Luminescence Quenching and Enhancement by Adsorbed Metal Ions with Ruthenium Diimine Complexes Immobilized on Silica Polyamine Composites

Summary

Luminescent ruthenium diimine complexes have been covalently bound to the surface of a silica polyamine composite (SPC) using peptide coupling agents. The loading of the complexes using this route is quite low (~0.01–0.04 mmol/g) leaving sufficient surface amines to coordinate added metal ions. When the composite particles containing the Ru complexes are exposed to solutions of Cu²⁺, Ni²⁺ or Zn²⁺, luminescence is quenched with efficiencies that follow concentration dependence and the relative binding affinities of the ions. When heavy metal ions such as mercury or lead are adsorbed onto the same surface, luminescence is enhanced by a factor of ~3.5. When the complexes are exposed to these metals in solution, no quenching or enhancement is observed. Both phenomena were shown to be the result of adsorption of the cations onto the polyamine surface by using the Stern-Volmer relationship. The mechanism of both quenching and enhancement is discussed and the options for further development of this novel metal sensing technique are presented.

Introduction

We have been studying the photophysical properties of a series of luminescent ruthenium diimine complexes with bisphosphine and other ancillary ligands. Our initial studies were aimed at improving the quantum yields and excited-state lifetimes of the triplet metal to ligand charge-transfer (³MLCT) emissions. The incorporation of phosphine and carbonyl ligands did indeed lead to much higher quantum yields (φ = 0.2 to 0.3; Ru(bpy)₃²⁺, φ = 0.06) in several cases while maintaining the long-lived excited-state lifetimes observed for the more conventional tris-diimine complexes (300–600 ns).^[1,2] These initial studies also reported the bioconjugation of the luminescent complexes and their incorporation into membrane model systems.^[2] In a subsequent study, the same peptide coupling chemistry used to make the bioconjugates was used to covalently bind the complexes to the surface of a silica polyamine composite (SPC) (vide infra).^[3] In both studies, the environment of the modified complexes had a profound influence on their photophysical properties. In the case of the lipid bioconjugates, shortening of the excited-state lifetime and blue shifting of the emission was noted only in the complexes that contained phosphines as ancillary ligands.^[2] In the case of the SPC bound complexes, the excited-state lifetimes increased 5–8 times over that observed in solution, but again only in the cases where the complex contained phosphines as ancillary ligands.^[3] The strong influence of the phosphine ligands was rationalized by proposing mixing of the metal to ligand charge transfer band(MLCT) with the π–π^* transitions of the diimine ligand as a result of phosphine participation in the ground state of the Ru atom, in complexes such as Ru(CO)(H)(trans-PPh₃)₂(η²-4,4′-dicarboxybipridine)][PF₆] (1) (Figure 1). This hypothesis is currently being explored computationally using time dependent density functional theory (TDDFT) methods.^[4] The related complex Ru(bpy)₂(η²-4,4′-dicar-boxybipridine)][PF₆]₂ (2) (Figure 1) showed no change in lifetime or emission wavelength when conjugated to a lipid or when bound to the SPC surface.^[2,3] This underlines the importance of the phosphine ligand in determining the photophysical properties of these luminescent Ru complexes.

Figure 1.

Structures of the luminescent complexes 1 and 2 discussed in this study.

Complex 1 was bound to the SPC surface via the carboxylate groups using the peptide coupling reagent HBTU, and complex 2 was first converted to an isothiocyanate by reaction with thiophosgene which reacts cleanly with the amine of the SPC (Figure 2). The loading of the complexes on the SPC surface was quite modest being ~0.013mmol/g for 1 and ~0.040mmol/g for 2.^[3]

Figure 2.

Proposed binding modes of 1 and 2 on the SPC surface.

Given that the amine loading on the SPC is 1.6 mmol/g, there are many uncoordinated amines on the surface that could potentially coordinate metal ions from aqueous solutions. The dramatic differences in the photophyscial properties of 1 and 2 observed after bioconjugation or covalent binding to the SPC surface raises the question as to how these properties would be affected by exposing the surface-bound complexes to aqueous solutions of ions that would bind to the unmodified amine sites on the SPC.

