*3.1. Probing Copper Coordination*

Amine 6 was obtained after a series of reactions starting from *<sup>L</sup>*-proline (Figure 2). The reaction between amine 6 with the corresponding aldehyde afforded ligands L1–L5. After purification, these compounds were characterized by 1H NMR, 13C NMR, 2D NMR, mass spectrometry, infrared spectroscopy, and X-ray diffractometry (see support information). The coordination of L1–L5 to CuCl2 in methanol resulted in the formation of complexes CuIIL1–L5 in high yields (71–91%). All complexes were characterized by proper microanalysis, EPR, infrared spectroscopy, mass spectrometry, cyclic voltammetry, and electronic spectroscopy at the UV–Vis region (Supplementary Materials). Some features of the characterization of the complexes deserve to be described since they indicate the coordination mode and the extent of dimerization. For instance, through FTIR spectroscopy (Figures S25–S29), it was evident that the coordination occurred via the nitrogen of azomethine [40,41], since the ν C=N vibration mode was red shifted by 4–7 cm<sup>−</sup>1. An increase of the νC–O energy indicated that the coordination was also occurring via a phenolate [34]. Hence, the presence of a band in the 640–470 cm<sup>−</sup><sup>1</sup> range corroborated that coordination occurred through oxygen and nitrogen, consistent with the ν M–O and M–N modes [42].

**Figure 2.** Synthesis of L1–L5 ligands from L-proline. Reaction conditions: (**a**) SOCl2(l), CH3OH; (**b**) BrCH2(Ph), K2CO3(s), CH3CN; (**c**) MgBr(Ph), THF; (**d**) NaN3(s), H2SO4 70%, CHCl3; (**e**) LiAlH4(s), THF; (**f**) salicylaldehyde and derivatives, Na2SO4(s), CH3OH.

Corroborating to the FTIR spectroscopy, a bathocromic shift was observed for the π→π\* bands of azomethine in the electronic spectroscopy, characterizing the coordination through this moiety [43]. In addition, the disappearance of the *n*→π\* azomethine transition indicated that the CuII coordination occurred on that position of the Schiff base [40]. Moreover, a broadband in the 550–770 nm region was consistent with the d–d transition of the metal center [43–45] of complexes CuIIL1 and CuIIL4 (636 and 650 nm, respectively). However, the other complexes exhibited a less evident d–d band, possibly due to a higher planar geometry in comparison to CuIIL1 and CuIIL4 (Figures S30–S33) [46].

In the cyclic voltammetry, for all ligands, the oxidation of phenol to quinone and the formation of radical cations was observed. These observations are in agreemen<sup>t</sup> with the redox behavior of other Schiff bases [47,48]. After coordination, these redox processes were displaced to higher potentials as an effect of electron depletion upon coordination by the metal [49].

The possibility of the formation of monomeric and dimeric structures was first inspected by high resolution mass spectrometry (HRMS) in which, for example, a peak corresponding to the monomer was present at *m*/*z* 538.1669, whereas the dimer was observed at *m*/*z* 1169.2410 for the CuIIL2 complex (Figure S58). However, since dimers can be formed in the gas phase depending on the solution concentration [50], we performed EPR analysis in solid and in solution. With these analyses, we evaluated the existence of equilibrium between monomers and dimers in solution that could be controlled by ligand substitution reactions.

### *3.2. Ligand Substitution and Electron Paramagnetic Resonance Measurements*

First, ligand substitution was followed by conductivity measurements to give us insights on the natural dissociation of the complexes. In dichloromethane (DCM) at 298 K, the conductivities were consistent with neutral compounds (values in the 2.00–6.00 μS cm<sup>−</sup><sup>1</sup> range). However, in acetonitrile (ACN) at 298 K, the conductivity was shown to constantly increase over time, somewhat reaching a plateau after half an hour (Figure 3A). The quasi-stabilized values for complexes CuIIL2, CuIIL3, and CuIIL4 were 63, 43, and 23 μS cm<sup>−</sup>1, respectively. These values indicate that the labilization of the chlorido ligand by acetonitrile was more pronounced for CuIIL2. The solvolysis of the chlorido ligand was facilitated when water was present in solution, as shown in Figure 3B, which shows

