3.1 IR spectra
The IR spectra of a series of the complexes were shown in Figure S2. In the IR spectrum of 1, the characteristic peak at 1673 cm−1 is attributed to the C=N stretching vibration, while the C=N stretching vibration peak of the ligand 1, 10'phen is at 1654 cm−1. The C=N stretching vibration peak of 1 is redshifted compared to the C=N stretching vibration peak of ligand 1, 10′phen, which indicates that the metal Co(Ⅱ) is coordinated with the ligand. In the IR spectrum of 2, the vibration peak at 1663 cm−1 is considered to be the stretching vibration of C=N, while the C=N stretching vibration of ligand 2, 2ʹbipy appears at 1665 cm−1. The C=N stretching vibration of 2 is blueshifted compared to the C=N stretching vibration peak of ligand 2, 2ʹbipy, which indicates that the metal Co(Ⅱ) is coordinated with the ligand. The main infrared data were listed in Table S3.
3.2 UVvis spectra
The UVVis spectra of a series of the compounds were shown in Figure S3. The ππ* transition and nπ* transition of the ligand 1, 10′phen appear at 256 nm and 325 nm. For 1, the two strong absorption bands at 220 nm and 261 nm were attributed to the ligandtoligand charge transition (LLCT). The strong absorption band at 376 nm can be attributed to the ligandtometal charge transition (LMCT). It further shows that the metal Co(Ⅱ) is coordinated with the ligand 1, 10′phen. The two strong absorption bands at 218 nm and 321 nm are the ππ* and nπ* transitions of 2, which can be attributed to the ligandtoligand charge transition (LLCT). The ππ* transition and nπ* transition of the ligand 2, 2ʹbipy appear at 264 nm and 316 nm. The strong absorption band at 411 nm can be attributed to the ligandtometal charge transition (LMCT). It further shows that the metal Co(Ⅱ) is coordinated with the ligand 2, 2ʹbipy. Due to the wavelength limitation of the instrument, the characteristic absorption peak of Co is not shown. The main UVvis spectra data were listed in Table S4.
3.3 Powder Xray diffraction (PXRD)
The experimental and simulated PXRD patterns of 1 and 2 are shown in Figure S4. As shown in the Figure S4a and Figure S4b, although the peak intensity of the experimental PXRD pattern and the simulated PXRD pattern are slightly different, the positions of the respective peaks are almost the same. This indicates that 1 and 2 are pure phase without impurities. At the same time, as shown in Figure S5a and Figure S5b, the crystallinity of 1 and 2 was calculated according to the jade 5 software, which were 54.62 % and 74.65 %, respectively.
3.4 Thermogravimetric analyses (TG)
To verify the thermal stability of the compound, thermogravimetric analysis was performed at a heating rate of 10°C/min under a N2 atmosphere with the temperature range from 30 to 800°C. The TG curve of 1 was shown in Figure S6a, the first stage of weight loss occurs at 30560 ℃, and the actual weight loss rate is 29.6% (theoretical weight loss rate: 30.1%), which corresponds to the loss of six 1,10'phen ligands. The second weight loss stage occurs at 560800 ℃, the actual weight loss rate is 8.6%, which corresponds to the collapse of the framework, and the final residue was metal oxide. The TG curve of 2 was shown in Figure S6b, the first stage of weightlessness occurs at 30150 ℃, and the actual weight loss rate is 1.7% (theoretical weight loss rate: 1.6%), which corresponds to the loss of two water and a hydroxide ion. The second stage of weight loss occurs at 150530 ℃, and the actual weight loss rate is 16.9% (theoretical weight loss rate: 16.0%), which corresponds to the loss of three DMF and three 2, 2ʹbipy. The third weight loss stage occurs at 530800 ℃, the actual weight loss rate is 6.0%, which corresponds to the part collapse of the framework, and the final residue was metal oxide. Comparing the weight loss of 1 and 2, the thermal stability of 2 is relatively better. The main thermogravimetric analyses data were listed in Table S5.
