Introduction

Recently, the most significant development of nanostructured materials with stimuli-responsive and switchable functions was achieved using supramolecular host–guest chemistry1,2,3. To bypass the synthetic hazards, the researchers focus on non-covalent interactions over the traditional organic covalent bonds4,5,6,7,8,9,10. The non-covalent interaction (e.g., dipole–dipole interactions, van der Waals interactions, and hydrogen bonding) played a leading role within the host–guest supramolecular approach where the small organic molecules (guest) held inside the nanocavity containing macrocycle host4,5,6,7,8,9,10. Inspired by the specificity of non-covalent interactions, we have researched developing host–guest-based advanced materials for water quality monitoring11,12,13.

The application of cucurbiturils (CBn, n = 6, 7, 8, 10; Fig. 1) to interact with small fluorescent molecules was realized in literature where cucurbiturils modulate the pKas (protonation states) of embedded guests such as benzimidazoles14. Consequently, the CBn-induced pKa shifts have been shown to establish pH-driven control over the sequestration and release of guest molecules14, 15. Guest retention and release in response to pH stimuli have also been utilized to construct stimuli-responsive nanostructured materials16.

Figure 1
figure 1

Schematic representation of pH-controlled TBI/CB7-loaded NPs. Chemical structures of the fluorescence probe TBI and the cucurbituril macrocycles, CB7, are also shown. Depending on the protonation/deprotonation of the nitrogen at position-3, two forms can exist of the coordinating ligand: TBI and TBIH+.

Full size image

From another perspective, the dyes encapsulated by CB7 have been adsorbed on the surface of metal nanoparticles17,18,19,20,21 for drug delivery18 magnetic resonance imaging20, and the fabrication of solar cell materials19. The iron-oxide magnetic nanoparticles (γ-Fe3O4) have been specifically utilized for chemical detection and sensing22. It was used for mercury sensing using water-soluble cyclodextrin and solvents as the trigger for switching11.

In this work, we demonstrated the pH-responsive iron oxide nanoparticles (NPs) surface coated with nanocontainer CB7 which encapsulates a benzimidazole-based fluorescent dye (2-thienylbenzimidazol) TBI, Fig. 123. Thus, the present supramolecular-based nanomaterials approach has much more potential to repeatedly use those solid materials by responding to pH triggers (as opposed to using a solvent)11.

The intensity measurements by fluorescence are not practical for the detection or separation application of undesired residues in water samples because the intensity is dependent on the concentration of the fluorophore. Thus, the time-resolved photoluminescence (PL) experiments were performed in the solid state of the TBI/CB7 host–guest supramolecular system on γ-Fe3O4 nanoparticles.

Results and discussion

Interactions of TBI with CB7 in solution

The encapsulation of TBI by CB7 is evidenced from UV–visible absorption measurements in water at pH 2 and 7 in Fig. 2 (see also pH-titration data below). CB7 encapsulates the protonated or neutral TBI molecules, causing a redshift in their absorption spectra from 315 to 320 nm.

Figure 2
figure 2

UV–visible absorption titration of TBI (20 μM with CB7) at different pH: (A) and (B) illustrate the binding with CB7 at pH 2 and 7, respectively. The inset shows the corresponding titration curve and the 1:1 binding fit (solid line) with the corresponding binding constant, K, for the data in (A) and (B). For clarity, the spectra at different CB7 concentrations are shown in different colors. The numbers are the corresponding maxima (in nanometers).

Full size image

Proton-NMR spectroscopy also confirmed the complexation of TBI protonated forms to CB7 at pD 2 (Fig. S1 in the Supporting Information). Previous reports on other benzimidazole derivatives demonstrated similar host-induced shifts on account for the shielding of the protons of the benzimidazole guest molecules when encapsulated in the electron-donating cavities of CB7 (e.g., fuberidazole, where the sulfur atom is replaced by oxygen)14, 24,25,26,27. Two pH values, pH 2 and 7, were selected for the binding experiments (Fig. 2A,B) based on the pH titration data in the absence (Fig. 3A) and the presence of CB7 (Fig. 3B). The complex affinities were also estimated at each pH (e.g., K = 2.8 × 106 M−1 at pH 2 and K = 2.4 × 105 M−1 at pH 7). pH and binding titration experiments were already conducted by monitoring the fluorescence emission of TBI upon varying the pH or the concentrations of CB725.