that in 2 min, the conductivity was already stabilized in limiting values of the 1:1 electrolyte range (55.0–90.0 μS cm<sup>−</sup>1) [51]. Interestingly, the conductivity of these complexes was more stable both in methanol and the methanol/water mixture (80/20 *v*/*v*), revealing a slower rate of ligand substitution in these solvents (Figures S53–S55). The observed values ranged from 45.0 to 53.0 μS cm<sup>−</sup><sup>1</sup> and 47.0 to 50.0 μS cm<sup>−</sup><sup>1</sup> in methanol and the methanol/water mixture, respectively, which were lower than that expected for the 1:1 electrolyte in methanol. The lower values might indicate the presence of a mixture of charged and neutral species in solution. The displaced chloride could be detected by the addition of silver nitrate solution as a white precipitate of AgCl.

**Figure 3.** Influence of the time in conductivity measurements of the complexes in (**A**) acetonitrile and (**B**) acetonitrile/water (80/20) mixture. Conductivity of CuIIL2 is shown as black squares, CuIIL3 is shown as red circles, and CuIIL4 is shown as blue triangles.

Considering the existence of a mixture of species in solution, EPR measurements were performed in the same solvents of the conductivity analyses to enable a better structural comprehension (Figure 4). In dichloromethane, all complexes presented values of gz higher than gx and gy (Table 1 and Table S7), suggestive of an axial symmetry [52–54], due to the presence of an unpaired electron in the dx2–y2 orbital. The axial symmetry supports the proposal of a square planar geometry of the complexes [42,55,56]. However, as shown in Table 1, the differences between the values of gx and gy are indicative of a distortion of the plane. In addition, the highest observed gz for CuIIL4 could mean that this complex has a tetrahedral distortion. In contrast, the Az is smaller for the CuIIL5 complex as a result of a greater distortion, indicating that aside from the difference between substituents in the ligands, another factor might be affecting the tetrahedral distortion of complexes CuIIL4 and CuIIL5.


**Table 1.** Electron Paramagnetic Resonance parameters for the CuII complexes of this work in dichloromethane at 298 K.

**Figure 4.** Comparison of the experimental EPR spectra of the complexes in dichloromethane at (**A**) 298 and (**B**) 77 K. Complex CuIIL1 is shown in purple, CuIIL2 is shown in red, CuIIL3 is shown in blue, CuIIL4 is shown in green, and CuIIL4 is shown in grey lines. The narrow line in the high magnetic field for CuIIL3 are a standard signal of CrIII:MgO sample (g = 1.9797).

In acetonitrile and acetonitrile/water (80:20) mixture, the values of gx and gy were more similar to each other, as an effect of ligand substitution, in which the nitrogen atom of ACN or the oxygen atom of water are coordinated to copper, forming a plane with higher symmetry than the previous N, N, O, Cl coordination plane. The increase in symmetry observed in the EPR measurements in ACN and ACN/H2O is in agreemen<sup>t</sup> with the chlorido displacement observed in the conductivity measurements. However, there was still no evidence of dimeric and monomeric species in equilibrium in these solutions.

Hence, the spinning radiuses of the molecules in solution were obtained from the spectral simulations using the EasySpin program [57] to obtain the rotational time correlation tcorr (Table 2). These values are associated withthe spinning velocity of a molecule in solution, expressing large values of tcorr when effective intramolecular interactions are generated. Curiously, the increasing order of tcorr was CuIIL4 > CuIIL1 > CuIIL3 > CuIIL2 > CuIIL5, revealing that complexes bearing a substituent group with oxygen (CuIIL2 and CuIIL5) are more effectively interacting in solution than CuIIL4. With the tcorr values, we were able to calculate the radiuses and volumes of rotation of the complexes using the Stokes–Einstein–Debye (SED) equation (tcorr = <sup>4</sup>πηa<sup>3</sup>/3KBT), where a is the molecular radius of rotation (Table 2). In general, all compounds have reduced their radius of rotation in acetonitrile and the acetonitrile/water (80:20) mixture. Therefore, the complexes were probably arranged mostly as dimeric structures in dichloromethane, and when coordinating solvents were present such as acetonitrile and water, the equilibrium between dimeric and monomeric species shifted to monomeric ones.


