particles [Fig. 3B]. The area under the absorption plot is minutely reduced with the successive additions of Hg(II) owing to the penetration of Hg(II) in the Stern layer36 of the giant clusters [Fig. 4A]. In contrast, rapid quenching of fluorescence of the AuAgF solution with the successive additions of Hg(II) along with unique red shifting explains the basis of the superiority of the emission property over the absorption property for the highly selective and sensitive Hg(II) sensing in the solution phase. A linear correlation exists between (I0 − I)/I0 and the concentration of Hg(II) over the range 0–10 μM (R2 = 0.98). The LOD at an S/N ratio of 3 for Hg(II) was found to be 6 nM, which is far lower than the United States Environmental Protection Agency (EPA)38 permissible level [Fig. 4]. TEM imaging reveals that Au(I)@Ag2/Ag3 particles are large (600 nm), spherical and core–shell in nature. Silver clusters comprise the shell and gold is present at the core. EDAX analysis as well as elemental mapping (which collect only surface information) of an individual giant spherical particle reveal a negligible atomic percentage of gold and an abundance of silver [Fig. S5, ESI†] although the [Ag(I)] added is only 1.7 times higher than [Au(III)]. This observation supports the formation of a Aucore–Agshell particle where gold is deeply buried beneath the silver atoms. Fluorescent tiny silver clusters obtain long-term stability by being present on the positively charged Au+ surface, as reported by Maretti et al.48 Such a positively charged core is responsible for synergistic evolution of large fluorescent particles due to the transfer of electron density from Ag(0). After the addition of Hg(II) to AuAgF solution, the fluorescence is greatly quenched along with red shifting. TEM imaging of the HgAuAgF solution reveals that the particles are then aggregated and individual giant spherical clusters cannot be identified. Moreover, the core–shell nature of the particle is not observed which confirms the rupture of the assembly.
Paper
Then the EDAX analysis of the HgAuAgF solution indicates an appreciable percentage of gold along with mercury and silver. This observation indicates that Hg(II) reacts with Ag(0) present at the surface (in the form of Ag2–Ag3 clusters) of the fluorescent particles and exposes the Au(I) core, rupturing the core– shell structure of the particle [Fig. 5]. By UV or blue light exposure, the AuAgF solution exhibits a very bright yellowish-orange spherical fluorescence image after being drop-cast on a glass slide. After the addition of Hg(II), the aggregated nature of the faint red particles with greatly reduced brightness is evident from the fluorescence image [Fig. 6]. The fluorescent particles are so robust that they can be obtained as a fluorescent solid by vacuum drying. The dry mass is redispersed in different water miscible organic solvents (e.g. acetone, acetonitrile, ethanol, methanol, isopropanol etc.) and quite remarkable fluorescence enhancement is noticed in non-aqueous solvents. Acetone is found to be the most suited solvent in this context bestowing >6 times enhancement in fluorescence intensity in comparison to water. Aaddition of Hg(II) to the water miscible non-aqueous solvent systems also quenches their fluorescence. Surprisingly, the extent of fluorescence quenching is virtually the same for water and all the water miscible solvents, broadening the applicability of the fluorescent core shell particles for Hg(II) sensing in different media. But, the successive bathochromic shift of the fluorescence spectral profiles of the giant fluorescent particles is observed exclusively in aqueous medium with the increasing quantity of Hg(II) [Fig. S6 A1, A2, ESI†].
Fig. 5 TEM images of (A1) AuAgF and (B1) HgAuAgF solutions. EDAX spectra of (A2) AuAgF and (B2) HgAuAgF solutions.
Fig. 4 (A) Absorption spectra and (B) fluorescence spectra of AuAgF after the addition of different concentrations of Hg(II). (B1) Degree of fluorescence quenching [inset: the linear detection limit for 0–10 μM of Hg(II)] and (B2) shift of emission maxima upon addition of [Hg(II)]. Conditions: AuAgF solution = 3 mL, λex = 400 nm.
This journal is © The Royal Society of Chemistry 2014
Fig. 6
Fluorescence images of (A) AuAgF and (B) HgAuAgF solutions.
Dalton Trans., 2014, 43, 11557–11565 | 11561
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Paper
Dalton Transactions
Not only water miscible solvents, but also water immiscible solvents have been employed for the preparation of fluorescent solutions. In the water pool of AOT/heptane reverse micelle (w0 = 10), fluorescent particles are again synthesized by the green chemical route under the sun. Then they are dried and redispersed in different water immiscible solvents e.g., ethyl acetate, heptane, tetrahydrofuran, toluene and chloroform. Like the aqueous solution, the efficient fluorescence quenching capability of Hg(II) is observed in water immiscible solvents. But, the extent of quenching is slightly less because of the coating of the reverse micelle. Again, the effect of different solvents on the size of the water pool in the reverse micelle explains the different extents of quenching unlike the water miscible solvents [Fig. S6 B1, B2, ESI†]. We have employed Hg(II) salts with various counter anions in the AuAgF solution and for all cases a similar red shifted quenched emission band is observed. This indicates that such a shift is indifferent to the nature of the counter anion and the alteration of fluorescence behavior is solely dependent on the Hg(II) ion. Fluorescent particles are also synthesized with S-lactoylglutathione in lieu of glutathione by employing a similar strategy as for AuAgF. A similar fluorescent solution is produced with an emission maxima of 570 nm. Addition of Hg(II) to such a fluorescent solution again shows the red shift along with fluorescence quenching like the AuAgF solution [Fig. 7]. Several reports50,51 are found where a blue shift is observed in the absorption maxima of silver nanoparticles after the introduction of Hg(II). A report by Wu et al.50 demonstrates the amalgam formed with Hg(II) and citrate capped silver nanoparticles causes blue shifting in the absorption spectra. Similarly, Ramesh and Radhakrishnan have reported a blue shift of a silver nanoparticle based polymer film in the presence of mercury.51 To the best of our knowledge, successive red shifts associated with quenching of fluorescence with Hg(II) is reported here for the first time. Being present on the Au(I) surface, the electron density of silver clusters is transferred towards gold rendering the silver to become positively charged. A zeta potential of +11.9 mV for the AuAgF fluorescent solution supports this fact. Now, the phenomenon of Hg(II) induced fluorescence quenching is ascribed to the 5d10(Hg2+)–4d10(Agδ+) metallophilic interaction, which is consistent with previous reports.36,37 It is important to mention in this context that such a d10–d10 interaction is also possible for
