Self‐Assembled Artificial Light‐Harvesting System Constructed Using Electrostatic Interactions in Aqueous Solution for the Sensing of Heavy Metal Cations

A supramolecular artificial light‐harvesting system (LHS) is successfully constructed in an aqueous environment using carbon dots (CDs) and α, β, γ, δ‐Tetrakis(1‐methylpyridinium‐4‐yl) porphyrin p‐Toluenesulfonate (TMPyP) via electrostatic interactions. Specifically, the CDs act as a donor, and TMPyP, which is loaded to the surface of the CDs, acts as an acceptor for the LHS. Energy transfer occurs from the CDs to the TMPyP in the assembled CDs‐TMPyP upon UV excitation. To demonstrate the applicability, the CDs‐TMPyP system is used as a sensor for Hg(II) in an aqueous solution. Subsequently, a functional material suitable for the detection and removal of Hg(II) is fabricated by integrating the CDs‐TMPyP into a polymer network.


DOI: 10.1002/adom.202203079
Such systems absorb solar energy and using highly effective energy transfer deliver excitation energy to a reaction center. Föster resonance energy transfer (FRET) plays an important part in the design of lightharvesting antenna systems. Therefore, appropriate energy donor and acceptor pairs need to be prepared. Up to now, much effort has been made to mimic natural light-harvesting antenna platforms using a variety of materials, [5] including organic molecules, [6] micelles, [7] dendrimers, [8] and porphyrin assemblies. [9] However, covalentbased LHS requires multiple synthetic steps, [10] which hinders their scalability and wider application. As a potential and simple alternative, supramolecular self-assembly provides a promising and simple technique for fabricating artificial light-harvesting materials. [11] In fact, a range of LHS, including supramolecular biomaterials, [12] organic-inorganic hybrid materials, [13] polymers, [14] and organogels [15] have been successfully fabricated using non-covalent assembly. However, most of these reported LHS are constructed in organic solvents, while some LHS have been constructed in aqueous solution. [15b,16] Nevertheless, the easy and simple fabrication of highly efficient light-harvesting assemblies in an aqueous solution is still challenging. In addition, the application of LHS constructed with supramolecular self-assembly has rarely been investigated. So far, only a few examples have been reported for LHS.
For example, Li and co-workers reported a supramolecular co-assembly based on a naphthyl-1,8-diphenyl pyridinium derivative and lower-rim dodecyl-modified sulfonatocalix[4]arene as the LHS for imaging of the golgi apparatus in PC-3 cells. [17] While Wang and co-workers developed a dendrimer-based artificial light-harvesting system for use as a photocatalyst. [8b] Herein, a novel LHS has been developed based on the electrostatic interaction and assembly between a water-soluble TMPyP and fluorescent CDs (Scheme 1). Due to the excellent overlap between the emission bands of the CDs and the absorption bands of TMPyP, an effective one-step FRET can occur from the CDs to the TMPyP. As such self-assembled and stable supramolecular nanoparticles were created when TMPyP was added to an aqueous solution of CDs, resulting in the noncovalent coating of donor CDs with a water-soluble energy acceptor the fluorescent TMPyP dye. Furthermore, the LHS constructed by electrostatic interaction can not only effectively capture light, but also www.advancedsciencenews.com www.advopticalmat.de Scheme 1. Schematic illustration of the LHS and its application for the detection of mercury ions. exhibited excellent selectivity, sensitivity and rapid response toward Hg(II) in aqueous solution. Based on the above system, we developed hydrogel and cellulose membrane materials, that can detect and remove Hg(II) simultaneously.
These results provide inspiration for the development of supramolecular assembled LHS toward imaging, sensing, diagnosis, and beyond. To the best of our knowledge, this is the first example documenting the use of carbon dots as energy donors to construct LHS using electrostatic interactions.