**Table 2.** Comparison of the values of the radiuses of rotation and approximate volumes of the complexes CuIIL1–CuIIL5 considering a spherical model.

In fact, when frozen solutions of the complexes in DCM were analyzed by EPR spectra and compared to simulated spectra (Figure 5), it was evident that a component attributed to molecular aggregates needed to be introduced for a better fit. These molecular aggregates had a magnetic

interaction, indicating that two CuII centers probably interacted with each other, causing the lines to broaden. This broadening was more pronounced for CuIIL2 and CuIIL5, which were the most effective complexes to form dimeric structures, as observed by the tcorr values. A similar trend was observed for the measurements performed in acetonitrile and in the acetonitrile/water (80:20) mixture, indicating equilibrium between monomeric and dimeric species in frozen solution. The stronger interaction observed for complexes CuIIL2 and CuIIL5 might indicate the presence of halogen-bonds between the oxygen from methoxy and ethoxy radicals with the chloride. Halogen bonds are more sensitive to steric effects and could be the reason for the higher volume observed for CuIIL2 in ACN than CuIIL5 [58]. Moreover, the dimerization of copper complexes in different solvents has already been observed by EPR measurements by other groups [56], corroborating our observations.

**Figure 5.** Comparison between the experimental and simulated EPR spectrum of CuIIL4 complex at 77 K in dichloromethane. The experimental EPR spectrum is shown as a black line, whereas the simulated spectrum is shown as a red line. Simulation spectra are composed by a monomeric and dimeric species that are shown in blue and green, respectively.

The difference in the gz observed for the different complexes may be a result of distinct aggregation structures. For instance, CuIIL1, CuIIL2, and CuIIL5 present values of gz0 higher than gx0 and gy0, which are similar to monomeric species, possibly due to the maintenance of an axial geometry even after aggregation. An opposite behavior was observed for complexes CuIIL3 and CuIIL4, which hadgz0 values lower than gx0 and gy<sup>0</sup> in the aggregate species. Therefore, these complexes (CuIIL3 and CuIIL4) may have formed aggregates with nonaxial geometry, suggesting that the unpaired CuII electron is not of the dx2–y2 orbital, possibly due to a trigonal bipyramidal geometry, as shown in Figure 6. Structural features behind thermodynamic differences for the interactions between one solvent molecule and the CuIIL1 monomer and dimer were evaluated by quantum chemical calculations. The dimers had very distinctive geometries in each electronic spin state, with the lowest-lying singlet states having a single Cl bridge between the two monomers (Figure 6B–D), which renders each monomer structurally different from the other, while the higher energy triplet state has two Cl bridging the two Cu(II) atoms (Figure S76). It was evidenced by the simulations that for all of the solvents considered as well as for the bare dimer in vacuum, the singlet state was always lower lying than the triplet state, corroborating the proposition of the structures in Figure 6.

**Figure 6.** Proposition of the aggregate structures of complexes CuIIL2 and CuIIL4 in frozen solutions of dichloromethane (**A**). Optimized structures for the CuIIL1dimer in its singlet state interacting with H2O (**B**), MeOH (**C**), and ACN (**D**).