Silica Polyamine Composites (SPC)

Before reporting our results for the impact of metal ion binding on the photophysical properties of 1 and 2, a brief introduction to the SPC technology is appropriate. These materials were originally developed as chelator materials for the recovery of toxic and valuable metals from mine waste streams and metal processing solutions.^[5] The materials are patented and are currently being produced and marketed by Johnson Matthey Ltd, UK. These are very robust materials and offer several advantages over conventional polystyrene resins, including more rapid mass transfer and matrix rigidity over a wide range of temperatures.^[6] The synthesis of the primary SPC is illustrated schematically in Figure 3, and these are subsequently modified with metal-selective ligands for particular industrial applications. Using a mixture of methyltrichlorosilane and 3-chloropropyltrichlorosilane in the initial silanization step resulted in materials with much higher metal ion capacities and lower cost relative to using only chloropropyltrichlorosilane. The ratio of 7.5:1 shown in the figure represents an optimized ratio used in the manufacture of these materials.^[7] The covalent binding of the luminescent Ru complexes to the SPC was motivated by their potential applications in heterogeneous catalysis and photo-promoted electron transfer.

Figure 3.

Schematic representation of the synthesis of the SPC, BP-1 used in this study.

Experimental Section

The methods used for evaluation of the luminescence spectra and the excited-state lifetime in solution and on the SPC are described in detail in the prior work.^[1–3] Observing luminescence and measuring lifetimes on the SPC required a special sample configuration. A microscope slide was fitted with two-sided black tape and the SPC particles were spread evenly over the surface of the tape. The slide was trimmed to exactly fit a one-centimeter cuvette, placed in the fluorimeter or laser beam and adjusted to get maximum luminescence with front-face illumination (Figure 4).

Figure 4.

Sample configuration for photophysical measurements of Ru complexes on SPC: sample on glass slide (left); close-up of particles (right).

It should be noted here that the particles are not evenly colored; this is a consequence of the uneven distribution of surface hydroxyl groups on amorphous silica gel and this heterogeneity is carried forward in the subsequent modification steps. It is not possible to measure quantum yields on immobilized complexes in the solid state. However, the intensity of the emission is sufficient to measure excited-state lifetimes (Figure 5a).

Figure 5.

a) Normalized emission of complex 1 due to the ³MLCT band on SPC; b) excitation spectrum of 1 showing contributions to the emission from the MLCT (470 nm) the phosphine ligand (350 nm) and the diimine ligand (300 nm).

The excitation spectrum (Figure 5b) is obtained by illumination at 635 nm and scanning the absorption region of the sample. In the case of 1, the spectrum reveals contributions to the excited-state from the phosphine ligand and the diimine ligand as well as from the ³MLCT. The excited-state lifetime is 720 ns in solution and 3,450 ns on the SPC.^[3] The emission frequency and the CO IR stretching frequency show only minor changes relative to their solution values (634 and 647 nm; 1944 and 1951 cm⁻¹ respectively).^[3] Complex 2, on the other hand, showed only a small change in excited-state lifetime relative to its solution value (220 and 270 ns respectively).

Luminescence quenching and enhancement experiments were conducted by placing ~50 mg of 1-BP-1 in a scintillation vial and covering the sample with 10 mL of metal-salt solution at a specific salt concentration. These mixtures were then shaken on a Precision Scientific 360 Orbital Shaker Bath at room temperature for 24 h. The salt solution then was decanted, replaced with 10 mL of DI water, and the mixture was sonicated for ~1 min; this procedure was repeated 3 times. After washing, the material was placed in an oven overnight at 60C and 0.5 atm. in order to remove trace water. The particles were then dispersed on the SEM tape (Figure 4).

Results and Discussion

There are many examples in the literature of a transition metal quenching the emission of a luminescent complex. This is usually accomplished by coordination of a transition metal to a ligand that, in turn, is attached to a ligand of a luminescent complex.^[8–10] This type of quenching is referred to as static quenching and involves a transient electron transfer to the transition metal.⁸ Dynamic quenching involves the collision of the quencher with the emitter, and in the case of triplet-state emitters, collision with another triplet state, such as oxygen, is involved. Enhancement of triplet emission is most commonly observed with aromatic molecules containing a heavy atom such as iodine.^[11] This could also be considered static quenching of fluorescence but enhancement of phosphorescence can also occur by collisions in solution with heavy atoms. In the case of a luminescent molecule immobilized on a surface the situation is less clear and this prompted us to examine the effects of coordinating transition metals and heavy atoms to the surface of BP-1 (Figure 3) having either 1 or 2 bound to the surface.

