Reactivity of [3+1+1] Uranyl-DGA Complex as Lewis-Acid Catalyst in Nucleophilic Acyl Substitution of Acid Anhydrides

Abstract

A UO₂²⁺ complex bearing N, N, N′, N′-tetraisopropyldiglycolamide (TiPDGA) and two DMF molecules was prepared to explore the catalytic activities of the Lewis-acidic U centre. The cationic complex, [UO₂(TiPDGA)(DMF)₂]²⁺, was obtained as a ClO₄⁻ salt under optimised reaction conditions with an appropriate mixing ratio between UO₂²⁺ and TiPDGA to maintain 1:1 stoichiometry, a non-coordinating ClO₄⁻ counteranion to reserve the coordination sites for substrate activation, and the presence of extra HClO₄ to suppress undesired hydrolysis of UO₂²⁺ competing with the expected complex formation. This UO₂²⁺ complex was characterised by IR, elemental analysis, X-ray crystallography, and ¹H NMR to confirm that the desired [3+1+1] equatorial coordination is actually formed in the solid state and is still maintained even after dissolution in CD₂Cl₂. [UO₂(TiPDGA)(DMF)₂]²⁺ was further subjected to nucleophilic acyl substitution reactions of acid anhydrides to assess its activity and capability as a Lewis-acid catalyst there. Although the observed reaction rates were not very rapid, some characteristic aspects to gain reaction- and substrate-selectivity appeared thanks to the equatorial coordination sphere sterically regulated by the tridentate auxiliary TiPDGA ligand and labile monodentate DMF molecules to activate an acid anhydride after ligand substitution.

1. Introduction

Although nuclear power is a promising energy source with an overwhelmingly large energy density without CO₂ emissions, at least during power generation, nuclear waste issues must be resolved appropriately and urgently for the sustainability of modern and future human societies and natural environments [1]. While high-level radioactive wastes (HLWs) at the backend stream of the nuclear fuel cycle are most frequently focused on in this connection [2,3,4,5,6], there is another issue of depleted uranium (DU) unavoidably afforded after the enrichment of ²³⁵U, the fissile, but less-abundant, isotope of U (0.72%; the rest is ²³⁸U), to fabricate nuclear fuels containing 3–5% ²³⁵U. As a result, a large amount of DU is left to be stored without any practical applications, except for a few niche uses (e.g., military purposes), despite the cost and labour spent in preceding processes such as mining, refining, and conversion.

Because isotope effects in chemistry are negligibly small, especially for U, the naturally occurring heaviest element, chemical applications should be directed to resolve the issues of DU. Uranium has a rich redox chemistry from +III to +VI [7], although lower valences such as +I and +II have also been found more recently [8,9,10,11]. While organometallic activities are usually available only in the low-valent U complexes up to +IV [12,13], the most accessible valency of U is +VI to form, in most cases, UO₂²⁺ (uranyl) under the ambient condition. Indeed, the functionalisation of UO₂²⁺ complexes towards molecular catalytic systems has been explored extensively as reviewed previously [14,15], where either the thermal or photochemical reactivity of UO₂²⁺ is employed.

UO₂²⁺ exhibits typical coordination chemistry as a hard Lewis acid [16], so that it exclusively prefers hard bases as ligands [17]. This is the reason why it is hard to expect organometallic behaviour from UO₂²⁺ for the direct activation and formation of organic bonds through oxidative additions and reductive eliminations usually observed in other complex catalysts including low-valent U [12,13]. In accordance with this, we believe that one of the feasible applications of UO₂²⁺ in a catalytic utility is a Lewis-acid catalyst. Several successful cases such as nucleophilic addition [18,19,20], the Diels–Alder reaction [21], acyl substitution [22], and hydrosilylation [23,24] catalysed by Lewis-acidic UO₂²⁺ species have already been reported. Figure 1a summarises the UO₂²⁺ catalysts employed in these former works.

Although these UO₂²⁺-based Lewis-acid catalysts actually promote aimed reactions, there is still some space to design molecular structures of them appropriately. To the best of our understanding, some vacancy in the equatorial plane of UO₂²⁺ must be reserved for the activation of a substrate of interest. Most ordinarily, the coordination site occupied by monodentate ligand(s), such as solvent molecules (L) or OPPh₃ as in Figure 1a, is offered for this purpose to avoid difficulty in preparing a coordinatively unsaturated UO₂²⁺ complex.

Considering the characteristic planar penta-coordination in the equatorial plane of UO₂²⁺, a tridentate planar ligand would be appropriate as an auxiliary scaffold to provide a regulated reaction space for the substrate activation. Due to the hardness of UO₂²⁺ described above, O-donating tridentate ligands should be preferred. For instance, N, N, N′, N′-tetraalkyldiglycolamide (TRDGA, Figure 1b), frequently employed in separation chemistry for trivalent lanthanides/actinides [25,26,27,28], meets the above requirement well. Furthermore, its non-charged form is expedient, especially in a reaction system such as nucleophilic acyl substitution, in which an acidic by-product is afforded [22].