Since all complexes behaved similarly in the EPR measurements in dichloromethane, acetonitrile, and the acetonitrile/water mixture at 298 K, only CuIIL2 was evaluated in methanol and the methanol/water in EPR measurements at 298 K. To provide a comparison between dimeric and monomeric species in solution, an analogous of complex CuIIL2 was synthesized with perchlorate as a counterion ([CuIIL2ClO4] Figure 1). The perchlorate ion is known for its high volume and would be expected to generate only monomeric species in solution. Therefore, this complex was compared with CuIIL2 in EPR measurements performed at 298 K in methanol and the methanol/water (80/20, *v*/*v*) mixture. Interestingly, the EPR spectrum of CuIIL2 in methanol (Figure 7) is visibly a mixture between two species. However, unexpectedly, [CuIIL2ClO4], despite presenting a profile of monomeric species in solution, exhibited a three times higher tcorr than CuIIL2, which could indicate that [CuIIL2ClO4] is in fact, dimeric. Indeed, the HRMS spectra exhibited *m*/*z* peaks corresponding to dimeric structures (1175.2834, Figure S59) and in the FTIR spectra, three bands associated with monodentated ClO4− species were observed at 1121, 1108 and 1027 cm<sup>−</sup>1, corroborating the hypothesis of [CuIIL2ClO4] complex dimerization. The addition of 20% water keeps the equilibrium in solution, as expected, due to the similarity of the conductivities of the complexes in methanol and the methanol/water mixtures. Therefore, it can be assumed that methanol and water do not fully displace the chloride. The unexpected dimerization of [CuIIL2ClO4] might strengthen the proposition of halogen bond formation in the CuIIL2 complex, suggesting a supramolecular structure with the possibility of use in dynamic catalysis [59].

**Figure 7.** EPR spectra of CuIIL2 and [CuIIL2ClO4] in methanol (i and ii) and CuIIL2 and [CuIIL2(ClO4)]in the methanol/water (80:20) mixture (iii and iv). The narrow line in the high magnetic field for CuIIL3 isa standard signal of the CrIII:MgO sample (g = 1.9797).

### *3.3. Urea Hydrolysis as a Model Reaction: Kinetics of NH3 Formation*

The self-organization of these complexes into dimeric or monomeric structures was shown to be dependent on the solvent and ligand exchange reactions, as shown in Figure 8. Due to the distinct behaviors of the complexes in acetonitrile/water and methanol/water, we suspected that hydrolytic catalysis could be tuned by influencing the equilibrium between monomeric and dimeric species. Hence, we decided to evaluate their potential as catalysts to hydrolyze urea, as a model reaction. Catalysis was performed primarily in the acetonitrile/water and methanol/water mixtures. In this reaction, ammonia is expected to be formed and aliquots of the reaction were analyzed by the Berthelot method [60] over the reaction times (5, 10, 20, 30, 40, 50, 60, 120, 240, and 480 s).

**Figure 8.** Equilibrium of dimeric and monomeric species dependent on water-association solvent.

Complexes CuIIL2, CuIIL3, and CuIIL4 were chosen to evaluate the influence of the complexes' aggregation in the reaction. In the acetonitrile/water mixture, the reaction was faster than the employed

method to detect ammonia formation (minimum reaction time: 5 s) and we could only observe the decrease in ammonia concentration over the reaction time (Figure 9A and Figures S63–S66). In contrast, in the methanol/water mixtures, the reaction profile changed, in which the increase in ammonia concentration was observed until a saturation level was reached (Figure 9B). The lability of the chlorido was smaller in methanol and the methanol/water (80/20) mixture when compared to the acetonitrile system, and indicates that the labilization of chlorido affects the path of the reaction. For instance, in acetonitrile/water, all complexes exhibited a decrease of the volume, possibly due to the formation of monomeric species in solution. Therefore, the equilibrium dimer/monomer is still present in methanol/water and it can be inferred that the presence of dimeric structures in solution is possibly slowing the reaction.

**Figure 9.** Ammonia quantification produced by CuIIL2 complex in (**A**) acetonitrile/water and (**B**) methanol/water mixtures up to 480 s at 308 K and at different urea concentrations: 5.2 mmol L−<sup>1</sup> (brown line, squares), 10.4 mmol L−<sup>1</sup> (green line, circles), 15.6 mmol L−<sup>1</sup> (blue line, triangles), and 20.8 mmol L−<sup>1</sup> (black line, inverted triangles).

