Synthesis and physico-chemical properties of TerP and GmAc-FluoroTerPWe synthesized the GmAc-grafted fluorescent hydrogel polymer after experimental optimization of constituent units to achieve a suitable balance between receptor groups, i.e., –OH, –COOH, and –CONH2/–CONH–, and physicochemical properties, i.e., storage and loss moduli, i.e., G′ and G″, respectively. An optimized receptor composition would be key for achieving the best photoluminescence-based sensing, harvesting of REMs in repetitive cycles, and examining the semiconducting property during post-adsorptive stage. Therefore, we have developed a two-stage optimization process as described in the Experimental, to obtain TerP and GmAc-FluoroTerP. We measured rheological properties of swelled TerP and GmAc-FluoroTerP hydrogels to ensure gelation properties in each case. As shown in Fig. 2, GmAc-FluoroTerP maintains its hydrogel behaviour up to 100 rad s−1, till when Gʹ > Gʺ is observed. In contrast, TerP retains such behaviour only up to 10 rad/s (Fig. 2b,d). TerP can sustain only up to 10% strain in the strain test, whereas GmAc-FluoroTerP is stable up to 80% strain (Fig. 2c,e). The rheological measurements clearly indicate the superior gelation behaviour of GmAc-FluoroTerP in comparison to TerP.Structural characterization of TerP1, TerP2, GmAc, and GmAc-FluoroTerP1We used a combination of 1H NMR, XPS, EPR, and FTIR to characterize the formation of GmAc-FluoroTerPs via grafting of GmAc within TerPs, in addition to the strategic introduction of C–N coupled new monomer-equivalent moiety, 2-(butyramidomethyl) succinic acid (BAMSA). The thermostability, crystallinity, and surface morphology of the materials were measured using TGA, XRD, and SEM/TEM, respectively.The 1H NMR spectra of TerP1, TerP2, and GmAc-FluoroTerP1 are shown in Figs. 3b, S1. The C–C coupled polymerization resulting in the formation of –H2C(β)–CH(CH2(α)COOH)CO–, and –H2C(β)–CH2(α)CO–/–H2C(β)–CH(CO–)– corresponding to IA, AM, and MBA, respectively, can be confirmed from the –CH2(α)–/ > CH– and –CH2(β)– chemical shifts. Notably, the 1H NMR peaks within 1.92–2.36 and 0.83–1.48 ppm in TerP1, 2.01–2.45 and 0.85–1.45 in TerP2, and 2.07–2.37 and 1.02–1.62 ppm in GmAc-FluoroTerP, respectively, indicate the saturation of acrylic functionalities of IA, AM, and MBA (Table S1)28,40. Overall, the 1H NMR spectra of TerP1 and TerP2 appeared to be identical. The XPS peaks of  –CH2–/ > CH– BE at 284.43 eV (Fig. 5) in GmAc-FluoroTerP further substantiate the C–C coupled polymerization (Table S2). The depletion of characteristic vinyl proton peaks within the chemical shift ranges of 5.78–6.32, 5.70–6.30, and 5.59–6.13 ppm for IA, AM, and MBA, respectively28,39, together with the loss of C–H def. and overtone of CH2=CRRʹ (R = –COO−/H/H and R′ = –CH2COO−/–CONH2/–CONHCH2– for IA/AM/MBA) at 910 and 1830 cm−1 (Fig. 7a), respectively, validate the completion of C–C coupled polymerization. The presence of AM and MBA in TerP1/TerP2 can be inferred from the 1H NMR peaks at 4.25/4.31 ppm, attributable to –CH2– of MBA, and the peaks within 6.61–6.75/6.58 and 7.52/7.51 ppm, respectively, attributable to N–H of AM and MBA. Similarly, the presence of AM and MBA in GmAc-FluoroTerP can be deduced from 1H NMR peaks at 4.45 ppm and within 6.40–6.80/7.59 ppm, attributable to –CH2– of MBA and N–H of AM/MBA, respectively. These components are also supported by the peaks observed in XPS of GmAc-FluoroTerP. Notably, the BEs of O1s and N1s of –CONH2/–CONH– were found to be 530.44/399.14 eV while the BE of C1s of C–N was observed at 288.13 eV28. The peaks at 1650/1630 cm−1 observed in the FTIR spectrum of TerP, can be attributed to amide-I peaks of –CONH– (Table S3). For GmAc-FluoroTerP, IR peaks were observed at 1654/1637 cm−1 and likely originated from amide-I peaks of –CONH–, thereby supporting the presence of AM/MBA41. The introduction of IA in GmAc-FluoroTerP is inferred from C1s/O1s BEs of –COOH, –COO−, –COO−, –COOH, and –COOH at 288.13, 286.35, 531.15, 530.44, and 531.85 eV, respectively. The presence of IA in TerP is supported by the observation of IR peaks corresponding to C = O str. of –(COOH)2, C = O asym. str. of –COO−, and C = O sym. str. of –COO− at 1740, 1568, and 1405 cm−1, respectively. The analogous IR peaks in GmAc-FluoroTerP were observed at 1737, 1561, and 1408 cm−1, and validate the presence of IA therein. The formation of a –CH2CH2CONH–CH2CH(CH2COOH)COOH, i.e. BAMSA moiety, in TerP from association of –CH2CH2COṄH– of AM with ĊH2CH(CH2COOH)COOH of IA can be inferred from 1H NMR peaks within 3.45–3.68 ppm28,40. These are attributable to N–CH2 group resulting from C–N coupling of AM and IA. The analogous peak for GmAc-FluoroTerP was observed at 3.49 ppm. The underlying conversion of a primary amide into a secondary amide can also be rationalized from the BE of N1s of –CONH– at 399.62 eV in GmAc-FluoroTerP, and the FTIR peaks corresponding to amide-II peak observed in TerP