Introduction

Metal nitride complexes are proposed as key intermediates in N2 activation and functionalization1,2. Uranium nitride complexes have attracted significant attention due to their potential applications in N2 fixation3,4,5, small molecule activation6,7,8,9,10,11,12, and next-generation nuclear fuels13,14,15,16. Since uranium nitride materials have been considered to be effective catalysts for the Haber–Bosch synthesis of NH3 from N217, a series of molecular uranium nitride complexes have been synthesized by N2 cleavage or azide reduction, which generally leads to uranium nitride complexes containing U(IV) centers and a few examples with U(III) and U(V) centers18,19,20,21,22,23,24,25,26,27,28,29,30,31,32. However, the synthesis of U(VI) nitrides lags far behind, largely due to the limited synthetic strategy and high chemical reactivity of U(VI) nitrides33,34,35,36,37,38,39.

Photolysis of metal azides is a typical method in transition metal chemistry to prepare metal nitrides40,41,42,43,44,45,46. However, despite continuous efforts, this method has been less successful in the synthesis of actinide nitride complexes. For example, Liddle and co-workers found that the U(VI) nitride species is unstable and easily decomposes to U(IV) amide through C–H activation under photolytic conditions, which is consistent with the direct photolysis of U(IV) azide species33. Kiplinger, Batista, and co-workers attempted to isolate a U(VI) nitride complex (I in Fig. 1) by the photolysis of (C5Me5)2U[N(SiMe3)2](N3), but could only isolate the intramolecular C–H activated product (C5Me5)(C5Me4CH2NH)U[N(SiMe3)234. Fortier and co-workers proposed that a U(VI) nitride (II in Fig. 1) was formed as a transient intermediate by the photolysis of the corresponding U(IV) azide complex, but the U(VI) nitride also could not be isolated38. Recently, Mazzanti and co-workers reported a photochemically robust U(VI) nitride [NBu4][U(OSi(OtBu)3)4(N)] (III in Fig. 1) by the photolysis of an anionic U(IV) azide analog39. This remains the limited example of the successful isolation of a U(VI) nitride complex via a photolysis route. These results demonstrate that U(VI) nitrides are accessible under photochemical conditions but require an appropriate ligand framework and specific conditions to stabilize this transient U(VI) nitride intermediate.

Fig. 1: U(VI) nitrides generated by the photolysis of uranium azides.
figure 1

Both I and II are proposed U(VI) nitride intermediates. Complex III is an isolated example of U(VI) nitride species. Complex IV is a U(VI) nitride stabilized by transition metals.

Full size image

The introduction of the main group element or another uranium center is an effective method to stabilize the U(VI) nitrides47,48,49. For instance, the Cummins group reported a borane-capped U(VI) nitride complex by chemical oxidation of the U(V) nitridoborate complex47. The groups of Hayton and Mazzanti demonstrated that bridged uranium(VI) nitrides with another uranium center are also feasible48,49. Recently, our group successfully isolated a U(IV) nitride supported by a multidentate N-P ligand (N(CH3)(CH2CH2NHPiPr2)2) via U-Rh synergistic N2 cleavage28. Thus we attempted to access U(VI) nitrides with the aid of transition metals by this platform. Herein, we report the synthesis of transition metal-stabilized U(VI) nitrides (IV in Fig. 1) by photolysis of U(IV)-azide precursors. The U(V) nitride intermediates are also successfully isolated by careful control of the irradiation time.

Results

Synthesis and structural characterization

Upon treatment of complex [{U[N(CH3)(CH2CH2NPiPr2)2](Cl)2(THF)}] (1)28 (Fig. 2) with 1 equiv. of [RhCl(COD)]2 (COD = cyclooctadiene) in tetrahydrofuran (THF) at room temperature (RT) overnight, the complex [{U{N(CH3)(CH2CH2NPiPr2)2}(Cl)2 [(μ-Cl)Rh(COD)]2}] (2a) was formed and was isolated as green crystals in 81% yield after recrystallization from toluene at −30 °C (Fig. 2). Its iridium analog, [{U{N(CH3)(CH2CH2NPiPr2)2}(Cl)2[(μ-Cl)Ir(COD)]2}] (2b), could also be prepared by the reaction of 1 with 1 equiv. of [IrCl(COD)]2 following the same procedure, and was isolated as orange crystals in 74% yield. The formation of complexes 2a and 2b is analogous to the frustrated Lewis pair reactivity observed previously in rare-earth organometallic chemistry50,51,52. Both 2a and 2b were characterized by nuclear magnetic resonance (NMR) spectroscopy (Supplementary Figs. 1 and 2), elemental analysis, and single-crystal X-ray diffraction.