Considering that complex [CuIIL2ClO4] was mostly dimeric in solution, it would be expected to observe a slower reaction rate of urea hydrolysis by this complex. In general, the behavior of [CuIIL2ClO4] was similar to the chloride complex (faster in acetonitrile/water and slower in methanol/water mixtures), but indeed, a four times lower conversion was observed (Figure 10). Moreover, the reaction in methanol/water only started to produce ammonia after 5 min of reaction, whereas the CuIIL2 complex was able to produce it after 1 min of reaction. Hence, it may indicate that the dimeric structure is not active toward urea hydrolysis.

**Figure 10.** Ammonia quantification produced by the [CuIIL2ClO4] complex in the acetonitrile/water (red line) and methanol/water mixture (black line) at 308 K using a 10.4 mmol L−<sup>1</sup> urea concentration. Only the positive portion of the error bars is shown in the graphic.

In order to verify how the solvent was affecting the equilibrium monomer/dimer, we performed the reaction of urea hydrolysis in other solvents (DMSO, THF, and ethanol), Figures S70–S72. We observed that a high reaction rate of urea hydrolysis was achieved in solvents that exhibited less pronounced hydrogen bonds with water (ACN, DMSO, and THF) than the organic solvents methanol and ethanol [61]. Therefore, we reasoned that the e ffect of preferential solvation [62] by the organic solvents in the aquation reaction of our complexes could be tuning the equilibrium dimer/monomer. The interaction between solvent molecules and the complex is probably occurring via the apolar sites of the solvent, since the complex has a neutral charge. This mode of solvation results in a secondary sphere organized in a way that the dipoles of the solvents are oriented to the bulk solution, and water can interact with these sites via hydrogen bonds (Figure 11A). The stronger hydrogen bond between methanol and water stabilizes the initial state of reaction (dimer–solvent), leading to a slower rate of ligand substitution due to the higher activation energy of the reaction in methanol/water (Figure 11B), thus forming less monomers than in ACN (or the DMSO, THF/H2O mixtures). These results are in contrast to the increased reaction rate observed in methanol and ethanol by the other groups [63] due to the stabilization of the transition state, which strengthens our supposition. Moreover, the reaction rate in the ethanol/water mixtures was even slower than in the methanol/water mixtures due to the longer chain of ethanol, resulting in a higher stabilization of the hydrogen bonds with water in the tertiary coordination sphere. Therefore, our analysis of the equilibrium monomer/dimer in solution verified the occurrence of only the monomer species in ACN/H2O by EPR assays, whereas in the MeOH/H2O mixtures, we detected the presence of dimeric structures by HRMS and EPR experiments, corroborating this hypothesis. The thermochemical data obtained from the DFT calculations of the CuIIL1 monomer or dimer interacting with a single solvent molecule support that the interaction between the dimeric structure is more stabilized in methanol and water than in acetonitrile. The enthalpic di fference between methanol and water was less than 1 kJ/mol, but an eight times higher enthalpy di fference was observed between acetonitrile and water (8.5 kJ/mol). Thus, the stabilization of the ground state in strong hydrogen-bond solvents and the competition between solvents is more pronounced in methanol/water systems, which could result in a lower substitution rate of the chloride in methanol/water mixtures in comparison to the acetonitrile/water mixtures, reinforcing our experimental data.

**Figure 11.** Preferential solvation shell of the complexes in methanol/water and acetonitrile/water mixtures ( **A**) and its e ffect in the stabilization of the ground state (dimeric species) (**B**).