Figure 3
figure 3

Ground state pH-titration of (A) 20 µM TBI only and (B) 20 µM TBI with 3 mM CB7. Inset showing corresponding pKa values. For clarity, the spectra at different pH are shown in different colors. The numbers are the corresponding maxima (in nanometers).

Full size image

Table S1 in the Supporting Information summarizes the different binding constants measured using NMR, UV–visible absorption, and fluorescence techniques. Compared to that binding constant measured by UV–visible, the value measured by NMR was lower by two orders of magnitude, which is expected because of the concentration differences (millimolar in NMR data versus micromolar in optical data). Binding constants of the excited states (PL method) to CB7 were also much lower by one-to-two orders of magnitudes (depending on the pH values) than those for the ground state’s binding affinities (UV method), which is unsurprising14. The pH-titration plot for the free TBI in Fig. 3A gave pKa values in the ground state because of the nitrogen protonation at position 3, which increased by ~ 3.3 units upon the addition of CB7 (Fig. 3B).

Because the protonation of CB7 occurs at pH 2, which affects the binding between the protonated TBI and CB7 by analogy to similar benzimidazoles14, a more accurate binding constant at pH 2 should be calculated from the corresponding pKa shift using the previously reported thermodynamic relation14, giving a value of K = 4.8 × 108 M−1. The relation is based on a four-state model, which involves (a) the uncomplexed and unprotonated TBI (guest), (b) the uncomplexed protonated TBIH+ (guest), (c) the unprotonated TBI/CB7 (host–guest) complex, and (d) the protonated TBIH+/ CB7 (host–guest) complex, which are all connected through a thermodynamic cycle as illustrated in Fig. 4.

Figure 4
figure 4

Four-state complexation model of the neutral and protonated TBI with CB7.

Full size image

The CB7-induced pKa shifts are unsurprising and agree with previous reports on other benzimidazole derivatives14, 24,25,26,27. To ensure a controlled release and retention of the guest from the host cavity, the binding affinity of the protonated and neutral forms of TBI to CB7 must be different. The present study confirms the three-order magnitudes difference. Consequently, it was decided to opt for this supramolecular approach to develop nanoparticles whose structures could be switched repeatedly in response to pH values while being monitored utilizing time-resolve PL measurements (see below).

Interactions of TBI with CB7NPs in the solid state

The fabrication of the nanocomposites is illustrated in Fig. 1. The γ-Fe3O4 nanomaterials were activated by conjugation to a host–guest complex of CB7 and TBI using similar approaches to those previously reported17. We mixed spherical CB7NPs (8 ± 1 nm in diameter) and TBI (CB7: TBI, 1:100) in water at pH 2 and room temperature for 24 h. The brown precipitate formed was collected with a magnet, washed several times with water to remove residual TBI, and analyzed by TGA and FTIR spectroscopy (Fig. 5).

Figure 5
figure 5

Thermo-gravimetric analysis (A) and FTIR spectra (B) of CB7NPs, TBI/CB7NPs, and TBI.

Full size image

The weight percentage of TBI interacting with CB7 on the surface of CB7NPs was determined by TGA. Figure 5a presents the weight losses of CB7NPs and TBI/CB7NPs. The more significant percentage loss occurs when TBI/CB7NPs are heated due to the loss of TBI. These data are consistent with the successful loading of CB7NPs with TBI. The TGA analysis of TBI/CB7NPs shows a composition of 90.95% iron oxide, 4.96% CB7, and 0.69% TBI corresponding to a 1:1 guest: host interaction (i.e., a 1:1 TBI: CB7 interaction, see Fig. S2 and Tables S2 and S3 in the Supporting Information for more details). The FTIR spectrum of TBI/CB7NPs is not simply the sum of its parts, which strongly suggests the presence of interactions between TBI and CB7NPs17. The spectrum of TBI/CB7NPs displays various peaks that can be attributed to ν(C–C) and ν(C–H) vibrations of CB7 as well as ν(C=C) and ν(C–H) vibrations of TBI. In addition, the FTIR spectrum of TBI/CB7NPs reveals a broadening of the peaks at 1733 and 1633 cm−1 corresponding to ν(C=O) vibration of the carbonyl portal of CB7. This observation suggests the presence of TBI in the CB7 cavity.