3.5 Crystal structures of the compounds 1 and 2
Compound 1 belongs to the monoclinic system with P21/c space group. The molecular structure includes a metallic Ag cluster anda Co cluster dimer with six 1,10'phen. The metal Co atom chelates with six nitrogen atoms (N1N6) from three 1,10'phen to form a sixcoordinate octahedral configuration, as shown in Fig.1a. The coordination mode of ligand 1, 10′phen is µ2ηN1ηN1, as shown in Fig.1c. The Ag clusters consist of eight silver ions and twelve iodide ions, as shown in Fig.1b.
In the Ag cluster, both Ag1 and Ag2 form a deformed triangular pyramid structure, Ag1 is coordinated by two µ2I (I2, I8) and one µ3I (I3), Ag2 is coordinated by two µ2I (I1, I2) and one µ3I (I4). Among them, Ag3, Ag4, Ag5, Ag6, Ag7 and Ag8 all form a twisted tetrahedral configuration, Ag3 is coordinated by two µ2I (I1, I12) and two µ3I (I3, I5), Ag4 is coordinated by three µ3I (I3I5) and one µ4I (I6), and Ag5 is coordinated by two µ2I (I7, I8), one µ3I (I4) and one µ4I (I6), Ag6 is coordinated to one µ2I (I10), two µ3I (I5, I11) and one µ4I (I6), Ag7 is coordinated to two µ2I (I7, I9), one µ3I (I11) and one µ4I (I6), Ag8 is coordinated to three µ2I (I9, I10, I12) and one µ3I (I11). The coordination environment of metallic Ag ions is shown in Table S6. The bond length range of AgI bond in compound 1 is about 2.7093.252 Å, which is similar to the bond length range of AgI bond 2.7373.181 Å reported in the literatures[40–42]. Six tetrahedral AgI4 and two triangular pyramid AgI3 are connected by sharing edges to form Ag8I12 building block of Ag cluster. As shown in Fig. 1d, the Ag8I12 building blocks are connected by I12 to form a onedimensional chain structure. As shown in Fig. 1e and 1f, a more stable interlaced layered structure is formed between the metal Ag cluster part and the metal Co cluster part through the electrostatic interaction between the anion and the cation.
Compound 2 belongs to the monoclinic system with C2/c space group. The molecular structure includes a metal Ag cluster, a metal Co cluster, a hydroxide ion, three free DMF and two lattice water. In the structure, the Co cluster is a cationic cluster with +2 valence. As shown in Fig. 2a, the metal Co atom and six nitrogen atoms from three 2, 2ʹbipyridine (N1N3, N1#1N3#1; #1: 1x, y, 0.5z) form a sixcoordinate octahedral configuration. As shown in Fig. 2c, the coordination mode of ligand 2, 2ʹbipyridine is µ2ηN1ηN1. Ag clusters are anionic clusters with 1 valence and consist of ten silver ions and eleven iodide ions as shown Fig. 2b. In which, Ag1 is coordinated by four µ4I (I1, I1#1, I2, I2#1) to form a tetrahedral configuration. Ag2 is coordinated by a µ2I (I4), a µ3I (I3) and a µ4I (I4) to form a deformed triangular pyramid configuration. Ag3 is coordinated by one µ2I (I4) and two µ3I (I5, I5#1) to form a triangular pyramid configuration. Ag4 is coordinated by a µ2I (I4) and a µ4I (I6) to form a V shape. Ag5 is coordinated by one µ3I (I3) and six µ4I (I1, I1#1, I2, I5, I6, I6#1) to form a deformed pentagonal biconical configuration. Ag6 is coordinated by one µ3I (I3) and four µ4I (I1, I2, I5, I6) to form a triangular double cone configuration. The coordination environment of metallic Ag ions is shown in Table S6. The bond length range of AgI bond in compound 2 is 2.7363.212 Å, which is similar to the bond length range of AgI bond (2.7373.181 Å) reported in the literatures[40–42]. Two V shaped AgI2, three triangular pyramid configuration AgI3, one tetrahedral configuration AgI4, two triangular biconical configuration AgI5 and two deformed pentagonal biconical configuration AgI7 are connected by the sharing edges to form a "kite Shaped" Ag10I11 building block. As shown in Fig. 2d, the building blocks are connected by shared edges Ag3I4 to form a onedimensional chain structure. Similarly, like the structure of 1, as shown in Fig. 2e and 2f, a more stable interlaced layered super molecular structure is formed between the metal Ag cluster part and the metal Co cluster part through the electrostatic interaction between the anion and the cation.