and GmAc-FluoroTerP at 1559 and 1541 cm−1, respectively. Accordingly, the transition of primary amide into secondary amide should undergo via a transition state containing free N-radical. To demonstrate the existence of N-radical, EPR of the solution mixture, containing monomers, GmAc, MBA, and initiator, was recorded just before gelation. The distinct peak at 330 mT (Fig. 4a) is evidence of the free N-radical42. While C–C coupled polymerization reaction may also be accompanied by the presence of C-radicals, Lee et al.43 and Doetschman and Mehlenbacher44 have suggested the resonance of C-radicals from acrylic moieties of acrylic monomers to be observed at much higher magnetic field (> 340 mT).Figure 31H NMR of (a) GmAc and (b) GmAc-FluoroTerP1.Figure 4EPR analyses of (a) solution mixture before gelation and (b) GmAc-FluoroTerP/Dy-GmAc-FluoroTerP/Yb-GmAc-FluoroTerP.We were keen on scrutinizing the structural integration of GmAc in GmAc-FluoroTerP. GmAc is composed of 1,3-linked β-D-galactopyranosyl (β-D-Galp) units substituted at O-2, O-4, and/or O-6 positions along with branches of two to five β-D-Galp residues linked together and attached to the fundamental β-D-Galp chain through 1,6-linkages29,31,45. Both fundamental and branches contain α-L-arabinofuranose (α-L-Araf), α-L-rhamnopyranose (α-L-Rhap), β-D-glucuronic acid (β-D-GlcpA), β-D-galacturonic acid (β-D-GalpA), and 4-O-methyl-β-D-glucuronic acid (4-O-methyl-β-D-GlcpA) residues29,45. Of different terminal units, α-L-Araf is the most abundant unit, whereas β-D-GlcpA, β-D-GalpA, α-L-Rhap, and α-L-Arap are the minor components of terminal unit23. The 1H NMR spectrum of GmAc (Fig. 3a) reveals the presence of various constituent sugars. Notably, H-1, H-2/H-3, H-4, H-5, and H-6 of β-D-Galp appeared at 4.54, 3.68, 3.95–4.19, 3.95, and 3.68–3.87 ppm16,28,45, respectively, while H-1, H-2, H-3, H-4, and H-5 peaks of α-L-Araf were observed at 5.33, 4.19, 3.95, 4.04, and 3.76 ppm, respectively45. The peaks observed at 4.72, 3.95, 3.41, 4.04, and 1.24 ppm can be assigned to H-1, H-2, H-4, H-5, and –CH3 of α-L-Rhap, respectively16,28,45. H-1, H-2, H-3, and H-5 of β-D-GlcpA are suggested by peaks at 4.49, 3.41, 3.56, and 3.95 ppm, respectively31,45,46,47, while the peaks at 4.49, 3.56, 3.68, and 3.95 ppm are assigned to H-1, H-2, H-3/H-5, and H-4 of β-D-GalpA45, respectively. For 4-O-methyl-β-D-GlcpA, the observed peaks at 4.49, 3.41, 3.56, 3.95, and 1.34 ppm can be attributed to H-1, H-2, H-3, H-5, and O–CH3, respectively45,48. The observation of 1H NMR peaks corresponding to specific protons of β-D-Galp, α-L-Araf, α-L-Rhap, β-D-GlcpA, β-D-GalpA, and 4-O-methyl-β-D-GlcpA clearly imply the inclusion of GmAc in GmAc-FluoroTerP. The peaks at 4.53, 3.96, and 3.89 ppm are attributed to H-1, H-4/H-5, and H-6 of β-D-Galp, respectively. Similarly, H-1, H-3, H-4, and H-5 of α-L-Araf are observed at 5.61, 3.96, 4.02, and 3.83 ppm, respectively, while H-1, H-2, H-5, and –CH3 of α-L-Rhap are observed at 4.69, 3.96, 4.02, and 1.12 ppm, respectively. The peaks at 4.53 and 3.96 ppm are assigned to H-1 of β-D-GlcpA/β-D-GalpA and H-5/H-4 of β-D-GlcpA/β-D-GalpA, respectively. The peak observed at 1.62 ppm is attributed to O–CH3 of 4-O-methyl-β-D-GlcpA. The –OH of GmAc resonated at O1s BE of 532.65 eV (Fig. 5b). The peaks corresponding to > CH– and –CH2– in plane bending, O–H in plane bending, and C–O and C–O–C str. of GmAc have undergone marginal shifting from 1413, 1378, and 1145/1022 cm−1 of GmAc to 1420, 1382, and 1140/1035 cm−1 in GmAc-FluoroTerP, respectively49. The grafting of GmAc through −CH2OH of α-L-Araf onto TerP is indicated by the notable shift of the characteristic −CH2OH peak of α-L-Araf from 3.76 ppm in GmAc, to the −CH2−O−CH2− attributable peak at 3.83 ppm in GmAc-FluoroTerP50. The grafting-driven formation of ether linkages in GmAc-FluoroTerP is inferred from C1s and O1s BEs of C–O–C at 285.09 and 531.15 eV, respectively21,28. The grafting is also supported by observed FTIR peaks at 2952/2854/1121 cm−1, attributable to the –CH2– asym. str./sym. str./def. of –CH2–O–CH2– type ether linkages in GmAc-FluoroTerP28,41. GmAc grafting is also evident from the
significantly elevated thermostability of GmAc-FluoroTerP compared to GmAc. Till a temperature of 169 °C (Fig. 7c), GmAc possesses the lowest thermostability because of maximum release of loosely bound water molecules. In contrast, TerP and GmAc-FluoroTerP resist weight loss facilitated by stable hydrogen-bonding. The rate of weight loss for GmAc is found to be fractionally higher than those of others, observed from the corresponding DTG peaks at 74, 82, and 75 °C for TerP, GmAc, and GmAc-FluoroTerP, respectively (Fig. S2). The adjacent –COOH and –CONH2 functional groups in GmAc-FluoroTerP facilitated the formation of anhydride and imide functionalities through elimination of H2O and NH3 along with their subsequent dissociation within 260–514 °C. This phenomenon is supported by the DTG