Fig. 2: Synthesis and reactivity of complexes 2, 3, 4, and 5.
figure 2

Complexes 3a and 3b with four azides were prepared by the reaction of complexes 2 with NaN3. By careful control of the irradiation time of complexes 3, the U(VI) nitrides 5 were formed via the U(V) nitrides intermediates 4. The nitride groups in complexes 5 react with acid or H2/acid to form NH4+ in good yield. The N atoms originating from N3− are shown in blue and transition metals (M) are shown in red.

Full size image

Treatment of 2a with 4 equiv. of NaN3 in THF at RT overnight led to the formation of a uranium(IV) azide complex [{U{N(CH3)(CH2CH2NPiPr2)2}(N3)2[(μ-N3)Rh(COD)]2}] (3a) as brown crystals in 70% yield by generating crystals from the concentrated reaction mixture (Fig. 2). The complex 2b also reacts with 4 equiv. of NaN3 affording the corresponding complex [{U{N(CH3)(CH2CH2NPiPr2)2}(N3)2[(μ-N3)Ir(COD)]2}] (3b) as red crystals in 65% yield. The 1H NMR spectra of complexes 3a and 3b exhibit a broad range of peaks from +67.46 to −50.01 ppm and +70.12 to −52.63 ppm, respectively, consistent with the paramagnetic U(IV) species (Supplementary Figs. 3 and 4).

The asymmetric structures of 2a and 2b feature one U atom and two Rh or Ir atoms, which are bridged by two Cl ligands (Fig. 3). The U–Namido and U–Namine bond lengths are comparable to those observed in the precursor 1. The bridged U–Cl distances are approximately 0.2 Å longer than the terminal U–Cl bonds, revealing a weak interaction between the U center and the bridged Cl atoms.

Fig. 3: Molecular structures of 2 and 3.
figure 3

a–d Solid-state structures of 2a (a), 2b (b), 3a (c) and 3b (d) by X-ray crystallography with 50% probability ellipsoids. Solvent molecules, hydrogen atoms, and the isopropyl moieties in PiPr2 are omitted for clarity. Uranium, green; rhodium, violet; Iridium, pink; phosphorus, orange; nitrogen, blue; chlorine, yellow-green; and carbon, grey.

Full size image

The molecular structures of complexes 3a and 3b were also confirmed by X-ray crystallography (Fig. 3). The salient features of complexes 3a and 3b are the two terminal uranium azide ligands and two sideways-bound azide units bridging Rh or Ir atoms. The distances of two-terminal U–Nazide bonds in 3a (2.307(5) and 2.304(5) Å) and 3b (2.295(8) and 2.311(8) Å) are significantly shorter than the bridged U–Nazide bond lengths (2.515(5) and 2.453(5) Å for 3a, 2.533(8) and 2.475(7) Å for 3b), but these all fall in the range of previously reported U(IV) azide complexes (2.142(5)−2.564(12) Å)23,24,34,36,38,53,54. The bond lengths of Rh–Nazide in 3a (2.091(5) and 2.103(5) Å) and Ir–Nazide in 3b (2.086(8) and 2.072(7) Å) are consistent with the values (2.076(10)−2.28(9) Å) previously reported in complexes with side-on bridged azide ligands55,56. The angles of bridged U-N-N (128.7(4)° and 125.2(4)° for 3a; 128.3(6)° and 124.6(6)° for 3b) are more distant from the terminal U-N-N (168.6(6)° and 160.5(5)° for 3a; 165.6(9)° and 161.4(8)° for 3b).

The reduction, photolysis, and thermolysis of uranium azide complexes are known to be effective methods for the synthesis of uranium nitrides with the release of N28,21,31,34,36,38,39. Accordingly, we attempted to synthesize uranium nitride complexes by these routes. Only unidentified products were formed by the reduction of 3 with KC8 or thermolysis of 3 at 70 °C for 2 days as suggested by the in-situ 1H NMR spectra (Supplementary Figs. 5–8). Photolysis of complex 3a in THF under UV light irradiation for 2 days resulted in the formation of [{U{N(CH3)(CH2CH2NPiPr2)2}(N3)[(μ-N3)Rh(COD)]2[(μ-N)Rh(COD)]}] (4a), which was isolated as brown crystals in 54% yield from toluene at RT (Fig. 2). In comparison with 3a, the 1H NMR spectrum of 4a exhibits a narrow spectral range of +23.37 ppm to −12.46 pm, which is consistent with a previously reported U(V) complex [{U(TrenDMBS)}2(μ-N)] (from +24.86 to −13.31 ppm) and reflects the 5f1 nature of these species (Supplementary Fig. 9)21. Exposure of complex 3b to UV light for 1 day resulted in the disappearance of 3b, and the formation of complex [{U{N(CH3)(CH2CH2NPiPr2)2}(N3)[(μ-N3)Rh(COD)]2[(μ-N)Rh(COD)]}] (4b) as brown crystals in 52% yield after recrystallization (Fig. 2). An in-situ 1H NMR study (Supplementary Fig. 10) showed that no paramagnetic uranium by-products were observed in this photolysis although this process involves the dissociation and rearrangement of complexes 3. Attempts to isolate the uranium by-products were unsuccessful.