Exposing complex 1 immobilized on BP-1 to solutions of CuSO₄ at intrinsic pH (~3.5) under equilibrium batch conditions resulted in a concentration dependent quenching of the ³MLCT band at 634 nm (Figure 6).

Figure 6.

Progressive quenching of the ³MLCT band of 1 on BP-1 by increasing concentration of Cu²⁺ (legend at the right indicates the dilution of the 1500 ppm Cu²⁺ stock solution).

Significant quenching is seen at a [Cu²⁺] = 30 ppm (gray trace). Full quenching is observed at a 1:1 ratio (1500 ppm). No quenching is observed with 1 in solution over a range of concentrations, and using Ni²⁺ and Zn²⁺ solutions with 1 on BP-1, quenching is observed but not as efficiently as with copper (Figure 7). The degree of quenching qualitatively follows the relative binding affinity of the three metals (Cu²⁺> Ni²⁺ ~ Zn²⁺) on the SPC BP-1.^[12]

Figure 7.

Normalized emission (E/E₀) of 1 on BP-1 in the presence of various concentrations of Cu²⁺, Ni²⁺ and Zn²⁺ and showing that no quenching is observed when 1 is in solution.

Complex 2 on BP-1 also shows a concentration dependent quench of its ³MLCT band when exposed to solutions of Cu²⁺ (Figure 8). It should be noted here that the degree of quenching at 30 ppm is approximately the same as for 1 on BP-1 (0.6 of the normalized emission, see Figures 6 and 7) even though the loading of 2 on BP-1 is significantly higher (0.040 versus 0.013 mmol/g). Thus, unlike the excited-state lifetime, the efficiency of quenching appears to be independent of the ancillary ligands. Quenching efficiency is slightly less when the particles are on the slide-tab (brown trace in Figure 8) than when they are equilibrated with the mobile particles.

Figure 8.

Quenching of the normalized ³MLCT emission of 2 on BP-1. The brown trace represents the emission when the particles on the slide are exposed to [Cu].

All of the data presented are consistent with static quenching where only Cu²⁺ ions that are in close proximity to the luminescent complex on the SPC surface contribute to quenching. If static quenching is the only operative quenching mechanism, a plot 1/Normalized emission intensity versus [Cu²⁺] should give a straight line with a positive slope according to the Stern-Volmer relationship.^[13] More importantly, this relationship predicts that if static quenching is the only operative mechanism then the excited-state lifetime will remain constant.^[13] This is because only the luminescent Ru complexes in close proximity to the quencher (Cu²⁺) will have their lifetimes affected. In the case of the SPC-Ru complex system discussed here, the low loading and the large number of amine sites are suggestive of static quenching. In the case of dynamic quenching, a decrease in lifetime reflects intermolecular collisions. A plot of 1/Normalized emission still gives a straight line but the excited-state lifetime increases in proportion to the rate of the collisions. If both mechanisms are operative, than the plot of 1/Normalized emission will be curved. The complete Stern Volmer relationship can be expressed in equations 1 and 2:

$${\mathrm{\tau }}_0/\mathrm{\tau }=1+{\mathrm{K}}_{\mathrm{D}}[\mathrm{Q}];{\mathrm{K}}_{\mathrm{D}}=t_0{k}_{\mathrm{q}}$$ (1)

τ and τ₀ are the lifetimes in the presence and absence of quencher Q, K_D is the quenching constant, which depends on the collision frequency, k_q

$$\begin{gathered} {\mathrm{E}}_0/\mathrm{E}=(1+{\mathrm{K}}_{\mathrm{D}}[\mathrm{Q}])(1+{\mathrm{K}}_{\mathrm{S}}[\mathrm{Q}]); \\ {\mathrm{K}}_{\mathrm{S}}=[\text{EQ}]/[\mathrm{E}][\mathrm{Q}] \end{gathered}$$ (2)

E₀ and E are emission intensities in the absence and presence of the quencher Q, and K_s is the static quenching constant which is determined by the mass action expression for the emitter-quencher complex in terms emission intensity and quencher concentration.

A plot of 1/Normalized emission intensity does give a straight line with a reasonable correlation coefficient (.86), and the lifetime is essentially constant in the [Cu²⁺] range of 0.0 to 1.3 mmol/L (Figure 8).