3.6 Electrochemical properties of 1 and 2
Figure 3 is the CV curve of carbon paste electrode and compounds 1 and 2 electrode in NaOH electrolyte solution. Compared with the carbon paste electrode, the redox peak of 1 and 2 appeared. Fig. 3a is the CV curve of 1, The oxidation peak is at 0.276 V and its current value is 2.696 × 10−7 A; the reduction peak is at 0.015 V and its current value is 2.407 × 10−7 A. Fig. 3b is the CV curve of 2. The oxidation potential is at 0.261 V and the current value is 8.018 × 10−6 A; the reduction potential is at 0.039 V and the current value is 1.347 × 10−5 A. According to equation 1, it can be detected whether the electrode reaction is Nernst reaction. When T = 298 K, it can be simplified to ΔEp = 59/z. When the value of ΔEp is close to 59/z, the electrode reaction is a reversible reaction. The ΔEp value of 1 is 251 mV, the ΔEp value of 2 is 222 mV, and its ΔEp value cannot be close to 59/z, so the electrode reactions of 1 and 2 are all irreversible reactions. In an irreversible reaction, the number of transferred electrons in the oxidation process can be calculated by equation 2. According to equation 1, the number of transferred electrons of 1 and 2 is calculated to be 1.5 and 1.5, respectively. Therefore, the compounds 1 and 2 transfer 1.5 electrons during the oxidation process.
ΔEp =2.3RT/zF (1)
│EpEp/2│=47.7/(1α) n (2)
Where ΔEp is the difference between oxidation potential and reduction potential (mV), R is 8.314 J/(mol·K), T is temperature (K), z is the number of transferred electrons, F is 96500 C/mol, Ep is the oxidation peak Potential (mV), Ep/2 is the halfpeak potential (mV), α is the electron transfer coefficient, which is approximately 0.5, and n is the number of transferred electrons.
The Influence of Different Scanning Speed on Cyclic Voltammetry
By studying the CV curves of 1 and 2 at different scan rates, the CV curve at the best scan rate was determined. It can be seen from Fig. 4 that with the increase of the sweep rate, the peak current of the reduction potential and the oxidation potential of 1 first increase to the maximum and then decrease. At a sweep rate of 0.05 V/s, the peak current of 1 reaches the maximum and the peak shape is complete. Therefore, the best sweep speed of the CV curve of 1 is 0.05 V/s. Fig. 5a and Fig. 5b are the CV curves of 2 at different sweep speeds. It can be seen from the figure that the reduction potential and the peak current of the oxidation potential of 2 increase with the increase of the sweep rate. When the sweep speed is 0.1 V/s, the peak current reaches the maximum and the peak shape is complete. Therefore, the best sweep rate for the CV curve of 2 is 0.1 V/s.
The Influence of Different Resting Time on Cyclic Voltammetry
Through the research on the CV curve of 1 and 2 under different resting time, the CV curve under the best resting time is determined. It can be seen from Fig. 6 that the peak currents of the oxidation potential and reduction potential of 1 gradually decrease and then almost remain unchanged when the resting time increases. When the rest time is equal to 5 s, the peak current of the CV curve of 1 reaches the maximum and the peak shape is complete. Therefore, the optimal resting time of the CV curve of 1 is 5 s. It can be seen from Fig. 7 that as the sweep rate increases, the peak current of 2 also decreases, and remains unchanged after decreasing to the minimum. When the sweep speed is 5 s, the peak current reaches the maximum and the peak shape is perfect. Therefore, the optimal resting time of the CV curve of 2 is 5 s.
The influence of different pH on cyclic voltammetry curve