After further photolysis of complex 4a under UV light for 2 days, a diamagnetic U(VI) species [{U{N(CH3)(CH2CH2NPiPr2)2}(N3){[(μ-N)Rh(COD)]2[Rh(COD)]}}] (5a) was isolated in 65% yield after recrystallization (Fig. 2). Similar to the Rh analogs, complex 4b could be further photolysed under UV light for 1 day, leading to the formation of the complex [{U{N(CH3)(CH2CH2NPiPr2)2}(N3){[(μ-N)Ir(COD)]2[Ir(COD)]}}] (5b) as brown crystals in 58% yield. The U(V) centers in complexes 4a and 4b were oxidized to U(VI) in complexes 5a and 5b in this photolysis process. Complexes 4 and 5 are heterobimetallic bridged nitrides featuring uranium and transition metals. Both 5a and 5b could be obtained by the photolysis of 3a and 3b under UV light for 4 and 2 days, respectively. These results suggest that the photolysis of 3 is controllable and that complexes 4a and 4b can be viewed as intermediates in these processes.

The solid-state structures of 4 and 5 were determined by single-crystal X-ray diffraction. The U centers in 4a and 4b are coordinated with a tridentate [N(CH3)(CH2CH2NPiPr2)2]2− ligand, three azide ligands, and a nitride ligand and have a distorted pentagonal bipyramidal geometry (Fig. 4). The nitride atom (N4) is bonded to one U atom and two Rh atoms. The U1-N4 distances are 1.963(5) Å in 4a and 2.011(4) Å in 4b, which are shorter than the U-N multiple bond lengths in a uranium-stabilized U(V) nitride complex with a U(V)=N=U(V) unit (2.0470(3) and 2.0511(3) Å)22. For further comparison, the U-Nnitride bond distances in the sodium- or borane-capped U(V) nitrides are 1.883(4) Å and 1.916(4) Å, respectively30,47. The Rh/Ir–N4 bond lengths in 4a and 4b are close to the Rh/Ir–Nazide bond lengths and are comparable to those observed in 3a and 3b, respectively.

Fig. 4: Molecular structures of 4 and 5.
figure 4

a–d Solid-state structures of 4a (a), 4b (b), 5a (c), and 5b (d) by X-ray crystallography with 50% probability ellipsoids. Solvent molecules, hydrogen atoms, and the isopropyl moieties in PiPr2 are omitted for clarity. Uranium, green; rhodium, violet; Iridium, pink; phosphorus, orange; nitrogen, blue; and carbon, grey.

Full size image

The centrosymmetric structures of 5 feature two bridged nitride atoms, which are capped by three Rh or Ir atoms (Fig. 4). The bond lengths of U1–N3 in 5a (1.975(5) Å) and 5b (2.005(6) Å) are comparable with the U(V)-Nnitride bond lengths found in complexes 4a and 4b, respectively. However, the U(VI)-Nnitride bond lengths in complexes 5 are longer than those in the U(VI)-Nimido bond in a trans-imido uranium complex (1.840(4)−1.866(2) Å)57 or in the U(VI)-Nnitride bond in a uranyl analog [OUN]+ (1.818(9) Å)48. Two resonance structures (Supplementary Fig. 33) are mainly responsible for the longer U(VI)-Nnitride bond lengths in complexes 5. The bond lengths of Rh-N3 (2.082(4) Å and 2.067(4) Å) and Ir–N3 (2.059(6) Å and 2.047(6) Å) in 5a and 5b are comparable to the Rh/Ir-N4 bond distances observed in 4a and 4b, respectively. Complexes 5 represent an example of a transition metal-stabilized U(VI) nitride species, which is generated by the photolysis of a uranium azide species.

