1. Introduction
Metal-organic porous materials represent a significant research field in materials science during the last decades, with numerous applications [
1]. However, in some cases, their stability is inadequate, and the material cannot withstand the necessary conditions for the application. To enhance the stability of the system, zirconium-based metal-organic materials can be used due to strength of the Zr−O bond [
2]. Among them, several porous gels (MOGs, metal-organic gels) have been prepared containing [Zr
6(O)
4(OH)
4]
12+ clusters connected by polycarboxylato ligands [
3,
4]. Gels based on Zr(IV) and polycarboxylate ligands [
5,
6] are complex materials that combine the properties of polymers with the capabilities provided by the presence of the zirconium atoms to generate an uncountable number of applications, such as sensor[
7], energy storage[
8], gas storage/removal[
9], etc.
Most of the MOGs reported are obtained through an energetically demanding route (i.e. employing heating procedures) and with environmentally unsustainable solvents such as N,N-dimethylformamide, N,N-dimethylacetamide or hydrochloric acid. In our research group we have prepared several zirconium based metal-organic gels containing the abovementioned [Zr
6(O)
4(OH)
4]
12+ SBU and polycarboxylic ligands through a rapid method (1-2 min) at room temperature, with green solvents and without employing modulators such as HCl and acetic acid [
10]. Among them a thermally and chemically robust MOG of general formula [Zr
6(O)
4(OH)
4(BTC)
2.13(HBTC)
2.81]·n solvent was synthesized, where BTC is the benzene-1,3,5-tricarboxylate anion (trimesate anion). The porosity and features of this material allowed us to use it as a stable catalyst in the continuous electroreduction of CO
2 [
10]. Now, the aim of the present work is to take advantage of the characteristics of this material to introduce different molecules with luminescent properties[
11] into its structure so that it can be used as a sensor [
12,
13]. Three fluorophores were selected (
Scheme 1): fluorescein (HFL), which contains a single carboxylic group capable of bonding with Zr metal atoms, naphthalene-2,6-dicarboxylic acid (H
2NDC)[
14], which has two carboxylic groups, and 4,4',4'',4'''-(porphine-5,10,15,20-tetrayl)tetrakisbenzoic acid (H
4TCPP) [
15,
16], a molecule with four bridging-capable carboxylic groups. These molecules shows a luminescent response in blue, yellow, and red, respectively, which is retained when anchored to the Zr(IV) metal centers. In principle, these fluorophores, in their anionic form, will replace some of the trimesate anions in the gel. Interestingly, the luminescence emission will undergo significative changes when this functionalized MOG is suspended in different solvents or in aqueous solutions containing different analytes. The change of the luminescent signal of each fluorophore under these conditions can be employed for sensing purposes [
17], but these changes are more specific when the three fluorophore molecules are present simultaneously in the MOG. The interaction of the adsorbed molecules modifies to a different extent the luminescent signals of each fluorophore present in the MOG to provide a characteristic colour for each analyte, resulting from the sum of the three emissions. However, the emission intensity is also affected by self-quenching and changes on the fluorophore density can generate significant variations on the emission features that hinder their use as chemical sensors. This fact altogether with the mechanical weakness that these MOGs, and most gels, typically exhibit[
18,
19] has made necessary to combine the original MOG with mechanically robust quartz microcrystals to form a functional composite sensor [
20]. The results will show that the MOG containing the three fluorophores incorporated into a matrix of quartz microcrystals provides the required mechanical strength as to be used to identify different solvents and analytes in aqueous solutions by measuring the relative intensities of the characteristic emission signals of each fluorophore upon UV excitation.