The changes in the solid UV spectra (e.g., a shift in the band gap from 1.9 to 2.1 eV) also confirm the loading of TBI on CB7NPs (when pH was also changed from 7 to 2), Fig. 6. The Tauc plots clearly show different band gap values between the TBIH+/CB7NPs (Eg = 2.10 eV) and only CB7NPs (Eg = 1.92 eV).

Figure 6
figure 6

Diffusive-reflectance spectra (DRS) of CB7NPs and TBIH+/CB7NPs, and their Tauc plots suggest direct optical bandgaps.

Full size image

As control experiments, the TBI/CB7 solid complex was prepared by grinding method (see “Methods” section). PL and TRPL were measured for all solid samples TBI (500 nm and 0.85 ns) and TBI/CB7 (452 nm and 2.7 ns), as shown in Fig. 7 and Fig. S3 in the Supporting Information. PL and excited-state lifetime measurements of solid samples for TBI upon inclusion inside CB7NPs (425 nm and 1.7 ns) were also measured at pH 2 (420 nm and 1.3 ns) and pH 7 (452 nm and 0.6 ns), as shown in Figs. 7 and S4 in the Supporting Information. Clear, distinct values are observed, which confirms the solid interactions. The blue-shift of ~ 50 nm and the increase in the lifetime value for TBI/CB7 when compared to the emission and lifetime decay from the free TBI (from 0.85 to 2.7 ns; Fig. S3) highlight the hydrophobic effects of the nonpolar cavity of CB7 in parallel to the results obtained in solution25. Similarly, an increase in the lifetime value (from 0.6 to 1.3 ns; Fig. S4) and a shift of ~ 25 nm was observed upon the protonation of TBI inside CBNPs (e.g., TBIH+/CB7NPs versus TBI/CB7NPs) because of similar polarity effects imposed by the more stable host–guest complex at lower pH values.

Figure 7
figure 7

Excited-state lifetime average values as calculated using the equation in the Method section (A) and excited-sate lifetime plots (B) of different solid samples: TBIH+/CB7NPs, TBI/CB7NPs, CB7NPs, TBI/CB7, and TBI. Lifetime measurements were performed at 298 K, λex = 375 nm, and λobs = 480 nm. IRF is the instrument response function shown in grey color.

Full size image

Noticeably, CB7NP (without TBI) has no pH-dependent emission spectra (Fig. S5 in the Supporting Information). Overall, the more accurate lifetime measurements demonstrated that the host–guest complex of CB7 renders the solid nanomaterial iron oxide more switchable through “non-covalent” interaction.

Thiophene was selected based on the expected soft–soft interactions between the sulfur atom and Hg2+23. The iron oxide magnetic nanoparticles were selected in particular to ease the collection of the magnetic materials by a magnet17. We, therefore, performed several experimental measurements to capture mercuric ions in the solid state. For example, we demonstrated in Fig. 8 that the interaction of TBI/CB7NPs with mercuric ions has caused a decrease in fluorescence intensity. The binding, in principle, can be regenerated upon changing the pH and using a magnet.

Figure 8
figure 8

Fluorescence spectra of TBI/CB7NP (A) and TBIH+/CB7NP (B) with Hg2+ at different concentrations (as indicated directly in the graphs) and 298 K.