Stability and reactivity of uranium-nitride complexes

The combination of late transition metal (Rh and Ir) in U(V) nitrides 4 and U(VI) nitrides 5 are different from the main group borane-capped U(VI) nitride47 or another uranium-stabilized U(VI) nitrides48,49. Consequently, the stabilities of the newly synthesized uranium nitrides 4 and 5 were investigated. Both 4 and 5 exhibits good stability in the solid-state or in THF solution for at least 2 weeks under a N2 atmosphere, probably because the active nitrides are stabilized by transition metals. Irradiation of U(V)-nitride-azide complexes 4a and 4b with UV light for 2 and 1 days resulted in the formation of U(VI)-nitride complexes 5a and 5b, respectively (Fig. 2). However, longer irradiation times than those used in the formation of 5 decompose the product to compounds which could not be isolated or identified.

The protonation of metal nitrides to produce NH3 is a critical step in N2 reduction and conversion. Thus, we examined the reactivity of 4 and 5 toward acid and H2 (Fig. 2). We found that substantial NH4Cl (in 68-82% yields) was formed by the reactions of uranium nitride complexes 4 and 5 with excess PyHCl in THF solution (Supplementary Figs. 14–17). The addition of 1 atm H2 to the crystalline solids or to the degassed THF solution of 4 and 5, led to a gentle color change from brown to red. Unfortunately, attempts to isolate the single crystals of the resulting products were unsuccessful. However, the addition of excess PyHCl to the reaction mixtures of 4 and 5 with H2 in THF solution led to the formation of NH4Cl in 41–67% yields, respectively (Supplementary Figs. 18–21). In comparison, no NH4+ was observed from the reactions of uranium azide complexes 3 with excess PyHCl (Supplementary Fig. 22). These results suggest that the nitride units in complexes 4 and 5 are the sources of the NH4+.

Magnetic studies and electronic absorption spectra

The variable-temperature magnetic data of complexes 3 and 4 in the solid-state were measured by a superconducting quantum interference device (SQUID) magnetometry (Fig. 5). The magnetic moments of 3a and 3b at 300 K are 3.90 μB and 3.54 μB, respectively, which are close to the theoretical value (3.58 μB) for the 5f2 uranium ion in the 3H4 ground state. With decreasing temperatures, the magnetic moments of 3a and 3b decrease persistently to 0.43 μB and 0.47 μB at 1.8 K, respectively, and with a trend to zero, which is characteristic of U(IV) complexes58,59. The effective magnetic moments of 4a and 4b at 300 K are 2.76 μB and 2.30 μB, respectively, which are very close to the theoretical value of 2.54 μB for one U(V) ion. The magnetic moments of 4a and 4b decrease smoothly to 0.91 μB and 1.03 μB at 1.8 K, respectively, with decreasing temperature. The temperature dependence and the magnitude of μeff values of complexes 4a and 4b are consistent with reported U(V) complexes30,59. These results suggest that the oxidation states of uranium centers in complexes 3 and 4 are U(IV) and U(V), respectively.

Fig. 5: Magnetic data.
figure 5

Variable-temperature effective magnetic moment data of 3a (blue), 3b (green), 4a (red) and 4b (pink).

Full size image

The ultraviolet/visible/near-infrared (UV/Vis/NIR) electronic absorption spectra of complexes 3-5 were recorded in THF at RT (Supplementary Figs. 27–32). These complexes exhibit similar absorption profiles with gradually decreasing molar absorption at high energy in the UV–Vis region but show different fingerprint features in the NIR region. Complex 3a has a moderately intense absorption peak at 413 nm (ε = 4265 M−1 cm−1), while complex 3b exhibits three moderate intense peaks at 330, 382, and 449 nm (ε = 3060–6518 M−1 cm−1). These absorptions can be assigned to ligand-to-metal charge-transfer transition. Complex 4a exhibits a moderately intense charge-transfer band at 413 nm (ε = 3731 M−1 cm−1) while 4b displays three peaks at 333, 378, and 446 nm (ε = 3159–7154 M−1 cm−1). Consistently, the charge-transfer transitions in 5a and 5b are observed at 451 nm (ε = 3384 M−1 cm−1) and 333, 382, and 451 nm (ε = 3081–5870 M−1 cm−1), which are similar to the spectra of 3a and 3b, respectively. In the NIR region, broad, but very weak absorption bands from 975 to 1200 nm (ε < 70 M−1 cm−1) were observed for 3a and 3b, which were assigned to f–f transitions and were consistent with the U(IV) complexes with 5f2 electronic configuration60,61,62. Complexes 4a and 4b exhibit similar weak absorption bands at 1030, 1310, and 1630 nm (ε < 90 M−1 cm−1) in the NIR region, which could be attributed to f-f transitions of U(V) complexes with 5f1 electronic configuration63,64. As expected for a U(VI) species with a 5f0 electronic configuration, the spectra of 5a and 5b show no characteristic absorption peak in the NIR region (900–1700 nm), consistent with previously reported U(VI) complexes [U(N)(TrenTIPS)]33 and [NBu4][U(OSi(OtBu)3)4(N)]39.