Results and Discussion
CDs were synthesized according to the reported procedure, [18] and TMPyP was obtained from commercial sources. The prepa-ration of CDs-TMPyP is illustrated in Scheme 1 and the detailed procedures can be found in the Experimental Section of the Supporting Information. To verify the assembly of CDs and TMPyP into nanoassemblies, we used transmission electron microscopy (TEM), and scanning electron microscope (SEM). As shown in Figure S1 (Supporting Information), the CDs were found to be well-dispersed quasi-spheroid nanoparticles with average diameters of 3.42 nm. High-resolution TEM (HR-TEM) images indicate that the CDs contain lattice fringes with a spacing of 0.21 nm, indicating the existence of a graphite-like structures. [19] In Figure S2 (Supporting Information), the X-ray diffraction image of CDs shows a diffraction peak ≈23°, demonstrating the (002) crystal face, which agrees well with the HR-TEM image. The average size of the CDs-TMPyP nanoparticles were determined to be 4.36 nm (Figure 1a; Figure S1, Supporting Information) under the optimal preparation conditions, confirming the successful preparation of particle-based nanoassemblies. The HR-TEM and element mapping results demonstrate that the CDs-TMPyP are composed of CDs and TMPyP owing to the identical lattice spacing with free CDs (Figure 1b) and even distribution of C, O, N, and S elements (Figure 1c). The size change induced by the formation of nanoassemblies was confirmed using dynamic light scattering (DLS) measurements. Compared to the hydrodynamic diameters of CDs and TMPyP (Figure 1d; Figure S4, Supporting Information), the hydrodynamic diameter of CDs-TMPyP nanoparticles was found to increase to ≈99.4 nm, indicating that the CDs and TMPyP form larger particle size assemblies in aqueous solution. The observed larger particle sizes obtained by DLS over TEM were attributed to the hydration of the nanoparticles. [19c,20] The SEM images revealed that CDs-TMPyP had a larger size than CDs ( Figure S3, Supporting Information), which was consistent with the results of TEM and DLS, verifying the successful assembly of CDs with TMPyP. A difference in particle size was observed between the SEM and the TEM that may be attributed to the difference in the sample preparation procedures. [21] Subsequently, the assembly behavior of CDs-TMPyP was investigated using X-ray photoelectron spectroscopy (XPS), Tyndall effect and Zeta potential measurements. In the XPS spectra (Figure 1e), the XPS spectrum confirms the chemical compositions of the CDs-TMPyP, which exhibit four characteristic peaks of C 1s (284.78 eV), N 1s (399.28 eV), O 1s (531.28 eV), and S 2p (167.38 eV). The HR XPS C 1s spectrum can be deconvoluted into three peaks at 283.2, 284.6, and 286.4 eV ( Figure S5, Supporting Information), due to C-C, C=C, and C=O bonds, respectively. [22] The HR XPS O 1s spectrum can be fitted by three peaks at 530, 531.8, and 533 eV that can be assigned to O-H, O=C, and C-O bonds, respectively. [23] For the N 1s, there are approximately three peaks at 397.8, 399.2, and 400.3 eV, corresponding to the C-N, -NH and C=N bonds, respectively. [24] The highresolution XPS spectrum of S 2p ( Figure 1f) reveals two dominant peaks at 161.4 and 163.1 eV, corresponded to S 2p3/2 and S 2p1/2, respectively. [25] These results confirm that the surface of the CDs-TMPyP contains hydroxyl, carboxyl, and amino functional groups. Additionally, a simple mixture of CDs with TMPyP in an aqueous solution displayed a noticeable Tyndall effect, demonstrating the formation of large aggregates ( Figure S6, Supporting Information). In contrast, neither CDs nor TMPyP displayed any obvious Tyndall effect, indicating that large aggregates were not formed with both CDs and TMPyP under the same conditions. Furthermore, the zeta potential of the CDs is −4.24 V (Figure 1g), and the zeta potential of TMPyP is +3.81 V (Figure 1h), indicating that CDs and TMPyP have opposite charges in aqueous solution, which confirms our premise that the CDs and TMPyP are associated by electrostatic attraction. When the CDs were mixed with TMPyP, the zeta potential of the CDs-TMPyP was −0.12 V (Figure 1i), verifying again the successful assembly of CDs with TMPyP, consistent with the results of TEM, DLS and Tyndall effect.