The CV curves of 1 and 2 were tested under the condition of phosphate buffer solution of different pH to study the effect of pH on the CV curves of 1 and 2. It can be seen from Fig. 8 that as the pH changes from 5 to 6, the peak current of the reduction potential and oxidation potential of 1 reaches the maximum; as the pH changes from 6 to 8, the peak current gradually decreases. Therefore, pH = 6 can be used as the optimal pH for the CV curve of 1. It can be seen from Fig. 9 that with the increase of pH, the peak current of 2 also increases, increases to the maximum and then decreases. When the pH is 8, the peak current reaches the maximum and the peak shape is good. Therefore, the optimal pH of the CV curve of 2 is 8.
3.7 Catalytic reduction of PNP by 1 and 2
After adding NaBH4 to the PNP solution, the color changed to bright yellow and a characteristic peak appeared at 400 nm. This is because PNP is ionized under alkaline conditions, and a large amount of H+ are released under the condition of the catalyst, and PNP is catalytically reduced to PAP. The characteristic absorption peak of PAP is at 298 nm. As shown in Fig. 10, as time increases, the characteristic peak at 400 nm gradually decreases, while the characteristic peak at 298 nm gradually increases. This means that the concentration of PNP is gradually decreasing, and the concentration of PAP is increasing. It can be seen that PNP is catalytically reduced to PAP under the combined action of 1 and 2 and NaBH4, and the reduction rates of 1 and 2 can reach more than 90%.
The effect of different concentrations of sodium borohydride
Because NaBH4 is a key parameter that controls the entire redox reaction, it is necessary to determine the optimal concentration of NaBH4. A series of catalytic reduction experiments were carried out under the conditions of NaBH4 concentration of 0.10.4 M, PNP concentration of 0.1 mM and 5 mg of the compound. As shown in Fig. 11, when the NaBH4 concentrations were 0.1 M, 0.2 M, and 0.4 M, the removal rates of PNP by the catalytic reduction of 1 were 95.7%, 94.4% and 97.4%, respectively. As the concentration of NaBH4 increases, the removal rate of catalytic reduction of PNP does not change significantly. Taking into account the length of the reduction time, the optimal concentration of NaBH4 for the catalytic reduction of PNP in 1 was determined to be 0.2 M. As shown in Fig. 12, when the NaBH4 concentration is 0.1 M, 0.2 M, and 0.4 M, the removal rate of PNP by the catalytic reduction of 2 is 30.1%, 92.9%, and 99.3%, respectively. As the concentration of NaBH4 increases, the removal rate of catalytic reduction PNP also increases, but the removal rate of 0.2 M and 0.4 M NaBH4 does not change much. Taking into account the utilization of the compound, the optimal concentration of NaBH4 for the catalytic reduction of PNP in 2 was determined to be 0.2 M.
The effect of different concentrations of PNP
In order to determine the optimal concentration of PNP in the catalytic reduction process, a series of catalytic reduction experiments were carried out under the condition that the NaBH4 concentration was 0.2 M and the 5 mg compound was unchanged. As shown in Fig. 13, when the PNP concentrations were 0.05 mM, 0.1 mM, and 0.15 mM, the removal rate of PNP by the catalytic reduction of 1 was 97.0%, 94.4%, and 90.1%, respectively. As the PNP concentration increases, the removal rate of catalytic reduction PNP decreases, but the removal rate does not change much. The time for catalytic reduction of 0.15 mM PNP solution of 1 is 70 min, which is relatively long, while the time for catalytic reduction of both 0.05 mM and 0.1 mM PNP solutions is 15 min. Taking into account the length of the reduction time and the utilization of the compound, the optimal PNP concentration for the catalytic reduction of 1 was determined to be 0.1 mM. As shown in Fig. 14, when the PNP concentration is 0.05 mM, 0.1 mM, and 0.15 mM, the