Theoretical studies

DFT calculations (B3PW91) were used to describe the bonding properties of these complexes, in particular complexes 4 and 5. To probe the validity of the computational approach, the geometry of each of the different complexes was compared with the experimental geometry. The two-terminal U–Nazide bonds are well reproduced (2.27/2.30 Å in 3a and 2.27/2.30 Å in 3b). The same is true for the two bridged U–Nazide bonds (2.49/2.52 Å in 3a and 2.51/2.53 Å in 3b). In the same way, the Rh-Nazide bonds are correctly described (2.11 Å) indicating that the computational method is appropriate to describe such complexes. The same results were obtained for complexes 4a/4b and 5a/5b with the same accuracy and a maximum deviation of 0.05 Å (see Supplementary Tables 11, 14, 17, and 20). It is interesting to note that the LUMOs of complexes 3a and 3b (Supplementary Figs. 43 and 50) imply mainly two natural localized molecular orbitals (NLMOs), one being an antibonding π interaction between N8 and N9 and one with a bonding interaction between N7 and U1 as found in the canonical molecular orbital (CMO) analysis. Moreover, the bonding of the azide ligand with both U and Rh leads to a non-symmetrical contribution of the two N-N bonds to the N-N-N hyperbond. Indeed, for a free azide, each N-N bond contributes to 50% to the hyperbond N-N-N while in complex 3a for instance the N7-N8 only contributes to 37% to the hyperbond. This clearly means that the coordination of the azide ligand to both U and Rh leads to a relocalization of the density between N8 and N9 which would ultimately lead to the N2 formation and loss from the azide. It is interesting to notice that according to these orbital energies, the calculated SOMO-LUMO gap is 78.0 kcal mol−1 explaining that photolysis is required to make complexes 3a/3b evolve and to form complexes 4a/4b. Moreover, it is interesting to note that the formation of the two latter complexes from 3a/3b is thermodynamically favorable (by 17.6 kcal mol−1 for 4a and 18.2 kcal mol−1 for 4b). The bonding situation was thus analyzed for complexes 4a and 4b and the bond lengths were found to be well reproduced.

Analysis of the unpaired spin densities (1.23 in 4a and 1.24 in 4b), clearly indicates the presence of a U(V) metal center in these two complexes. At the NBO level, the U1-N4 bond is found to be a double bond with polarized bonds toward N (68% for σ and 72% for π in 4a and 72/75% in 4b, Supplementary Table 10). These bonds involve fd hybrid orbitals on uranium (78% 5f-20% 6d for σ and 59% 5f-39% 6d for π in 4a but 59% 5f-33% 6d and 57% 5f-33% 6d in 4b). Although polarized, these bonds appear to be relatively covalent as demonstrated by the U1-N4 Wiberg Bond Index (WBI) of 1.96 in 4a (1.83 in 4b). The associated Rh1-N4 bond was only observed at the second-order donor-acceptor level but U-N-Rh interactions are found in the NLMO as a 3-centers bond, the associated WBI being 0.45. On the other hand, the Ir1-N4 bonds in 4b are similar to the Rh-N ones in 4a with a WBI of 0.49. The NBO decomposition of the NLMOs associated with the U-N4 bonding interactions (Supplementary Table 13) shows some delocalization toward the Rh centers (resp. Ir, Supplementary Table 19), in line with some 3c-2e bonds. This is similar to what was found by scrutinizing the molecular orbitals (MO, Fig. 6), the U1-N4-Rh(Ir) bonds in 4a and 4b are 3c-2e bonds. Similar to complexes 3a and 3b, the LUMOs of complexes 4a and 4b (Supplementary Figs. 45 and 52) also display some antibonding interactions in the azide ligands. According to the orbital energy, the SOMO-LUMO energy gap is 80.0 kcal mol−1 consistent with the need of photolysis to make complexes 4a/4b to evolve. For the transformation of 4a/4b into complexes 5a/5b, the reaction is thermodynamically favored by 43.5 kcal mol−1 (5a) and 49.6 kcal mol−1 (5b).

Fig. 6: Molecular orbitals.
figure 6

U-N-Rh(Ir) bonding molecular orbitals of complexes 4a (a), 4b (b), 5a (c) and 5b (d).