Subsequently, UV-vis and fluorescence spectroscopic methods were used to evaluate the characteristics of the synthesized CDs and TMPyP. The spectra of the CDs showed a UV absorption band at 280 nm and a fluorescence emission peak at 480 nm ( Figure S7, Supporting Information). Meanwhile, the fluorescence quantum yield of the CDs was determined to be 10.62% (Table S1, Supporting Information). Additionally, the emission band was gradually red-shifted when the excitation wavelength was varied ( Figure S8, Supporting Information). While, the absorption spectra of TMPyP exhibited a typical Soret band and four Q bands ( Figure S9, Supporting Information). The band at 420 nm was assigned to the Soret band arising from the transition of a 1u ( ) -e g * ( ), and the other four absorption maxima (520, 550, 600, and 650 nm) were attributed to the Q bands corresponding to the a 2u ( ) -e g * ( ) transitions. [26] The water-soluble fluorescent dye TMPyP was chosen as an acceptor ( Figure S10, Supporting Information), since the absorption band of TMPyP overlaps with the fluorescence band of the CDs. The fluorescence behavior of the CDs-TMPyP assembly was then investigated. As shown in Figure 2a, the fluorescence emission spectra of CDs-TMPyP excited at 400 nm showed a typical FRET phenomenon: the fluorescence emission peak of CDs donors at 480 nm decreased, while that of TMPyP acceptors at 700 nm increased with an increase of the TMPyP concentrations, which is consistent with the efficient energy transfer from the donor to the acceptor. [15b,16b,d] The fluorescence quantum yield of the CDs-TMPyP system was determined as 6.44% (Table S1, Supporting Information). Moreover, fluorescence lifetimes were determined to confirm the light-harvesting properties. First, the fluorescence lifetimes of the CDs were determined to be 1 = 3.10 ns and 2 = 12.39 ns by fitting the decay curve to a double exponential decay ( Figure S11 and Table S1, Supporting Information). Then on assembly with TMPyP, the lifetimes of the CDs-TMPyP decreased to 1 = 2.39 ns and 2 = 9.54 ns due to the electrostatic attraction between the CDs and TMPyP, suggesting that the energy can be transferred from the donor CDs to the acceptor TMPyP in the CDs-TMPyP LHS. Moreover, the antenna effect and energy transfer efficiency were investigated to quantitatively estimate the efficiency of the LHS. Based on the fluorescence quenching rate of the CDs-TMPyP system, energy transfer efficiency was calculated to be 38% at a volume ratio of donor/acceptor ratio of 167:1 (Figure 2b). Furthermore, the antenna effect was calculated to be 14.1 at a high donor/acceptor = 1000:1 ratio. In addition, since the CDs-TMPyP are assembled by electrostatic attraction, the fluorescence behavior of the CDs-TMPyP in different pH (pH 4.03 and 8.63) buffers was investigated. It was shown that with an increasing concentration of TMPyP, the fluorescence intensity of the CDs at 480 nm decreased gradually, whereas the fluorescence emission of the TMPyP at 700 nm increased (Figure 2c,e). The energy transfer efficiency was calculated to be 44.2% (pH 4.03) and 44.4% (pH 8.63) respectively at a volume ratio of donor/acceptor = 167:1. Meanwhile, the antenna effect was calculated to be 12.56 (pH 4.03) and 13.8 (pH 8.63) at a donor/acceptor ratio of 200:1, respectively (Figure 2d,f). These results are consistent with the results for the CDs-TMPyP in an aqueous environment, indicating that the CDs-TMPyP light-harvesting system constructed by electrostatic attraction displays excellent stability. In addition, since porphyrins have multiple proton donors and acceptors, the interactions with CDs can occur through intermolecular hydrogen bonds. [27] Therefore, the assembly of uncharged porphyrins and CDs was investigated in pH 4.03 buffer, using 4,4′,4′′,4′′′-(Porphine-5,10,15,20-tetrayl) tetrakis(benzoic acid) (TCPP) and meso-Tetraphenylporphyrin (TPP). With an increasing concentration of TPP (TCPP), the fluorescence intensity of the CDs at 480 nm decreased visibly, while the fluorescence emission of TPP (TCPP) at 700 nm increased weakly (Figures S12 and S13, Supporting Information), and the energy transfer efficiency was calculated to be 25.2% (27%) at a volume ratio of donor/acceptor of 167:1.