In this study, we aimed to prepare a [3+1+1] UO₂²⁺ complex, [UO₂(TRDGA)(L)₂]²⁺ (Figure 1b), and to assess its catalytic activity in the nucleophilic acyl substitution of acid anhydrides as a benchmark reaction. While several UO₂²⁺-TRDGA complexes have been reported previously [26,29,30], such a [3+1+1] equatorial coordination has not been attained yet. Instead, a non-planar hexa-coordination with TRDGA (R = iPr, TiPDGA) and two NO₃⁻ or a planar, but fully chelated, hexa-coordination by two TRDGA molecules (R = Me) were observed. To achieve the expected [UO₂(TRDGA)(L)₂]²⁺ of Figure 1, a reaction condition including the selection of a counteranion must be optimised.

2. Results and Discussion

2.1. Synthesis and Characterisation of UO₂²⁺-Diglycolamide Complex

Yellow prismatic crystals of the desired UO₂²⁺-TiPDGA complex, [UO₂(TiPDGA)(DMF)₂](ClO₄)₂⋅CH₂Cl₂, were obtained in 80% yield after the reaction between [UO₂(DMF)₅](ClO₄)₂ and TiPDGA from a 1:1 stoichiometric mixture of CH₂Cl₂ solution and recrystallisation by slow diffusion of Et₂O vapor to the mother liquor. This compound was characterised by ¹H NMR, IR, elemental analysis, and X-ray crystallography as summarised in the Materials and Methods Section.

The molecular structure of [UO₂(TiPDGA)(DMF)₂]²⁺ is illustrated in Figure 2. A typical trans-dioxo structure of UO₂²⁺ was maintained in this complex, where the mean U≡O bond distance and the O≡U≡O angle are 1.75 Å and 179.5(2)°, respectively. This axial UO₂²⁺ is further surrounded by additional ligands such as TiPDGA and DMF molecules in its equatorial plane to form pentagonal bipyramidal coordination geometry around the U centre. All O atoms of TiPDGA are bound to U in a tridentate manner to form two 5-membered chelate rings, which may play critical roles to stabilise this complex even under presence of excess DMF molecules arising from [UO₂(DMF)₅]²⁺ [17], a source of UO₂²⁺ used in this work.

The interatomic distances from the U centre to O atoms of the amide (O_amide) and ether moieties (O_ether) of TiPDGA are 2.35 Å (mean) and 2.520(3) Å, respectively, implying a stronger coordination of O_amide to U compared with that of O_ether in accordance with the stronger Lewis basicity of the former compared with that of the latter [31]. Both the U−O distances between UO₂²⁺ and TiPDGA displayed in Figure 2 are significantly shorter than those found in UO₂(TiPDGA)(NO₃)₂ reported by Kannan et al. [26] (U−O_amide = 2.40 Å (mean) and U−O_ether = 2.592(3) Å). In this another UO₂²⁺-TiPDGA complex, two NO₃⁻ ions are involved in the equatorial coordination around UO₂²⁺. Due to the steric demands occurring in the equatorial plane of UO₂(TiPDGA)(NO₃)₂, these NO₃⁻ ions exhibit different coordination manners; one is bidentate, while the other is monodentate, implying that there is no space wide enough to allow the bidentate coordination of both NO₃⁻ ions. The same trend in difference between the bond distances of O_amide and O_ether from U was also found between our [UO₂(TiPDGA)(DMF)₂]²⁺ of Figure 2 and [UO₂(TRDGA)₂]²⁺ (R = Me) [29,30]. As evidenced by U−O_amide = 2.42 Å (mean) and U−O_ether = 2.63 Å (mean), the equatorial plane of the latter 1:2 complex was somewhat expanded despite the less bulky methyl substituents on the amide N atoms to maintain the planar hexa-coordination around UO₂²⁺. In contrast, any steric constraints observed in the former related systems do not occur in the current [UO₂(TiPDGA)(DMF)₂]²⁺. This is because only one TiPDGA ligand is included in the coordination sphere, and also because both DMF molecules are bound to U in a monodentate manner (see below). Indeed, the shorter bond distances of U−O_amide and U−O_ether resulted, which would be associated with the stronger coordination of TiPDGA. This structure was attained by employing the non-coordinating ClO₄⁻ as a counteranion and the stoichiometric loading of TiPDGA to UO₂²⁺ in the reaction mixture.