Full size image

Finally, the bonding in complexes 5a/5b was analyzed. The optimized geometries compare well with the experimentally observed geometries. In both complexes, the uranium centers are at the oxidation state +VI and have no residual spin densities. The two double U-N bonds are similar to those found in complexes 4 with a WBI of 1.91 for 5a (1.80 for 5b), and the NLMO, as well as the MO, clearly indicate 3c-2e U-N-Rh(Ir) bonds (Fig. 6). The slight decrease observed of the WBI and therefore of the covalency while going from complexes 4 to 5 might be explained by the strength of the U-Rh interaction in the two systems. Indeed, in 4 where the U-M distance is around the sum of the van der walls radii, the U…Rh WBI is 0.10 whereas in 5 where the distance is shorter, and the WBI is larger (0.22). The stronger U-Rh interaction would thus decrease the U-Nnitrido interaction. At the NBO level, the U1-N3 and U1-N3’ bonds are found to be U-N double bonds polarized toward N and these bonds also involve fd hybrid orbitals on uranium (Supplementary Table 10).

Discussion

In summary, transition metal stabilized U(VI) nitride complexes 5a and 5b were generated by photolysis of U(IV) azide precursors, 3a and 3b, respectively. The U(V) nitride intermediates 4a and 4b were successfully isolated by controlling the duration of the irradiation. They were sufficiently stable to be fully characterized and could be further photolyzed to the U(VI) nitrides 5 under UV irradiation conditions. Therefore, the photolysis of the azide units in complexes 3a and 3b is controllable and is a stepwise process. The excellent stabilities of both the U(V) nitride intermediates 4 and the U(VI) nitride products 5 are probably due to the active nitrides being stabilized by transition metals. A DFT study allowed us to analyze the bonding of these complexes, indicating the presence of U-N multiple bonds and U-N-Rh(Ir) 3c-2e interactions. Moreover, DFT calculations showed that the photolysis, which is necessary to break the N-N bond of the azide ligands, suggests excitation from an f-electron of the metal to the LUMOs of complexes 3 and 4, and this primarily displays an antibonding N-(N2) interaction. This study implies that a multiple N-P ligand is an effective platform with which to capture unstable U(VI) nitrides with the aid of TMs65.

Methods

General considerations

Experiments were performed under argon or nitrogen atmosphere using standard Schlenk-line and glove-box techniques. Solvents were dried and degassed with a solvent purification system before use. See the supplementary information for detailed experimental procedures, crystallographic analyses (Supplementary Figs. 34–41 and Supplementary Tables 1–9), and computational details (Supplementary Dataset, Supplementary Figs. 42–56, and Supplementary Tables 10–22).

Synthesis of 2a and 2b

A solution of complex 1 (72 mg, 0.1 mmol, 1 equiv.) in THF was added dropwise to a THF solution of [RhCl(COD)]2 (49 mg, 0.1 mmol, 1 equiv.) under an atmosphere of nitrogen. The mixture was stirred overnight at RT, and then the solvents were removed under reduced pressure and the residues were extracted with toluene. The filtrate was concentrated to 1 mL and placed at −30 °C for 24 h. Green crystals of 2a (93 mg, 81%) suitable for X-ray diffraction were obtained. Orange crystals of 2b (98 mg, 74%) were obtained following the same procedure in which complex 1 (72 mg, 0.1 mmol, 1 equiv.) was treated with 1 equiv. [IrCl(COD)]2 (67 mg, 0.1 mmol, 1 equiv.). 2a: 1H NMR (THF-d8, 400 MHz, ppm) δ 25.93 (s, 2H), 20.98 (s, 6H), 17.66 (s, 2H), 16.78 (s, 2H), 14.16 (s, 2H), 13.35 (s, 2H), 11.06 (s, 2H), 10.35 (s, 12H), 8.71 (s, 6H), 5.15 (s, 6H), 2.80 (s, 12H), −1.19 (s, 2H), −3.12 (s, 2H), −4.04 (s, 2H), −39.61 (s, 1H), −42.76 (s, 1H), −46.13 (s, 1H). 31P{1H} NMR (THF-d8, 162 MHz, ppm) δ 78.94 (d, 1JP-Rh = 392 Hz), −87.20 (br). Anal. Calcd. for C33H63Cl4Rh2N3P2U.toluene: C, 38.69; H, 5.76; N, 3.38. Found: C, 38.12; H, 5.88; N, 3.36. 2b: 1H NMR (THF-d8, 400 MHz, ppm) δ 22.79 (s, 2H), 19.36 (s, 6H), 16.40 (s, 3H), 15.35 (s, 3H), 11.84 (s, 6H), 9.55 (s, 2H), 8.85 (s, 7H), 8.29 (s, 3H), 7.50 (s, 6H), 5.87 (s, 2H), 4.82 (s, 1H), 3.65 (s, 1H), 3.24 (s, 1H), −0.47 (s, 4H), −1.73 (s, 4H), −3.29 (s, 4H), −10.87 (s, 1H), −31.83 (s, 1H), −39.26 (s, 2H), −40.94 (s, 2H), −49.69 (s, 2H). 31P{1H} NMR (THF-d8, 162 MHz, ppm) δ 63.59, −150.90. Anal. Calcd. for C33H63Cl4Ir2N3P2U.toluene: C, 33.83; H, 5.04; N, 2.96. Found: C, 33.67; H, 5.20; N, 2.92.