Moreover, the antenna effect was calculated to be 6.04 (4.04). The above results indicated that compared with TPP (TCPP), TMPyP exhibits a higher and more stable antenna effect, and the CDs-TMPyP system could be employed as a promising light harvesting material in an aqueous solution.
Since TMPyP can be used as a fluorescent probe for the detection of heavy metal ions in an aqueous environment. [28] The fluorescence behaviour of TMPyP was investigated upon the addition of different metal ions. As shown in Figure S14a (Supporting Information), only the addition of Hg 2+ caused the fluorescence quenching of TMPyP, and the titration experiments exhibited a very sensitive fluorescence response toward Hg 2+ for TMPyP ( Figure S14b, Supporting Information). To demonstrate the effective sensing of Hg 2+ via the CDs-TMPyP assembly, the fluorescence behavior of CDs-TMPyP upon the addition of Hg 2+ was also investigated (Figure 3a), with an increasing concentration of Hg 2+ , the fluorescence intensity of the TMPyP (acceptor) at 700 nm decreased gradually, while the fluorescence emission of the CDs (donor) increased when excited at 400 nm, and the fluorescence color changed from brown to green. It is clear that this "reverse" FRET process is turned on by Hg 2+ ions, and that a rationtric change of CDs-TMPyP occurs only in the presence of Hg 2+ ( Figure S14c, Supporting Information), suggesting that as a fluorescent probe the CDs-TMPyP assembly exhibited excellent selectivity toward Hg 2+ ions. In fact, CDs-TMPyP did not produce any observable response toward many metal ions including Mg 2+ , Na + , Mn 2+ , Ni 2+ , Fe 3+ , Al 3+ , Zn 2+ , Ca 2+ , Cr 3+ , Cu 2+ , Co 2+ , or K + (Figure 3d). A pH titration confirmed that the CDs-TMPyP exhibited stable fluorescence emission over a wide pH range from 5-12 ( Figure S15, Supporting Information), indicating that CDs-TMPyP is suitable for applications under physiological conditions. Under acidic aqueous conditions (pH 1-4), the pyrrole groups of the porphyrin skeleton can be protonated, resulting fluorescence fluctuations of CDs-TMPyP. Furthermore, fluorescence lifetime experiments indicated that the decay curve of CDs-TMPyP followed a double exponential decay with a fluorescence lifetime of 1 = 2.39 ns and 2 = 9.54 ns, with the addition of Hg 2+ , the lifetimes of CDs-TMPyP-Hg 2+ increased to 1 = 3.11 ns and 2 = 12.43 ns due to the chelation of TMPyP with Hg 2+ (Figure 3b; Table S1, Supporting Information). In addition, with the addition of Hg 2+ , the DLS of CDs-TMPyP changed from 99.4 to 165.1 nm ( Figure S16, Supporting Information), indicating that Hg 2+ removed TMPyP from the surface of CDs due to chelation, which results in the CDs fluorescence recovery (Figure 3a). Subsequently, the binding energies of CDs-Hg 2+ and CDs-TMPyP were calculated using density function theory (DFT). Since there is no known specific structure for the CDs, graphene oxide sheets were used to simulate the structure of CDs, and the functional groups on the CDs were symmetrically distributed on the edge of graphene oxide. The binding energy of the CDs-Hg 2+ is 2.088 V, while the binding energy of the CDs-TMPyP is 3.612 V (Figure 3c). The thermodynamic mechanism of the interaction between CDs, Hg 2+ , and TMPyP was investigated using isothermal titration calorimetry, as shown in Figure S17 (Supporting Information), the titration of CDs into TMPyP released significant energy. The association