As displayed in Figure 2, the remaining coordination sites were occupied by two DMF molecules to saturate the equatorial plane of UO₂²⁺, where the coordination of DMF molecules to U are as strong as those of O_amide of TiPDGA as evidenced by the similar bond distances (U(1)−O(6): 2.325(4) Å, U(1)−O(7): 2.343(4) Å). These U−O bonds are even shorter than those of [UO₂(DMF)₅]²⁺ (2.39 Å (mean) in ClO₄⁻ salt, 2.37 Å (mean) in BF₄⁻ salt) [17,32] and UO₂(salophen)DMF (2.410(3) Å) [33]. This implies that the equatorial coordination sphere of [UO₂(TiPDGA)(DMF)₂]²⁺ is sterically less constrained. Indeed, the sum of the equatorial bond angles around the U centre (Σ_eq) is 360.3(1)° to maintain a nearly ideal planarity of the typical pentagonal equatorial coordination around UO₂²⁺. Additionally, one CH₂Cl₂ molecule and two ClO₄⁻ ions are included in this crystal structure (Figure S1) as a crystalline solvent and non-coordinating counteranions of the positively charged [UO₂(TiPDGA)(DMF)₂]²⁺, respectively. Some of these isolated components were disordered even at 183 K.

In the above synthetic procedure, we added 0.1 vol% of conc. HClO₄(aq) (10 μL) to the CH₂Cl₂-based mother liquor (10 mL) to obtain [UO₂(TiPDGA)(DMF)₂](ClO₄)₂⋅CH₂Cl₂ of Figure 2. When this operation was skipped, another crystalline phase, [UO₂(TiPDGA)(DMF)(H₂O)](ClO₄)₂⋅H₂O (Figure S2), was obtained. Although the hydrolysis of UO₂²⁺ to give OH⁻ was once considered, all isolated O atoms (O(7), O(16)) were assigned to H₂O molecules because of electro-neutrality in its crystal lattice. The bond distances between U and O atoms at the apical positions were 1.76 Å (mean). The tridentate coordination of TiPDGA through amide and ether O atoms and their bond lengths (U−O_amide: 2.33 Å (mean), U−O_ether: 2.526(2) Å) almost remain unchanged from those of [UO₂(TiPDGA)(DMF)₂]²⁺ shown in Figure 2. The coordination of DMF seems to be stronger than that of H₂O in terms of bond lengths (U−O_DMF: 2.333(3) Å vs. U−O_H2O: 2.379(3) Å). Additionally, a yellow chunk material also deposited together with the crystalline [UO₂(TiPDGA)(DMF)(H₂O)](ClO₄)₂⋅H₂O. Although the identity of this compound has not been determined yet due to its insolubility to any organic solvents, it would be of a hydrolytic product of UO₂²⁺. This undesired reaction can be suppressed under an acidic condition, so that presence of the extra HClO₄ plays a crucial role to provide [UO₂(TiPDGA)(DMF)₂]²⁺ as a predominant product. For the simplicity of the reaction system, we thereafter employed the DMF solvate, [UO₂(TiPDGA)(DMF)₂](ClO₄)₂⋅CH₂Cl₂ of Figure 2, to explore its catalytic activity in the nucleophilic acyl substitution of acid anhydrides, the main topic of this work.

Using ¹H NMR spectroscopy, we examined whether or not the molecular structure of [UO₂(TiPDGA)(DMF)₂]²⁺ shown in Figure 2 would still be maintained in a CD₂Cl₂ solution. As a result, the NMR signals arising from TiPDGA and DMF of this sample solution (see the Materials and Methods Section, Figure S3) exhibited remarkable downfield shifts compared with those of their free forms (free DMF: 7.96 (s, 2H, CHO), 2.91 (s, 3H, N(CH₃)₂), 2.82 (s, 3H, N(CH₃)₂); free TiPDGA: 4.22 (s, 4H OCH₂), 3.92 (septet, 2H, CH(CH₃)₂), 3.45 (septet, 2H, CH(CH₃)₂), 1.41 (d, 12H, CH(CH₃)₂), 1.18 (d, 12H, CH(CH₃)₂)). Therefore, both TiPDGA and DMF remain coordinated to UO₂²⁺ even in the CD₂Cl₂ solution to maintain the original structure of [UO₂(TiPDGA)(DMF)₂]²⁺ displayed in Figure 2.

2.2. Nucleophilic Acyl Substitution Reactions Catalysed by [UO₂(TiPDGA)(DMF)₂]²⁺

In accordance with our former work [22], we first examined the catalytic activity of [UO₂(TiPDGA)(DMF)₂]²⁺ in the reaction between acetic anhydride (Ac₂O) and 2-phenylethanol (PhetOH) to afford 2-phenylethyl acetate (PhetOAc) (Scheme 1). Concentrations of the reactants and product at different time were determined by ¹H NMR spectroscopy. Figure 3a illustrates the typical progress of the reaction. Both reactants were congruently consumed. Along with this, the production of PhetOAc was observed and reached 90% yield at 10 h (entry 1,

Table 1. Summary of nucleophilic acyl substitution reactions of acid anhydrides ((RCO)₂O + R′OH → RCOOR′ + RCOOH) catalysed by [UO₂(TiPDGA)(DMF)₂]²⁺ in CD₂Cl₂ ^a.