Noticing the strong e ffect of the solvent in the equilibrium monomer/dimer, we suspected that this could enable the tuning of the catalytic behavior by an allosterism (or upregulation).In order to check this possibility, we evaluated complexes CuIIL2, CuIIL3, and CuIIL4 in di fferent urea concentrations (Figure 9 and Figure S68). It should be noted, however, that the increase in urea concentration also increased the water content of the mixture, which could influence the monomer/dimer equilibrium. Noticeably, the complexes presented di fferent behaviors upon an increase in urea (and water) concentrations. For instance: CuIIL3 had a sigmoidal behavior, whereas the behavior for CuIIL2 was linear (Figure 12). By keeping the water constant at 20%, the water e ffect in the monomer/dimer equilibrium decreased, and essentially, this e ffect was more pronounced for CuIIL2, observing a 2-fold decrease of the reaction rate when the water concentration was constant. This result indicates that water has a positive

effect in catalysis for the CuIIL2 complex due to the shift of inactive dimeric species into active monomeric species. In contrast, the CuIIL3 complex and CuIIL4, already have their equilibrium shifted to monomeric species and therefore, do not present a strong influence of water in catalysis, even though a slight positive allosteric effect of water is also observed.

**Figure 12.** Initial rate of urea hydrolysis reaction versus urea concentration performed by CuIIL3 (**A**) and CuIIL2 (**B**). The black line and squares are relative to the increase in water content of the reaction from 0 to 40%, whereas the red line and circles are relative to the water content kept constant at 20%.

Since other aspects of the reaction could be affecting the reaction kinetics such as the activity of water and reaction pH, we decided to verify their significance in our systems. For instance, the composition of a solvent mixture containing water will change the activity of water (aw) [64] and the increase in aw could impact the reaction mechanism and kinetics of a hydrolytic reaction. For instance, the water activity of the methanol/water mixtures used in our systems ranged from 0.26 to 0.65 (Figure S76) (determined by the equation from Zhu et al. [65]), whereas the acetonitrile/water (10% volume) mixture had a water activity of 0.9. Therefore, it would be reasonable to ascribe the observed differences of reaction rates to the different water activity of these systems. However, the catalytic reaction performed in ACN containing 2% (aw~0.4) (Figure S74) still showed a high reaction rate of urea hydrolysis in this solvent mixture. This finding reinforces the proposed idea of solvent effect in the reaction kinetics, since acetonitrile aids in the equilibrium shift to monomeric species.

In addition to the water activity, the pH in our reactions was not controlled and the increase in pH during the reaction could affect the coordination sphere and the reaction rate. Taking into consideration this possibility, we performed a reaction including a pH indicator (phenol red) and we did not observe a significant effect in the pH due to the produced ammonia over the reaction time in the UV–Vis spectrum (Figure S73A). To remove any possible interference produced by the indicator, we performed a reaction control without the pH indicator and added it after the reaction reached a saturation level of ammonia. Again, no color change was observed (Figure S73B), indicating that in the methanol/water and acetonitrile/water mixtures (Figure S73C), the produced ammonia did not seem to severely affect the reaction pH. In fact, the pH of the buffered solutions in solvent mixtures has been previously analyzed by other groups, and they noticed that the NH3/NH4+ buffer presented a lower pH in the solvent mixture than the observed one in pure water [66]. An in situ FTIR experiment (Figure S75) showed an increase of a band in the 3000 cm<sup>−</sup><sup>1</sup> region, indicative of the formation of ammonium ions. After 90 s, we observed that this band oscillated between a minimum and a maximum, suggestive of the equilibrium between ammonia and ammonium. Therefore, due to the fact that the pH of the reaction is not expressively altered during the reaction, we suspect that the main interaction of ammonium is with the hydroxide formed in the hydrolysis reaction of ammonia.

In order to evaluate the impact of the pH in our catalytic reactions, we decided to use buffered aqueous solutions of urea at three different pHs: 3, 6, and 8, and in this case, we observed a profile indicative of a catalyzed reaction at both basic and acidic pHs (Figure S73D). Interestingly, the reaction rates using buffered solutions were lower than the non-buffered one, which might indicate an inhibition of the hydrolysis by the buffers.

Since the in situ reversibility of the system was not observed, the best classification of the effect of water in the urea hydrolysis reaction is "upregulation" or "regulation" [12]. We believe that the strong water dependence of this system will be able to be explored in the future as water sensors.