Cu(II) and Fe(III) explaining the quenching behavior of them.37 Guo and Irudayaraj37 have shown that the fluorescence of a denatured bovine serum albumin passivated silver cluster is quenched due to such a metallophilic interaction, but without any red shift. But, what causes the red shifting of emission maxima in our case in the presence of Hg(II)? It is important here to state that the evolution of fluorescent Au(I)core–(Ag2/Ag3)shell hydrosol is synergistic and irradiation of Au(III) or Ag(I) in the presence of glutathione under the sun only produces a fluorescent solution [Fig. S7, ESI†]. Sunlight irradiation of Au(III)/glutathione or Ag(I)/glutathione individually does not generate fluorescent particles. But, ageing of the irradiated Au(III)/glutathione (AuF) exhibits fluorescence to some extent after ∼6 days. Then, the fluorescence intensity is enhanced gradually with ageing. After ∼30 days of ageing [Fig. 8A], it becomes constant and further ageing causes almost no further change in fluorescence. A similar type of observation is noticed by Luo et al. at the time of synthesis of Au(0)@Au(I)–thiolate core–shell nanoclusters.52 But, the fluorescence intensity of AuF is 16 times lower than the AuAgF solution. Moreover, the emission maxima is 55 nm more red shifted than the AuAgF solution (λem 564 nm for AuAgF and 620 nm for AuF). Again, addition of Hg(II) to AuAgF solution causes a red shift (a maximum of ∼55 nm is possible by introducing Hg(II) to AuAgF solution) and fluorescence quenching. TEM imaging and EDAX spectral profile reveal the released Au destroys the core shell structure. It may be thought that Hg(II) generates AuF from AuAgF by removing Ag clusters from the surface. To investigate the fact, we have successively added dilute KCN solution (cyanide dissolution test53) to the fluorescent Au(I)core–(Ag2/Ag3)shell AuAgF solution. A gradual decrease of fluorescence intensity is observed due to the removal of silver from the shell with the formation of AgCN. But, no red shift [Fig. 8B] is there indicating that the generation of AuF solution is not responsible for such Hg(II) induced red shifting. A possible reason for the red shift of HgAuAgF is due to the aggregation of fluorescent clusters. With the increase of the n value in the Agn (n = number of atoms), a red shift of emission maxima has already been documented.54 Again, Hg(II) induced aggregation of silver particles is also found in the literature.55 In our case, TEM and fluorescence images support the Hg(II)
Fig. 7 (A) Effect of Hg(II) [10 × 10−6 M] with different counter anions on the fluorescence of 3 mL AuAgF. (B) Fluorescence spectra of (a) fluorescent solution (3 mL) obtained in presence of S-lactoylglutathione and (b) after addition of 10 × 10−6 M HgCl2.
Fig. 8 (A) Alteration of fluorescence behavior of 3 mL AuF solution upon ageing. (B) Effect of KCN on 3 mL AuAgF solution. Conditions: λex = 400 nm.
11562 | Dalton Trans., 2014, 43, 11557–11565
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Dalton Transactions
stimulated aggregation of the particles in aqueous medium. Metallophilic 5d10(Hg2+)–4d10(Agδ+) interaction and Hg(II) induced aggregation are attributed to be pivotal for such a surprising red shift associated with the fluorescence quenching. A continuous decrease of fluorescence intensity is observed with the increase in temperature for both AuAgF and HgAuAgF solutions which relates to pronounced Brownian motion56 of the particles. Subsequent cooling generally brings back the lost fluorescence, but in the present cases, recovery is not complete as expected [Fig. 9]. Cooling does not revert the original fluorescence intensity of AuAgF solution due to anomalous and irreversible aggregation of some silver clusters. This fact is substantiated from the IT/IRT vs. temperature (T ) studies [Fig. 9A1 in the inset] where heating and cooling cycles have different slopes (IT = Fluorescence intensity at temperature T; IRT = fluorescence intensity at room temperature). To some extent a similar observation has been noted for the HgAuAgF solution where less pronounced dislocation of Hg and more pronounced aggregation of some silver clusters occur simultaneously. Similar fluorescence reversal like the original AuAgF solution has been observed for the HgAuAgF solution during cooling above 30 °C. The cooling of the pre-heated HgAuAgF solution could not bring back dislocated Hg onto the giant particles but presumably some aggregated silver particles return to cluster form below 30 °C. Then enhanced fluorescence becomes much more pronounced than what was observed for AuAgF during cooling below 30 °C. This fluorescence reversal is demonstrated again from the IT/IRT vs. temperature (T ) studies [Fig. 9B1 in the inset] and heating–cooling cycles show similar slopes for HgAuAgF solution below 30 °C. We have performed a time dependent fluorescence study [Fig. 10] to monitor the Hg(II) induced quenching of fluorescence. As soon as Hg(II) is added to the fluorescent AuAgF solution, fluorescence is quenched without any red shift. It is due to a heavy metal effect associated with spin–orbit coupling.42 Then small increments in the fluorescence are observed up to 3 hours without any virtual change in the emission maxima. This is due to the combined effects of metallophilic interaction37 of mercury with silver clusters and the heavy metal effect. Then, again, successive quenching of florescence is observed with red shifting of the emission maxima. After
Paper
Fig. 10 Kinetics of Hg(II) [10 × 10−6 M] induced fluorescence quenching of 3 mL AuAgF solution. Conditions: λex = 400 nm.
ageing the reaction mixture for 9 hours, a constant emission peak (∼55 nm red shifted) is observed. Aggregation of silver clusters is responsible for this red shift phenomenon. The TEM study demonstrates that particle size is decreased greatly after 9 hours of incubation. These particles with smaller size (comprised of Hg, Ag, Au and capping agents) are not responsible for the red shift in the fluorescence spectra. However, the red shift is observed due to the aggregation of emissive silver clusters54 by Hg(II) present with the disintegrated particles. It is important to state in this context that no plasmon band for silver clusters is observed before and after Hg(II) addition signifying that nanocrystals are not formed.10 Standard reduction potentials of the species involved in the AgAuF and HgAuAgF solutions are as follows: GSSG þ 2e þ 2Hþ ! 2GSH 2Hg2þ þ 2e ! Hg2 2þ Hg2 2þ þ 2e ! 2Hg Agþ þ e ! Ag
E 0 ¼ 0:23 V E 0 ¼ 0:92 V
E 0 ¼ 0:85 V
E 0 ¼ 0:80 V
Au3þ þ 3e ! Au E 0 ¼ 1:50 V Au3þ þ 2e ! Auþ
E 0 ¼ 1:41 V
Auþ þ e ! Au E 0 ¼ 1:69 V
Fig. 9 (A) Effect of temperature on the fluorescence of the AuAgF solution, (A1) IT/IRT vs. temperature (indicating complete fluorescence reversal is not achieved). (B) Effect of temperature on the fluorescence of the HgAuAgF solution, (B1) IT/IRT vs. temperature (indicating complete fluorescence reversal is achieved below 30 °C).