Full size image

Moreover, the decay-associated spectra (DAS) analysis (Supporting Information, Fig. S6) revealed three excited-state lifetime components confirming the complex formation between TBI/CB7NP and mercuric ions in the solid state by analogy to the results in solution and the absence of nanoparticles25. In solution or solid-state, adding mercuric ions decreased the fluorescence intensity of TBI or TBI/CB7 in the water25 or on top of NP while keeping the excited-state lifetime unchanged (Table S4 in the Supporting Information). This advanced kinetic analysis unfolded different electronic states of TBI/CB7NPs and that only the shortest-living species (whose emission appears around 450 nm) was affected by mercury binding compared to others.

The repeated use of TBI/CB7NP for capturing mercuric ions is established by controlling the pH of the media, as manifested in the change in the fluorescence intensity. At neutral pH, the fluorescence intensity of TBI/CB7NP decreases when it binds a low concentration of Hg2+ in the solid state. However, TBIH+/CB7NP does not bind mercuric ions at acidic pH, and the fluorescence intensity is restored. The Hg2+-bound composite is easily separable with a magnet.

The long-term aim of the present research is to develop materials that are easy to handle, built, portable, and more importantly, to prove the concept of utilizing TBI/CB7 as a motif to significantly improve the turnover of the solid nanomaterials back to their initial state in response to pH. Figure 1 summarizes the flaws in the supramolecular approach employed by the TBI/CB7-functionalized γ-Fe3O4 NPs. At neutral pH, the ligands are primarily neutral and bind much less CB7 when compared to the protonated/cationic form (see the binding constants measured in Fig. 2). This explains the increase in the lifetime value of the CB7-modified NPs at low pH because of polarity effects. A subsequent increase in the pH of the tested water samples (that contain mercuric ions) substantially brings the ligand back to its neutral form. Our designed resetting mechanism is novel and has yet to be realized in literature, and it can be expanded to other ligands and analytes. The selected ligands are non-toxic, rendering the device disposable, as well.

Conclusion

We modified magnetic NPs to capture mercuric ions from water samples at neutral pH. Specifically, the time-resolved optical behaviors of the CB7-modified NPs reveal a significant improvement in the turnover of their structures in response to pH because of the different binding affinities of CB7NP towards the protonated and neutral coordinating TBI ligand. Additionally, neutral and not protonated nanocomposites bind mercuric ions in water samples, enforcing the resetting mechanism. The most hazardous pollutant, mercury, in water resources is becoming of significant concern in many places where water contamination is critical in the quest for long-term economic and social stability.

Overall, the current approach is unique regarding the controllable separation of mercury ions using an external magnet and in response to pH through preferential binding of the host to guest molecules on the top of magnetic surfaces.

Methods

Samples

CB7 (purity > 99.9%) was produced by Sigma-Aldrich and used without any further purification. As instructed by Sigma-Aldrich, the calculated concentrations have considered the presence of 20% water in the supplied CB7 vials. The description for preparation and characterization of the TBI23 in the present study is included in the Supporting Information. Millipore water was used.

Synthesis of CB7NPs

An aqueous solution (1 mL) of CB7 (n = 3 × 10−5 mol) was added to a colloidal suspension of NPs (4 mL, nFe = 7 × 10−4 mol) and transferred to a 10 mL microwave vessel with a crimp cap. The solution was heated by microwave irradiation of 2.45 GHz in a microwave reactor (CEM Discovery, CEM Inc. USA). The power was modulated to reach a temperature of 50 °C in 1 min and to maintain that temperature for 30 min. The maximum power applied was 300 W. Stirring was initiated at 50 °C during the heating cycle. Two heating cycles were used to prepare CB7NPs. The NPs were washed with water and precipitated by using a magnet. Iron concentration was deduced from UV–visible absorption data.

TBI/CB7 complex synthesis

TBI/CB7 complex in the solid state was prepared by a grinding method, in which an equal amount of the two components were mixed and grounded for 20 min with acetone.

