3.1 Characterization of Coordination Polymers
Single-crystal X-ray diffraction analysis indicates Cu-dpe(I) belongs to the monoclinic crystal system with space group C2/c (Table S1). The crystal structure of Cu-dpe(I) contains Cu(I) metal centers, dpe ligands, and monovalent H2PO4- anions. In a triangular coordination geometry, each Cu(I) center is coordinated with two nitrogens from the dpe ligand and one oxygen from the phosphate group, as shown in Fig. 1a. Each dpe ligand bridges two Cu+ ions equatorially to build the extended one-dimensional (1D) chain. H2PO4- anions are located in the interspaces of the anionic 1D chain running along the b axis. The P−OH distance ranges from 1.53 to 1.58 Å, and the P=O distances are in the range of 1.49-1.51 Å. There are two hydrogen bonds (H-bonds) between every two orthophosphates. The O-H…O hydrogen bond distance is in the range of 2.4-2.6 Å (Fig. S1). As shown in Fig. 1b, the protonated oxygen atoms are constructed in a three-dimensional (3D) network through the weak hydrogen bonding interactions.
The crystal structure of Cu-dpe(II) contains Cu(II) metal centers, one dpe ligand, one HPO42- anion and one water molecule. Cu(II) ion adopts five-coordination geometry, each Cu(II) center is coordinated with two nitrogens from two dpe ligands and three oxygens from two phosphate groups and a water molecule, as shown in Fig. 1c. The whole structure is extended into a two-dimensional layered structure. The layers are filled with guest water molecules. These water molecules stabilize the layered structure of the whole molecule by piling up with each other.
The powder X-ray diffraction (PXRD) pattern of Cu-dpe(I) and Cu-dpe(II) at room temperature is identical to the simulated pattern from the single crystal structure (Fig. 2a). Cu-dpe(I) is an orange, rod-shaped crystal and Cu-dpe(II) is a green block-like crystal, as shown in the inset of figure 2b and 2c, respectively. Thermogravimetric analysis (TGA) proﬁle of Cu-dpe(I) shows no clear weight loss below 200 °C but a gradual decrease in weight above 200 °C, as shown in Fig. S2. The first ca. 6 % weight loss is due to the release of guest water molecules (calcd. 6.53%) from 80 °C to 200°C. The second weight loss from 250 to 310 °C is around 26 %, which attributed to the absence of free dpe ligands in the structure. The thermal stability of Cu-dpe(II) is worse than that of Cu-dpe(I). From 64 °C to 95 °C, a weight loss of about 17%, which is mainly attributed to the remove of the guest water molecules (calcd. 13.06%) and coordinate water molecules (calcd. 4.35%). When the temperature rises to 150 °C, the main framework of the compound gradually collapses. Infra-red (IR) spectroscopy and Raman of Cu-dpe(I) and Cu-dpe(II) was performed in the figure S3. In IR, the characteristic band shown at 1073 cm-1 is confirmed by the antisymmetric and symmetric Cu–O(P) stretching, while the stretching vibrations of P-OR group is in the range of 1000 to 400 cm-1 [28, 29].The stretching vibrations of C-N on the aromatic ring of dpe is at 1583 and 1420 cm-1 . A strong Raman band located at 1635 cm-1 is observed, which is the stretching vibrations of v(C=N) .
Complexes with a d10 metal center have attracted extensive interest due to their excellent photoluminescence properties. In addition, N-pyridyl ligands form a charge transfer interaction with the metal center relatively easily due the presence of both n- and π-electrons [32,33]. UV-vis absorption spectra of Cu-dpe(I) and Cu-dpe(II) (Fig. S4) in the wavelength of 200 to 1000 nm were observed. The dpe has a wide absorption band from 200 to 420 nm, which attributed to the π–π* transition of large conjugated π-electron systems . There is a broad metal-to-ligand charge transfer (MLCT) (d→π*) in the visible region around 430 nm . As a result, Cu-dpe(I) exhibites a broader absorption band and a significant red-shift compared to the free dpe ligands. The absorption intensity of Cu-dpe(II) extends to the near infrared (NIR) range compared with the dpe ligand, and the complex has better light absorption performance in the range of 200-1000 nm. The adsorption band of Cu-dpe(II) shows a significant blue shift compared with the ligand, which is different from that of Cu-dpe(I). This can be attributed to metal-to-ligand charge transfer (MLCT) [13,36]. In addition, Cu-dpe(I) with a band gap of approximately 2.05 eV and Cu-dpe(II) with a band gap of approximately 3.03 eV are obtained from the Tauc plot, as shown in Fig. S4b. CPs with a band gap from 1.91 eV to 3.15eV are appropriate candidates as photocatalysts [37,38]. Therefore, Cu-dpe(I) and Cu-dpe(II) can be considered as ideal photocatalytic materials.