Entry	R b	R′ c	Yield (%) d
1	Me	2-phenylethyl	90
2	Me	2-phenylethyl	9 e
3	Me	Me	100
4	Me	Et	88
5	Me	nBu	88
6	Me	Ph	95
7	Me	tBu	78
8	Et	2-phenylethyl	85
9	iPr	2-phenylethyl	48
10	tBu	2-phenylethyl	10
11	Ph	2-phenylethyl	6
12	tBu	Me	14

^a Typical condition: 0.5 M acid anhydride + 0.5 M nucleophile + 10 mM catalyst (2.0 mol%), 10 h, if not specified. ^b R of acid anhydrides. ^c R′ of nucleophiles. ^d Determined by ¹H NMR peak integrals. ^e No catalyst loaded.

). Although even under the absence of [UO₂(TiPDGA)(DMF)₂]²⁺, the same product was generated, and its yield was only 9% at 10 h (entry 2, Table 1). Consequently, we confirmed that this nucleophilic acyl substitution reaction was catalysed by [UO₂(TiPDGA)(DMF)₂]²⁺. The reaction rate of the current system was slower than those reported in the former time, where the same reaction was completed within 1 h to reach >95% yield [22]. Such a difference in the kinetic aspect would be ascribed to the equatorial plane sterically regulated by TiPDGA, the auxiliary tridentate ligand employed in this work. Based on the loading ratio of [UO₂(TiPDGA)(DMF)₂]²⁺ (10 mM) with the substrates (0.5 M each) and the product yield (90% at 10 h) in entry 1 of Table 1, the turnover number (TON) of the current catalytic system was 45.

We have further surveyed the capabilities of [UO₂(TiPDGA)(DMF)₂]²⁺ as a Lewis-acid catalyst in the current reaction system in the use of different nucleophiles and different acid anhydrides. First, several aliphatic alcohols and phenol were tested as nucleophiles to be reacted with Ac₂O (entry 3–6, Table 1). As a result, the corresponding acetate esters were generally afforded in high yields (88−100%) comparable with that of PhetOAc of entry 1 (90%), implying that a nucleophile would not largely affect the reaction mechanism.

An exception was found in the use of tBuOH (entry 7, Table 1). While tBuOAc was obtained in 78% yield at 10 h as expected, isobutylene ((CH₃)₂C = CH₂) was also generated as another product after the E1 elimination of tBuOH. Note that both the tBuOAc yield and its production rate in the current system were higher and faster than those observed in our former system catalysed by [UO₂(OPPh₃)₄]²⁺ (Figure 1, 63% yield at 72 h) [22]. Furthermore, the yield of isobutylene in the current system was less than 3%, which was much different from 32% yield in the use of the [UO₂(OPPh₃)₄]²⁺ catalyst [22]. Because H₂O generated as a by-product of the E1 elimination of tBuOH readily consumes Ac₂O, the suppression of this side reaction is highly preferred. The steric control of the UO₂²⁺ equatorial plane with TiPDGA illustrated in Figure 1 would also be favourable to prevent the extra activation of bulky tBuOH for the E1 path, probably through its O-coordination to the U centre of the catalyst.

In contrast to the nucleophiles, the selection of acid anhydrides ((RCO)₂O) has significant impact on the efficiency of the current catalytic acyl substitution. As shown in entry 1 and 8–11 of Table 1, yields of the 2-phenylethyl esters of different acids remarkably decreased with an increase in length or bulkiness of R in the parent acid anhydrides. Most typically, pivalic (R = tBu) and benzoic (R = Ph) anhydrides resulted in their corresponding esters only in 10% and 6% yields, respectively. Especially for the latter case, it is hard to know whether the reaction was catalysed by the UO₂²⁺ complex or not. Even when MeOH was used as a sterically least-hindered nucleophile (entry 12, Table 1), the yield of its pivaloyl ester (14%) was only slightly higher than that of the 2-phenylethyl analogue (10%, entry 10). A series of these trends implies that the acid anhydride must be activated by the [UO₂(TiPDGA)(DMF)₂]²⁺ catalyst to initiate the reaction, and that such an initial process is significantly controlled by steric effects. Details of the mechanistic aspects will be explored in the next section.

2.3. Reaction Kinetics and Catalysis Mechanism

From the data in Table 1, we had already gained some qualitative insights about the mechanistic aspects of the current reaction system. As mentioned above, an acid anhydride decides the yield of its product, while a nucleophile has only a small impact on the reaction efficiency. Herein, a more detailed and quantitative assessment of its reaction kinetics and catalysis mechanism was performed by taking the acetylation of PhetOH with Ac₂O as a model reaction. The results are displayed in Figure 3b–d.