This journal is © The Royal Society of Chemistry 2014
The XPS study reveals that in AuAgF solution Ag and Au are in the zero and +1 oxidation states, respectively. As the reduction potential of Hg(II) to Hg(0)/Hg(I) is quite high and the solution contains excess GSH which is prone to oxidation, Hg(0)/Hg(I) formation is feasible. Hg(0) can make an amalgam57 with silver clusters resulting in aggregation. There is also a possibility that Hg(II)/Hg(I) is reduced to Hg(0) by oxidizing Ag(0) due to galvanic replacement.58 The shift of binding energy towards lower value indicates the oxidation of Ag clusters. Au(I) is exposed to the solution after Hg treatment [evident from TEM images and EDAX analysis] and gets reduced to its Au(0) state to some extent. It is also supported by the high reduction potential value of the Au(0)/Au+ system.
Dalton Trans., 2014, 43, 11557–11565 | 11563
View Article Online
Paper
Dalton Transactions
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
The shift of binding energy of gold towards a lower energy value indicates the reduction of Au+ species. XPS spectra of HgAuAgF solution under freeze drying condition supports the +2 oxidation state of Hg.25 So, Hg(0)/Hg(I) present in the solution reduces Au+. As a result, Hg(0)/Hg(I) is oxidized to Hg2+ due to galvanic replacement. The probable electrochemical reactions are as follows: 2Hg2þ þ 2GSH ! Hg2 2þ þ GSSG þ 2Hþ Hg2 2þ þ 2GSH ! 2Hg þ GSSG þ Hþ 2Ag þ Hg2 2þ ! 2Agþ þ 2Hg
ing to our procedure and the as obtained results are summarized in Table 1. Values obtained from our proposed method show good agreement with the observed values determined by AFS, implying that the present sensing application works reliably for environmental samples. Usually large aggregated metal particles cause metal enhanced fluorescence45,46,59,60 while ultrasmall6 (≤2 nm) metal particles exhibit fluorescence. In the present work, fluorescent tiny silver particles on a Au(I) surface make the core–shell particles robust as well as fluorescent. Hg(II) selectively kills the fluorescence with unique red shifts, making the giant clusters an important solution based sensing platform for the detection of Hg(II).
2Ag þ Hg2þ ! 2Agþ þ Hg
Conclusions
2Auþ þ 2GSH ! 2Au þ GSSG þ 2Hþ Auþ þ Ag ! Au þ Agþ 2Auþ þ 2Hg ! 2Au þ Hg2 2þ 2Auþ þ Hg ! 2Au þ Hg2þ To test the practical application of our proposed method, several water samples were collected and spiked with Hg(II) with concentrations of 0, 20, 40, and 80 nM. Our proposed method and Atomic Fluorescence Spectroscopy (AFS) have been employed to test them. Tap water, river water, and lake water samples have been used as test samples. Filtration by qualitative filter paper and then centrifugation for 20 min at 14 000 rpm has been carried out for the river and lake water samples before analysis. The total concentrations of mercury in river water and lake water samples have been measured by AFS to be less than 0.1 nM before spiking. The samples spiked with various concentrations of Hg(II) have been employed for detection purposes with four replicate measurements accordTable 1 Determination of Hg(II) in environmental water samples with our proposed strategy and AFS
Hg(II) (nM) Sample Tap water 1 Tap water 2 Tap water 3 Tap water 4 River water 1 River water 2 River water 3 River water 4 Lake water 1 Lake water 2 Lake water 3 Lake water 4
Added 0 20 40 80 0 20 40 80 0 20 40 80
Proposed method (Meana ± SDb) c
20.1 ± 0.24 40.3 ± 0.39 80.3 ± 0.48 c
20.3 ± 0.18 39.7 ± 0.30 80.0 ± 0.36
c
19.6 ± 0.21 40.1 ± 0.52 79.5 ± 0.60
a
Acknowledgements The authors are thankful to the UGC, DST, NST, and CSIR, New Delhi, India, and the IIT Kharagpur for financial assistance. The authors are also thankful to Miss. Isozaki of Tokyo University of Science, Tokyo, Japan, for XPS measurement.
AFS (Mean ± SD)
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
PAPER
Cite this: Dalton Trans., 2014, 43, 11557
View Journal | View Issue
Fluorescent Au(I)@Ag2/Ag3 giant cluster for selective sensing of mercury(II) ion† Mainak Ganguly,a Chanchal Mondal,b Jaya Pal,a Anjali Pal,b Yuchi Negishic and Tarasankar Pal*a Highly stable Au(I)core–Ag(0)shell particles have been synthesized in aqueous solution via a green chemistry pathway utilising sunlight irradiation. The shell of the particles is composed of fluorescent Ag2 and Ag3 clusters which make the large core–shell particles highly fluorescent. The Au(I) core of the particles offers long-term stability to the silver clusters, which are otherwise unstable in solution at room temperature, by the transfer of electron density from the shell. Successive additions of Hg(II) ions to the fluorescent solution cause efficient and selective quenching of the fluorescence with gradual red shifting of the emission peak. The metallophilic 5d10(Hg2+)–4d10(Agδ+) interaction as well as Hg(II) stimulated aggregation have been ascribed to causing the fluorescence quenching and red shift. The fluorescent Au(I)core–Ag(0)shell particles are a highly selective and sensitive sensing platform for the detection of Hg(II) down to 6 nM in the presence of various metal ions. The detection limit is far below the permissible level as determined by
Received 21st April 2014, Accepted 6th May 2014
the EPA. Interferences due to Cu(II) and Fe(III) have been eliminated using Na2-EDTA and NH4HF2, respect-
DOI: 10.1039/c4dt01158a
ively. The fluorescent particles are successfully transferred to various solvent systems making Hg(II) determination also possible in non-aqueous media. Finally, the temperature dependent fluorescence change
www.rsc.org/dalton
with and without Hg(II) provides information about the metallophilic interaction.