Synthesis of 3a and 3b

A solution of complex 2a (57 mg, 0.05 mmol, 1 equiv.) in THF was added to a mixture of NaN3 (13 mg, 0.2 mmol, 4 equiv.) in THF under an atmosphere of nitrogen. The mixture was stirred overnight at RT and then the mixture was filtered. The filtrate was concentrated to 1 mL and placed at −30 °C for 24 h, producing brown crystals of 3a suitable for X-ray diffraction (41 mg, 70%). Red crystals of 3b (44 mg, 65%) were obtained following the same procedure by treatment of 2b (66 mg, 0.05 mmol, 1 equiv.) with 4 equiv. NaN3 (13 mg, 0.2 mmol, 4 equiv.). 3a: 1H NMR (THF-d8, 400 MHz, ppm) δ 67.46 (s, 1H), 59.45 (s, 1H), 19.80 (s, 6H), 7.00 (s, 12H), −3.17 (s, 12H), −8.97 (s, 6H), −10.60 (s, 6H), −12.85 (s, 6H), −15.62 (s, 1H), −22.34 (s, 2H), −25.67 (s, 1H), −31.81 (s, 2H), −42.28 (s, 2H), −44.68 (s, 3H), −49.98 (s, 1H), −50.01 (s, 1H). 31P{1H} NMR (THF-d8, 162 MHz, ppm) δ 76.79 (d, 1JP-Rh = 392 Hz), 77.58 (d, 1JP-Rh = 396 Hz). Anal. Calcd. for C33H63N15P2Rh2U.THF: C, 35.61; H, 5.74; N, 16.84. Found: C, 35.80; H, 5.73; N, 16.44. 3b: 1H NMR (THF-d8, 400 MHz, ppm) δ 70.12 (s, 1H), 61.77 (s, 1H), 21.97 (s, 6H), 5.05 (s, 12H), −2.46 (s, 12H), −9.88 (s, 6H), −11.40 (s, 6H), −13.63 (s, 6H), −15.90 (s, 2H), −23.74 (s, 2H), 33.15 (s, 1H), −33.19 (s, 1H), −42.70 (s, 2H), −46.20 (s, 3H), −52.59 (s, 1H), −52.63 (s, 1H). 31P{1H} NMR (THF-d8, 162 MHz, ppm) δ 64.60, 64.41. Anal. Calcd. for C33H63Ir2N15P2U.THF: C, 31.15; H, 5.02; N, 14.73. Found: C, 31.17; H, 4.98; N, 13.99.