When the nucleophile concentration, [PhetOH], was varied from 100 mM to 480 mM (Figure 3b), little differences in the reaction progress were observed unless PhetOH was consumed significantly. Indeed, the initial rate (d[PhetOAc]/dt = k_ini) of this series ranged from 22.7 μM⋅s⁻¹ to 27.5 μM⋅s⁻¹ despite 4.8-fold variation of [PhetOH]. This indicates that the nucleophile is not involved in the rate-determining step of this catalytic reaction.

We also examined the dependency of the reaction kinetics on the concentration of the other reactant, [Ac₂O]. As a result, the reaction kinetics became faster and faster with an increase in [Ac₂O] (Figure 3c). Indeed, k_ini in this series linearly increased with an increase in [Ac₂O]. Consequently, the rate of this catalytic reaction is the first-order of [Ac₂O] with a slope equal to 4.0 × 10⁻⁵ s⁻¹ (right panel of Figure 3c), implying that one Ac₂O molecule is involved in the rate-determining step of this reaction, most probably the activation of Ac₂O by the UO₂²⁺ catalyst.

The activation of Ac₂O in this system would be attained through its coordination to the U centre of [UO₂(TiPDGA)(DMF)₂]²⁺. However, there would be no vacancies large enough to accept an additional coordination in the equatorial plane of this UO₂²⁺ complex as displayed in Figure 2 and Figure S4. Therefore, some of the coordinating DMF molecules should be replaced with Ac₂O. If this assumption is correct, the reaction kinetics of the current catalytic system should be controlled by the concentration of DMF additionally loaded to the reaction system ([DMF]_add). As a matter of fact, the reaction rate significantly decreased with an increase in [DMF]_add (Figure 3d). Therefore, the above hypothesis was found to be correct.

To precisely assess the dependency of k_ini on [DMF] in this series, it is further necessary to know its net concentration in the sample solution, because the DMF molecules may be released from [UO₂(TiPDGA)(DMF)₂]²⁺ to provide extra free DMF in the reaction system. To resolve this problem, the UV-vis absorption spectra of a CH₂Cl₂ solution dissolving [UO₂(TiPDGA)(DMF)₂]²⁺ were first recorded at different [Ac₂O], where PhetOH was absent. As a result, little variation was observed as shown in Figure 4a. It was confirmed that the substitution of DMF with Ac₂O does not proceed predominantly. Note that the result observed in this UV-vis titration experiment only describe the thermodynamic stability of the preferred coordination of DMF to the U centre superior to that of Ac₂O, and that it never rules out any possibility of the ligand substitution from DMF to Ac₂O to form a reaction intermediate in the catalytic process of Figure 3.

We also collected the UV-vis absorption spectra of the CH₂Cl₂ solution dissolving [UO₂(TiPDGA)(DMF)₂](ClO₄)₂⋅CH₂Cl₂ (10 mM) at different [PhetOH], where Ac₂O was absent. In contrast to the little dependency on [Ac₂O] in Figure 4a, significant spectral variation was observed with an increase in [PhetOH] as shown in Figure 4b. Due to the absence of any remarkable absorption of PhetOH at λ > 340 nm (Figure S5), the substitution of the coordinated DMF by PhetOH was confirmed. We analysed this spectral series with HypSpec software (ver. 1.1.33) [34] to assess the thermodynamic stabilities of the occurring species. Prior to this, the factor analysis [35] for the data set in Figure 4b revealed that this spectral variation consisted of two principal components having meaningful eigenvalues (Figure S6), suggesting that two UO₂²⁺ complexes including the parent [UO₂(TiPDGA)(DMF)₂]²⁺ were present in the titration series observed in Figure 4b. Consequently, the following ligand substitution equilibrium (Equation (1)) with a logarithmic stability constant (log β_sub) should be considered.

[UO₂(TiPDGA)(DMF)₂]²⁺ + PhetOH ⇄ [UO₂(TiPDGA)(DMF)(PhetOH)]²⁺ + DMF
log β_sub

(1)

Although the second substitution of the remaining DMF molecule would also be possible, its contribution in Figure 4b should be considered negligibly small based on the above factor analysis of Figure S6. Equation (1) is further divided into the following equilibria, Equations (2) and (3), to describe formation of [UO₂(TiPDGA)(DMF)₂]²⁺ and [UO₂(TiPDGA)(DMF)(PhetOH)]²⁺, respectively.

[UO₂(TiPDGA)]²⁺ + 2 DMF ⇄ [UO₂(TiPDGA)(DMF)₂]²⁺
log β₂₀

(2)

[UO₂(TiPDGA)]²⁺ + DMF + PhetOH ⇄ [UO₂(TiPDGA)(DMF)(PhetOH)]²⁺
log β₁₁

(3)

log β_sub = log β₁₁ – log β₂₀

(4)

Herein, the desolvated UO₂²⁺ complex, [UO₂(TiPDGA)]²⁺, is not an actual species occurring in the reaction system but needs to be taken into account for the expediency of the analysis.