Introduction Noble metal nanoclusters with a size smaller than 2 nm are gaining importance day by day in various fields like sensing,1,2 labeling,3 catalysis,4 antibacterial activity5 etc. The clusters possess molecule-like properties leading to the generation of fluorescence emission.6 Gold clusters exhibit interesting luminescent properties and have been extensively studied for their exciting optical and structural properties.7,8 In spite of poor photostability towards oxidation, many reports are found showing a stronger fluorescence signal of silver clusters compared to gold confined in the same matrix.9,10 To date, several compounds have been reported for synthesizing Ag clusters.10–22 Polyelectrolytes such as poly(amidoamine) dendrimer, poly(methacrylic acid), and poly(acrylic acid) have been widely employed for obtaining fluorescent Ag clusters in aqueous medium.16,17 Thiolate passivated Ag clusters a Department of Chemistry, Indian Institute of Technology, Kharagpur-721302, India. E-mail: [email protected]; Fax: (+) 91-03222-255303 b Department of Civil Engineering, Indian Institute of Technology, Kharagpur-721302, India c Department of Applied Chemistry, Tokyo University of Science, Tokyo-1628601, Japan † Electronic supplementary information (ESI) available: MALDI-TOF mass analysis, LCMS, elemental mapping, fluorescence spectra. See DOI: 10.1039/ c4dt01158a
This journal is © The Royal Society of Chemistry 2014
have been reported only recently.11–14 The inherent biocompatibility with biomolecules has encouraged researchers nowadays in this context for in vivo applications.18–22 Highly fluorescent and emission tunable Ag clusters encapsulated by DNA have been prepared by Dickson and coworkers.19,20 However, these DNA-encapsulated Ag clusters are not stable under high salt concentration preventing their extensive and extended applications. Although mercury poisoning is a serious threat to human health, global mercury emissions continue to rise with an alarming pace.23 Solid waste incineration and the combustion of fossil fuels are significant sources of mercury emissions.24 The long atmospheric lifetime of mercury is responsible for contamination across vast quantities of land and water.25 The problem becomes worse as bacteria convert elemental and ionic mercury to methyl mercury, which is a potent neurotoxin to the food chain. Serious sensory, motor, and cognitive disorders in human beings are the most severe outcomes of mercury poisoning.26–28 In environmental and health monitoring, detection of Hg(II) is undoubtedly important. As a consequence, different sensing platforms for Hg(II) detection such as small fluorescent organic molecules, various polymers, liposomes, DNAzymes, proteins, oligonucleotides, inorganic materials etc. have been developed.29–34 Noble metal fluorescent clusters of gold and silver to detect Hg(II) have become a hot field of
Dalton Trans., 2014, 43, 11557–11565 | 11557
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Paper
research.2,35,36 Guo and Irudayaraj identified the 5d10(Hg2+)– 4d10(Ag+) metallophilic interaction to be responsible for selective fluorescence quenching of silver clusters by Hg(II).37 In our present study, we have synthesized Au(I) supported fluorescent Ag clusters (AuAgF) in solution via a green chemical approach under solar light irradiation. The ultra small fluorescent Ag2 and Ag3 clusters, responsible for the fluorescence, cover the Au(I) moiety rendering highly fluorescent Au(I)core–Ag(0)shell giant clusters. The fluorescence intensity is greatly reduced specifically after the introduction of Hg(II) ions, causing successive red shifting of the emission maxima. The release of gold upon removal of fluorescent silver clusters by Hg(II) from the Au(I) surface is documented here. The rupture of the core–shell particle and subsequent fluorescence quenching make the core–shell particle a highly selective as well as sensitive Hg(II) sensor down to the detection limit of 6 nM, well below the EPA permissible level.38
Experimental section Materials and instruments All the reagents were of AR grade. Triple distilled water was used throughout the experiments. Chloroauric acid (HAuCl4), silver nitrate (AgNO3), glutathione (GSH), (R)-S-lactoylglutathione, ethylenediaminetetraacetic acid disodium salt dehydrate (Na2-EDTA), ammonium hydrogen difluoride (NH4HF2) and all the metal salts were purchased from Sigma-Aldrich. Dioctyl sodium sulfosuccinate (AOT) was obtained from Wako Pure Chemical Industries Ltd. All the solvents were purchased from Merck. All glassware was cleaned with freshly prepared aqua regia, subsequently rinsed with copious amount of distilled water and thoroughly dried before use. UV-vis absorption spectra were recorded on a SPECTRASCAN UV 2600 digital spectrophotometer (Chemito, India). The fluorescence measurements were performed at room temperature using a LS55 fluorescence spectrometer (Perkin Elmer, USA). A fluorescence microscope (Olympus DP72) was used to take fluorescence images. Particle morphology was observed using a field emission scanning electron microscope (Supra 40, Carl ZEISS Pvt. Ltd). X-ray photoelectron spectroscopy (XPS) analysis was done with a VG Scientific ESCALAB MK II spectrometer (UK) equipped with Mg Kα excitation source (1253.6 eV) and a five-channeltron detection system. For the measurement of ζ potential, a Malvern Nano ZS instrument employing a 4 mW He–Ne laser operating at a wavelength of 633 nm was used. TEM analyses were carried out with a H-9000 NAR, Hitachi instrument, using an accelerating voltage of 300 kV. Atomic fluorescence measurements were performed on an atomic fluorescence spectrometer (AFS-9700, Beijing, China). Synthesis of fluorescent solution A milky white suspension was observed as soon as 2.0 mL 10−2 M HAuCl4 was added to 60 mL 16 × 10−4 M aqueous glutathione. After solar light irradiation along with vigorous stirring
11558 | Dalton Trans., 2014, 43, 11557–11565
Dalton Transactions
for ∼6 hours, 3.4 mL 10−2 M AgNO3 was added. Then the mixture was kept under the sun for a further ∼3 hours with vigorous stirring. A straw yellow solution (AuAgF) is obtained with strong fluorescence (λem = 564 nm). Aliquots of fluorescing 3 mL AuAgF and HgAuAgF [partially fluorescence quenched by 10 × 10−6 M Hg(II) after ageing for 9 hours] solutions were used thoroughout the course of the study. Synthesis of fluorescent particles by AOT mediated route For the preparation of the fluorescing species in water immiscible organic solvents, 0.1 M AOT in heptane was introduced. In the 6 mL AOT–heptane solution, GSH, aqueous chloroauric acid and silver nitrate were added to obtain the same final concentrations as mentioned for AuAgF. w0 i.e., the ratio of the number of water and AOT molecules was maintained to be 10. The solution was exposed to the sun for ∼15 hours with vigorous stirring to achieve a fluorescent yellow solution. The fluorescent solution was vacuum dried and redispersed in various water immiscible organic solvents.
Results and discussion The water soluble Au(I)@Ag2/Ag3 giant cluster has been synthesized by sunlight irradiation in the presence of AgNO3, HAuCl4 and GSH precursors. Thus, the green chemically evolved yellow solution (AuAgF) exhibits an intense fluorescence (quantum yield 6% using quinine sulfate in 0.1 M H2SO4 as reference and highly stable with unaltered emissive property with respect to intensity and emission maximum for months) with the emission maxima ∼564 nm when the solution is excited at ∼400 nm. Consequently, a large Stokes shift of 164 nm is observed. It is intriguing to note that the presence of both gold and silver are essential to obtain such giant fluorescent Au(I)core–Ag(0)shell clusters. The absence of any one of the metals i.e., gold or silver, evolves a non-fluorescent solution proving the synthetic tactic to be synergistic. The Stokes shift can be further increased by increasing the concentration of Hg(II). In other words, a gradual red shift of the emission maxima is witnessed along with strong and successive quenching of the fluorescence intensity. The fluorescence of the AuAgF solution is quenched by the Hg(II) ion. The Hg(II)–AuAgF mixed solution reaches a stable fluorescence condition after a waiting time of 9 hours. The quenched fluorescence is associated with the concentration dependent red shifting of the emission maxima. Under the experimental conditions employed, we introduced 10 × 10−6 M Hg(II) into 3 mL AuAgF solution and the solution exhibits a 55 nm red shifted peak after 9 hours. This becomes the representative test solution (HgAuAgF) for investigation. Here the quenching is 73%. In the case of other metal ions, virtually no alteration of the fluorescence intensity is found except for Cu(II) and Fe(III). Fig. 1A and 1A1 display the comparative effect of different metal ions on the fluorescence intensity of the as produced fluorescent solution [I0 = fluorescence intensity of AuAgF,
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Dalton Transactions
Fig. 1 (A) Fluorescence spectral profile and (A1) bar diagram showing the effect of different metal ions on the fluorescence of the AuAgF solution. (B) Fluorescence spectral profile and (B1) bar diagram showing the effect of different metal ions on the fluorescence of the HgAuAgF solution. Conditions: AuAgF solution = 3 mL; [Mn+] = 10 × 10−6 M; λex = 400 nm.