peaks at 242/335/361 °C. Beyond that, decomposition of polymer backbone begins for GmAc-FluoroTerP, as indicated by DTG peak at 727 °C. Overall, the elevated thermostability of GmAc-FluoroTerP compared to TerP beyond 663 °C can be attributed to the strong association of GmAc with TerP.Figure 5(a–c) C1s, N1s, and O1s of GmAc-FluoroTerP; (d–f) C1s, N1s, and O1s of Dy-GmAc-FluoroTerP; and (g–i) C1s, N1s, and O1s of Yb-GmAc-FluoroTerP.The XRD of GmAc shows sharp peak at 2θ = 19.07° (Fig. 7e) that is indicative of a highly crystalline natural polymer. Both TerP and GmAc-FluoroTerP contain polymer specific peaks at 2θ = 22.48 and 23.66°, respectively. In contrast, the prevalence of organized Miller planes in the crosslinked TerP and GmAc-FluoroTerP moieties can be inferred from the prominent XRD peaks at 2θ = 11.06 and 10.13°. The FESEM photomicrograph of TerP (Fig. 8a) showed featureless dense morphology, whereas distinct phase separated heterogeneous morphology were observed for GmAc-FluoroTerP (Fig. 8b).Characterization of Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerPThe adsorption of Dy(III) and Yb(III) by GmAc-FluoroTerP can be inferred from the appearance of distinct Dy(III) and Yb(III) specific peaks around 1297/1335 and 195 eV, in the wide scan survey spectra of Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerP, respectively (Fig. 6a). Despite C1s, N1s, and O1s peaks, all three survey plots contain Na1s and Na1s Auger peaks around 1071 and 978/497 eV, justifying the presence of Na(I)-salts introduced during neutralization of IA by concentrated NaOH. From EPR analyses (X-band studies) of Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerP, g-values have decreased from 1.91 of GmAc-FluoroTerP to 1.85 and 1.87 in Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerP, respectively (Fig. 4b), suggesting the alteration in the number of electrons within the polymer moiety.Figure 6(a) High resolution survey plot, (b) Dy3d of Dy-GmAc-FluoroTerP, and (c) Yb4d of Yb-GmAc-FluoroTerP.The chemical conjugation of Dy(III) with –COOH/–CONH2/–CONH–, –COO−, and –COOH functionalities can be inferred from the shifting of O1s BEs from 530.44, 531.15, and 531.85 eV of GmAc-FluoroTerP to 530.99, 531.65, and 532.77 eV (Fig. 5e), respectively, in Dy-GmAc-FluoroTerP. These shifts are supported by the lowering of Dy 3d3/2 and 3d5/2 peaks from 1335.6051 and 1298.6052 eV of Dy(III)-compounds to 1335.36 and 1297.49 eV (Fig. 6b) in Dy-GmAc-FluoroTerP, respectively. Notably, the spin–orbit coupling of Dy 3d3/2 and 3d5/2 peaks of 37.87 eV ensure the presence of Dy as Dy(III), even after adsorption onto GmAc-FluoroTerP. In Dy-GmAc-FluoroTerP, the shifting of N1s peaks is observed to 399.56/399.95 eV (Fig. 5f) because of –CONH2 of AM and –CONH– of BAMSA comonomer, suggesting the involvement of both N-atoms of amide and imide in ionic bonding with Dy(III). The prevalence of both ionic and coordinate bonding in Dy-GmAc-FluoroTerP can be inferred from the newly appearing Dy–O peaks at 528 and 476 cm−1 (Fig. 7b)53. The shifting of C=O asym. and sym. str. of –COO− from 1561 and 1408 cm−1 to 1610/1554/1542 and 1445/1407 cm−1, respectively, in Dy-GmAc-FluoroTerP, are further evidence of multiple modes of coordination of the GmAc-FluoroTerP with Dy(III). Other observations that substantiate ionic and coordinate bonding between Dy(III) and GmAc-FluoroTerP include the shifts of Δν = 203/147/109/135/97 cm−1 corresponding to M (monodentate)/I (ionic)/BB (bidentate bridging)/BC (bidentate chelation) coordination (Table S4). We relied on the reports by Zeleňák, et al. and Mccluskey, et al. for inferring metal ion complexation with acrylate donors from FTIR data54,55. The substantial increase in the thermostability of Dy-GmAc-FluoroTerP above 100 °C compared to GmAc-FluoroTerP (Fig. 7d), also reflects the multiple modes of coordination. The FESEM photomicrograph of Dy-GmAc-FluoroTerP showed mostly bulk penetrations with modest surface depositions of Dy(III)-complexes (Fig. 8c). The sharp XRD peaks of Dy-GmAc-FluoroTerP, observed at 10.80, 22.63, and 39.20° (Fig. 7f) are in agreement with the structures captured by FESEM.Figure 7(a, b) FTIR, (c, d) TG, (e, f) XRD plots of TerP/GmAc/GmAc-FluoroTerP and GmAc-FluoroTerP/Dy-GmAc-FluoroTerP/Yb-GmAc-FluoroTerP.Figure 8FESEM images and/or EDX plots (inset) of (a) TerP, (b) GmAc-FluoroTerP, (c) Dy-GmAc-FluoroTerP, and (d) Yb-GmAc-FluoroTerP and (e–i) TEM images of GmAc-FluoroTerP at varying magnification (2 µm–200 nm).The coordination of Yb(III) with –COOH/–CONH2/–CONH–, –COO−, and –COOH functionalities can be inferred by the shifting of O1s BEs to 531.03, 531.87, 532.63 eV (Fig. 5h) in Yb-GmAc-FluoroTerP. The XPS of Yb-GmAc-FluoroTerP is indicative of the coordinative attachment of Yb(III) ions with the O-donors of GmAc-FluoroTerP. Such coordination is substantiated from the reduction in satellite of Yb 4d3/2, Yb 4d3/2, satellite of Yb 4d5/2, and Yb 4d5/2 BEs