2. Results and Discussion
To develop a fluorescent sensor based on these materials, three fluorophores whose fluorescence emission closely matches the three primary colours (blue, yellow, and red) were selected: naphthalene-2,6-dicarboxylic acid (H
2NDC), fluorescein (HFL) and 4,4',4'',4'''-(porphine-5,10,15,20-tetrayl)tetrakisbenzoic acid (H
4TCPP). Their combination within the same porous matrix provides a material that, upon excitation with the same wavelength, changes the emission of each fluorophore in a non-equal way, resulting in colour changes visible to the naked eye. The fast gelation conditions obtained under the above-described synthetic conditions facilitates the incorporation of these complex mixtures of fluorophores by the coordination of their deprotonated counterparts to the zirconium metal centers avoiding any possible segregation that could take place under slower thermodynamically driven reaction conditions. As result, this approach provided MOGs with deep blue (421 nm;
Zr-BTC-NDC), yellow (530 nm;
Zr-BTC-FL) and red (680/712 nm;
Zr-BTC-TCPP) emissions (
Figure 1 and
Figure S5a). It is worth noting that the
Zr-BTC gel also produces a fluorescence response at the violet spectrum (393 nm) but it is too weak to be used effectively for sensing purposes. The fluorescence maximum of the fluorophores anchored to the Zr-BTC system undergoes a slight redshift compared to the free molecule dissolved in water (380 nm for H
3BTC [
21], 426 nm for H
2NDC [
22,
23], 500 nm for HFL [
24], and 643/706 nm for the Q(0,0) and Q(0,1) emissions of H
4TCPP [
25,
26,
27]).
Upon exposure of the three fluorophores incorporated
Zr-BTC-NDC-FL-TCPP gel to different solvents, the visual inspection of the samples reveals changes of the fluorescence emissions (
Figure 2a). However, when quantitative measurements are attempted, there is not reproducibility as any manipulation of the gel, as those required to place the sample in the fluorimeter, leads to deformations/densifications of the monolith that also affect the intensity of the fluorescence emissions due to self-quenching effects (
Figure S5b). The corresponding aerogel was also prepared by CO
2 supercritical drying, and it retains the fluorescence (
Figure 2b) but the resulting monoliths are so fragile that again there is no option to obtain reproducible luminescence measurements.
Therefore, an alternative strategy was approached in which, instead of generating macroscopic gel monoliths, nanometric particles of the gel were grown within a matrix of SiO
2 microcrystals intended to protect the fluorescent metal-organic nanoparticles from external mechanical stress. SiO
2 (quartz) was selected because of its transparency towards visible light and a significant portion of UV (> 200 nm). The preparation of this mixture was performed using the same conditions as for metallogels but adding the micrometric quartz particles (90% quartz and 10% metallogel) under vigorous stirring for 12 h. The resulting product was thoroughly washed with methanol using a soxhlet apparatus until the washing liquid remained colourless and did not show any fluorescence emission upon exposure to an UV lamp (365 and 254 nm). Finally, the product was dried at room temperature. SEM/EDX images showed that SiO
2 microcrystals were surrounded by far smaller agglomerates of particles of the
Zr-BTC-NDC-FL-TCPP gel (
Figure 3a,
Figures S3 and S4). Under these conditions the fluorescence response remained stable during manipulation, enabling quantitative measurements. The big size difference between the quartz microcrystals and the
MOG nanoparticles prevents the compaction of these particles when pressure is applied. The packing of the quartz microparticles, even under pressure creates voids that allow the far smaller
MOG particles to accommodate within them and withstand the pressure without affecting the fluorescence emission. In addition, composite material exhibits chemical stability comparable to that of the pristine (non-silica containing) metallogel. The preliminary visual inspection of the single component MOG-SiO
2 composites provided the expected red, green, and blue emissions for
Zr-BTC-TCPP-SiO2,
Zr-BTC-NDC-SiO2 and
Zr-BTC-FL-SiO2, respectively (
Figure 3b). Since, unless neat MOG, MOG-SiO
2 composites exhibited a reproducible fluorescence response, a fine-tuning of the fluorophore concentrations in
Zr-BTC-NDC-FL-TCPP-SiO2 was performed until the combination of the three emissions provided a white emission (
Figure 3b). The best results were obtained using 1.66 mmol ZrCl
4, 2.716 mmol H
3BTC, 0.030 mmol HFL, 0.282 mmol H
2NDC and a surprisingly low value of only 2.5·10
-3 mmol H
4TCPP due to its very effective self-quenching capacity. The immersion of the composites in different solvents produced fluorescence emission colour changes related to the non-equivalent changes in the response of each fluorophore, that were appreciable visually (
Figure 3b) and registered in the spectra below discussed. The possible leaching of the fluorescence active molecules and zirconium atoms, when exposed to these solvents and also to different molecules dissolved in water, was checked by fluorescence analysis and elemental analysis of the washing liquid. The results showed that leaching was negligible, and the anchorage of the active molecules is strong enough for their involvement in sensing applications. Emission colour changes were also observed in the presence of different organic molecules (benzyl alcohol, caffeine, fructose, glucose, imidazole, phenol and urea) dissolved in water when the samples were immersed in these solutions. Despite in most cases the
MOG-SiO2 remained stable allowing its reutilization, it was noticed that carboxylic molecules or other molecules with functional groups that can strongly coordinate to the zirconium metal center create a significant leaching of the fluorophore molecules (mainly fluorescein as deduced from resulting green/yellow fluorescence observed in the mother liquid).