We investigated the solid-state fluorescence of Cu-dpe(I), Cu-dpe(II) and dpe ligand at room temperature to further understand its photoluminescent property. The maximum emission peaks are observed at 366 nm for dpe ligand (λex = 332 nm) while 388 nm for Cu-dpe(I) (λex = 247 nm) and 399 nm for Cu-dpe(II) (λex = 247 nm) (Fig. S5). In contrast to the dpe ligand, Cu-dpe(I) shows a red-shift (22 nm) and Cu-dpe(II) shows a red-shift (33 nm) while they also shows an intense emission intensity. The red shift of the emission peak may be due to the metal-to-ligand charge transfer [16, 39]. An intense emission intensity is originate from the coordinated N-donor ligands to Cu(I/II) centers, which could enhance the conformational rigidity of the ligands in compound and thereby reducing the energy loss [19, 40]. This effect thus results in a higher fluorescence intensity in Cu(I/II) complex.
3.2 Photocatalytic Degradation of MB
Based on the above characterizations, Cu-dpe(I) and Cu-dpe(II) are a potential semiconductor catalytic material. Cu-dpe(I) have d10 metal center and conjugated organic bonding group, it is expected to be used as photosensitive materials. The surface areas of Cu-dpe(I) and Cu-dpe(II) are 1.290 m2/g and 1.317 m2/g respectively, which are quite limited. Therefore, the degradation of dye should be attribute to photocatalytic degradation rather than adsorption. Several studies have shown that photocatalytic degradation can be performed using non-porous CPs. Accordingly, we evaluated the photocatalytic degradation of MB for these two compounds. The change of MB photocatalytic degradation rate with reaction time under different conditions was studied (Fig. 3a). The blank experiments show that MB degrades slowly under visible light, but the degradation efficiency is very low and can be ignored. Under the visible light, with the only addition of Cu-dpe(I) or Cu-dpe(II) is not conducive to the photocatalytic degradation of MB, and the degradation rate is approximately 20%. H2O2 has photocatalytic activity to degrade dyes under visible light, but the degradation effect is not good and only 40% of MB is degraded. Therefore, H2O2 has photocatalytic activity for MB under visible light. Further, when the catalyst and H2O2 are both added to the aqueous solution, the degradation efficiency of MB is significantly improved under visible light. The photocatalytic degradation efficiency of Cu-dpe(I) is 88.1%, and that of Cu-dpe(II) is 97.2% within 120 minutes. The combination of Cu- dpe(I/II) and H2O2 showe good degradation effect on MB. This indicates that the catalyst has good photocatalytic activity, and the degradation effect of Cu-dpe(II) is slightly higher than that of Cu-dpe(I).
In order to further compare the photocatalytic degradation efficiency, the kinetic behavior of the photocatalyst was investigated, and the photodegradation reactions were all in accordance with the pseudo-first-order reaction kinetic model (Fig. 3b). The relationship between ln(C0/C) and time is shown in the Fig. 3b. Among them, the rate constants K of degradation reactions of Cu-dpe(I) and Cu-dpe(II) containing H2O2 to degrade MB were 0.0147 min-1 and 0.0286 min-1, respectively, which were higher than those of the other three groups (MB, MB + H2O2, MB + catalyst). This indicates that the reaction rate of catalyst with H2O2 is higher than that of any other three components.
It seemed that the degradation mechanism should be associated with photo-Fenton-like process. We did free radical trapping experiments to investigate the main actives in the photocatalytic reaction. (Fig. 3c). The main actives were removed by introducing CCl4 (e- trapping agent), isopropyl alcohol (IPA, ·OH trapping agent), N2 (·O2- trapping agent) and EDTA (h+ trapping agent). The results show that ·OH is the key active group for the degradation of MB for Cu-dpe(I). For Cu-dpe(II), ·OH plays a major role in the degradation process, ·O2- plays the secondary role in the reaction rate, followed by e- has a slight promotion effect on the degradation reaction. H2O2 is the precursor of ·OH, which is an effective and highly active oxidizing substance. In the photocatalytic process of coordination polymers, ultraviolet and visible light can induce the coordination polymers to produce oxygen or nitrogen-metal charge transfer by driving electrons from the highest occupied molecular orbital (HOMO) to the lowest occupied molecular orbital (LUMO), which strongly requires an electron to return to its stable state. Thus, H2O2 or H2O captures an electron, and then reduces to form ·OH which can effectively decompose MB . While e- can be transferred to the surface of catalyst to form ·O2- with O2, and further capture H2O to form ·OH. Therefore, inhibition of e- and ·O2- in Cu-dpe(II) will lead to poor degradation effect . As shown in Fig. 3e, Cu-dpe(I) and Cu-dpe(II) retained effective MB degradation after three cycles. The complexes possess good structural stability after three cycles of MB photodegration.