from 205.9/198.8/192.4/188.1/184.656 eV of Yb(III)-salts to 192.99/191.36/188.73/185.25/184.01 eV, respectively, in Yb-GmAc-FluoroTerP (Fig. 6c). Also, like Dy(III) in Yb-GmAc-FluoroTerP, the shifting of N1s peaks to 399.61/400.27 eV (Fig. 5i) implies the interaction of both N-atoms of amide and imide with Yb(III). In fact, both ionic and coordinate bonds in Yb-GmAc-FluoroTerP are confirmed by the observation of new Yb–O57,58,59 peaks at 618, 550, 460 cm−1, and the shifting of C=O asym. and sym. str. of –COO− to 1614/1556/1541 and 1447/1405 cm−1, respectively, in Yb-GmAc-FluoroTerP. The Δν = 209/151/109/136/94 cm−1 related to M/I/BB/BC coordination in Yb-GmAc-FluoroTerP and superior thermostability of polymer backbone above 100 °C in Yb-GmAc-FluoroTerP provide additional support for the stronger binding of Yb(III). The superficial depositions of Yb(III)-complexes in FESEM photomicrograph of Yb-GmAc-FluoroTerP (Fig. 8d) are consistent with the sharp XRD peaks observed at 23.72 and 40.55°.Fluorescence of GmAc-FluoroTerP and sensing of Dy(III) and Yb(III)We first investigated the fluorescence behaviour of GmAc-grafted terpolymer hydrogel in comparison to its constituent monomeric units namely, IA, AM, and GmAc, as well as the polymeric components CoP and TerP. Optical images of these constituents are shown in Fig. 9A. TerP and GmAc-FluoroTerP display strong blue fluorescence upon UV illumination, in contrast to negligible fluorescence of the monomeric components. The intrinsic fluorescence of the synthesized TerP and GmAc-FluoroTerP were confirmed by these measurements. We scrutinized the arrival of fluorophore in TerP by synthesizing the CoP hydrogel with lower amounts of initiators while maintaining other conditions of the synthesis. The lack of fluorescence of CoP suggests the in situ generated C–N coupled third comonomer (BAMSA) to be the fluorophore. This comonomer is present in TerP and GmAc-FluoroTerP. We have earlier confirmed the incorporation of this comonomer by NMR, XPS, FTIR, and EPR analyses.Figure 9(A) Optical fluorescence images of IA, AM, GmAc, CoP, TerP, and GmAc-FluoroTerP in different solvents and solid state, (B) fluorescence spectra of GmAc-FluoroTerP in different solvents (CHCl3, DMSO, and DW), (C) fluorescence of GmAc-FluoroTerP in mixed solvents displaying the properties of AIEE, and (D) solid state fluorescence spectra of TerP, GmAc-FluoroTerP, Dy-GmAc-FluoroTerP, and Yb-GmAc-FluoroTerP, and CIE plots of GmAc-FluoroTerP in unary and binary solvents (inset of CIE plots: zoomed portion of CIE plots).The elevated fluorescence intensity of GmAc-FluoroTerP as compared to TerP, can be attributed to hydrogen bonding-driven fluorescence emission of the fluorophore. The fluorescence spectrum of GmAc-FluoroTerP was recorded in polar protic (DW), polar aprotic (DMSO), and weakly polar (CHCl3) solvents. The λmax of GmAc-FluoroTerP appear at 247/273, 271, and 275/333 nm (Fig. S3) in CHCl3, DMSO, and DW, respectively. The variable intensities of UV–visible spectra can be related to the variable concentrations and extinction coefficients of GmAc-FluoroTerP across solvents. The observed bathochromic shifting with the increasing polarity of solvent can be ascribed to the gradual decrease in energy difference between ground and excited states. Using these excitation wavelengths, the emission maxima of GmAc-FluoroTerP were observed at 434, 421, and 445 nm (Fig. 9B), respectively, in CHCl3, DMSO, and DW. We measured the solid-state fluorescence of TerP and GmAc-FluoroTerP, and observed a greater fluorescence intensity of GmAc-FluoroTerP, analogous to the solution-state measurements (Fig. 9D). The blue-light emission from GmAc-FluoroTerP in solid and solutions states are further characterized by CIE 1931 plots (Fig. 9D). We also tested the robustness of the blue fluorescence of the GmAc-FluoroTerP hydrogel in solid and solution state, by assessing it as a proverbial security ink (Fig. 10A). Any writing, barcode, QR code etc., written by aqueous suspension of GmAc-FluoroTerP is invisible to the naked eye, but visible under UV-illumination. Even upon drying of the hydrogel, the fluorescent ink continues to perform adequately based on the compatible fluorescence of the solid form of GmAc-FluoroTerP.Figure 10(A) Anticounterfeiting applications of GmAc-FluoroTerP (from left to right: illuminated IITGN written by security ink in UV, illuminated blank slide under UV lamp, and control plate having no writing); (B) turn-off sensing of GmAc-FluoroTerP in presence of Dy(III) and Yb(III), and (C) I/I0 plots of GmAc-FluoroTerPs in presence of various REMs.We next investigated the fluorescence of GmAc-FluoroTerP in solid-state, unary solution-state, and in binary solvents of different polarities. The gradual addition of DMSO in DW solution of GmAc-FluoroTerP resulted in a maximum fluorescence intensity at 2:1 (v/v) mixture of the two solvents (Fig. 9C). The identical observation occurred for the reverse study, i.e., gradual addition of DW in DMSO solution of GmAc-FluoroTerP. We subsequently plotted relative intensity as a function of vol.