constant (K) of 2.28 × 10 5 corresponds to a very intense binding interaction. Additionally, the enthalpy change (ΔH) of binding was −20.87 kJ mol −1 and the stoichiometric ratio of binding (N) was 0.76, confirming that the interaction of the CDs with TMPyP was an enthalpy-driven reaction. Whereas, the K of the CDs and Hg 2+ was 2.70 × 10 4 , the ΔH of binding was −2381.76 kJ mol −1 and the N was 0.28. The negative ΔS indicates that the uniformity of the system energy was reduced, which might be attributed to the well-organized self-assembly by a weak interaction-driven reaction, such as electrostatic interactions or hydrogen bonding. The above results indicate that the binding energy of the CDs to TMPyP is much greater than that of the CDs to Hg 2+ . This is consistent with the results of DFT calculations. suggesting that when Hg 2+ was added to the CDs-TMPyP system, Hg 2+ would preferentially bind to TMPyP rather than the CDs. In order to confirm the mechanism of the interaction between TMPyP and Hg 2+ , the UV-Vis absorption spectra of TMPyP during the addition of Hg 2+ were investigated. The absorption peak of TMPyP was red-shifted after the addition of Hg 2+ , and a FRET-like effect appeared at 500-600 nm (inset of Figure S18, Supporting Information), which can be elucidated by the ligand-to-metal charge-transfer transition between Hg 2+ and TMPyP molecules. [29] The color change can be recognized by naked eyes ( Figure S19, Supporting Information), and the CDs-TMPyP gradually changed color from brown to green. After 5 min, the color remained almost unchanged, demonstrating an equilibrium for TMPyP-Hg 2+ complexation. Specifically, when Hg 2+ binds with the TMPyP ring, the lone pair of electrons of the nitrogen atom in the TMPyP ring will transfer to the vacant d orbital of Hg 2+ , which results in fluorescence quenching of TMPyP. [30] Moreover, the absorption and color remains almost unchanged after the addition of 15 μM, confirming equilibration of binding and optimum formation of TMPyP-Hg 2+ . The ratio of absorbance at the Soret band before and after the red shift can be employed to monitor the interaction of TMPyP with Hg 2+ . [31] In addition, for a concentration range from 0 to 15 μM, a linear correlation of the fluorescence intensity ratio was found upon the increasing concentration of Hg 2+ ( Figure S20, Supporting Information). The R 2 was calculated to be 0.999, indicating good linear ratiometric fluorescence response toward Hg 2+ . The emission ratio of the two wavelengths (I 480 /I 700 ) was employed for the calculation of the detection limit of 6.2 × 10 −7 M (LOD = 3 /k), which was below the minimum standard required for Hg 2+ in wastewater, as such the CDs-TMPyP nanosensor displayed high sensitivity for Hg 2+ . [28] The CDs and TCPP (TPP) are associated using hydrogen bonding, therefore the fluorescence emission of the CDs-TCPP (TPP) after the addition of Hg 2+ was investigated. As shown in Figure S21 (Supporting Information), compared to CDs-TMPyP, with an increasing Hg 2+ concentration over a range from 0 to 25 μM, the fluorescence emission of the CDs at 480 nm increased gradually, while the fluorescence intensity of TCPP (TPP) at 700 nm also increased when excited at 400 nm. These results indicated that the CDs-TCPP (TPP) could not maintain the ratiometric change with the addition of Hg 2+ , indicating that CDs-TCPP (TPP) cannot be used as a ratiometric sensor for Hg 2+ detection.