3.3 Luminescence Sensing Properties
Copper-based LCPs, including Cu(II) and Cu(I), is an interesting alternative fluorescence complex. Cu(II) absorbs light of the longer wavelengths in the visible to near-infrared spectral region , indicates a low possibility of overlap mechanism for detection of metal ions . Cu(I), on the other hand, have a d10 electronic configuration and thus no d-d transitions. This could be effective to improve the selectivity of LCPs toward Fe3+ through a possible overlap mechanism. However, there are rare examples of the Cu(I) complexes for the detection of Fe3+ ions  and no reported with an overlap mechanism.
The above results clearly indicate Cu-dpe(I) is a promising photoluminescent CPs and thus maybe useful in the detection of metal ions. We first measured the fluorescence response of Cu-dpe(I)(lex = 381 nm) and Cu-dpe(II) (lex = 322 nm) to various solvents (Acetonitrile, Methanol, Ethanol, DCM, Acetone, DMF, THF, and H2O) (Fig. 4a, 4b). It shows that Cu-dpe(I) has an obvious fluorescence response to a pure water. However, Cu-dpe(II) has no specific fluorescence response to many kinds of solutions. Therefore, it is not suitable for use as a fluorescent probe. Due to excellent fluorescence properties of Cu-dpe(I), we further explored the reaction of Cu-dpe(I) in aqueous solutions for the detection of metal ions. We measured the emission spectra of Cu-dpe(I) in aqueous solutions containing trace amounts of various metal ions. As shown in Fig. 4c, Fe3+ exhibited a drastic quenching effect on the luminescence of Cu-dpe(I), while no significant influence was observed for other metal ions.
The concentration-dependent luminescence behavior of Fe3+ ion was performed to further assess the sensitivity of Cu-dpe(I) to Fe3+. As shown in Fig. 4d, the fluorescence intensity of Cu-dpe(I) in the aqueous solution gradually decreased as the Fe3+ concentrations increased from 20 mM to 200 mM. The emission intensity was almost completely quenched when the concentration of Fe3+ was up to 200 mM. The Stern–Volmer equation (I0/I = Ksv[M] + 1)  was used to analyze the corresponding quenching coefficient of Fe3+. Here, I0 is the original fluorescence intensity, I is the fluorescence intensity, and [M] is the molar concentration of Fe3+ (mmol L−1). A linear relationship was observed when Fe3+ ion ([M]) at low concentrations (inset of Fig. 4e). As a result, the quenching constant Ksv was estimated to be 1.25 ´ 104 M−1, which is comparable to those of other reported Fe3+ quenching coefficients (Table S3) [16, 44-48]. In addition, we measured the fluorescence behaviors of Cu-dpe(I) in both metal ions solutions of Mn+ with and without Fe3+ (M = K+, Na+, Pb2+, Cr3+, Cu2+, Sr2+, Ni2+, Zn2+, Co2+, and Cd2+) to further understand the selectivity of Cu-dpe(I) towards Fe3+. As shown in Fig. S6, the fluorescence intensity of Cu-dpe(I) was sharply decreased when Fe3+ ions were added into the solutions. Furthermore, a control experiment was conducted with Mn+/Fe3+ mixed ions. These results indicate Cu-dpe(I) has excellent selectivity and sensitivity for Fe3+ ions in aqueous solution.
Currently, the luminescent quenching of Fe3+ ions can be attributed to the collapse of the frameworks, cation exchange, and competitive absorption between the Fe3+ ions and the LCPs (overlap mechanism) [22, 49]. As the PXRD results shown in Fig. S7, there are no changes in the crystal structure of Cu-dpe(I) were observed after being immersed in the Fe3+ ions solution for 24 hours, suggesting the quenching is not related to the structural collapse. Moreover, it is hard for a neutral LCPs of Cu-dpe(I) to capture Fe3+ ions through a cation exchange mechanism . Therefore, the resonance energy transfer is another mechanism maybe responsible for the quenching effect . If the emission spectrum of the fluorophore (donor) has a certain degree of overlap with the absorption of the analyte (acceptor), resonance energy transfer from the donor to the acceptor can be observed . As shown in Fig. 4f, a large overlap between the absorption spectrum of Fe3+ ions solution and the excitation wavelength of compound Cu-dpe(I) is observed, while not for the other metal ion solutions. We thus confirm that this high selectivity for Fe3+ ions is contributed by the competitive absorption.