% for both DW/DMSO and DMSO/DW ratios. The Gaussian fitting of both the plots indicated the most probable value for having the highest relative emission intensity at DMSO/DW = 61.67 vol.% ratio and DW/DMSO = 150 vol.% (Fig. 9C). The enhanced fluorescence intensity in the mixed solvents could be due to aggregation induced enhanced emission (AIEE)34. We conducted TEM measurements on GmAc-FluoroTerP to assess aggregate formation. TEM images of GmAc-FluoroTerP in 2:1 (v/v) mixture of DW:DMSO are shown in Fig. 8e–i, and clearly indicate nanoscale flower-like self-assemblies60. DLS measurement of the GmAc-FluoroTerP in DW revealed a single hydrodynamic radius (Rh′) at 299 nm (Fig. 11). In contrast, three distinct Rh′ values at 205, 446, and 2644 nm were observed in the DW + DMSO mixture (2:1 (v/v) mixture of DW:DMSO), indicating various molecular aggregates formed under those conditions.Figure 11DLS analyses of (a/b) TerP/GmAc-FluoroTerP in DW, (c) GmAc-FluoroTerP in mixed solvent, i.e., DW/DMSO in 2:1 (v/v) mixture, and (d/e) Dy-GmAc-FluoroTerP/Yb-GmAc-FluoroTerP in DW.From among the two synthetic monomers used for synthesizing GmAc-FluoroTerP, IA has a larger O-content while AM is a rich N-donor. With the goal of modulating the O-donor: N-donor ratios in GmAc-FluoroTerP, we varied the IA:AM ratio from 1:5 in GmAc-FluoroTerP1, to 5:1 in GmAc-FluoroTerP2. Both GmAc-FluoroTerP1 and GmAc-FluoroTerP2 were studied for ligand-selective quenching of REMs. Interestingly, the fluorescence intensities of GmAc-FluoroTerP1 and GmAc-FluoroTerP2 were dramatically reduced upon the gradual addition of Dy(III) and Yb(III) (Fig. 10B). None of the other REMs tested had a significant fluorescence quenching effect. To assess the ligand-selective turn-off sensing of Dy(III)/Yb(III) by GmAc-FluoroTerP1/GmAc-FluoroTerP2, 10 mg L−1 DW solutions of Ce(III), Sm(III), Eu(III), Gd(III), Pr(III), Nd(III), Dy(III), and Yb(III) were added into DW solutions of GmAc-FluoroTerP1/GmAc-FluoroTerP2, and emission intensities of all the solution mixtures were recorded at the excitation wavelength of 333 nm. Based on the observed emission spectra, the relative intensities (I/I0) of GmAc-FluoroTerP1 and GmAc-FluoroTerP2 reduced significantly only for Dy(III) and Yb(III), respectively (Fig. 10C). We used the fluorescence turn-off assay using GmAc-FluoroTerP sensor to determine LOD of Dy(III) and Yb(III) as 0.13 nM and 60.8 pM, respectively (Fig. S4). These LODs are significantly superior to those reported previously for both the REMs61,62,63. Quenching of fluorescence intensity after addition of Dy(III) and Yb(III) was also justified from the significant lowering of < t > value from 2.967 ns (Fig. 12A) of GmAc-FluoroTerP to 2.046 and 1.599 ns in Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerP, respectively64. Considering the preference for Dy(III) and Yb(III) for a particular configuration of the GmAc-FluoroTerP, the observed selectivity in the fluorescence turn-off sensing could be rationalized by application of Pearson’s HSAB principle. The preference for hard-hard and soft–soft interactions compared to cross-combinations, as per the HSAB principle, align with the preference of Dy(III) to bind with the softer N-donating GmAc-FluoroTerP1. The relative hardness/softness of Dy(III) and Yb(III) can be inferred from the differences in their ionic radii, 91.2 and 86.8 pm for Dy(III) and Yb(III), respectively, implying softer Dy(III) relative to Yb(III)65. Accordingly, Yb(III) is strongly bonded with O-donating GmAc-FluoroTerP2. To the best of our knowledge, the direct application of HSAB principle in developing a ligand-selective multi-atomic polymeric sensor material is yet to be reported.Figure 12(A) TCSPC measurements (B) logic gate experiment towards turn-off quenching of fluorescence intensity of GmAc-FluoroTerP in presence of Dy(III) and Yb(III), and (C) schematic representation of logic gate experiment of fluorescence quenching.To address the practical challenges of selective sensing of Dy(III) and Yb(III) by GmAc-FluoroTerPs, the sensing experiments were carried out in binary, ternary, quinary, and complex mixtures. The binary solutions were prepared by separately mixing Dy(III) and Yb(III) with one of M(III) [M(III): Ce(III), Sm(III), Eu(III), Gd(III), Nd(III), and Pr(III)]. Ternary and quinary solutions were prepared strategically to cover the permutation-combination of all the REMs with Dy(III)/Yb(III). Accordingly, the fluorescence quenching data were obtained for a total of 44 solutions [22 for Dy(III) and 22 for Yb(III)] and the relative intensity plots were prepared (Fig. 10C).From Fig. 10C, the I/I0 value of GmAc-FluoroTerPs (i.e., control) solution remained almost unaltered upon addition of Ce(III), Sm(III), Eu(III), Gd(III), Nd(III), and Pr(III). However, I/I0 