Encouraged by the fluorescence properties of CDs-TMPyP, the Hg 2+ sensing behavior of CDs-TMPyP loaded hydrogel and cellulose membrane were investigated. The hydrogel and cellulose membrane were synthesized according to reported methods, [18] and the preparation of the hydrogel is illustrated in Figure 4a and the detailed procedures can be found in Experimental Section of the Supporting Information. The fluorescent hydrogel displayed two obvious peaks, which provide great potential for the applications of Hg 2+ sensing ( Figure S22, Supporting Information). As shown in Figure 4b, the fluorescence intensity of the hydrogel at 700 nm decreased with increasing Hg 2+ ion concentrations (0-100 μM), while at 480 nm the fluorescent intensity increased, indicating that the fluorescent hydrogel exhibits excellent sensing performance for Hg 2+ . The fluorescence behavior of the hydrogel may be attributed to energy transfer between the CDs and TMPyP. In addition, the fluorescence emission of the cellulose membrane at 700 nm decreased with increasing concentrations of Hg 2+ (0-50 μM). However, the intensity was unchanged at 480 nm (Figure 4c), which also provided a reliable reference signal for the detection of Hg 2+ . Since the fluorescence intensity of Hg 2+ detected using CDs-TMPyP loaded on the hydrogel is ratiometric, the adsorption capabilities of the hydrogel were assessed. The influence of the pH on Hg 2+ uptake by the fluorescent hydrogel was evaluated. The adsorption capacity for Hg 2+ enhanced as the pH increased and a maximum was reached at pH 5.0 ( Figure S23, Supporting Information), which was attributed to the electrostatic interactions between the fluorescent hydrogel and Hg 2+ . [32] The fluorescent hydrogel is protonated and positively charged at low pH, resulting in a reduction of the adsorption capacity. [33] In order to understand the adsorption mechanism, the impact of contact time on the Hg 2+ adsorption ability was also evaluated. The uptake ability for Hg 2+ was initially rapid and subsequently slowed until equilibrium was reached after ≈1 h ( Figure S24, Supporting Information). The fitting of kinetic models are given in Table S2 and Figure S24 (Supporting Informtion), insert. Clearly, a pseudo-second-order model can be used to fit the experimental data and the correlation coefficient R 2 was calculated to be 0.999, suggesting excellent correlation between Q cal and Q exp . Therefore, the fitting results indicated that a chemisorption process dominated the adsorption process. [34] Furthermore, the fitting parameters of the Freundlich and Langmuir adsorption isotherms of the fluorescent hydrogel are illustrated in Figure S25, and Table S3 (Supporting Information), respectively. The adsorption fitted the Langmuir model (R 2 = 0.994) better at 298 K compared to the Freundlich model (R 2 = 0.962), indicating that the adsorption of Hg 2+ is by monolayer coverage. [35] The maximum uptake ability of the fluorescent hydrogel for Hg 2+ was calculated to be 124.6 mg g −1 based on the fitting results.

Conclusion
In summary, an efficient LHS was constructed in an aqueous solution using a facile supramolecular self-assembly approach, CDs and hydrophilic fluorescent dye TMPyP, were just mixed to generate an artificial light-harvesting platform that was based on a highly efficient FRET process from donor (CDs) to acceptor (TMPyP). At the donor/acceptor ratio of 1000:1 in aqueous solution, AE and Φ ET were 14.1% and 38%. Meanwhile, the AE and Φ ET achieved 12.56% and 44.2% in acid buffer solution (pH 4.03) and 13.8% and 44.4% in basic buffer solution (pH 8.63) at the donor/acceptor ratio of 200:1, respectively, which indicates that the CDs-TMPyP light-harvesting system exhibits excellent stability. Interestingly, the prepared CDs-TMPyP light-harvesting system also exhibited excellent selectivity, sensitivity, and rapid sensing response toward Hg 2+ . Simply loading CDs-TMPyP with a hydrogel, resulted in a system suitable for the efficient detection and removal of Hg 2+ from aqueous solution, this hydrogel sensor method represents a useful way to construct platforms capable of detecting and adsorbing different metal ions, and provides useful tools for environmental remediation.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.