As demonstrated by the ¹H NMR spectroscopy (Figure S3), we already confirmed that [UO₂(TiPDGA)(DMF)₂]²⁺ is present as a predominant species in the CD₂Cl₂ solution, indicating that the stability of this complex is high enough. Based on this fact, we derived log β₁₁ of Equation (3) from the spectral variation in Figure 4b, where log β₂₀ of Equation (2) was fixed to an arbitrary value (e.g., log β₂₀ = 10, 15, 20) large enough to guarantee the predominance of [UO₂(TiPDGA)(DMF)₂]²⁺ in the sample solution under the absence of PhetOH. Consequently, the estimated log β₁₁ was always smaller by 0.28 than log β₂₀ regardless of any above assumptions of log β₂₀, i.e., log β_sub of Equation (1) is equal to −0.28. Such a preferential solvation of DMF compared with that of PhetOH should be ascribed to difference in the Lewis basicity between them, evidenced by the donor numbers (DN) of DMF (26.6) and alcohols (~20) [31].

The calculated speciation diagram of [UO₂(TiPDGA)(DMF)₂]²⁺ and [UO₂(TiPDGA)(DMF)(PhetOH)]²⁺ at different [PhetOH] (Figure S7) suggests that a mole fraction of [UO₂(TiPDGA)(DMF)(PhetOH)]²⁺ is 0.85 at [PhetOH] = 0.1 M, for instance. At the same time, one of DMF molecules initially present in [UO₂(TiPDGA)(DMF)₂]²⁺ is released in the same extent to provide an extra concentration of DMF ([DMF]_sub). Accordingly, the total concentration of free DMF ([DMF]_total) in each sample solution of Figure 3d) was the sum of [DMF]_add and [DMF]_sub. When k_ini at different [DMF]_add was plotted against [DMF]_total⁻¹, the linear relationship between these parameters was yielded (right panel of Figure 3d). Finally, k_ini was found to be the first-order of [DMF]_total⁻¹ with a slope equal to 7.8 × 10⁻¹ μM²⋅s⁻¹.

In summary, the overall rate equation of the current catalytic acyl substitution studied in Figure 3 can be written as follows:

d[PhetOAc]/dt = k_ini = k₅[Ac₂O][DMF]⁻¹[U]

(5)

where [U] denotes the concentration of the UO₂²⁺ catalyst, [UO₂(TiPDGA)(DMF)₂]²⁺. From the dependencies of k_ini with [Ac₂O] and [DMF]_total in Figure 3, the rate constant (k₅) of Equation (5) was estimated to be 1.6 × 10⁻⁴ s⁻¹ at 27.5 °C. Based on this rate equation and the additional mechanistic insight obtained above, the catalytic cycle of the acyl substitution reaction of Ac₂O with PhetOH mediated by [UO₂(TiPDGA)(DMF)₂]²⁺ was proposed as shown in Figure 5.

As a preliminary step of this catalytic system, the ligand substitution reaction between [UO₂(TiPDGA)(DMF)₂]²⁺ and [UO₂(TiPDGA)(DMF)(PhetOH)]²⁺ (Equation (1)) is rapidly equilibrated in a reaction mixture, where the second ligand substitution to form [UO₂(TiPDGA)(PhetOH)₂]²⁺ is unlikely to occur, probably due to steric hindrance between the neighbouring PhetOH molecules. Ac₂O enters the equatorial plane of UO₂²⁺ to replace one of the monodentate ligands. This is the rate-determining step of this catalytic reaction as defined by Equation (5). It is hard to know which [UO₂(TiPDGA)(DMF)₂]²⁺ or [UO₂(TiPDGA)(DMF)(PhetOH)]²⁺ is an actual activator of Ac₂O, and which DMF or PhetOH is replaced by Ac₂O for its activation. Because of these circumstances, the monodentate ligands of the intermediate species in Figure 5 are simply denoted by “L”. The bidentate coordination of Ac₂O instead of both Ls, as observed in our former system of [UO₂(OPPh₃)₄]²⁺ [22], is unlikely to occur because k_ini was independent of [PhetOH], as shown in Figure 3b, despite the predominant formation of [UO₂(TiPDGA)(DMF)(PhetOH)]²⁺ at the initial period of the reaction as shown in Figure S7 based on log β_sub of Equation (1). The positively polarised carbonyl C of Ac₂O after its coordination to the U centre readily undergoes a nucleophilic attack from PhetOH to afford PhetOAc and AcOH. These reaction products are immediately removed from UO₂²⁺ due to the weakly donating natures of esters (DN ~17) and carboxylic acids (DN = 10.5) [31]. Finally, the UO₂²⁺-TiPDGA complex returns to the initial state to repeat this catalytic cycle. The Lewis basicity of Ac₂O is actually not very strong as evidenced by DN = 10.5 [31]. This should be considered one of the reasons why the reaction kinetics currently observed are commonly slow. In connection with this, the equatorial plane of UO₂²⁺ sterically constrained by tridentate TiPDGA would be another main cause of the slower reaction kinetics compared with that catalysed by [UO₂(OPPh₃)₄]²⁺ that we reported previously [22].