Synthesis of 4a and 4b

A solution of complex 3a (59 mg, 0.05 mmol, 1 equiv.) in THF (5 mL) was transferred into a Schlenk tube under an atmosphere of nitrogen and then irradiated with a 40 W UV lamp for 2 days. The solvents were removed under reduced pressure and the residues were extracted with toluene. The filtrate was concentrated to 0.5 mL and placed at RT for 48 h, giving brown crystals of 4a suitable for X-ray diffraction (24 mg, 54%). Brown crystals of 4b were obtained following the same procedure by exposing a THF solution of complex 3b (68 mg, 0.05 mmol, 1 equiv.) to UV light for 1 days (28 mg, 52%). 4a: 1H NMR (THF-d8, 400 MHz, ppm) δ 23.37 (s, 1H), 15.28 (s, 2H), 12.58 (s, 1H), 9.79 (s, 1H), 9.49 (s, 1H), 9.08 (s, 3H), 8.33 (s, 3H), 8.18 (s, 1H), 6.89 (s, 1H), 6.69 (s, 1H), 6.23 (s, 2H), 5.80 (s, 2H), 5.62 (s, 2H), 5.53 (s, 1H), 5.26 (s, 2H), 5.04 (s, 2H), 4.88 (s, 3H), 4.82 (s, 1H), 4.47 (s, 1H), 4.40 (s, 6H), 3.77 (s, 6H), 3.68 (s, 1H), 2.95 (s, 2H), 1.28 (s, 6H), 0.99 (s, 2H), 0.66 (s, 2H), 0.46 (s, 3H), 0.02 (s, 2H), −0.39 (s, 2H), −0.77 (s, 2H), −2.89 (s, 1H), −4.05 (s, 1H), −5.45 (s, 1H), −6.08 (s, 1H), −8.05 (s, 4H), −10.79 (s, 1H), −12.46 (s, 1H). 31P{1H} NMR (THF-d8, 162 MHz, ppm) δ 94.65 (br), 76.79 (d, 1JP-Rh = 392 Hz). Anal. Calcd. for C41H75N13P2Rh3U: C, 36.24; H, 5.56; N, 13.40. Found: C, 36.59; H, 5.63; N, 13.35. 4b: 1H NMR (THF-d8, 400 MHz, ppm) δ 24.73 (s, 1H), 16.26 (s, 1H), 15.08 (s, 1H), 12.34 (s, 1H), 11.32 (s, 1H), 10.70 (s, 1H), 10.42 (s, 1H), 9.53 (s, 3H), 8.26 (s, 1H), 7.58 (s, 1H), 6.66 (s, 2H), 6.47 (s, 3H), 6.17 (s, 2H), 5.96 (s, 2H), 5.33 (s, 4H), 5.15 (s, 2H), 5.09 (s, 3H), 4.85 (s, 2H), 4.73 (s, 6H), 4.55 (s, 6H), 4.46 (s, 1H), 4.35 (s, 2H), 4.19 (s, 2H), 1.56 (s, 3H), 0.96 (s, 6H), 0.90 (s, 2H), −0.40 (s, 2H), −0.40 (s, 2H), −0.48 (s, 2H), −3.20 (s, 1H), −5.36 (s, 1H), −6.87 (s, 1H), −9.53 (s, 4H), −10.95 (s, 1H), −12.17 (s, 2H), −12.48 (s, 1H). 31P{1H} NMR (THF-d8, 162 MHz, ppm) δ 73.08, 68.63. Anal. Calcd. for C41H75N13P2Ir3U.1.5toluene: C, 35.05; H, 4.97; N, 10.32. Found: C, 35.14; H, 4.99; N, 10.44.

Synthesis of 5a and 5b

Method A: A solution of complex 4a (34 mg, 0.025 mmol, 1 equiv.) in THF (5 mL) was transferred into a Schlenk tube under an atmosphere of nitrogen and then irradiated with a 40 W UV lamp for 2 days. The mixture was filtered and the filtrate was concentrated to 0.5 mL and placed at RT for 48 h, brown crystals of 5a suitable for X-ray diffraction were obtained (21 mg, 65%). Brown crystals of 4b were obtained following the same procedure by exposing a THF solution of complex 5b (41 mg, 0.025 mmol, 1 equiv.) to UV light for 1 days (23 mg, 58%). Method B: A solution of complex 3a (59 mg, 0.05 mmol, 1 equiv.) in THF (5 mL) was transferred into a Schlenk tube under an atmosphere of nitrogen and then irradiated with a 40 W UV lamp for 4 days. The mixture was filtered and the filtrate was concentrated to 0.5 mL and placed at RT for 48 h, brown crystals of 5a suitable for X-ray diffraction were obtained (16 mg, 38%). Brown crystals of 5b were obtained using the same procedure by exposing a THF solution of complex 3b (68 mg, 0.05 mmol, 1 equiv.) to UV light for 2 days (17 mg, 32%). 5a: 1H NMR (THF-d8, 400 MHz, ppm) δ 6.36 (s, 2H), 5.88 (s, 2H), 4.91-4.95 (m, 5H), 3.34-3.42 (m, 6H), 2.84–2.96 (m, 6H), 2.35–2.42 (m, 12H), 2.15–2.20 (m, 6H), 1.54–1.57 (m, 12H), 1.10–1.19 (m, 24H). 31P{1H} NMR (THF-d8, 162 MHz, ppm) δ 41.48 (d, 1JP-Rh = 372 Hz). Anal. Calcd. for C41H75N8P2Rh3U: C, 38.21; H, 5.87; N, 8.69. Found: C, 38.24; H, 5.97; N, 8.50. 5b: 1H NMR (THF-d8, 400 MHz, ppm) δ 6.36 (s, 2H), 5.88 (s, 2H), 4.56–4.57 (m, 5H), 3.36–3.43 (m, 6H), 2.88–2.94 (m, 6H), 2.40–2.48 (m, 12H), 2.18–2.22 (m, 6H), 1.51–1.58 (m, 12H), 1.23–1.27 (m, 24H). 31P{1H} NMR (THF-d8, 162 MHz, ppm) δ 29.30. Anal. Calcd. for C41H75N8P2Ir3U: C, 31.63; H, 4.86; N, 7.20. Found: C, 32.63; H, 4.89; N, 7.36.