value of GmAc-FluoroTerP1 fell sharply after addition of Dy(III). For binary mixtures of Dy(III) with M(III), the quenching effects were easily inferred. Though the quenching effects were not as spectacular as the unary solution, the performance of GmAc-FluoroTerP could be easily visualized in the binary mixtures. Since M(III) individually had no fluorescence quenching phenomenon, therefore, the quenching effect in binary solutions mostly originates from Dy(III). Nevertheless, due to competitive attachment of other M(III) with GmAc-FluoroTerP1, the efficiency of sensing dropped slightly. We observed analogous behaviour for ternary and quinary solutions. Notably, in all cases, fluorescence quenching could be connected to the presence of Dy(III). GmAc-FluoroTerP1 is thus capable of selectively sensing Dy(III) from its unary, binary, ternary, quinary, and complex mixtures with other REMs. We observed a similar pattern of behaviour for sensing of Yb(III) by GmAc-FluoroTerP2.GmAc-FluoroTerP response to Dy(III) and Yb(III) as a logic gateWe used EDTA as a control to confirm the complexation of Dy(III)/Yb(III) with GmAc-FluoroTerP on the basis of fluorescence quenching. EDTA solution does not possesses any fluorescence properties at 445 nm i.e., the emission maxima of GmAc-FluoroTerP in DW and therefore, polymer-EDTA solution exhibited the undisturbed fluorescence intensity with the mere GmAc-FluoroTerP solution. However, no fluorescence quenching in Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerP can be observed in presence of EDTA solution. Since, EDTA is a stronger chelating ligand compared to GmAc-FluoroTerP, it binds with Dy(III) and Yb(III) faster than that of GmAc-FluoroTerP, leaving the unaltered fluorescence of GmAc-FluoroTerP. Therefore, the complexation reaction between Dy(III)/Yb(III) with GmAc-FluoroTerP can be captured as a truth table (Table 2) corresponding to the logic gate (Fig. 12B,C). In both the cases, NOT and NAND gates with respect to EDTA and Dy(III)/Yb(III) associated closely the truth table inferred herewith. To the best of our knowledge, the use of logic gate concept for identifying metal-polymer complexation as the basis of fluorescence quenching of hydrogel by REMs is still unexplored.Table 2 Truth table for logic gate.Harvesting of Dy(III) and Yb(III) by poly[itaconic acid]/PIA, poly[acrylamide]/PAM, and GmAcThe individual harvesting potential of the ingredients of GmAc-FluoroTerP, i.e., IA, AM, and GmAc, were studied first to ensure their possible contribution in the harvesting potential of GmAc-FluoroTerP towards Dy(III) and Yb(III). Since IA and AM are water-soluble, the adsorption experiments could not be carried out by using these ingredients. Therefore, their homopolymers (i.e., PIA and PAM), synthesized by using the same crosslinking agent (i.e., MBA) and initiators (i.e., NaHSO3 and K2S2O8), were studied for harvesting Dy(III) and Yb(III) from their unary solutions. While the PIA hydrogel displayed good adsorption capacity of 60.82 and 58.45 mg g−1 for Dy(III) and Yb(III) adsorption, respectively, it was found to be very fragile. Therefore, it was extremely challenging to desorb metal ions from PIA and study the re-application of PIA for multi-cycle adsorption–desorption process. Though PAM hydrogel had better mechanical stability than that of PIA, presence of amide functionality alone was unable to produce complexation with REMs. Hence, very low adsorption capacity of 47.95 and 47.13 mg g−1 for Dy(III) and Yb(III) adsorption, respectively, were observed (Fig. S5). For GmAc, though we observed higher adsorption capacity than those of PIA and PAM in the first cycle of adsorption (85.28 and 83.62 mg g−1 for Dy(III) and Yb(III) adsorption, respectively), isolation of GmAc particles after adsorption was not possible, once again highlighting the constraints in studying desorption and reusability. Overall, GmAc particles are unsuitable for use as adsorbent. GmAc-FluoroTerP addresses the disadvantages and constraints surrounding use of the individual ingredients. This hydrogel can adsorb REMs efficiently, desorb the already-adsorbed REMs via simple alteration of pHa, maintain enough network integrity to exhibit multi-cyclic adsorption–desorption performances and is compatible with post-adsorptive applications.Harvesting of Dy(III) and Yb(III) by GmAc-FluoroTerPs from unary mixtureOur motivation for developing a hydrogel-sensor for Dy(III) and Yb(III) was based on leveraging the extensive adsorptive interactions that could be exerted by the polymeric receptor. We next studied the adsorptive interaction of GmAc-FluoroTerPs with Dy(III) and Yb(III). The pHPZC values of GmAc-FluoroTerPs varied within 5.61–5.85 (Fig. 2f) depending upon the relative population of O- versus N-based ligands present in polymer network. Both the GmAc-FluoroTerPs are expected to be rendered anionic at pHa > 5.85 due to the deprotonation of –COOH of IA. Hence, Dy(III) and