Complexes 1 and 2 on BP-1 are triplet-state emitters and would therefore be expected to exhibit an enhancement of emission in the presence of heavy atoms via spin-orbit coupling. Third row transition and post-transition metal atoms, such as mercury and lead, have been shown to have reasonable affinity for the SPC matrix and possess the large orbital angular momentum associated with significant spin-orbit coupling phenomena.^[14] Indeed, when mercury nitrate solution is added to the SPC particles, there is a concentration-dependent enhancement of the emission at 634 nm (Figure 10). A 50% enhancement is observed in the presence of 10 ppm Hg²⁺ (Figure 9, blue-green trace 1:50 dilution). In the presence of 500 ppm Hg²⁺, an enhancement of 3.3-fold is observed (Figure 10, blue trace, undiluted stock solution). The excited-state lifetime remains unchanged over the entire concentration range examined indicating a static enhancement mechanism as for the quenching observed with Cu.

Figure 10.

Enhancement of ³MLCT emission by addition of [Hg²⁺] to 1 on BP-1 particles (legend at right gives [Hg²⁺] in terms of the dilution of a 500-ppm stock solution.

Figure 9.

Stern-Volmer plots of 1/Normalized emission versus [Cu] and excited-state lifetime versus [Cu].

Under similar conditions solutions of Pb(NO₃)₂ induce an 3.7 enhancement of the ³MLCT emission in the presence of 500 ppm [Pb²⁺] (Figure 11). As with the transition metals, very little change in the emission intensity is observed with 1 and Pb(NO₃)₂ in solution (Figure 11).

Figure 11.

Enhancement of ³MLCT emission for 1 on BP-1 for various concentrations of [Hg²⁺] and [Pb²⁺] and for 1 in solution with [Pb²⁺] = 10, 50 and 100 ppm.

Solutions of UO₂(NO₃)₂ also showed an enhancement of ³MLCT emission when equilibrated with 1 on BP-1. A 2.7-fold enhancement of the normalized emission was observed with 250 ppm of UO₂(NO₃)₂. It appears that the enhancement is general for heavy atoms, and like the quenching observed with transition metal ions, the enhancement is static in nature as the excited-state lifetime remains constant over the entire concentration range studied (Figure 12).

Figure 12.

Plots of Normalized emission versus [Hg] and excited-state lifetime versus [Hg].

Conclusions and Future Work

The most interesting aspect of these results is that complex 1 can exhibit luminescence quenching and enhancement that is dependent on the nature of the metal in the immediate environment. In the case of the transition metals, the excited triplet state likely undergoes a transient electron transfer to a LUMO orbital on a neighboring metal and then decays to the ground state by radiationless decay (Figure 13).^[14] In the case of the heavy metals, the large spin-orbit coupling increases the efficiency of the intersystem crossing, leading to increased population of the triplet state (Figure 13).^[14,15]

Figure 13.

a) Transient electron transfer; b) the heavy atom effect: account for the changes in the luminescence of the complexes on the SPC.

The other interesting aspect of these results is that the metal ions inducing the quenching or enhancement are not bound to the ligand system of the luminescent molecule. This raises the question of how close the SPC surface-adsorbed ion would be to 1 or 2 in order to see an effect on the emission. This is a difficult question to answer with an amorphous solid. EXAFS measurements could provide average distances between Cu and Ru sites.

As far as using these systems for metal sensing is concerned, several obstacles need to be overcome. First, the kinetics are slow using the current SPC particles. It takes ~8 h for maximum quenching to be achieved with a 25 ppm Cu solution using 1-BP-1. In addition, greater sensitivity is required to detect environmentally significant concentrations of the metal ions of interest. Although we have made nano-versions of the Ru-SPC systems, it is doubtful if the increased surface area would lead to sufficient improvements in the sensitivity and the adsorption rates needed. An attractive alternative to the use of particles is a thin-film device where the luminescent complexes would be dispersed in a relatively nonpolar polymer and where the charged Ru centers, modified with a lipophilic tail,^[2] would be expected to reside primarily on the surface. This thin film could then be coated with a thin film of polyamine or ligand-modified polyamine. This approach eliminates the problem of achieving much higher covalent binding of the Ru complex to the SPC surface and at the same time promises to allow for much more rapid adsorption of metal ions to the surface than diffusion into porous particles (Figure 14).

Figure 14.

Schematic diagram of thin of PMMA containing luminescent molecules with lipophilic tails, spray coated with poly(allylamine).

Thin-film devices using polyamines are well known and efforts to translate this approach to a metal sensing device are underway in our laboratory.^[16,17]