TBI/CB7NPs preparation and characterizations

CB7NPs (nCB7 = 3 × 10−4 mol) and TBI (3 × 10−3 mol) were mixed in water (2 mL) and stirred for twenty-four hours at room temperature and pH = 7 to form inclusion complexes on the surface of NPs. The product was precipitated using a magnet and washed several times with water to afford TBI/CB7NPs. The presence of TBI was confirmed using FTIR spectroscopy. TBI/CB7NPs suspension (prepared above) was placed in water (2 mL) and stirred for 1 h at room temperature and pH = 2. The product was precipitated using a magnet and washed several times with water (pH 2) to afford TBIH+/CB7NPs.

Hg2+-TBI/CB7NP preparation

TBI/CB7NPs (nCB7 = 3 × 10−4 mol) and Hg(OAc)2 salt (nHg2+  = 1 × 10−3 mol) were mixed in water (2 mL) and stirred for 1 h at room temperature and pH = 7 to form metal complexes on the surface of NPs. The product was precipitated using a magnet and washed several times with water to afford Hg2+-TBI/CB7NPs.

Thermogravimetric analysis (TGA)

Solid samples (10 mg) under N2(g) flux were characterized with a SDT Q600 TA Instruments analyzer at a heating rate of 5 °C/min over a temperature range of 35–700 °C.

Spectroscopy

UV–Visible absorption spectra were measured on a Cary-300 instrument (Varian). To estimate the binding constant, the total guest’s concentrations must remain unchanged while changing the concentration of the host molecules. The total concentration of the host is then plotted against the absorption or fluorescence intensity at a given wavelength. A Varian 400 MHz spectrometer measured NMR spectra in D2O in ppm against TMS reference. The addition of carefully chosen amounts of HCl (DCl) or NaOH (NaOD) controls the pH values of the solutions (± 0.2 units) as recorded using a pH meter (WTW 330i equipped with a WTW SenTix Mic glass electrode). The absorption spectra of the solid samples were obtained by using the Kubelka–Munk conversion (K–M = (1 − R)2/2R) of the recorded diffusive-reflectance spectra at room temperature for the solid samples on an FS5 spectrometer (Edinburgh, UK) equipped with an SC-30 (integrating sphere) as the sample holder. The specular reflection of the sample surface light was removed from the signal by directing the incident light at the sample at an angle of 0°; only the diffusive reflected light was measured. Polytetrafluoroethylene (PTFE) polymer was used as the reference. The bandgap energy (Eg) values of the solid samples from the DRS spectra were calculated using Eg = 1240 eV nm l−1, where l is the absorption edge (in nm). The solid-state photoluminescence (PL) measurements were carried out for the suspension samples after they were vacuum-dried over the demountable corvettes under a fume hood for twenty-four hours. The time-resolve photoluminescence (TRPL) spectra were collected using time-correlated single-photon counting (TCSPC) on a LifeSpec II spectrometer (Edinburgh Instruments) by using EPL-375 picosecond diode laser (λex = 375 nm, repetition rate = 5 MHz, and instrument function = 30 ps) for excitation in the solid state. The monitored emission maxima were at 480 nm. The time-resolved emission (intensity of ~ 1000–3000 counts/s) was collected (up to 10,000 counts/s) by a red-sensitive high-speed PMT (Hamamatsu, H5773-04) detector. Marquardt–Levenberg algorithm opted to analyze the collected data utilizing the iterative reconvolution method to minimize χ2. The contribution of each lifetime, τi with an amplitude αi in the multiexponential model, to the steady-state intensity was adjusted using the formula

$${f}_{i}=\frac{{\alpha }_{i}{\tau }_{i}}{\sum_{j}{\alpha }_{j}{\tau }_{j}} ,$$

where the sum in the denominator is over all the decay times and amplitudes. The average excited-state lifetime is then calculated by

$$\overline{\tau }=\sum_{i}{f}_{i}{\tau }_{i}$$

For the decay-associated spectra (DAS) measurements, emission decays collected every 10 nm over the entire emission spectra of the solid samples with a dwell time of 10 s at each wavelength were globally fitted to a tri-exponential model function and then convoluted with an instrument response function (IRF) of ~ 30 ps. The time-resolved data were specifically analyzed using the Edinburgh FAST software.

Consent to participate

The participants consented.