I = fluorescence intensity of AuAgF after individual addition of different metal ions]. It is worth noting that Cu(II) as well as Fe(III) result in fluorescence quenching like Hg(II). But, the phenomenon of quenching by Cu(II) and Fe(III) is not as efficient as Hg(II). Besides, the gradual red shift with increased concentration is unique for mercury, showing promise for designing a solution based sensing platform for mercury. Now, to investigate the effect of various metal ions on the Hg(II) induced quenched fluorescence (HgAuAgF), we have found that it is almost indifferent to other interfering metal ions except for Cu(II) and Fe(III). For both the Cu(II) and Fe(III) cases, the fluorescence of HgAuAgF is further decreased [Fig. 1B and 1B1]. In this context, it can be stated that Fe(III) exhibits a significant quenching capability towards the AuAgF solution. But, the quenching effect with Fe(II) is insignificant. Some efforts39,40 have been devoted to distinguish between Fe(II) and Fe(III). The distinction between Fe(II) and Fe(III) has been achieved fluorometrically even at very low concentration (nM level) exploiting the as-prepared fluorescent giant cluster (AuAgF solution) [Fig. S1, ESI†]. The quenching capability of Hg(II) is clearly much higher than Fe(III) and Cu(II). Nevertheless, we have introduced Na2EDTA and NH4HF2 to eliminate the interference due to Cu(II) and Fe(III), respectively. It may be noted that these two masking agents cause no observable interference to the fluorescence of AuAgF. When we introduced excess Na2-EDTA to the aqueous fluorescent AuAgF solutions containing the three quencher metal ions [Cu(II), Fe(III), Hg(II)] individually, appreciable turn on fluorescence was noticed only for Cu(II) owing to a reasonable stability constant of the Cu-EDTA complex. Surprisingly, the stability constants for Cu(II), Fe(III) and Hg(II) with EDTA are 18.8, 25.7 and 21.5, respectively. This indicates that the affinity of Cu(II) to the fluorescent giant cluster is
This journal is © The Royal Society of Chemistry 2014
Paper
comparatively low unlike Fe(III) and Hg(II). Complexation of Cu(II) with EDTA regenerates the lost fluorescence of the AuAgF solution although the stability constant of the EDTA complex is comparatively low for Cu(II) [Fig. 2A]. Quenching may also be due to free Cu(II) ions in solution exerting spin–orbit coupling.41,42 It may also be due to the formation of a weak complex43 of copper with glutathione decreasing the capping capability for the Au(I)@Ag2/Ag3 giant cluster. The fluorescence re-appears when the copper forms an EDTA complex leaving free GSH. To eliminate the interference of Fe(III), NH4HF2 has been added after metal ion addition [Fig. 2B]. Decent ‘turn on’ fluorescence is observed in the solution in which fluorescence was originally quenched by Fe(III). The HF2− restores the fluorescence due to the formation of highly stable [FeF6]3−.44 Besides, for both the Na2-EDTA and NH4HF2 treatments, Hg(II) stimulated quenched fluorescence remains indifferent. Consequently, the discrimination of the three metal ions [Hg(II), Fe(III), Cu(II)] causing quenching has been achieved with the addition of Na2-EDTA and NH4HF2. So, after separate incubation of different metal ions in AuAgF solution followed by the addition of excess Na2-EDTA and NH4HF2, Hg(II) exclusively quenches (with successive bathochromic shifts) the fluorescence of AuAgF solution qualifying the AuAgF solution as a potential sensing platform [Scheme 1] for Hg(II) [Fig. 2C]. In other words, fluorescence of AuAgF is altered only with Hg(II), Fe(III) and Cu(II) as indicated in Fig. 1. The interference due to Fe(III) and Cu(II) has been eliminated by masking them with NH4HF2 and Na2-EDTA, respectively [Fig. 2]. Other cations, including Cd(II) and Zn(II), have no effect on the fluorescence property of AuAgF. Again, they do not affect the
Fig. 2 Effect of addition of (A) Na2-EDTA and (B) NH4HF2 on the fluorescence of AuAgF solution after the quenching of fluorescence by Cu(II), Fe(III) and Hg(II). (C) Comparative account of fluorescence alteration after incubation of AuAgF solution with different metal ions followed by addition of Na2-EDTA + NH4HF2. Conditions: AuAgF solution = 3 mL; [Mn+] = 10 × 10−6 M; λex = 400 nm.
Dalton Trans., 2014, 43, 11557–11565 | 11559
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Paper
Scheme 1 Giant fluorescent Au(I)core–(Ag2/Ag3)shell clusters as a highly selective sensing platform for Hg(II).
fluorescence of HgAuAgF indicating that they are inert in the context of Hg(II) detection. The lifetime measurement of the AuAgF solution implies that three different components 5.3 ns (24%), 0.3 (46%) ns and 75.5 ns (30%) are present with an average lifetime of 24 ns. After the addition of Hg(II), the three different components change to 6.6 ns (26.4%), 0.6 ns (33.7%) and 89 ns (39.8%) with an average lifetime of 71 ns [Fig. S2, ESI†]. So, the particles stay in the excited state for a longer time after being attached to Hg(II) ions which is unusual. Dynamic quenching reduces the lifetime while static quenching keeps the lifetime unaltered. We mention later that Au(I)core–Ag(0)shell particles are changed in the presence of Hg(II). The longer stability in the excited state of particles of the HgAuAgF solution may reduce the radiative decay rate by non-radiative transfer of energy to Hg(II) or solvent as reported by Lakowicz et al.45 As a result, dissipation of energy takes place as heat or lossy surface waves.45,46 The formation and fluorescence of silver clusters are well documented in the literature. The emissive nature of Ag clusters originates from the fact that electrons can travel from the submerged and quasi-continuum 5d band to the lowest unoccupied conduction band of Ag clusters, termed as an interband transition.2,6 The MALDI-tof analysis [Fig. S3, ESI†] detects the formation of Ag2 and Ag3 particles that are responsible for the origin of strong fluorescence. As MALDI is a hard ionization technique and fragmentation is associated with the MALDI laser desorption process, we have also performed LCMS mass analysis in ES+ and ES− mode confirming the presence of such tiny fluorescent Ag2 and Ag3 clusters [Fig. S4(a) and S4(b), ESI†].