3. Materials and Methods

3.1. Materials

Caution! All U isotopes are α-emitters. Therefore, standard precautions to handle radioactive materials should be followed. Perchlorate compounds are potentially explosive, so that such chemicals should be handled with great care. All chemicals except for U sources were of reagent grade and used as received, if not specified.

3.2. Synthesis of N, N, N′, N′-Tetraisoprolyldiglycolamide (TiPDGA)

Diglycolic dichloride (0.695 mL, 5.85 mmol, TCI) reacted with diisopropylamine (1.20 mL, 13 mmol, TCI) under the presence of triethylamine (2.00 mL, 14 mmol, Wako) in CH₂Cl₂ (30 mL, Wako) at 0 °C for 2 h, where the round-bottom-flask reaction vessel was equipped with a CaCl₂ drying tube. An amount of 1.2 M HCl(aq) (3 mL) was loaded to the reaction mixture, followed by stirring for 5 min. After drying the separated organic layer over MgSO₄, the filtrate was concentrated by a rotary evaporator. The residue was triturated with n-hexane (8 mL) under gentle heating to provide the colourless solid of TiPDGA (0.345 g, 20% yield). ¹H NMR (CDCl₃, δ/ppm vs. TMS): 4.22 (s, 4H OCH₂), 3.92 (septet, 2H, CH(CH₃)₂), 3.45 (septet, 2H, CH(CH₃)₂), 1.40 (d, 12H, CH(CH₃)₂), 1.18 (d, 12H, CH(CH₃)₂). IR (ATR, diamond prism, cm⁻¹): 1644 (ν_{C = O}).

3.3. Synthesis of [UO₂(TiPDGA)(DMF)₂](ClO₄)₂·CH₂Cl₂

UO₃ (0.268 g, 0.936 mmol) from our in-house stock was dissolved in 70% HClO₄ (180 μL, 2.1 mmol), followed by heating to concentrate it until white fume occurred. The cooled yellow residue of UO₂(ClO₄)₂·xH₂O was dissolved in triethyl orthoformate (10 mL). Under vigorous agitation at RT, DMF (305 μL, 3.96 mmol) was added dropwise to this mixture to afford a bright yellow precipitate. The deposit was filtered out to obtain UO₂(DMF)₅(ClO₄)₂ (0.508 g, 0.609 mmol, 65% yield). This yellow compound further reacted with TiPDGA (0.294 g, 0.953 mmol) in CH₂Cl₂ (10 mL), where 70% HClO₄ (10 μL) was also loaded to avoid the hydrolysis of UO₂²⁺. After stirring for 2 h, this mixture was concentrated by slow evaporation in the dark. The yellow deposit was recrystallised from CH₂Cl₂ under Et₂O vapor to give prismatic yellow crystals of [UO₂(TiPDGA)(DMF)₂](ClO₄)₂·CH₂Cl₂ (0.744 g, 80% yield).

When no 70% HClO₄ was added to the CH₂Cl₂-based reaction mixture, the DMF-H₂O mixed adduct, [UO₂(TiPDGA)(DMF)(H₂O)](ClO₄)₂ H₂O, was crystallised together with a yellow insoluble chunk precipitate instead of [UO₂(TiPDGA)(DMF)₂](ClO₄)₂·CH₂Cl₂.

3.3.1. Characterisation of [UO₂(TiPDGA)(DMF)₂](ClO₄)₂·CH₂Cl₂

¹H NMR (CD₂Cl₂, δ/ppm vs. TMS): 8.96 (s, 2H, CHO of DMF), 5.96 (s, 4H OCH₂), 5.34 (s, 0.5H, CH₂Cl₂), 4.13 (septet, 2H, CH(CH₃)₂), 3.96 (septet, 2H, CH(CH₃)₂), 3.41–3.39 (s × 2, 12H, N(CH₃)₂ of DMF), 1.78 (d, 12H, CH(CH₃)₂), 1.46 (d, 12H, CH(CH₃)₂). IR (ATR, diamond prism, cm⁻¹) 1644 (ν_C=O of DMF), 1607 (ν_C=O of TiPDGA), 934 (ν₃ of UO₂²⁺). Anal. calcd for C₂₃H₄₈Cl₄N₄O₁₅U: C 27.61, H 4.84, N 5.60; found: C 26.86, H 4.53, N 5.45. Crystallographic data for [UO₂(TiPDGA)(DMF)₂](ClO₄)₂·CH₂Cl₂. CCDC 2386158, F_w = 1000.48, 0.15 × 0.21 × 0.24 mm³, monoclinic, P2₁/n (#14), a = 11.1680(6) Å, b = 27.1647(11) Å, c = 13.5279(6) Å, β = 106.148(8)°, V = 3942.1(3) ų, Z = 4, T = 183 K, D_calcd = 1.686 g·cm⁻³, μ = 4.452 cm⁻¹, GOF = 1.061, R (I > 2σ) = 0.0527, wR (all) = 0.0931.