The three emission maxima present in the multifluorophore
Zr-BTC-NDC-FL-TCPP-SiO2 sample provides the opportunity to use their fluorescence spectra for the univocal identification of chemical species based on a stimuli-response interaction with the MOG matrix. The eye sensitivity is greater for green and lower for blue and red colours [
28], which means that although perceiving through our eyes a white emission, it does not mean the intensity of the emission from the three fluorophores is equal. However, as our goal is to develop a quantitative sensoric system, we must try to get these emissions intensities as close as possible to achieve similar sensitivity for each fluorophore. In this sense, the excitation spectrum for each fluorophore was measured and the excitation wavelength was fixed at 325 nm, the value in which three fluorophore emissions are closer (
Figure 4).
The quantitative analysis of intensity of each fluorescence emission could in principle provide a sensing opportunity for the solvents and for the analytes dissolved in water if the mass concentration of the analyte is kept fixed (0.1 mg/mL). However, as the fluorescence signal's intensity depends greatly on the excitation intensity of the UV source and the amount of sample, among other parameters, a different approach was used employing relative intensities. As we have three fluorophores incorporated simultaneously, their main emissions NDC (I
1; 387 nm), FL (I
2; 525 nm) and TCPP (I
3; 690 nm) and three relative intensity values I
1/I
2, I
1/I
3 and I
2/I
3 can be employed for the analysis. These measurements (
Figure 5) must be performed while the solid sample is immersed in the solvent or aqueous solution which implies a specific setup for the experiment, A quartz cuvette originally designed for liquid measurement was used. The
Zr-BTC-NDC-FL-TCPP-SiO2 composite (0.1 g) immersed in the solvent or aqueous solution (10 mL) was transferred to this cuvette. The particles were forced to deposit on the quartz window using a centrifuge. The whole procedure was repeated three times for every solvent or analyte to verify data's reproducibility (
Table 1,
Tables S1 and S2). A detailed description of the procedure can be found in the materials and methods section and in a video provided as
supplementary material. Depicting these relative intensity values in a 3D diagram the solvents and analytes can be easily distinguished (Figure 6).
Figure 5.
Fluorescence emission of Zr-BTC-NDC-FL-TCPP-SiO2 when immersed in different solvents (a) and molecules dissolved in water, 0.1 g/mL (b).
Figure 5.
Fluorescence emission of Zr-BTC-NDC-FL-TCPP-SiO2 when immersed in different solvents (a) and molecules dissolved in water, 0.1 g/mL (b).
Figure 5.
Depiction of the relative intensity fluorescence emission values in a three-dimensional map for the solvents (a) and molecules dissolved in water (b). I1, I2, and I3 correspond to the intensity at the maximum of the peak corresponding to NDC, FL and TCPP fluorophores, respectively.
Figure 5.
Depiction of the relative intensity fluorescence emission values in a three-dimensional map for the solvents (a) and molecules dissolved in water (b). I1, I2, and I3 correspond to the intensity at the maximum of the peak corresponding to NDC, FL and TCPP fluorophores, respectively.
A comparable method could be developed using samples that contain a single fluorophore[
7,
29,
30], but the results for different solvents/molecules would spread in just one dimension providing less differentiated results or less specificity in the identification. Another option would be to employ three sets of samples with just one fluorophore in each one, but strict control on the sample amount and measurement setup would be required.