11560 | Dalton Trans., 2014, 43, 11557–11565
Dalton Transactions
Not only mass, but also absorption and emission spectra support the formation of silver clusters. No plasmon band is observed for Ag2–Ag3 clusters as reported by different groups.10,47 The emission spectrum matches closely with similar tiny clusters reported by Maretti et al.48 Usually, silver clusters are unstable and their dissociation energy is similar to the excitation energy. The Au+ surface in our case provides a support to stabilize the clusters. Zheng et al.6 reported that Ag2–Ag8 clusters are highly emissive due to their interband transition. In the present case, Ag2–Ag3 clusters are likely to be responsible for the fluorescence of the giant clusters. The Au(I) surface helps to stabilize the Ag clusters48 which are otherwise unstable. Recently, aggregation induced emission of Au@Ag clusters has been reported.49 Gold clusters become more emissive in the presence of silver.49 In our case, no emission is observed for gold if the process is repeated without silver. It indicates that the (Ag2/Ag3)shell on Au(I)core is responsible for the intense fluorescence. XPS analysis [Fig. 3A] of the highly fluorescent AuAgF solution under freeze drying condition indicates the formation of Au(I) as well as Ag(0) [85.81 eV and 89.46 eV binding energy for Au(I) 4f7/2 and Au(I) 4f7/2; 369.20 eV and 375.20 eV binding energy for Ag(0) 3d5/2 and Ag(0) 3d3/2].21 In the HgAuAgF solution, the binding energies of both Au and Ag are shifted towards ∼1 eV lower energy indicating the tendency of reduction and oxidation, respectively. Besides, the binding energy of S(II) 2P3/2 is also decreased by ∼1 eV after the addition of mercury owing to the strong Hg–S affinity. Moreover, 101.32 and 105.14 eV binding energies of Hg in the HgAuAgF solution corrspond to Hg(II) 4f7/2 and Hg(II) 4f5/2, respectively.25 In other words, the +2 oxidation state of
Fig. 3 (A) XPS spectra of AuAgF: wide angle (a) and high resolution for the elements Au (b), Ag (c), S (d); (B) XPS spectra of HgAuAgF: wide angle (a) and high resolution for the elements Au (b), Ag (c), Hg (d), S (e).
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Dalton Transactions
mercury is unaltered in the presence of fluorescent Au(I) @Ag(0)–thiolate core–shell particles [Fig. 3B]. The area under the absorption plot is minutely reduced with the successive additions of Hg(II) owing to the penetration of Hg(II) in the Stern layer36 of the giant clusters [Fig. 4A]. In contrast, rapid quenching of fluorescence of the AuAgF solution with the successive additions of Hg(II) along with unique red shifting explains the basis of the superiority of the emission property over the absorption property for the highly selective and sensitive Hg(II) sensing in the solution phase. A linear correlation exists between (I0 − I)/I0 and the concentration of Hg(II) over the range 0–10 μM (R2 = 0.98). The LOD at an S/N ratio of 3 for Hg(II) was found to be 6 nM, which is far lower than the United States Environmental Protection Agency (EPA)38 permissible level [Fig. 4]. TEM imaging reveals that Au(I)@Ag2/Ag3 particles are large (600 nm), spherical and core–shell in nature. Silver clusters comprise the shell and gold is present at the core. EDAX analysis as well as elemental mapping (which collect only surface information) of an individual giant spherical particle reveal a negligible atomic percentage of gold and an abundance of silver [Fig. S5, ESI†] although the [Ag(I)] added is only 1.7 times higher than [Au(III)]. This observation supports the formation of a Aucore–Agshell particle where gold is deeply buried beneath the silver atoms. Fluorescent tiny silver clusters obtain long-term stability by being present on the positively charged Au+ surface, as reported by Maretti et al.48 Such a positively charged core is responsible for synergistic evolution of large fluorescent particles due to the transfer of electron density from Ag(0). After the addition of Hg(II) to AuAgF solution, the fluorescence is greatly quenched along with red shifting. TEM imaging of the HgAuAgF solution reveals that the particles are then aggregated and individual giant spherical clusters cannot be identified. Moreover, the core–shell nature of the particle is not observed which confirms the rupture of the assembly.
Paper
Then the EDAX analysis of the HgAuAgF solution indicates an appreciable percentage of gold along with mercury and silver. This observation indicates that Hg(II) reacts with Ag(0) present at the surface (in the form of Ag2–Ag3 clusters) of the fluorescent particles and exposes the Au(I) core, rupturing the core– shell structure of the particle [Fig. 5]. By UV or blue light exposure, the AuAgF solution exhibits a very bright yellowish-orange spherical fluorescence image after being drop-cast on a glass slide. After the addition of Hg(II), the aggregated nature of the faint red particles with greatly reduced brightness is evident from the fluorescence image [Fig. 6]. The fluorescent particles are so robust that they can be obtained as a fluorescent solid by vacuum drying. The dry mass is redispersed in different water miscible organic solvents (e.g. acetone, acetonitrile, ethanol, methanol, isopropanol etc.) and quite remarkable fluorescence enhancement is noticed in non-aqueous solvents. Acetone is found to be the most suited solvent in this context bestowing >6 times enhancement in fluorescence intensity in comparison to water. Aaddition of Hg(II) to the water miscible non-aqueous solvent systems also quenches their fluorescence. Surprisingly, the extent of fluorescence quenching is virtually the same for water and all the water miscible solvents, broadening the applicability of the fluorescent core shell particles for Hg(II) sensing in different media. But, the successive bathochromic shift of the fluorescence spectral profiles of the giant fluorescent particles is observed exclusively in aqueous medium with the increasing quantity of Hg(II) [Fig. S6 A1, A2, ESI†].
Fig. 5 TEM images of (A1) AuAgF and (B1) HgAuAgF solutions. EDAX spectra of (A2) AuAgF and (B2) HgAuAgF solutions.
Fig. 4 (A) Absorption spectra and (B) fluorescence spectra of AuAgF after the addition of different concentrations of Hg(II). (B1) Degree of fluorescence quenching [inset: the linear detection limit for 0–10 μM of Hg(II)] and (B2) shift of emission maxima upon addition of [Hg(II)]. Conditions: AuAgF solution = 3 mL, λex = 400 nm.
This journal is © The Royal Society of Chemistry 2014
Fig. 6
Fluorescence images of (A) AuAgF and (B) HgAuAgF solutions.