3.3.2. Characterisation of [UO₂(TiPDGA)(DMF)(H₂O)](ClO₄)₂⋅H₂O

Crystallographic data of [UO₂(TiPDGA)(DMF)(H₂O)](ClO₄)₂⋅H₂O. CCDC 2386159, F_w = 878.49, 0.28 × 0.31 × 0.43 mm³, monoclinic, P2₁/n (#14), a = 13.5908(4) Å, b = 12.1091(3) Å, c = 20.5865(5) Å, β = 100.8350(10)°, V = 3327.57(15) ų, Z = 4, T = 296 K, D_calcd = 1.754 g·cm⁻³, μ = 5.106 cm⁻¹, GOF = 1.099, R (I > 2σ) = 0.0289, wR (all) = 0.0703.

3.4. Methods

NMR spectra were recorded by a JNM ECX-400 spectrometer (¹H: 399.78 MHz, JEOL, Japan). IR spectra were measured by a FT/IR-4700 (JASCO, Japan) equipped with a diamond ATR apparatus. CHN elemental analysis was performed using a CHN Corder MT-6 (Yanaco, Japan). UV-vis absorption spectra were acquired by a V-770 spectrophotometer (JASCO, Japan). The structural characterisation of UO₂²⁺ complexes was carried out by single-crystal X-ray diffraction. A single crystal of each sample was mounted on a Kapton capillary, and if necessary, located in the cryogenic N₂ stream at the specified temperature. Intensity data were collected using a R-AXIS RAPID diffractometer (Rigaku, Japan) with a graphite-monochromated Mo-Kα radiation (λ = 0.71075 Å). The obtained diffraction data were analysed by the Olex2 v.1.5 software package [36] suited with SHELX [37]. The structure was solved by SIR92 and expanded using Fourier techniques. All nonhydrogen atoms were anisotropically refined by SHELXL 2017/1 [38]. Hydrogen atoms were refined as riding on their parent atoms with U_iso(H) = 1.2U_eq(C). The final cycle of the full-matrix least-squares refinement of F² was based on the observed reflections and parameters and converged with the unweighted and weighted agreement factors, R and wR, respectively.

3.5. Catalytic Reaction

The nucleophilic acyl substitution of acid anhydrides catalysed by [UO₂(TiPDGA)(DMF)₂]²⁺ was studied as follows. In an NMR tube, [UO₂(TiPDGA)(DMF)₂](ClO₄)₂·CH₂Cl₂ was dissolved in CD₂Cl₂ (700 μL), followed by loading a nucleophile and an acid anhydride appropriately. The NMR spectra of each sample were recorded every 5–30 min to monitor the progress of the reaction. During each experiment, the temperature of the sample was maintained at 27.5 °C with the temperature-controlled air flow in the NMR instrument.

4. Conclusions

In this study, we prepared a new UO₂²⁺ complex, [UO₂(TiPDGA)(DMF)₂]²⁺, bearing N, N, N′, N′-tetraisopropyldiglycolamide (TiPDGA) and two DMF molecules to form a [3+1+1] coordination on the equatorial plane. The former auxiliary ligand provides a sterically regulated reaction field arising from its tridentate coordination manner, while the latter solvent molecules are reserved for ligand substitution with an acid anhydride to be activated after the coordination to the U centre at the initial stage of the reaction. Although several related UO₂²⁺-TRDGA complexes have been reported previously, the molecular structures of them are not very suitable for providing a Lewis-acidic reaction field required in this work. In contrast, the [3+1+1] equatorial coordination in the obtained [UO₂(TiPDGA)(DMF)₂]²⁺ was attained by controlling preparation conditions such as the stoichiometric loading of TiPDGA and employing the non-coordinating counteranion, ClO₄⁻. The addition of extra HClO₄ is also important to suppress the undesired hydrolysis of UO₂²⁺. The catalytic activity of [UO₂(TiPDGA)(DMF)₂]²⁺ in the nucleophilic acyl substitution reactions of acid anhydrides was also examined. While the reaction rates were not very fast, some specific features, such as the suppression of the E1 side reaction of tBuOH and the sterically constrained selectivity in the activation of acid anhydrides, were observed. Based on the kinetic evaluation under different conditions, the reaction mechanism of this catalytic system was proposed.