Yb(III) adsorption were carried out at pHa = 7.0, to facilitate the electrostatic interaction between negatively charged GmAc-FluoroTerPs and REM cations. The adsorption isotherm data were fitted to different isotherm models, such as Langmuir, Freundlich, BET, Sips, and Henry38. The equilibrium adsorption data fits best with Langmuir model based on the highest adjusted R2/F and the lowest χ2 values21. The maximum AC, i.e., qmax varied as 125.94/119.61, 125.57/109.22, 125.73/113.67, and 108.50/112.07 mg g−1 for Dy(III)/Yb(III) at 288, 298, 308, and 318 K (Fig. 13a,d and Table S5), respectively. Considering that the Langmuir model depends on the presumption of monolayer adsorption only, the adsorption of Dy(III)/Yb(III) on to the surface of GmAc-FluoroTerP is assumed to be monolayered. The better fitting of kinetics data with a pseudo second order kinetics model, compared to fitting with pseudo first order kinetics model (Fig. 13b,e), and measurement of activation energy of 31.18/42.55 kJ mol−1 indicate chemisorption of Dy(III)/Yb(III) by cognate GmAc-FluoroTerP. The chemisorption is justified by formation of ionic and/or co-ordinate bonding, as inferred previously from FTIR and XPS analyses. The exothermic and spontaneous nature of chemisorption are supported by the negative values of ΔH0/ΔG0 and the positive values of ΔS0 throughout the working temperatures (Fig. 13c,f and Table S6).Figure 13(a/d) Langmuir fitting of Dy-GmAc-FluoroTerP/Yb-GmAc-FluoroTerP, (b/e) pseudo second order fitting of Dy-GmAc-FluoroTerP/Yb-GmAc-FluoroTerP, and (c/f) Arrhenius fitting of Dy-GmAc-FluoroTerP/Yb-GmAc-FluoroTerP.Harvesting of Dy(III) and Yb(III) by GmAc-FluoroTerPs from binary and complex mixtures of lanthanoidsThe competitive adsorption of Dy(III) and Yb(III) from binary [i.e., Dy(III) + Ce(III), Dy(III) + Sm(III), Dy(III) + Eu(III), Dy(III) + Gd(III), Yb(III) + Ce(III), Yb(III) + Sm(III), Yb(III) + Eu(III), and Yb(III) + Gd(III)] and complex mixtures [i.e., Complex1: Dy(III) + Ce(III) + Sm(III) + Eu(III) + Gd(III) and Complex2: Yb(III) + Ce(III) + Sm(III) + Eu(III) + Gd(III)] were carried out to envisage any selective adsorption of these metal ions in presence of other lanthanoids. To carry out isothermal adsorption experiment from binary mixture, 25 mL of 20 ppm Dy(III)/Yb(III) solution was mixed with 25 mL of 20 ppm M(III) solutions [M(III) = Ce(III), Sm(III), Eu(III), Gd(III), Nd(III), and Pr(III)], to obtain 50 mL mixture of 10 ppm Dy(III)/Yb(III) and 10 ppm M(III). Similarly, the complex mixtures were prepared by mixing 10 mL each of 50 ppm Dy(III)/Yb(III) with 10 mL each of 50 ppm M(III). Therefore, the resultant complex mixture contained 10 ppm equivalent of 5 REMs [Complex1: four M(III) + Dy(III) and Complex2: four M(III) + Yb(III)]. To each of these resultant solutions, 0.015 g of dried GmAc-FluoroTerP was added, stirred on magnetic stirrer to achieve equilibrium, and the residual concentration of metal ions, i.e., unadsorbed metal ions, were evaluated by ICP-MS. The %adsorption was calculated by using Eq. 3:$$\% adsorption = \frac{{C_{0} – C_{t} }}{{C_{t} }} \times 100\%$$
(3)
here C0 and Ct represent concentration of REMs at t = 0 and t, respectively.The %adsorption of Dy(III) and Yb(III) manifested the antagonistic effects experienced by the polymer in presence of other REMs owing to the presence of other common ions. Thus, the %adsorption values of Dy(III)/Yb(III) reduced from 91.20 ± 1.50%/93.50 ± 1.75% of unary solutions to 65.25 ± 1.35%/69.88 ± 1.45%, 62. 25 ± 2.25%/67.35 ± 3.75%, 60.12 ± 2.75%/62.65 ± 1.75, 58.55 ± 2.50/57.48 ± 3.76%, and 34.25 ± 4.78%/43.95 ± 5.05% (Fig. 14) in presence of Ce(III), Sm(III), Eu(III), Gd(III), and complex mixtures, respectively. The %adsorption of both Dy(III) and Yb(III) were found to be > 90%, 57–65%, and 34–44% for unary, binary, and multi-component adsorption processes, respectively. On the other hand, the %adsorption of other REMs were found to vary within 50–55% for binary mixtures and 27–35% for multi-component mixtures. Therefore, no significant selectivity could be inferred towards adsorption of Dy(III) and Yb(III) in presence of either one REM or multiple REMs.Figure 14Assessment of selectivity of GmAc-FluoroTerP towards adsorption of Dy(III) and Yb(III) from unary, binary, and complex mixtures of REMs: %adsorption of (a) Dy(III), (b) Yb(III), and (c) other REMs.Conductivity measurement of GmAc-FluoroTerP before and after metal ion loading: post-sensing and post-harvesting applicationThe successful detection of REMs is only one part of a broader process that facilitates use of the harvested elements. The already-harvested REMs could be easily desorbed by simple change of the ambient pH to 2.0, and the recovered GmAc-FluoroTerP possessed almost identical properties to that of the initial state, confirmed by XRD peaks of GmAc-FluoroTerP, obtained after REM-desorption (Fig. S6). Instead of a typical recycling test, we planned