Dalton Trans., 2014, 43, 11557–11565 | 11561
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Paper
Dalton Transactions
Not only water miscible solvents, but also water immiscible solvents have been employed for the preparation of fluorescent solutions. In the water pool of AOT/heptane reverse micelle (w0 = 10), fluorescent particles are again synthesized by the green chemical route under the sun. Then they are dried and redispersed in different water immiscible solvents e.g., ethyl acetate, heptane, tetrahydrofuran, toluene and chloroform. Like the aqueous solution, the efficient fluorescence quenching capability of Hg(II) is observed in water immiscible solvents. But, the extent of quenching is slightly less because of the coating of the reverse micelle. Again, the effect of different solvents on the size of the water pool in the reverse micelle explains the different extents of quenching unlike the water miscible solvents [Fig. S6 B1, B2, ESI†]. We have employed Hg(II) salts with various counter anions in the AuAgF solution and for all cases a similar red shifted quenched emission band is observed. This indicates that such a shift is indifferent to the nature of the counter anion and the alteration of fluorescence behavior is solely dependent on the Hg(II) ion. Fluorescent particles are also synthesized with S-lactoylglutathione in lieu of glutathione by employing a similar strategy as for AuAgF. A similar fluorescent solution is produced with an emission maxima of 570 nm. Addition of Hg(II) to such a fluorescent solution again shows the red shift along with fluorescence quenching like the AuAgF solution [Fig. 7]. Several reports50,51 are found where a blue shift is observed in the absorption maxima of silver nanoparticles after the introduction of Hg(II). A report by Wu et al.50 demonstrates the amalgam formed with Hg(II) and citrate capped silver nanoparticles causes blue shifting in the absorption spectra. Similarly, Ramesh and Radhakrishnan have reported a blue shift of a silver nanoparticle based polymer film in the presence of mercury.51 To the best of our knowledge, successive red shifts associated with quenching of fluorescence with Hg(II) is reported here for the first time. Being present on the Au(I) surface, the electron density of silver clusters is transferred towards gold rendering the silver to become positively charged. A zeta potential of +11.9 mV for the AuAgF fluorescent solution supports this fact. Now, the phenomenon of Hg(II) induced fluorescence quenching is ascribed to the 5d10(Hg2+)–4d10(Agδ+) metallophilic interaction, which is consistent with previous reports.36,37 It is important to mention in this context that such a d10–d10 interaction is also possible for
Cu(II) and Fe(III) explaining the quenching behavior of them.37 Guo and Irudayaraj37 have shown that the fluorescence of a denatured bovine serum albumin passivated silver cluster is quenched due to such a metallophilic interaction, but without any red shift. But, what causes the red shifting of emission maxima in our case in the presence of Hg(II)? It is important here to state that the evolution of fluorescent Au(I)core–(Ag2/Ag3)shell hydrosol is synergistic and irradiation of Au(III) or Ag(I) in the presence of glutathione under the sun only produces a fluorescent solution [Fig. S7, ESI†]. Sunlight irradiation of Au(III)/glutathione or Ag(I)/glutathione individually does not generate fluorescent particles. But, ageing of the irradiated Au(III)/glutathione (AuF) exhibits fluorescence to some extent after ∼6 days. Then, the fluorescence intensity is enhanced gradually with ageing. After ∼30 days of ageing [Fig. 8A], it becomes constant and further ageing causes almost no further change in fluorescence. A similar type of observation is noticed by Luo et al. at the time of synthesis of Au(0)@Au(I)–thiolate core–shell nanoclusters.52 But, the fluorescence intensity of AuF is 16 times lower than the AuAgF solution. Moreover, the emission maxima is 55 nm more red shifted than the AuAgF solution (λem 564 nm for AuAgF and 620 nm for AuF). Again, addition of Hg(II) to AuAgF solution causes a red shift (a maximum of ∼55 nm is possible by introducing Hg(II) to AuAgF solution) and fluorescence quenching. TEM imaging and EDAX spectral profile reveal the released Au destroys the core shell structure. It may be thought that Hg(II) generates AuF from AuAgF by removing Ag clusters from the surface. To investigate the fact, we have successively added dilute KCN solution (cyanide dissolution test53) to the fluorescent Au(I)core–(Ag2/Ag3)shell AuAgF solution. A gradual decrease of fluorescence intensity is observed due to the removal of silver from the shell with the formation of AgCN. But, no red shift [Fig. 8B] is there indicating that the generation of AuF solution is not responsible for such Hg(II) induced red shifting. A possible reason for the red shift of HgAuAgF is due to the aggregation of fluorescent clusters. With the increase of the n value in the Agn (n = number of atoms), a red shift of emission maxima has already been documented.54 Again, Hg(II) induced aggregation of silver particles is also found in the literature.55 In our case, TEM and fluorescence images support the Hg(II)
Fig. 7 (A) Effect of Hg(II) [10 × 10−6 M] with different counter anions on the fluorescence of 3 mL AuAgF. (B) Fluorescence spectra of (a) fluorescent solution (3 mL) obtained in presence of S-lactoylglutathione and (b) after addition of 10 × 10−6 M HgCl2.
Fig. 8 (A) Alteration of fluorescence behavior of 3 mL AuF solution upon ageing. (B) Effect of KCN on 3 mL AuAgF solution. Conditions: λex = 400 nm.
11562 | Dalton Trans., 2014, 43, 11557–11565
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
Dalton Transactions
stimulated aggregation of the particles in aqueous medium. Metallophilic 5d10(Hg2+)–4d10(Agδ+) interaction and Hg(II) induced aggregation are attributed to be pivotal for such a surprising red shift associated with the fluorescence quenching. A continuous decrease of fluorescence intensity is observed with the increase in temperature for both AuAgF and HgAuAgF solutions which relates to pronounced Brownian motion56 of the particles. Subsequent cooling generally brings back the lost fluorescence, but in the present cases, recovery is not complete as expected [Fig. 9]. Cooling does not revert the original fluorescence intensity of AuAgF solution due to anomalous and irreversible aggregation of some silver clusters. This fact is substantiated from the IT/IRT vs. temperature (T ) studies [Fig. 9A1 in the inset] where heating and cooling cycles have different slopes (IT = Fluorescence intensity at temperature T; IRT = fluorescence intensity at room temperature). To some extent a similar observation has been noted for the HgAuAgF solution where less pronounced dislocation of Hg and more pronounced aggregation of some silver clusters occur simultaneously. Similar fluorescence reversal like the original AuAgF solution has been observed for the HgAuAgF solution during cooling above 30 °C. The cooling of the pre-heated HgAuAgF solution could not bring back dislocated Hg onto the giant particles but presumably some aggregated silver particles return to cluster form below 30 °C. Then enhanced fluorescence becomes much more pronounced than what was observed for AuAgF during cooling below 30 °C. This fluorescence reversal is demonstrated again from the IT/IRT vs. temperature (T ) studies [Fig. 9B1 in the inset] and heating–cooling cycles show similar slopes for HgAuAgF solution below 30 °C. We have performed a time dependent fluorescence study [Fig. 10] to monitor the Hg(II) induced quenching of fluorescence. As soon as Hg(II) is added to the fluorescent AuAgF solution, fluorescence is quenched without any red shift. It is due to a heavy metal effect associated with spin–orbit coupling.42 Then small increments in the fluorescence are observed up to 3 hours without any virtual change in the emission maxima. This is due to the combined effects of metallophilic interaction37 of mercury with silver clusters and the heavy metal effect. Then, again, successive quenching of florescence is observed with red shifting of the emission maxima. After
Paper
Fig. 10 Kinetics of Hg(II) [10 × 10−6 M] induced fluorescence quenching of 3 mL AuAgF solution. Conditions: λex = 400 nm.