removal rate of PNP by the catalytic reduction of 2 is 100%, 92.9%, and 62.8%, respectively. With the increase of PNP concentration, the removal rate of PNP by the catalytic reduction of 2 decreases. When 2 is used for catalytic reduction of 0.05 mM PNP solution, the reduction time is 2 min, which does not make full use of the compound. When catalytic reduction of 0.15 mM PNP solution, the reduction time is longer and the reduction rate is lower. Taking into account the utilization of the compound and the reduction removal rate, the optimal PNP concentration for the catalytic reduction of 2 was determined to be 0.1 mM.
The influence of the amount of different catalysts
In order to make full use of the reducibility of the compound, the optimal dosage of the compound in the catalytic reduction of PNP was determined. A series of catalytic reduction experiments were carried out under constant conditions of NaBH4 concentration of 0.2 M and PNP concentration of 0.1 mM. As shown in Fig. 15, in the case of 3 mg, 5 mg, and 10 mg of 1, the removal rates of catalytic reduction of PNP are 83.4%, 94.4%, and 97.7%, respectively. With the increase of the amount of 1, the removal rate of catalytic reduction PNP increases, but the degree of removal rate changes little. Taking into account the length of the reduction time and the utilization of the compound, the optimal dosage of 1 was determined to be 5 mg. As shown in Fig. 16, in the case of 3 mg, 5 mg, and 10 mg of 2, the removal rate of PNP by the catalytic reduction of 2 is 48.6%, 92.9%, and 100%, respectively. As the amount of 2 increases, the removal rate of catalytic reduction of PNP increases. When 10 mg of 2 is used for catalytic reduction of PNP solution, the reduction time is very short, and it is useless to make full use of the compound. When 3 mg of 2 is used for catalytic reduction of PNP solution, the reduction time is longer and the reduction rate is lower. Therefore, considering the utilization and reduction rate of the compound, the optimal dosage of 2 was determined to be 5 mg.
3.8 Kinetic Study of Catalytic Reduction Process
In order to analyze the kinetic behavior of the catalytic reduction of PNP solution by 1 and 2, we used the quasifirstorder kinetic equation 3 and the quasisecondorder kinetic equation 4 to linearly fit the catalytic reduction process of PNP, and we obtained Fig. 17 and other related parameters Table 1.
ln (Ct /C0 ) = k1t (3)
1/Ct  1/C0 = k2t (4)
Where Ce (mol/L) represents the equilibrium concentration, Ct (mol/L) represents the concentration at a certain moment, k1 (min−1) and k2 (L/(mol/min)) represent the rate constant respectively, t (min) Is the reaction time.
It can be seen from Fig. 17 that the R2 values of the pseudofirstorder kinetic equation fitted by the catalytic reduction of PNP by 1 and 2 are 0.9952 and 0.9901, respectively, which are closer to 1 than the R2 value of the pseudosecondorder kinetic equation. This indicates that the catalytic reduction of PNP solution by 1 and 2 conforms to the quasifirst order kinetic model.
Table 1
Correlation data of pseudofirstorder and pseudosecondorder kinetics
Compounds

Pseudo firstorder kinetics model

Pseudo secondaryorder kinetics model

k1 (min−1)

R2

k2 (g/(mg‧min))

R2

1

0.1797

0.9952

10.862

0.905

2

0.1044

0.9901

5.5104

0.9592

3.9 Study on the mechanism of catalytic reduction of PNP
In order to better understand the process of the catalytic reduction of PNP solution by 1 and 2, we speculated on the mechanism of the catalytic reduction process. It can be seen from the above that the compounds 1 and 2 exhibits good redox properties. Therefore, the mechanism of the catalytic reduction of PNP by 1 and 2 is: ⅰ) The silver cluster in compound 1 and 2 oxidizes BH4− to BO2− and releases H+; ii) H+ further decatalytically reduces the PNP solution[43]. It becomes a PAP solution and emits hydrogen gas.