an innovative post-harvesting application on the REM-adsorbed-GmAc-FluoroTerP by use of CV and EIS experiments. Despite using three electrode techniques and solution-phase analyses, glass slides and solid hydrogel have been utilized to measure the electrical properties. The open circuit potential of GmAc-FluoroTerP was found to be 0.197 V. Therefore, the entire EIS experiments were conducted at 0.197 V as initial potential. The conductivity of glass slides was found to be 1.21 × 10−6 S cm−1 and hence was considered as negative control. The Nyquist and Bode plots are presented in Fig. 15a,b. From the fitting parameters of Nyquist plots, the conductivity of GmAc-FluoroTerP is determined to be 6.86 × 10−4 S cm−1 (Fig. 15c), placing it in the category of semiconductors66. Importantly, conductivity of GmAc-FluoroTerP is found to be higher than that of TerP (5.75 × 10−4 S cm−1), insinuating the involvement of lone pair electrons of –OH/–COOH of GmAc towards conductance. Since GmAc-FluoroTerP exhibits selective sensing of Dy(III) and Yb(III), we studied the relative change in conductivity of GmAc-FluoroTerP after loading of Dy(III), Yb(III), Eu(III), Sm(III), Gd(III), and Ce(IV). Interestingly, the conductivity changes from 6.86 × 10−4 S cm−1 of GmAc-FluoroTerP to 9.74 × 10−4, 12.16 × 10−4, 9.08 × 10−4, 8.34 × 10−4, 7.69 × 10−4, and 5.72 × 10−4 S cm−1 for Dy(III), Yb(III), Eu(III), Sm(III), Gd(III), and Ce(IV) doping, respectively (Fig. S7). Interestingly, all the conductivity values fall within the range of semiconductors. The semiconducting nature of the hydrogels were further envisaged by calculating the band gap (Eg) of the hydrogels. The Eg of GmAc-FluoroTerP was calculated to be 1.37 eV, which is perfectly suitable for semiconductors. Interestingly, the Eg values reduced to 0.77 and 0.49 eV for Dy(III)- and Yb(III)-GmAc-FluoroTerPs, respectively, that suggests the enhancement in semiconducting phenomenon after REM-doping. The increase in conductivity after complexing with lanthanoids, except in case of Ce(IV), can be attributed to the availability of free electrons introduced within GmAc-FluoroTerP from metal ions. The sole exception of Ce(IV) among the currently studied set of elements can explained from the inappropriate coupling of anionic [Ce(SO4)4]4− with the anionic GmAc-FluoroTerP at pH > pHPZC. Notably, the greatest increase in conductivity for Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerP indicates the maximum diffusion of metallic electron with the microstructures of GmAc-FluoroTerP possibly due to the strongest complexation. The greater conductivity aligns with the selective sensing of Dy(III) and Yb(III) by GmAc-FluoroTerP among the other lanthanoids. From CV analyses, the specific capacitance (Cp) of GmAc-FluoroTerP is found to reduce from 1341.34/2018.01 F g−1 to 1009.68/1160.39 and 946.42/820.82 F g−1 in Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerP, respectively, at 100/50 mV/s scan rates (Fig. 15d–f). Such reduction of Cp with increasing scan rate in CV has already been reported by Zhang et al. (2019)67. Such a decrease in Cp values after doping of Dy(III) and Yb(III) harmonizes the increased conductivity of GmAc-FluoroTerP after doping. Overall, the GmAc-FluoroTerP possesses semiconducting properties along with a fair degree of specific capacitance. The conductivity of GmAc-FluoroTerP increases after introduction of lanthanoid cations to make the loaded GmAc-FluoroTerPs as the stronger semiconductors. The maximum increase in conductivity of Dy-GmAc-FluoroTerP and Yb-GmAc-FluoroTerP indicates the strongest complexation of GmAc-FluoroTerP with Dy(III) and Yb(III).Figure 15(a) Nyquist and (b) Bode plots of TerP/GmAc-FluoroTerP/Dy-GmAc-FluoroTerP/Yb-GmAc-FluoroTerP; (c) conductivity values of TerP, GmAc-FluoroTerP, and REM-doped GmAc-FluoroTerP; and (d–f) CV plots of GmAc-FluoroTerP/Dy-GmAc-FluoroTerP/Yb-GmAc-FluoroTerP at various scanning rates.Comparison of our resultsWe studied the existing literature for harvesting of Dy(III) and Yb(III) by use of various nano particle incorporated natural polymer based materials, magnetically active nano materials such as multi-walled carbon nanotubes/graphene oxide/graphene oxide modified organic bentonite, and several microporous membrane-based materials. We have summarized the details and performance of such materials in Table 3. Existing reports have covered a range of initial concentrations (0.2–500/10–173 ppm for Dy(III)/Yb(III)), temperatures (i.e., 293–343/298–320 K for Dy(III)/Yb(III)), and pHa (i.e., 2.0–9.0/1.4–7.0 for Dy(III)/Yb(III)). The maximum adsorption capacity for GmAc-FluoroTerP reported in this work is highest compared to all previous reports.Table 3 Comparison of Reported Adsorbent Performance for Dy(III) and Yb(III).
Ligand-selective turn-off sensing, harvesting and post-adsorptive use of Dy(III) and Yb(III) by intrinsically fluorescent flower-shaped Gum Acacia-grafted hydrogels