ageing the reaction mixture for 9 hours, a constant emission peak (∼55 nm red shifted) is observed. Aggregation of silver clusters is responsible for this red shift phenomenon. The TEM study demonstrates that particle size is decreased greatly after 9 hours of incubation. These particles with smaller size (comprised of Hg, Ag, Au and capping agents) are not responsible for the red shift in the fluorescence spectra. However, the red shift is observed due to the aggregation of emissive silver clusters54 by Hg(II) present with the disintegrated particles. It is important to state in this context that no plasmon band for silver clusters is observed before and after Hg(II) addition signifying that nanocrystals are not formed.10 Standard reduction potentials of the species involved in the AgAuF and HgAuAgF solutions are as follows: GSSG þ 2e þ 2Hþ ! 2GSH 2Hg2þ þ 2e ! Hg2 2þ Hg2 2þ þ 2e ! 2Hg Agþ þ e ! Ag
E 0 ¼ 0:23 V E 0 ¼ 0:92 V
E 0 ¼ 0:85 V
E 0 ¼ 0:80 V
Au3þ þ 3e ! Au E 0 ¼ 1:50 V Au3þ þ 2e ! Auþ
E 0 ¼ 1:41 V
Auþ þ e ! Au E 0 ¼ 1:69 V
Fig. 9 (A) Effect of temperature on the fluorescence of the AuAgF solution, (A1) IT/IRT vs. temperature (indicating complete fluorescence reversal is not achieved). (B) Effect of temperature on the fluorescence of the HgAuAgF solution, (B1) IT/IRT vs. temperature (indicating complete fluorescence reversal is achieved below 30 °C).
This journal is © The Royal Society of Chemistry 2014
The XPS study reveals that in AuAgF solution Ag and Au are in the zero and +1 oxidation states, respectively. As the reduction potential of Hg(II) to Hg(0)/Hg(I) is quite high and the solution contains excess GSH which is prone to oxidation, Hg(0)/Hg(I) formation is feasible. Hg(0) can make an amalgam57 with silver clusters resulting in aggregation. There is also a possibility that Hg(II)/Hg(I) is reduced to Hg(0) by oxidizing Ag(0) due to galvanic replacement.58 The shift of binding energy towards lower value indicates the oxidation of Ag clusters. Au(I) is exposed to the solution after Hg treatment [evident from TEM images and EDAX analysis] and gets reduced to its Au(0) state to some extent. It is also supported by the high reduction potential value of the Au(0)/Au+ system.
Dalton Trans., 2014, 43, 11557–11565 | 11563
View Article Online
Paper
Dalton Transactions
Published on 06 May 2014. Downloaded by North Dakota State University on 26/05/2015 18:02:21.
The shift of binding energy of gold towards a lower energy value indicates the reduction of Au+ species. XPS spectra of HgAuAgF solution under freeze drying condition supports the +2 oxidation state of Hg.25 So, Hg(0)/Hg(I) present in the solution reduces Au+. As a result, Hg(0)/Hg(I) is oxidized to Hg2+ due to galvanic replacement. The probable electrochemical reactions are as follows: 2Hg2þ þ 2GSH ! Hg2 2þ þ GSSG þ 2Hþ Hg2 2þ þ 2GSH ! 2Hg þ GSSG þ Hþ 2Ag þ Hg2 2þ ! 2Agþ þ 2Hg
ing to our procedure and the as obtained results are summarized in Table 1. Values obtained from our proposed method show good agreement with the observed values determined by AFS, implying that the present sensing application works reliably for environmental samples. Usually large aggregated metal particles cause metal enhanced fluorescence45,46,59,60 while ultrasmall6 (≤2 nm) metal particles exhibit fluorescence. In the present work, fluorescent tiny silver particles on a Au(I) surface make the core–shell particles robust as well as fluorescent. Hg(II) selectively kills the fluorescence with unique red shifts, making the giant clusters an important solution based sensing platform for the detection of Hg(II).
2Ag þ Hg2þ ! 2Agþ þ Hg
Conclusions
2Auþ þ 2GSH ! 2Au þ GSSG þ 2Hþ Auþ þ Ag ! Au þ Agþ 2Auþ þ 2Hg ! 2Au þ Hg2 2þ 2Auþ þ Hg ! 2Au þ Hg2þ To test the practical application of our proposed method, several water samples were collected and spiked with Hg(II) with concentrations of 0, 20, 40, and 80 nM. Our proposed method and Atomic Fluorescence Spectroscopy (AFS) have been employed to test them. Tap water, river water, and lake water samples have been used as test samples. Filtration by qualitative filter paper and then centrifugation for 20 min at 14 000 rpm has been carried out for the river and lake water samples before analysis. The total concentrations of mercury in river water and lake water samples have been measured by AFS to be less than 0.1 nM before spiking. The samples spiked with various concentrations of Hg(II) have been employed for detection purposes with four replicate measurements accordTable 1 Determination of Hg(II) in environmental water samples with our proposed strategy and AFS
Hg(II) (nM) Sample Tap water 1 Tap water 2 Tap water 3 Tap water 4 River water 1 River water 2 River water 3 River water 4 Lake water 1 Lake water 2 Lake water 3 Lake water 4
Added 0 20 40 80 0 20 40 80 0 20 40 80
Proposed method (Meana ± SDb) c
20.1 ± 0.24 40.3 ± 0.39 80.3 ± 0.48 c
20.3 ± 0.18 39.7 ± 0.30 80.0 ± 0.36
c
19.6 ± 0.21 40.1 ± 0.52 79.5 ± 0.60
a
Acknowledgements The authors are thankful to the UGC, DST, NST, and CSIR, New Delhi, India, and the IIT Kharagpur for financial assistance. The authors are also thankful to Miss. Isozaki of Tokyo University of Science, Tokyo, Japan, for XPS measurement.
AFS (Mean ± SD)
